2000 HSR&D Annual Meeting: Abstract 200-267
*200. Impact of Filing a PTSD Disabilty Claim on Gulf War Veterans’ VA Outpatient Health Care Utilzation
Lawrence Armand Fortier, Minneapolis VA Medical Center, Center for Chronic Diseases Outcomes, Minneapolis, MN, M Murdoch, Minneapolis VA Medical Center, Center for Chronic Diseases Outcomes, DB Nelson, Minneapolis VA Medical Center, Center for Chronic Diseases Outcomes, S Nugent, Minneapolis VA Medical Center, Center for Chronic Diseases Outcomes, N Sayer, Minneapolis VA Medical Center, M Spoont, Minneapolis VA Medical Center
Objectives: The process veterans must follow to become service-connected (SC) for military-related disorders is reportedly stressful and may be especially difficult for those with posttraumatic stress disorder (PTSD). This increase in stress may lead to increased use of VA health care resources during the application process. Our objective was to describe the outpatient health care utilization of Gulf War era (GWE) veterans filing disability claims for PTSD and to assess the association between utilization and eventual SC status.
Methods: Historical cohort study of all GWE veterans filing PTSD disability claims between Aug. 1990 and May 1998. Outpatient visits were identified using VA administrative databases.
Results: In the period studied, 2,422 GWE veterans filed disability claims for service-related PTSD. Of these, 80% were male, and most had served in the Army (63%). Just 35% had claims approved (compared to an overall approval rate of 55%). Excluding visits made solely for the purpose of becoming service-connected, 70% of the cohort (n = 1,694) logged in at least one VA outpatient visit. In the year prior to filing a disability claim, these "VA users" averaged 1.68 clinic stops per month, 0.48 mental health visits per month, and 0.16 primary care visits per month. After filing a claim, but before a rating decision was made, the mean monthly number of clinic stops, mental health visits, and primary care visits made by VA users increased between 144% and 156%. Once a rating decision was made, health care utilization for both successful and unsuccessful claimants returned to rates similar to those in the pre-claim period. When adjusting for gender we found significant differences in monthly utilization rates in the different time periods (i.e., period prior to filing a claim, period after filing a claim but before a rating decision was made, and period after the rating decision was made). As in the unadjusted analysis, outpatient visits made after filing a claim but before receiving a rating decision exceeded visits in the other two periods by factors ranging from 1.3 to 1.4. We also found differences in monthly rates between gender but no SC status effects. Relative to women, male GWE veterans had 1.30 times as many clinic visits per month, 1.42 times as many primary care visits per months, and 1.40 times as many mental health care visits per month.
Conclusions: The increase in outpatient visits and, particularly, in mental health visits seen during the claims development process supports our contention that the process is stressful to veterans. That similar increases and falls in health care utilization were seen in both those veterans who ultimately became service-connected and those who did not argues against malingering. The gender discrepancy in utilization is also concerning in that it may represent greater barriers to VA care for women veterans with PTSD.
Impact: Future research is needed to develop a claims process that is as innocuous to veterans as possible and to ensure that both male and female veterans have adequate access to VA resources when filing disability claims.
*201. Iplementation of VA Guidelines for Self-monitoring of Blood Glucose (SMBG): Effect on Glucose Control and Pharmacy Costs
Joy L. Meier, PharmD, VA Northern California Health Care System, Martinez, CA, ALM Swislocki, VA Northern California Health Care System, P Bartlebaugh, VA Northern California Health Care System, JR Lopez, VA Northern California Health Care System, RH Noth, VA Northern California Health Care System, D Siegel, VA Northern California Health Care System
Objectives: Recent VA Medical Advisory Panel Guidelines for Pharmacologic Management of Diabetes Mellitus (DM) recommend that stable Type II diabetics controlled on oral agents or diet therapy perform SMBG twice-weekly. Data from VA Northern California Healthcare System (VANCHCS), as well as other VA facilities, indicate that SMBG testing is commonly done more frequently than recommended. SMBG represents a major cost of diabetes care, yet optimal testing frequency is controversial. The objective of this research project was to assess the impact of implementation of Diabetes Clinical Practice Guidelines for SMBG on glucose control and pharmacy costs in Type II diabetics.
Methods: Researchers analyzed prescriptions for blood glucose monitoring strips from the computer database filled over both a six-month baseline period and a six-month post-implementation period and related these data to oral hypoglycemics filled over similar time periods using Access™ and Excel™. All patients receiving insulin were excluded from analysis. Providers and pharmacists were educated on SMBG testing frequency recommendations by means of written instruction and interactive presentations at primary care team meetings. Pharmacists changed BG strip prescriptions for diabetics performing SMBG and on diet or oral hypoglycemic therapy to 50 strips for 90 days and sent an instructional letter with the first prescription. Pharmacists and providers verbally instructed these diabetics to perform SMBG according to VA guidelines when possible. Researchers compared patient's baseline average testing frequency and HbA1c with those obtained during a six month interval beginning two months following implementation. Average monthly cost savings were calculated as the difference in the average monthly expenditures for the baseline vs. post-implementation period.
Results: At baseline, 913 of 1213 SMBG diabetics on oral hypoglycemics had HbA1c tested (mean 7.83 +/- 1.34%SD); their frequency of SMBG was 1.36 +/- 0.95 strips/pt/day. Post-implementation, 974 of 1278 diabetics had HbA1c tested (mean7.86 +/- 1.54%; p=.63 vs baseline); frequency of SMBG decreased by 46% to 0.74 +/-0.5 strips/pt/day (p<0.0001). At baseline, 154 of 254 SMBG diabetics on diet therapy had HbA1c tested (mean 6.85 +/- 0.97%); their frequency of SMBG was 1.07 +/- 0.90 strips/pt/day. Post-implementation, 177 of 282 diabetics on diet therapy had HbA1c tested (mean 6.78 +/- 1.20%; p=.56 vs baseline); frequency of SMBG decreased by 35% to 0.70 +/-0.51 strips/pt/day (p<0.0001). Average monthly cost savings were $8800, or $6.37/pt/month.
Conclusions: In VANCHCS, we successfully decreased the frequency of SMBG in Type II diabetics, resulting in substantial cost savings without affecting glucose control. Our current SMBG testing frequency in this population more closely aligns with the Diabetes CPG recommendations.
Impact: The VA could realize considerable cost-savings if these guidelines were followed nation-wide.
*202. Improved Sexual Function Following Castration in Lower SES Metastatic Prostate Cancer Patients: Can it be Real?
Nicholas Slimack, Northwestern University and Chicago VA Healthcare System, Chicago, Illinois, VA Cancer Prostate Cancer Outcomes Study, Northwestern University and Chicago VA Healthcare System
Objectives: Metastatic prostate cancer disease management requires consideration of quality of life (QOL) in treatment decision making. One prior study has reported worse emotional functioning after 3-months of castrating hormonal therapy (Moinpour C, JNCI 1998). Sexual dysfunction, pain, urinary tract symptoms, and psychosocial adjustment difficulties occur in patients with metastatic disease who undergo castration. To investigate the long term QOL impact of castration, we prospectively examined 23 newly diagnosed metastatic prostate cancer patients from the VA, many of whom are from low SES backgrounds. This cohort also represents a low-literacy population that has not been studied in prior QOL research that use self-administered questionnaires.
Methods: Newly diagnosed metastatic prostate cancer patients from urology and medical oncology clinics at 6 VA medical centers (Long Beach, Chicago (Westside/Lakeside), Durham, Shreveport, Gainesville) were interviewed at baseline (n=23), 3 months (n=23), and 12 months (n=12). Questions were administered by a research assistant. Participants were on average 73 years of age, 50% were married, 80% African-American, and 38% had not completed high school. All patients received hormonal/radiation therapy subsequent to enrollment. A cancer-specific QOL instrument (the EORTC multi-dimensional QOL instrument) was used, as it has been validated in the setting of low literacy prostate cancer patients (Knight S, J Urol 1998). Prostate cancer specific questions also were used.
Results: Findings for the QOL subscales suggested decrements in role, emotional, and global physical functioning 3-months after treatment, which were no longer apparent at 12-months. A symptomatic evaluation suggested that pre-treatment diarrhea symptoms decreased at 3-months, while nausea/vomiting and constipation problems persisted at both 3 and 12 month evaluations. Within the prostate-specific symptom items, patients experienced improved urinary function at 3-months but decreased at 12 months. In these castrated men, reports of improvements in sexual function and satisfaction were noted at both 3- and 12-months follow-up.
Conclusions: Among men with metastatic prostate cancer, decrements in emotional functioning following castration may be short-lived. Despite castration, sexual functioning was reported to improve, raising questions about comprehension and validity of prostate cancer specific QOL questions in this population. Prostate cancer patients may be unwilling to frankly discuss sexual function with an interviewer.
Impact: The results reported here raise concern about the best way to treat metastatic prostate cancer patients in terms of their QOL. The findings are inconsistent with other studies of metastatic patients receiving hormonal therapy. Further follow-up analysis and additional patient accrual will reveal any trends in the QOL dimensional aspects. A recent study by Moinpour and her colleagues found that at 3 month follow-up, the patients receiving flutamide had worse emotional functioning than those receiving placebo. The Moinpour et al. study also reported that the patients in the flutamide group had more diarrhea at 3 months than the placebo group.
HSR&D Funded: IRR-95-120.1.
*203. Improvement in Diabetic outcomes with the Implementation of a Proficiency Testing Program
Philip Foulis, MD, MPH, James A. Haley Veteran’s Administration Hospital, Tampa, FL, SD Barnett, Birch and Davis Associates, ED Felber, Total Therapeutic Management, Inc., RJ Schrot, James A. Haley Veteran’s Administration Hospital
Objectives: We evaluated the success of a voluntary home glucose monitoring proficiency testing program among diabetic patients by measuring length of hospital stay and percent change in hemoglobin A1c. Data was compared between individuals that entered the proficiency testing program and those that elected not to participate in it. The American Diabetes Association has established the goal of attaining a hemoglobin A1c less than 7% to retard long term vascular complications of diabetes. This program was implemented to verify the accuracy of home glucose testing among diabetic veterans, since this is a critical component to achieve the American Diabetic Association’s goal.
Methods: A voluntary proficiency testing program was initiated in January, 1996 to determine the accuracy of a patient’s finger stick glucose compared to the serum value in the laboratory. All patients receiving glucose strips (n = 3650) were notified of the program. Seventeen percent of the patients (n = 620) participated. Patients performed a finger stick glucose and the laboratory simultaneously assayed a serum glucose. Patient proficiency was assessed by comparing the variation of the finger stick glucose to the laboratory result. Data from a random selection of non-participants (n = 2972) was also analyzed. Analysis variables included age, race, sex, hemoglobin A1c values, and frequency of hospitalizations and length of stay. Comparisons were made between participants and non-participants, and stratified comparisons were made between pre-proficiency and post-proficiency testing. This determined whether the intervention of proficiency testing had an impact on the patient’s hemoglobin A1c and length of stay during hospitalization.
Results: Hemoglobin A1c levels were compared in the pre-proficiency versus the post-proficiency groups. There was a significant (p < 0.001) reduction of 0.45% in hemoglobin A1c results after participating in proficiency testing. Patient’s length of stay during each hospitalization was compared between the participants and non-participants. Participants in the program had a statistically significant shorter length of stay of 8.2 days (Confidence Interval 6.7 to 9.7) when compared to non-participants stay of 10.3 days (Confidence Interval 9.5 to 11.1).
Conclusions: Our findings suggest that participation in a proficiency testing program for home glucose monitoring improves hemoglobin A1c levels and decreases hospital stay. Patients receive useful information regarding their ability to obtain accurate glucose measurements while actively participating in their own care. Health care providers can utilize patients’ glucose results more confidently in patient management.
Impact: Improved hemoglobin A1c levels and decreased hospital days should reduce long-term complications in diabetic veterans and improve utilization of health care resources.
*204. Increasing Depression Treatment Adherence in Primary Care: Patient Treatment Preference
Nicole Maureen Hasenberg, VA PSHCS Seattle Division, Seattle, WA, EF Chaney, VA PSHCS Seattle Division, SC Hedrick, VA PSHCS Seattle Division, R Koperski, VA PSHCS Seattle Division
Objectives: Recent treatment guidelines implementation research has stressed the role of patient treatment preference in influencing treatment adherence. We assessed factors that affect patient treatment preference for medication and/or psychotherapy depression treatment as part of Project MOOD, the first randomized trial of collaborative care for depression in VA primary care. We hypothesized that current comorbid illness, previous mental health history, current and previous antidepressant medication experience, and views about depression causation would impact treatment preference.
Methods: From January 1998 to March 1999, Project MOOD enrolled and performed detailed interviews on 356 patients with major depression and/or dysthymia in the Seattle VA General Internal Medicine Primary Care Clinic. Unipolar depressed patients were excluded only if they were currently in mental health specialty care, were acutely psychotic or suicidal, or had primary substance abuse or PTSD. We asked patients about their treatment preferences as part of treatment planning.
Results: The majority of patients (58%) wanted both medication and counseling; 23% preferred counseling only; 14% wanted medication only; 3% had no preference and 2% did not want either treatment. The factors, which were significantly related to preference, were antidepressant medication experience and understanding of depression causation. The odds of a patient who is currently taking antidepressant medication (and is still depressed) preferring medication or both to counseling alone was 5.8 times the odds of a patient who is not currently taking antidepressant medication. The odds of a patient who believes that his/her depression is due to a medical illness preferring medication or both to counseling alone was 3.1 times the odds of a patient who believes depression is a personal problem. The odds of a patient who has taken antidepressant medication in the past preferring medication or a combination of both to counseling alone is 2.2 times the odds of a patient who has not previously taken medication.
Conclusions: Patients were more willing to consider medication or combined therapy if they were currently taking or had previously taken antidepressant medication, or if they believed that depression is a physical illness.
Impact: This information can assist health planners in allocating mental health treatment resources to primary care. It can also be used in designing depression treatment patient education materials.
HSR&D Funded: 95-097.B.
*205. Influence of Organizational Change on Learning Needs of VA Employees for the New Millennium
Karen M. Sanders, HH McGuire VA Medical Center, Richmond, VA, Richmond, Virginia, MA Jacobs, HH McGuire VA Medical Center, Richmond, VA, K Cooperman, Medical College of Virginia, Richmond, VA
Objectives: The VA health care system is an organization in transition. However, the effect of this change process on VA employees has not been well studied. We used data from serial educational needs assessments to not only assess educational needs, but to gain insight into our employees’ psychological adaptation to this rapidly changing environment. Objectives: 1) To determine and compare the learning needs of VA employees before and after implementation of Service Lines in one medical center 2) To identify critical educational needs as perceived by service line managers 3) To compare identified needs perceived by VA employees and service line managers.
Methods: An educational needs survey was developed by a panel of experts on the VISN 6 Subcouncil for Education/Research. The survey was distributed to staff in 1997 and 1998, prior to and following the implementation of Service Lines. In 1998, a focused structured interview was used to obtain information from service line chiefs regarding educational needs of their staff. Kendall’s tau was used to examine trends and compare rankings.
Results: Eight hundred and twenty-four employees (42 percent) responded in 1997. The top learning needs included Computer Skills (54 percent), Dealing with Change (40 percent), Continuing Education (34percent), and Stress Reduction (33 percent). Seven hundred and thirty-two (48 percent) of the employees responded in 1998. The top training needs as defined by our staff were PC Skills (40 percent), Opportunity for Advancement (36 percent), Stress Reduction (34 percent), Continuing Education (29 percent), Cross Training/New Skills (23 percent), and Retirement Counseling (21 percent). There were significant differences between the 1997 and 1998 rankings (Kendall’s Tau b 0.43, p.01). Service line chiefs ranked the top training needs for their staff as computer training, customer service training and supervisory training for new service line chiefs.
Conclusions: Computer training remains the top educational need identified by both groups. Stress reduction and continuing education remain high educational priorities for staff for both years. While respondents identified Dealing with Change as a top training need in 1997, they did not perceive it as a top training need in 1998. A major medical center educational effort was directed toward dealing with change during the transition into service lines, which may have helped staff and consequently reflected its lower ranking. It is noteworthy that staff identified cross-training and opportunities for advancement as top needs, which may reflect a healthy acknowledgment of the need for new skills in today’s environment. Service line chiefs were consistent as a group in their identification of educational needs. Customer service and related issues were identified as staff training needs. Interestingly, new service line chiefs (physicians) also identified that supervisory training was a needed component for them to successfully function in their new administrative roles.
Impact: In response to major structural changes in VHA, as well as technological advancements, educational needs of employees are shifting rapidly. National training initiatives must consider the dynamic nature of these needs and employ interventions that would meet these changing needs.
*206. Joint VHA/VBA Project to Reengineer Organizational Process of Providing C & P Examinations
Maudie C. Foy, RN, MS, ANP-CS, Jerry L. Pettis Memorial VAMC, Loma Linda, CA, DC Evans, Jerry L. Pettis Memorial VAMC, AL Dew, Jerry L. Pettis Memorial VAMC, K Allen, Los Angeles Regional Office (VBA), G Manar, Los Angeles Regional Office (VBA)
Objectives: To improve a local facility’s ability to meet joint VHA/VBA Compensation & Pension examination quality standards and timeliness goals
Methods: A Joint VHA (Pettis VAMC) and VBA (LARO) venture was piloted to train a nurse practitioner (NP) to serve as the nexus between them. The NP underwent a three-month training program at the LARO to understand the adjudication process and the problems encountered with incorrectly prepared medical disability examinations. The NP facilitated four continuous quality improvement cycles empowering the C & P team to develop novel solutions to streamline the process. The following steps were accomplished: 1) The NP daily reviews all C & P exam requests to ensure the minimum number of appropriate clinic requests are sent. 2) The NP performs a 100% review of all completed exams to make sure the reports are accurate for rating purposes. 3) The NP trained all clerical staff and clinicians on what is required for each type of exam. 4) The NP empowered the clerical staff’s work process by helping them create new Excel spreadsheets to track and case manage each exam request.
Results: The C & P process to fix inadequate examinations was streamlined from six to three steps. A previously dysfunctional system that averaged 55 days to process medical disability examinations was transformed into one that averages 28 days from time a request is received to electronic release of completed examinations to the appropriate VBA regional office. The percentage of "inadequate/insufficient exams" has fallen from 3% to less than one percent.
Conclusions: A novel approach to a difficult bureaucratically snarled system (C & P examinations) was developed. We pioneered, via joint effort by facility administration and the LA Regional Office, a new methodology, that of a NP serving as clinician/administrator of a local C & P program and simultaneously the link between them, that transformed the quality and timeliness of the C & P examination process. The change was based on the recognized fact that there is a knowledge, language and cultural gap between VBA and VHA. Cross training of clinical personnel can fill that gap and make the system work as it was intended. The nurse practitioner is viewed by both partners as "one of us" (part of their individual cultures) and thus is able to smooth the movement of C & P examinations through the usual bureaucratic process.
Impact: A VBA rating-trained nurse practitioner guided re-engineering of a local C & P process whereby one of the key changes resulted in the NP being seen as a staff member of both a VHA Medical Center and a VBA Regional Office; this has transformed a local facility’s ability to offer superior customer service in timely, cost-efficient manner
*207. Journal Reading Habits of Internists
Sanjay Saint, MD, MPH, Ann Arbor VA HSR&D, Ann Arbor, MI, DA Christakis, University of Washington, S Saha, Portland VAMC, JG Elmore, University of Washington, DE Welsh, Ann Arbor VA HSR&D, TD Koepsell, University of Washington
Objectives: Dissemination of research findings relies heavily on medical journals. However, little is known about the current reading habits of physicians and their perceptions of the quality of the medical journals they read. We thus undertook a national survey of internists to answer the following questions: 1) How much time do internists currently spend reading journals? 2) Which journals do they read? 3) What are their impressions of those journals? and 4) What proportion of the time do internists read only the abstract rather than the entire article? In addition, we were curious whether the answers to these questions differed based on an internist's training in clinical epidemiology and study design.
Methods: A random sample of 208 internists from the AMA physician masterfile and 208 internists with formal training in epidemiology and study design were the subjects of this study. Participants were asked to estimate the number of hours they spent reading journals per week and the percentage of articles for which they limited their reading to only the abstract. Respondents were also asked which of nine prominent internal medicine journals they subscribe to and read regularly, and to rate the quality of each journal. Multivariable methods were used to adjust for possible confounding factors.
Results: Of the 399 eligible participants, 264 returned surveys (response rate 66%). Overall, respondents reported spending an average of 4.4 hours per week reading medical journal articles and reported reading only the abstract for 63% of articles. Internists with and without epidemiology training reported spending similar amounts of time reading the medical literature per week (4.1 vs 4.7 hours; P=0.5) and reading only the abstract (67% vs 59%; P=0.06). Compared with internists without training in epidemiology, epidemiology-trained internists were significantly more likely to subscribe to and read certain journals (Lancet, Journal of General Internal Medicine) and less likely to subscribe to and read others (Hospital Practice, Southern Medical Journal). For seven of the nine journals, significant differences were found between internists with and without epidemiology training. Internists trained in epidemiology gave higher quality ratings to JAMA, Journal of General Internal Medicine, Lancet, and New England Journal of Medicine, and lower quality ratings to American Journal of Medicine, Hospital Practice, and Southern Medical Journal when compared to internists without epidemiological training.
Conclusions: The internists in our survey, regardless of training in epidemiology, rely heavily on abstracts. Internists with and without training in clinical epidemiology differ significantly, however, with regard to the journals they subscribe to and read, and their perceptions of individual journal quality.
Impact: Journals serve a critical role in the dissemination of new medical evidence. Internists may rely heavily on an article's abstract perhaps due to their desire for concise summaries of research findings. It is notable, however, that physicians likely to be producers of clinical research by virtue of their training have divergent views about the quality of several medical journals compared with their epidemiology-untrained colleagues who are likely to be the main consumers of clinical research.
*208. Low Incidence of Unsuspected Anaerobic Bacteremia: Are Routine Anaerobic Blood Cultures Still Indicated?
Eduardo Ortiz, MD, MPH, Salt Lake City VA Medical Center and University of Utah Health Sciences Center, San Diego, CA, MA Sande, Salt Lake City VA Medical Center and University of Utah Health Sciences Center
Objectives: The incidence of anaerobic bacteremia appears to have peaked in the 1970s, with a subsequent decrease occurring during the past two decades. Despite this decrease, most clinicians routinely order anaerobic blood cultures in the work-up of patients with suspected bacteremia. The value of routine anaerobic blood cultures in guiding treatment decisions or improving clinical outcomes is unclear. To evaluate this issue, we analyzed data on all anaerobic isolates in our institution from 1994 to 1996.
Methods: We conducted a retrospective review of microbiology laboratory data followed by selective chart review at the Salt Lake City Veterans Affairs Medical Center. We determined the number of bacterial blood cultures processed through the laboratory from January 1, 1994, to December 31, 1996, and the number of anaerobic, aerobic, and fungal isolates. For patients with a positive anaerobic blood culture result, we performed chart reviews to identify whether risk factors for anaerobic bacteremia were present before the cultures were drawn. We also assessed the impact of a positive result on treatment.
Results: There were 6891 sets of bacterial blood cultures (13,782 bottles) processed through the laboratory, yielding 1626 patients with positive results. Anaerobic isolates were recovered from 36 (2.2%) patients in 48 bottles. Aerobic isolates were recovered from 1550 (95.3%) patients, and fungal isolates were recovered from 40 (2.5%) patients. Seven (0.4%) patients had true anaerobic bacteremia, 28 (1.7%) patients had isolates determined to be contaminants, and 1 patient's chart was not available for review. All 7 (100%) patients with anaerobic bacteremia had a known or suspected source of anaerobic infection before the cultures were drawn. They include 2 patients with necrotic diabetic foot ulcers, 1 patient with an epidural abscess, 1 patient with a bowel resection, 1 patient with a ruptured appendix, 1 patient with liver disease, ascites, and peritonitis, and 1 patient with lymphocytic leukemia and neutropenic enterocolitis. Antibiotic changes were made in 4 of the 7 patients after the positive anaerobic results were known. Anaerobic coverage was added in 2 patients, and antibiotics were discontinued in 2 patients. Antibiotic changes appear to have led to clinical improvement in one patient.
Conclusions: We conclude that clinically important anaerobic bacteremia is rare, and that it usually occurs in patients who have a suspected source of anaerobic infection or risk factors for anaerobic bacteremia. Routine use of anaerobic blood cultures rarely results in clinically important diagnostic or therapeutic gains.
Impact: Anaerobic blood cultures should be selectively ordered in patients at increased risk for anaerobic infections. Health care facilities should assess their incidence of clinically important anaerobic bacteremia and use this information to guide local policies. Further evaluation is needed to determine the feasibility of ordering selective anaerobic blood cultures, the utility of anaerobic blood cultures in detecting facultative anaerobes, and the role of anaerobic blood cultures in improving clinical outcomes.
*209. Lung Volume Reduction Surgery is Associated with Improvements in Health-related Quality of Life
Hugh F Huizenga, White River Jct VAMC and Dartmouth Medical School, White River Junction, VT, SG Fihn, VA Puget Sound Health Care System and University of Washington, K Kierland, VA Puget Sound Health Care System
Objectives: to evaluate the effect of lung volume reduction surgery (LVRS) on health-related quality of life for patients with chronic obstructive pulmonary disease.
Methods: prospective cohort study of all patients undergoing lung volume reduction surgery at a single institution. All patients underwent preoperative evaluation of pulmonary function and health-related quality of life as measured by the SF-36, QWBII, and St. Georges Respiratory Disease questionnaire. Follow-up measurements were obtained 6 months post-operatively.
Results: 15 patients have undergone surgery. 6 month follow-up data are available for 10 patients. There has been one peri-operative death. Compared to pre-operative values, LVRS resulted in decreases in residual volume ( 6.50 liters vs 4.58, 293 % predicted vs 203 % predicted, p<.005) and total lung capacity (10.46 vs 8.8 liters, 148% predicted vs 124.2,p<.002). There were also statistically and clinically significant improvements in health-related quality of life as measured by the SGRQ impacts scale( 54.8 to 33.3, p<0.03) and total scale (66.2 to 53.15 p<0.01) and by the SF-36 role physical (4.54 to 50.0, p<0.04, and vitality scales(29.1 to 54.1, p <0.01) No significant differences in health utility state were detected as measured by the QWB (0.64 pre-op vs 0.65 post-op) Improvements in 6 minute walk distance, Forced Expiratory Volume in 1 second (FEV1) and Forced Vital Capacity (FVC) were measured but did not attain statistical significance
Conclusions: In this small group of selected patients, lung volume reduction surgery appeared to be associated with an improvement in self-assessed health-related quality of life at six months post-operatively. Further follow-up will determine whether these improvements are sustained over time
Impact: Chronic obstructive pulmonary disease is a common problem among veterans. For carefully selected patients, lung volume reduction surgery results in short-term improvements in health related quality of life
HSR&D Funded: IIR96-024
*210. Marketing Smoking Cessation: is a Free Patch as Good as Free Soap?
Anne M. Joseph, MD, MPH, Minneapolis VAMC, CCDOR, University of Minnesota, Minneapolis, MN, Jenny Frissell, Minneapolis VAMC, CCDOR, Sean Nugent, Minneapolis VAMC, CCDOR
Objectives: It is a persistent challenge to recruit smokers to cessation interventions, despite demonstrated efficacy of these treatment modalities. The objective of this controlled trial is to test whether inclusion of a single free patch in a mailed intervention to attend a free smoking cessation class improves 1) response to the invitation, 2) attendance to the class, and 3) smoking cessation rates.
Methods: The subjects include ~8600 veterans in the seven county metro area of Minneapolis-St. Paul not currently using the VA for health care (smoking status unknown). To date, 798 subjects have been randomly assigned to receive an invitation to attend a one session class that included a sample patch (letter + patch group; L+P), and 795 subjects have received an identical invitation without a patch (letter group; L). Prior data suggest that the prevalence of smoking in this group is approximately 20%; about 160 per group. Both groups are offered a free course of patches as part of the class protocol.
Results: There were 12 respondents in the L+P group, compared to 14 in the L group, p=0.53. Attendance among respondents in the L+P group was 92%, compared to 93% in the L group, P=0.52. One week post-quit date self-reported abstinence was 12% in the L+P group and 14% in the L group, P=0.81.
Conclusions: These early results do not suggest that inclusion of a single free patch as a marketing tool for a smoking cessation program improves response rates, or improves early indicators of successful cessation among respondents to the invitation. They do suggest that a population based outreach program can result in a significant rate of application and attendance to a smoking cessation program.
Impact: These data have relevance to population based approaches to recruitment to smoking cessation interventions that are conducted independently of primary care providers. Providers often complain that they do not have sufficient time to assess and refer smokers for treatment, and this method may serve to select smokers in the preparation stage of change and give them easy access to smoking interventions.
*211. Medicare HMO Enrollment among VA Health Service Users and Non-users
Robert O. Morgan, PhD, Houston Center for Quality of Care and Utilization Studies, Houston VAMC, Houston, TX, BA Virnig, University of Minnesota, CA DeVito, Miami VAMC
Objectives: High rates of HMO enrollment by dually eligible (VA, Medicare) veterans raise concerns about continuity, quality, and increasing costs of care for veterans. However, there is little information on how the HMO enrollment rates among veterans who use the VA health care system compare to those for veterans who do not. It is unclear whether VA system use influences a veteran's likelihood of enrolling in a Medicare HMO, and/or whether HMO enrollment is related to veterans' health status and plans for future VA use. The objective of this analysis is to examine whether veterans who use the VA health care system are more or less likely to be enrolled in Medicare HMOs compared to similar, non VA-using veterans, and whether HMO enrollees differ from non-enrollees in self-reported health status and their anticipated use of VA health care.
Methods: Data were obtained from the South Florida Health and Health Services Use Survey conducted in early 1999. Logistic regression, multiple regression, and contingency table analyses were used to examine factors associated with enrollment in an HMO, and the relationship between HMO enrollment and health status and likelihood of future VA use.
Results: The survey had an overall response rate of 61%. Data used here are for the 554 male VA-users and 610 male veteran "non-users" who completed the survey. Non-users were veterans with no VA use in the prior five years. The overall HMO enrollment rate was high (43%). Controlling for income, age, and distance from VA facilities, VA-users were less likely to belong to an HMO than were non-users (p<=.01). This difference was pronounced among veterans with lower household incomes (<$25,000/year), with little difference in HMO enrollment between higher income VA-users and non-users (VA-user status by income interaction, p<=.01). Both income and, surprisingly, distance had significant inverse relationships to HMO enrollment (p<=.001 for both). HMO enrollees in general reported better health and fewer health limitations than non- enrollees (p<=.05 and p<= .02, respectively), with low income, non HMO-enrolled, VA-users reporting the worst health and most limitations. VA non-users in general were less likely than VA-users to report plans for future use (p<=.0001). Among VA-users, HMO-enrollees were less likely than the non-enrollees to report plans for future VA use (p<=.001).
Conclusions: HMO enrollment is high among Medicare eligible, South Florida veterans. However, for lower income Medicare eligible veterans, use of the VA system is inversely related to the likelihood of HMO enrollment. Among VA-users, HMO enrollees appear to be healthier than non-enrollees and anticipate being less likely to use VA health care services in the coming year.
Impact: Our findings help illuminate the relationships between income, health, use of the VA health care system, and the availability of managed care as a health care alternative. In the current setting of dramatic changes in the national health care picture it is crucial that the VA be able to understand these relationships in order to effectively allocate resources and continue its mission of providing health care to veterans.
HSR&D Funded: HSRD IIR 95-079.
*212. Monitoring and Assessing the Relationship Between Veteran Enrollment and VA Healthcare and Quality
Eileen Ciesco, PhD, and Gregg A. Pane, MD, MPA, VHA National Performance Data Feedback Center, Durham/Washington, NC/DC, GA Pane, VHA Office of Policy and Planning, J Schaefer, VHA National Performance Data Feedback Center, ME Martindale, VHA Office of Policy and Planning, K Greaney, VHA National Performance Data Feedback Center, S Thornberry, VHA Office of Policy and Planning
Objectives: The objective of this analysis is to monitor and assess the impact of veteran enrollment for VA healthcare upon its quality and performance indicators. This is the first analysis to link VA healthcare enrollment and quality indicators. Public Law 104-262, The Veterans Health Care Eligibility Reform Act of 1996, required that VA provide needed, quality care (most inpatient and outpatient care) in a timely manner only to the extent that funds were available to support their care.
Methods: The Office of Policy and Planning (OPP) provided a database to the National Performance Data Feedback Center (NPDFC) that contained information on veteran enrollment through May 28, 1999. The NPDFC merged the enrollment data with selected quality measures. These measures included patient satisfaction survey data, clinical indicators that make up the prevention index and chronic disease index, and follow-up care for inpatients discharged with a mental illness diagnosis. Performance on these measures, by priority group, was comparead between 1997 and 1998. All data presented are unadjusted.
Results: For 1997 and 1998 patient-reported problem rates (by customer service standards) for ambulatory care were consistently lower than problem rates for inpatient care. The percentage of inpatients and outpatients rating the overall quality or care as very good or excellent incrased only slightly in FY 1998 over FY 1997 from 65 percent to 66 percent for inpatients and from 62 percent to 65 percent for outpatients. Priority 7 enrollees in each year reported the highest percentages of overall quality of care as very good or excellent, while Priority 1 enrollees, in both settings in FY 1998, reported the lowest percentages. There is some variation by priority group over time from FY 1997 to FY 1998, with outpatients having larger increases. Other quality indicators improved across all years, e.g., the Chronic Disease and Prevention Indices and the percent of patients receiving outpatient psychiatric care within 30 days following hospitalization for mental illness.
Conclusions: In general, quality indicators have been rather stable since FY 1997. Indicators related to outpatients seem to have improved more than those related to inpatient care. This might be related to the expansion of outpatient and preventive healthcare in PL 104-262. Small differences arise, e.g., Priority 1 veterans report lower overall quality of care than the other priorities, but there appears to be little change for this group of veterans from FY 1997 to FY 1998. Priority 7 veterans, many of whom are new enrollees, seem to have the highest quality rates.
Impact: VA has only begun to assess the impact of the entire FY 1999 period, the first year of official enrollment. However, this continuous monitoring is enabling VA to better assess impacts of the law, identify problems as they occur, and implement actions to correct any quality problems. Thus far VA has been able to provide a comprehensive package of inpatient and outpatient care services to a VA healthcare system open to all priorities of care.
*213. Multidisciplinary Rehabilitation vs Medical Care: using Meta-analysis to Determine the Clinical Effectiveness of Physical Medicine
Ron L. Evans, University of Washington, Seattle, WA, Seattle, WA, RT Connis, American Association of Anesthesiologists
Objectives: Research studies in physical medicine have not demonstrated the effectiveness of inpatient rehabilitation services, primarily due to differences in methodological approaches which have led to inconsistent findings. To address this problem, research studies comparing the clinical effects of rehabilitation with medical care were evaluated for 3 uniformly available outcomes: survival, functional ability, and discharge location
Methods: Published trials were obtained from citations in Medline and Nursing and Allied Health Abstracts covering the period from 1978 to 1998 using the descriptors: rehabilitation, physical medicine, activities of daily living, and functional ability. We used meta-analysis of completed clinical trials to test the hypotheses that rehabilitation results in better health care outcomes. To be included studies had to prospectively evaluate a multidisciplinary program of physical rehabilitation by comparing outcomes with a control group and report results in a manner that allowed quantitative analysis of uniformly available variables: eg, mortality, function, and residence. Eleven studies met this rather stringent criteria.
Results: Results indicated that rehabilitation services are significantly associated with better rates of short term survival (86% versus 81%, chi square=7.30, p < 0.01) and improved function during hospital stay (Fisher Combined Test, chi square=67.12, p < 0.01), but significance was not observed at one year. Also, rehabilitation patients returned home and remained there more frequently than controls (Odds ratio=2.08, chi square=40.15, p< 0.001).
Conclusions: We concluded that patients who participate in rehabilitation: 1] function better at hospital discharge, 2] have a better chance of short term survival, and 3] return home more frequently than non-participants. Contrarily, long term survival and function were unaffected by the relatively short term treatment [3-5 weeks].
Impact: The sustaining benefit of returning home may justify provision of inpatient rehabilitation. However, the lack of other long term benefits suggests that services may need to be continued at home or in subacute care settings to optimize their effectiveness. Future clinical research might include social and behavioral outcomes so that results of rehabilitation research can be applied to clinical settings and can more accurately reflect psychosocial objectives.
HSR&D Funded: IIR 94.125.
*214. Nurse Exercise Counseling for Primary Care Patients: Physical Activity and Fitness Changes After 6 Months
Patricia M. Dubbert, PhD, VAMC Jackson, MS and Univ MS Sch Med, Jackson, MS, KM Cooper, VAMC Jackson, MS, K Kirchner, VAMC Jackson, MS and Univ MS Sch Med, D Bilbrew, VAMC Jackson, MS, K Kerr, VAMC Jackson, MS, EF Meydrech, Univ MS Sch Med
Objectives: Recent reviews and public health guidelines recommend physical activity counseling by providers in primary care (PC) settings. This study examined the effects of nurse counseling interventions on physical activity and fitness in elderly patients in VA PC clinics.
Methods: In a randomized trial comparing 3 levels of nurse physical activity counseling, 60-80 year old patients with chronic medical conditions but stable health and who could benefit from walking were referred by PC providers. After completing baseline (BL) measures, all patients received theory-based individualized counseling from the nurse with a goal of walking at least 20 minutes 3 times a week. Patients were then randomized to receive personal phone calls directly from the nurse (PC), calls from the nurse alternating with automated calls delivering a message from the nurse (APC), or no further contact with the nurse (NC) over the next 6 months. They were also instructed to keep weekly walking diaries and mailed these to the data collector. Patients returned for followup (FU) assessments at 6 months, repeating BL measures including a submaximal treadmill test or 6 minute walking test, 7 day physical activity recall interview, estimate of minutes walking per day, body weight, and resting heart rate. Group (PPC, APC, NC) X Time (BL, FU) repeated measures ANOVAs were used for statistical analysis.
Results: 212 patients were randomized and 186 returned for 6 month FU. Randomized patients were mean age 68.5 (SD 4.7) years, 29% minority, 56% rural, and most had >= 2 chronic medical conditions including arthritis, obesity, hypertension, diabetes, heart disease, and lung disease. Attrition at FU was not different by group and primarily due to illness or transportation problems. At FU, participants reported increased daily energy expenditure on the physical activity recall (p = .02) and increased minutes walked per day (p < .01) with no between group differences. Many patients refused the FU treadmill test and there were no other significant findings for this measure. Those who did not do the treadmill (n=80) completed a FU 6 minute walk with improved performance (p=.04) with no differences between groups. Body weight did not change. Resting heart rate was lower at FU for APC and NC but not PPC (between groups p=.03). Change in estimated daily energy expenditure was significantly correlated with change in 6 minute walk performance (p<.05).
Conclusions: Results suggest that, 6 months after initiation of nurse counseling, patients in all 3 intervention groups had increased physical activity and fitness.
Impact: To our knowledge, this is the first trial to demonstrate increased fitness as well as physical activity increases in elderly PC patients following exercise counseling for a home-based program. Regular physical activity can decrease risk of disability and improve quality of life for many elders. Nurse counseling interventions are an effective method of physical activity promotion in elderly PC patients.
HSR&D Funded: NRI 95-022.
*215. Organizational Characteristics Contributing to Successful Outcome in a Multi-Hospitial Collaborative Safety Improvement Project
Peter D. Mills, PhD, Director, Center for Learning and Improvement of Patient Safety, Veterans Affairs, White River Junction, VT, WB Weeks, Director, Center for Learning and Improvement of Patient Safety, Veterans Affairs, A Kabcenell, Insititute for Healthcare Improvement
Objectives: To explore which characteristics of health care organizations or improvement teams were associated with successful performance in a multi-hospital safety improvement effort focused on reducing adverse drug events.
Methods: Prior to beginning the project, we used telephone interviews to asses the following variables in twenty-three hospitals participating in a collaborative quality improvement effort designed to reduce adverse drug events: 1. whether the project was part of the strategic plan of the hospital 2. leadership support 3. specificity of the team’s plan to reduce adverse drug events 4. team membership makeup 5. prior experience with the measures 6. prior experience with quality improvement efforts 7. size of the facility and 8. area of focus. We created a composite indicator by combining results from variables 1-6, above. Outcome scores, defined by the extent to which each team met their initial objectives, were calculated at the end of seven months
Results: The following variables were significantly correlated with the outcome score: Leadership support (r = .44, p = .046) Physician participation on the team (r = .441, p = .035) Area of Focus = reduction of adverse drug events associated with narcotics (r = -.585, p = .003) Area of Focus = reduction of adverse drug events associated with insulin (r = .559, p = .006) Area of Focus = reduction of adverse drug events associated with infusion pumps (r = .413, p = .05). Measures of Skewness and Kurtosis were between (-2) and 2 for all variables. When these variables are entered into a stepwise regression analysis, the variables measuring the use of narcotics and the total prework score predict 57 percent of the variance in the final outcome score. No other measured variables were significantly correlated with outcome scores
Conclusions: These exploratory data suggest that leadership support in the organization and physician participation on the improvement team may be an important part of a successful safety improvement project. In addition, choice of topic may be related to success. Finally, teams that report being well-prepared to take on the improvement project, with clear aims and measures, a well balanced team, prior experience and early support from all levels, had more positive results.
Impact: These findings suggest that management support and physician involvement may be critical success factors in quality improvement efforts. As VA embarks on additional quality improvement efforts, the organization should ensure that teams are properly staffed and have adequate support to succeed.
*216. Outcome and Treatment Preferences in Schizophrenia: Do Stakeholders Share Goals?
Ellen Fischer, PhD, Center for Mental Healthcare and Outcomes Research, North Little Rock, AR, M Shumway, University of California at San Francisco
Objectives: Although effective treatment for schizophrenia requires sustained engagement in care, treatment drop-out and non-adherence are common. Theory and limited empirical evidence suggest that the likelihood of sustained involvement in care will increase when stakeholders (the individual with schizophrenia, his/her mental healthcare provider, and involved family members) agree on treatment goals. This HSR&D-funded study is designed to assess (1) the extent and nature of agreement on outcome priorities and treatment preferences within stakeholder sets, (2) the extent of provider and family-member awareness of consumer priorities and preferences, and (3) the magnitude of the association between within-set agreement/within-set knowledge and involvement in care.
Methods: Data on priorities and preferences are being collected through interviews with 60 stakeholder sets, each comprised of a consumer with schizophrenia, a member of the consumer's family, and the consumer's primary mental healthcare provider. Priorities and preferences are elicited separately for outcome and treatment domains, using category rating and direct importance rating techniques. Family members and providers are also asked to rank the 7 outcome and 9 treatment domains in the order they believe the consumer in their set would choose. Kendall's tau is used to quantify and to dichotomize within-pair agreement. Cutoffs of tau>=0.43, p<=0.12 for outcome priorities and tau>=0.37, p<=0.10 for treatment preferences were used to dichotomize agreement within pairs.
Results: Analysis of data from the first 13 stakeholder-sets suggest that within-set agreement on outcome priorities is limited. Agreement was highest between consumers and family members (38% of pairs in agreement (PIA); average tau=0.21), followed by family members and providers, and consumers and providers (PIA: 23%, 23%; average tau=0.22, 0.14, respectively). Agreement among all 3 within-set pairs was observed in only 8% of sets, while no within-set agreement (no pairs in agreement) was observed in 54% of sets. Providers appeared to have more accurate knowledge of consumer outcome priorities than did family members (PIA: 31% v 15%; average tau: 0.23 v 0.10). Within-set agreement was greater for treatment preferences. Agreement was highest for family-provider pairs, followed by consumer-family and consumer-provider pairs (PIA: 75%, 50%, 33%; average tau=0.50, 0.26, 0.16, respectively). Results for all 60 sets will be presented at the HSR&D Meeting.
Conclusions: Preliminary analyses suggest that there is limited within-set agreement on outcome priorities and treatment preferences. Provider and family understanding of consumer priorities and preferences also appears limited. If relatively low rates of goal-sharing are confirmed and are shown empirically to adversely affect treatment adherence, interventions to increase agreement on goals will offer a promising strategy for reducing the efficacy-effectiveness gap in schizophrenia care.
Impact: Schizophrenia, prevalent among users of the VA healthcare system, has a pervasive adverse impact on the lives of affected veterans. Treatment for schizophrenia represents a substantial cost burden for the VA. This study is a necessary first step in a translational research agenda to improve the quality and outcomes of care for veterans with schizophrenia, increase consumer satisfaction with care, and reduce costs through reductions in avoidable inpatient care.
HSR&D Funded: IIR 98-094.
*217. Pain Management and Behavioral Outcomes in Patients with Dementia
Martha Buffum, RN, DNSc, CS, VAMC San Francisco/UCSF, San Francisco, CA, M Brod, Lewin-Tag, Inc., San Francisco/UCSF, L Sands, University of California, San Francisco (UCSF), C Miaskowski, University of California, San Francisco (UCSF), A Stefani, VAMC San Francisco/UCSF
Objectives: Dementing illness often co-exists with painful medical conditions associated with aging, such as degenerative joint disease, osteoarthritis, skin ulcers, back pain, or headaches. The AHCPR Clinical Practice Guideline on Acute Pain Management recommends pain assessment as a standard of practice for all patients. Elderly persons with dementia have special needs for assessment, management, and evaluation. When patients are unable to verbalize pain, objective measurement of their discomfort must include behavioral observations such as agitation, confusion, and signs of discomfort. No research relates systematic pain assessment with problematic behaviors or treatment of pain based on this assessment in patients with dementia. This research is intended (1) to elucidate the relationships between pain, discomfort, agitation, and confusion; and (2) to determine the influence that pain management has on decreasing the outcomes of discomfort, agitation, and confusion.
Methods: This study uses a randomized, double-blind, placebo-controlled, cross-over design with each treatment arm lasting two weeks; crossover occurs without washout. The arms are 1) acetaminophen 650 mg qid with placebo qid PRN and 2) placebo qid with acetaminophen 650 mg qid PRN. Sixty-six patients are being recruited over a 3-year period. Patients must: be >= 55 years; have severe dementia; have a documented painful condition that would respond to acetaminophen; unable to report pain reliably; have >= 1 episode of agitation per day. An RA (blind to treatment) tests each patient for discomfort, agitation, and confusion at baseline and two times per day on two days of the week for four weeks. Nurses rate patients for agitation each shift. Repeated measures analyses will examine effect of treatment on agitation, confusion, and discomfort. Study sites are Jewish Home for the Aged; Golden Gate Healthcare Center; Palo Alto VA Health Care System, Nursing Home Care Units at Menlo Park and Livermore.
Results: In FY99, screening 243 patients revealed 42 persons (17%) were eligible: 14 completed the study, 5 are enrolled, 6 are consented, 11 are being consented, 5 refused, and 1 died prior to enrollment. To date, 19 patients completed the study and 4 dropped out. Major activities during this year have included recruitment, data collection, and data entry.
Conclusions: Project work is ongoing.
Impact: Our study has potential for impacting the aging veteran population with dementia. If managing undetected pain can alleviate behavioral problems, more demented persons may be able to live more comfortably and experience less pain in extended care settings. Study findings will impact both male and female veterans and non-veterans in long term care settings. Positive findings demonstrate the need for pain assessment, management, and evaluation in the nursing home.
HSR&D Funded: NRI 951922.
*218. Patient Ratings of Satisfaction with Health Care: a Study of Veterans Served by VA Primary and Eye Care Clinics
Xinhua Steve Ren, PhD, Center for Health Quality, Outcomes, and Economic Research, Bedford VA Hospital, Bedford, MA, L Kazis, Center for Health Quality, Outcomes, and Economic Research, Bedford VA Hospital, A Lee, Center for Health Quality, Outcomes, and Economic Research, Bedford VA Hospital, W Rogers, The Health Institute, New England Medical Center, S Pendergrass, Veterans Health Administration, VISN 16
Objectives: As the Veterans Health Administration (VHA) places high priority on becoming a performance based organization, there is an increasing need to quantify and refine its outcome measurement system. In this presentation, we examine the relationship between patient satisfaction and health status, two domains of VHA outcome measurement system, in the context of sociodemographic characteristics such as age, gender, race, and marital status. We also provide a knowledge base on the levels of patient satisfaction with different services (e.g., primary vs. eye care) across different VHA facilities.
Methods: We analyzed data from the VA Primary and Eye Care Patient Survey conducted in VISN 16 between April and July, 1999. Patients were randomly selected from primary (N=5,415) and eye (N=3,922) care clinics in each of the 10 facilities within VISN 16. Participants in the study were mailed a health status (the Veterans SF-12V) and patient satisfaction (CSS) questionnaire. We used t statistics to test the mean differences in the CSS and SF-12V scales between primary and eye clinics as well as across different facilities. We conducted Pearson’s product moment correlations and ordinary least squares regression (OLS) to examine the relationship between health status and patient satisfaction.
Results: Consistent with previous findings, patients who were younger, female, unmarried, and nonwhite were less likely to be satisfied with the care they received, whereas patients with better health status were more likely to be satisfied with the care they received. This pattern is consistent across primary and eye care patients. After adjusting for sociodemographics and health status, patients from eye care were more satisfied with timeliness and accessibility of care (p < 0.001), information and education (p < 0.001), and emotional support (p < 0.001) than patients from primary care; whereas primary care patients were more satisfied with respect for patient preferences (p < 0.001) and continuity of care (p < 0.001) than eye care patients. Patient ratings of satisfaction also varied across different facilities with some facilities doing better in some areas and some facilities doing worse in other areas.
Conclusions: The study results indicate that there exists variation in patient ratings of satisfaction with health care between primary and eye care clinics as well as across different facilities. After adjusting for sociodemographics such as age, gender, race, and marital status, patients with poorer health status were less likely to be satisfied with the care they received from the VHA than patients with better health status.
Impact: Results have implications for the use of patient satisfaction scores as a marker for quality of care, i.e., different clinics and different facilities should target specific areas for improvement. Results also have implications for the use of health status as an important adjuster in future work. Both physical and mental health were correlated with the satisfaction scales, suggesting that while comparisons across facilities should consider adjusted satisfaction scores, such comparisons should also take into account that better health status may be a result of higher quality of care.
*219. Perception of Barriers to Providing Mental Health and Substance Abuse Care in Community Based Outpatient Clinics
Marisue Cody, PhD, HSR&D Center for Mental Healthcare and Outcomes Research, North Little Rock, AR, JE Kirchner, HSR&D Center for Mental Healthcare and Outcomes Research
Objectives: Initial evaluations of Community Based Outpatient Clinics (CBOCs) have found that they increased the provision of physical health care, but had no impact on the use of specialty mental health services. The purpose of this study is to assess the barriers for mental health and substance use services of elderly patients within CBOCs.
Methods: We selected three study sites (Monroe, Louisiana; Texarkana, Texas; and Lawton Oklahoma) based on the following criteria: location within VISN 16, staffed by Veteran Affairs personnel, located within 1 ½ to two hours driving time from the Parent VAMC, minimal or no existing mental health or substance use services provided at the CBOC, willingness to participate in this effort, similar number of unique patients, and location within towns with population size similar to other CBOC sites. Following site selection, we conducted a comprehensive process evaluation of existing mental health services and existing barriers to mental health care at these sites.
Results: All CBOC study sites reported travel distance as a barrier to getting mental health care. The travel distance from the Lawton CBOC to the Oklahoma City VAMC is 85 miles; from the Monroe CBOC to Shreveport VAMC, travel distance is 101 miles; and from the Texarkana CBOC to Shreveport VAMC, travel distance is 70 miles. These rural areas also have fewer specialty mental health/substance abuse community services. Additional barriers noted by CBOC study sites included difficulty reaching homeless veterans, staffing shortages, patient waiting time, and patient compliance. VA system policies also present barriers to care. These include VA hospitals having to follow state and county/parish mental health code, VA system developing Process of Care Mandates without funding to support new programs, the VA system’s commitment to reduce cost by 30%, and the VA’s integration of medical, surgical, and mental health care.
Conclusions: This assessment supports the need to integrate mental health care at CBOCs. We will be testing a model that places an APN with prescribing privileges and specialty training in mental health. The role of the APN in the integrated CBOCs will be to fully optimize the primary care management of veterans with mental health and alcohol abuse seen at CBOCs.
Impact: Within CBOCs there is a wide variation in the presence of mental health staff. Currently VISN directors are in a dilemma. They are being forced to make urgent and significant decisions concerning the best method to provide mental health services to veterans served through CBOCs. Thus, it is critical that we provide policy makers information concerning the clinical effectiveness and cost effectiveness of integrated models. If APNs are able to provide appropriate and guideline concordant MH/SA care with the same or better outcomes for elderly veterans seen by psychiatrists or psychologists in primary care settings, this could be a less costly alternative to integrating MH/SA services within VAs nationally.
HSR&D Funded: pending.
*220. Perceptions of Access to VA and non-VA Medical Care among Elderly African-American and White Veterans with Chronic Knee and Hip Pain
Said A. Ibrahim, MD, MPH, Louis Stokes Dept. VAMC & CWRU, Cleveland, OH, Cleveland, OH, Christopher J. Burant, MA, CWRU, Cleveland, OH, Laura A. Siminoff, PhD, CWRU, Cleveland, OH, C. Kent Kwoh, MD, Louis Stokes Dept. VAMC & CWRU, Cleveland, OH
Objectives: Access to medical care is an important determinant of health outcomes, and a major factor in observed ethnic disparities in health care outcomes. Patient perceptions of access to medical care impact utilization of non-acute care. Previous reports have suggested that veterans from ethnic minorities and socio-economically disadvantaged communities have lower perceptions of access to VA care. In a VA-funded study of racial/cultural variations in the management of osteoarthritis, we sought to examine how elderly, AA and white veterans with moderate to severe chronic knee and/or hip pain compare with respect to their perceptions of access to VA and non-VA medical care.
Methods: 228 AA (46%) and white (54%) patients (age >50) were interviewed. Patients with moderate to severe knee or hip pain for more than six months were eligible for the study. Trained research staff interviewed patients with regard to their perceptions of access to VA and non-VA care. Responses were dichotomized into "Difficult" vs. "Not Difficult" categories. Information on non-VA (additional) health care insurance was also acquired.
Results: AA and white patients in this cohort were comparable with respect to age (67 ± 9 vs. 68 ± 9, p= 0.53), severity of arthritis [measured by Lequesne scale (mean score 11 vs. 12, p=0.09)], WOMAC scale for activity difficulty (mean score 17 vs. 18, P=0.25), GDS depression (5 vs. 5, p=0.60), perceived disability due to pain (66% vs. 63%, p=0.50), moderate or higher impact of pain on quality of life (70% vs. 71%, p=1.00), and living alone (37% vs. 30%, p=0.26). However, AA were more likely to report lower income (41% vs. 22%,p=0.04) and employment (6% vs. 14%, P= 0.05). With respect to the overall access to medical care question, 80% of whites and 79% of AAs responded "Not Difficult," (P= 0.85). With respect to the access to non-VA care question, 55% of whites and 45% of AAs responded "Not Difficult," (P= 0.11). White patients were much more likely than AAs patients to report access to non-VA insurance (75% vs. 62%; P= 0.04).
Conclusions: In this cohort of male, elderly veterans with moderate to severe chronic knee and/or hip pain, there were no significant differences between AA and white veterans with respect to perceptions of access to VA and non-VA medical care. However, AA patients were much less likely than whites to report having non-VA insurance. Lower access to non-VA insurance among AAs may reflect reported lower income and employment status. However, small sample size effect can not be ruled out as an explanation for the observed lack of difference in AA/white patient perceptions on access to VA and non-VA medical care.
Impact: Improving access to VA care for all veterans is a priority. Understanding potential barriers to access may help the VA improve access to care for all veterans—particularly those from ethnic minorities and socio-economically disadvantaged communities.
HSR&D Funded: ECV-0003.
*221. Perceptions of Global Quality of Life in Patients with Chronic Knee or Hip Pain: Differences between African-American and White Veterans
Said A. Ibrahim, MD, MPH, Louis Stokes Dept. VAMC & CWRU, Cleveland, OH, Cleveland, OH, Christopher J. Burant, MA, CWRU, Cleveland, OH, Laura A. Siminoff, PhD, CWRU, Cleveland, OH, C. Kent Kwoh, MD, Louis Stokes Dept. VAMC & CWRU, Cleveland, OH
Objectives: African-American (AA) patients are less likely than whites to utilize total joint replacement. The reasons for this variation are poorly understood. Since ethnic differences in how patients with knee/hip pain perceive their quality of life may impact their decision-making with respect to treatment options, we compared veterans with chronic knee or hip pain with respect to their perceptions of global quality of life (QOL).
Methods: 228 AA (46%) and white (54%) patients (age >50) were interviewed. Eligible patients had to have moderate to severe knee or hip pain for more than six months. Patients who have had prior surgery on a knee or hip were excluded. The primary outcome measure was global QOL rated by a single question and dichotomized in this analysis as either excellent, very good, or good, fair and poor. Logistic regression models were used to adjust for potential confounders including clinical, demographic and psycho-social variables.
Results: AA and white patients in this cohort were comparable with respect to age (67 ± 9 vs. 68 ± 9, p= 0.53), severity of disease measured by Lequesne scale (mean score 11 vs. 12, p=0.09), WOMAC scale for activity difficulty (mean score 17 vs. 18, P=0.25), history of trauma (53% vs. 45%, p=0.23), GDS depression (5 vs. 5, p=0.60), importance of pain to health (85% vs. 77%, p=0.17), perceived disability due to pain (66% vs. 63%, p=0.50), moderate or higher impact of pain on QOL (70% vs. 71%, p=1.00), and living alone (37% vs. 30%, p=0.26). However, AA were more likely to report lower income (41% vs. 22%,p=0.04) and employment (6% vs. 14%, P= 0.05). AA patients were less likely to rate their quality of life as excellent or very good than white patients (14% vs. 26%, p=0.03). After adjusting for all the above potential confounders, AA remained less likely to rate their QOL as excellent or very good (adjusted odds ratio 0.30, 95% CI 0.12 to 0.74).
Conclusions: Among this group of male veterans with chronic knee or hip pain, AA patients are less likely to report very good or excellent QOL, even though the severity of knee/hip pain and functional impairment was similar in each group. Ethnic differences in global quality of life perceptions may not explain the observed variations in total joint replacement.
Impact: This study suggests that factors other than patient perceptions of quality of life need to be explored to explain ethnic variations in the utilization of joint replacement for the management of osteoarthritis. With the aging of the US population in general and the increasing representation of ethnic minorities in the VA population, better understanding of the reasons for racial variations in management of chronic diseases is important.
HSR&D Funded: ECV-0003
*222. Predicting Change in Functional Status in VA Long Term Care: Do Robust Estimates Improve Prediction?
Carter Rakovski, Center for Health Quality Outcomes Economics Research, Veterans Administration Medical Center, Bedford MA, Bedford, MA, AK Rosen, Center for Health Quality Outcomes Economics Research, Veterans Administration Medical Center, Bedford MA, BH Chang, Center for Health Quality Outcomes Economics Research, Veterans Administration Medical Center, Bedford MA, J Wu, Center for Health Quality Outcomes Economics Research, Veterans Administration Medical Center, Bedford MA, D Berlowitz, Center for Health Quality Outcomes Economics Research, Veterans Administration Medical Center, Bedford MA, A Ash & M Moskowitz, Health Care Research Unit, Section of General Internal Medicine, Evans Memorial Department of Medicine, Boston Medical Center, Boston MA
Objectives: An important outcome measure for assessing quality in long-term care is change in functional status (FS). To assess quality across facilities, the VA needs to measure the severity of its patients. We developed a risk adjustment model to predict change in FS in the VA long-term care population using ordinary least squares regression (OLS). OLS assumes homoscedasticity (i.e., that the variance of the residuals is the same for each observation). We question this assumption because FS is measured on a scale from 3 to 15, and observations in the middle of the scale can vary more than observations at the ends of the scale, since patients in the middle have more room to change in either direction. If homoscedasticity is violated, then the standard errors of the regression coefficients are incorrect. As a result, faulty conclusions about the significance of variables and their inclusion in a model might be drawn based on erroneous p-values. This study compares the results of regression analysis using OLS versus a method that is robust to heteroscedastic data.
Methods: Our sample consisted of 16,995 residents who resided in VA long-term care facilities between 4/1/96 and 10/1/96. The Patient Assessment File (PAF) was used to evaluate changes in FS. Baseline and follow-up FS were measured using 3 ADLs: eating, transferring, and toileting (each rated on a scale of 1 to 5). Change in FS was measured as the difference between baseline FS and follow-up FS (-12 to +12). Patient characteristics related to change in FS (e.g. age, time between assessments, medical conditions, and diseases) were used as risk factors, along with baseline ADLs, to predict change in FS. We examined plots to identify any heteroscedasticity among the residuals. We calculated robust standard errors estimates using the widely accepted White's estimator, which allows for unequal variances. These estimates are compared with the variance estimates based on OLS. The significance of individual covariates based on these two variance estimates is also compared.
Results: The plots show evidence of heteroscedasticity, namely, variability in the dispersion of observations for different values of the independent variable and predicted values. There was a range in variance among the residuals for patients with different levels of baseline FS. The variance of observations was 50% less at the ends of the ADL summary scale compared to the middle. Among the 23 coefficients estimated, the difference between the ordinary and robust standard errors ranged from 0.3% to 13.9% with an average of 1.8%. One covariate, which was significant in OLS, became non-significant at the p<.05 significance level when a robust standard error was used.
Conclusions: Robust estimates of standard errors give more accurate p-values. The homoscedastic assumption is likely to fail when using scale-based variables such as FS. These assumptions should be checked when using OLS linear regression to build risk adjustment models to measure quality.
Impact: Risk adjustment is commonly done when examining outcomes of care. When estimating risk-adjusted models to predict scale-based outcome measures, homoscedasticity should be checked to assure proper variable selection.
HSR&D Funded: IIR-96-065.
*223. Prevalence and Predictors of HIV Testing among Homeless Women - Good News
Brooke Herndon MD, Greater LA VA Healthcare System, Sepulveda, CA, A Kilbourne, Pittsburgh VA Healthcare System, MM Wang, Greater VA Healthcare System, ML Lee, Greater VA Healthcare System, S Asch, Greater VA Healthcare System, L Gelberg, UCLA, Dept. Family Medicine
Objectives: The potential for HIV testing to lead to improved health outcomes has never been greater, yet little is known about utilization of this service among populations at highest risk for HIV. The aim of this study is to determine the prevalence and predictors of HIV testing in a population of urban, poor, predominately minority women.
Methods: Women were interviewed in shelters and meal programs in Los Angeles County to create a probability cluster sample (n=974). Bivariate analysis of patient characteristics, contact with health and social services, and indications for HIV testing identified variables with p<0.1, and these were entered into a multiple logistic regression model. Additionally, stepping algorithms with random sub-sampling for cross-validation were used to develop a more parsimonious multiple logistic regression model. Samples were weighted and cluster analysis was performed on final models.
Results: Approximately 70% of the sample reported receiving an HIV test in the past year. After adjustment for cluster effects, only a history of drug abuse or dependence and having a regular source of care were significant predictors of HIV testing in the past year. Sixty-five percent of respondents reported currently having a regular source of care. Neither income, public benefits, health insurance status, nor having a case manager predicted receipt of HIV testing. Similarly, none of the indications for HIV testing were statistically significant predictors. Of note, fifteen percent of respondents reporting a pregnancy in the past year also reported that they did not receive an HIV test. There were no significant negative predictors of HIV testing in the past year.
Conclusions: The rates of HIV testing in this probability sample of women are among the highest reported for any population studied and are consistent with prior research indicating that individuals at highest risk are more likely to be tested. However, there is a significant minority of at-risk women with specific indications for testing, including pregnancy, who are not being reached. Policies focused on providing a regular source of care are likely to have the greatest impact on serotesting prevalence.
Impact: VA facilites that focus on providing continuity of care will likely achieve the highest rates of HIV testing among their most disenfranchised female clients. Intensive outreach will be needed to increase the serotesting prevalence among women at highest risk for HIV.
*224. Primary Care: Comparison of Geriatric Mental Health Delivery System
Edwin J. Olsen, MD, Miami VAMC, Miami, FL, UN Durai, Chicago Westside VAMC, J Kirchner, Little Rock, Arkansas VAMC, D Krahn, Wisconsin VAMC, M Llorente, Miami VAMC, D Oslin, Philadelphia VAMC
Objectives: The VA is committed to improving the quality of life for elders by helping to ensure that high quality mental health services are readily accessible in primary care settings. The VA and SAMHSA have developed this important initiative to compare the effectiveness of service delivery models which treat MH/SA problems in primary care as opposed to specialty MH/SA settings. The primary objective is to determine if there are statistically significant differences over a full range of access, clinical, functional, and cost variables between service delivery models that are referring consumers to specialty mental health services outside the primary care setting (referral model) and those that are providing such services in an integrated fashion within the primary care setting itself (integrated model). The following secondary issues will be addressed: Consumer insurance status, Social support and stigma, Consumer satisfaction, Cultural competence, Clinician satisfaction and behavior, and Organizational/ environmental influences.
Methods: The Program is designed as a multi-site study to compare the effectiveness of integrated and referral models of care. There are 5 VA Sites, 6 SAMHSA Sites, Harvard Coordinating Center and the Miami VA Coordinating center. Over the past year, the sites have joined together to design the protocol. Elderly consumers receiving care in primary care clinics will be given a short screen to assess whether they meet the initial eligibility requirements. If referred to the study, consumers will then be asked to take a battery of assessment instruments that will determine whether they possess any of the multi-site target conditions. Subjects possessing any of these target conditions will be randomized into either integrated or referral care models and will be subsequently reassessed at 3, 6 and 12 months. The assessment/outcomes battery is designed to determine the presence and severity of mental health and substance abuse problems.
Results: After the initial year of planning and consensus development, the study design has been completed and will be implemented in February 2000. The 11 study sites have committed to recruiting a total of more than 1,900 consumers with depression and 670 with alcohol abuse. Each of the study sites will enroll roughly 150 older adults over the age of 65. The multisite study sample consists of elderly consumers who come from a diversity of geographic, cultural, and socioeconomic backgrounds. The sites have agreed on Target Conditions, Inclusion Criteria, Exclusion Criteria, and Timing of Consume Assessments, Consumer Flow, Informed Consent and Randomization. The Clinical models of care have been characterized and the Intervention Protocol established. Conclusions: The process of designing a multisite study has been arduous; however the scientific design has been adhered to through consensus development. A consumer Flow Diagram for the study has been completed, as well as assessment instruments. Pilot screening and assessment data have been analyzed and will be presented.
Impact: This study represents a cooperative venture among multiple Federal Agencies. The design meets scientific rigor and will provide valuable information in developing the best mental health care delivery systems for the elderly in Primary Care Settings.
*225. Principles of Effectiveness Trials: A Proposal and Implementation Examples from Cooperative Study #430
Mark S. Bauer, Providence VA Medical Center; Brown Univ, Providence, RI, E Dawson, Providence VA Medical Center, WO Williford, Perry Point Cooperative Studies Coord Ctr
Objectives: To propose a series of 9 principles for the construct of controlled effectiveness trials and to illustrate their implementation with examples from VA Cooperative Study #430, "Reducing the Efficacy-Effectiveness Gap for Bipolar Disorder."
Methods: This presentation derives from review of the literature and of the experience of the Executive Committee for CSP #430 in developing a controlled effectiveness trial for a structured intervention to improve outcome in bipolar disorder
Results: New treatment interventions typically are tested under the highly controlled conditions of an "efficacy study." Application of these findings usually results in disappointing impact in general clinical practice. Recognition of this "efficacy-effectiveness gap" for most interventions (Institute of Medicine, 1985) has led to greater emphasis on testing interventions in "effectiveness studies" under conditions that duplicate as closely as possible the conditions in the target practice venues to which the results will be applied. Commonly, researchers interested in effectiveness trials have focused on broadening subject samples to resemble the wider population of interest. However, sampling is only one of several areas that must receive attention for optimal design of effectiveness studies. We propose and discuss nine specific principles across four distinct aspects of effectiveness studies: sample selection, structure and implementation of the intervention, collection of research data, and data analysis. These are discussed with respect to supporting the internal and/or external validity of the study.
Conclusions: Effectiveness trials can be successfully designed and implemented with attention to the nine principles presented above. Examples from the ongoing CSP #430 are discussed.
Impact: Attention to effectiveness principles will result in greater public health impact of research dollars spent on controlled trials.
HSR&D Funded: C SP #430; DEV97-015.
*226. Assessing the Attributes of the Quality of Dying: Perspectives of Patients, Families, and Health Care Providers
Karen E. Steinhauser, Durham VAMC, Durham, NC, Elizabeth C. Clipp, RN PhD, Durham VAMC and Duke University Medical Center, Maya McNeilly, PhD, Duke University Medical Center, Nicholas Christakis, MD, PhD, MPH, University of Chicago School of Medicine, Lauren McIntyre, PhD, Purdue University, James A. Tulsky, MD, Durham, VAMC and Duke University Medical Center.
Objectives: Health care providers and the public have targeted considerable resources towards improving the quality of dying through such interventions as hospice care, education of medical personnel, and promotion of advance directives. However, despite these efforts, no well-established standardized tools exist to assess the quality of dying. In particular, no tools explicitly acknowledge the variability in patients’ and families’ perspectives. Consequently, attempts to evaluate the effectiveness of hospice care have suffered. Our primary objective was to survey patients, families and providers to assess the importance they assigned to various components of care at the end of life. The survey items were generated from qualitative definitions of a "good death" collected in focus groups discussions and follow up interviews with a variety of end-of-life care participants. These data will be used to develop an instrument to assess the quality of dying.
Methods: We mailed 500 surveys (per group) to national samples of physicians, non-physician providers (nurses, social workers, chaplains, and hospice volunteers), patients with serious illness and recently bereaved family members. The health care provider samples were generated from professional associations memberships and represented the ethnic and gender composition of those professions. Samples of patients and families were stratified by ethnicity and gender; we over-sampled women and African American participants.
Results: We received very high response rates from each of the national samples. Seventy three percent (73%) of physicians; 86% of non-physician providers; 70% of patients and 64% of family members returned completed surveys. Questions in the survey fell into six domains of the quality of dying: pain and symptom management, treatment decisions, preparation, completion, affirmation of the whole person, and contributing to others. These domains differ from the physiological, psychosocial and spiritual divisions found in traditional instruments, and therefore, offer a new conceptualization of a "good death." We present similarities and differences in the importance various health care professionals assign to attributes of care at the end of life. We also document some of the contradictions regarding values versus behaviors in care at the end of life. While many health care providers report an extensive use of aggressive cure-directed treatments, they also report a desire for the withdrawal of such techniques among terminally ill patients. Finally, participants stressed the very individual nature of the dying experience. "Good deaths" are crafted through a process of communication about patients’ and families’ values and preferences regarding care at the end of life.
Conclusions: Control of physical symptoms at the end of life is important but only a point of departure for improving the quality of dying. Participants suggested a "good death" also includes attention to psychosocial and spiritual issues. Instruments designed to assess the quality of dying must include these multiple domains.
Impact: The results of this study will be used to develop an instrument to measure the quality and effectiveness of interventions to improve end-of-life care, such as palliative care units or hospice programs. It will allow us to create instruments that focus on aspects of care patients and families consider most important.
HSR&D Funded: IIR96-066.
*227. Proposal to Create a Telephone Monitoring Follow up System for Home Bound, Aide and Attendance, Fee Basis or Other Eligible Patients on Warfarin Therapy
Alfreda Leshia Kinnon, PharmD, Atlanta VA Medical Center, Decatur, GA
Objectives: Background: Warfarin represents one of the most commonly prescribed medications at the Atlanta VAMC. Due to a narrow therapuetic index, routine monitoring is essential to minimize potential hemorrhagic complications. Because of scheduling limitations, anticoagulation monitoring appointments may not coincide with other clinic visits and ultimately results in no-shows for follow up. This is especially a concern for patients who live a substantial distance form the VA or those for whom frequent travel is difficult due to multiple comorbid conditions. Patient non-compliance with anticoagulation monitoring represents a substantial barrier to treatment and may adversely impact patient morbidity and mortality.
Methods: Approved VA sites would draw prothrombin time labs. Results would be called in and faxed to a designated pharmacist(s). Pharmacist would call patient regarding adjustment of warfarin dose. Results of therapuetic intervention would be doucmented in CPRS utilizing an anticoagulation monitoring template. Eligible patients: Undergoe detailed educational review of warfarin to include: indication,drug identification,side effects,drug-drug,drug-food interactions. Demonstrate understanding of prothrombin time and importance of frequent monitoring. Therapeutic anticoagulation per recommendations of the American College of Chest Physicians for 3 consecutive months. Documented patient consent to stated guidelines concerning compliance with monitoring and addtional follow up when applicable. Target Patient Population: Multiple comorbid conditions (i.e. severe COPD or impaired mobility) Fee basis recipients who prefer to have coordinated health care managed at VA.
Results: An outcome and cost effectiveness analysis of patients on warfarin monitored in this program will be performed.
Conclusions: A program such as this one will result in cost savings for the VA and will improve therapeutic anticoagulation rates, thereby reducing complications associated with hypertherapeutic or hypotherapeutic anticoagulation.
Impact: Potential Benefits: Improve compliance with anticoagulation monitoring Decrease adverse events, complications, hospitilizations and ER visits associated with non-therapeutic anticoagulation due to appointment non-compliance. Increase accessibility to allow for timely lab draws Reduce patient waiting time,travel time. Postively impact customer satisfaction Reduce non-urgent walk-in primary care visits Reduce travel expenditures and related health care costs.
*228. Publication Bias: Can we measure it?
John E. Cornell, GRECC/STVHCS-ALMD, Medicine/UTHSC-SA, San Antonio, Texas, CD Mulrow, Ambulatory Care/STVHCS-ALMD, Medicine/UTHSC-SA, MF Luther, Research and Development/STVHCS-ALMD
Objectives: Our objective is to assess the impact study heterogeneity has on measures of publication bias in quantitative systematic reviews. Graphical and statistical techniques typically used to assess publication bias masure the degree of asymmetry in funnel plots. We examine the assumptions that support funnel plot asymmetry as a true indicator of publication bias by emirically testing whether heterogeneity in study characteristics can explain asymmetry in funnel plots.
Methods: Data from a systematic review of newer pharmacotherapies for the treatment of depression was used. "Publication bias" had been detected among the double-blind, placebo control comparisons in this review. English and non-English literature was searched using the Cochrane Collaboration Depression, Anxiety, and Neurosis Group's specialized registry. Forty-four RCTs of six weeks or greater duration that evaluated one of 32 specific newer antidepressants or herbal treatments were included. The measure of treatment effectiveness was based on significant clinical improvement in depressive symptoms: i.e., a 50% reducion in HAMD scores. Analyses were based on a modified intent-to-treat principle. Publication bias was measured by examination of the funnel plot and appliction of Eggar's test. Eggar's test is based on a meta-regression of the standard error onto the treatment effect estimate. The slope of the meta-regression model is used to indicate the degree to which "publication bias" may be present in the included RCTs. We extended Eggar's meta-regression by including multiple variables indicative of study heterogeneity and tested whether these covariates reduced the slope (bias parameter) associated with "publication bias".
Results: Significant "publication bias" was identified among the comparisons of newer pharmacotherapies with a placebo control regardless of the measure of effect that was used: risk difference (RD, p<0.023), relative benefit (RB, p<0.001), or odds ratio (OR, p<0.001). Study level covariates that we added to the Eggar's meta-regression model were baseline severity of depression reported in the trials, the proportion of total drop-outs in the treatment and placebo groups, and the proportion of placebo responders. A logit transformation was applied to all measures expressed as proportions. The extended model produced a consistnt set of results across the three effect measures (RD, RB, OR). Introduction of the covariates reduced the value of the bias parameter by 50% or better. Treatment dropouts and the placebo response were the most important covariates that affected the bias parameter.
Conclusions: Asymmetry in a funnel plot is substatially influenced by the underlying heterogeneity among studies. Graphical or statistical tests of "publication bias" based on asymmetry in a funnel plot may reflect study heerogeneity rather than true publication bias.
Impact: These results provide one of the first empirical tests of the impact of study heterogeneity on measures of publication bias. The results have important implications for the use and interpretation of publication bias for researchers and policymakers who either conduct or interpret quantitative systematic reviews.
HSR&D Funded: HFP 98-002.
*229. Qualitative Analysis of Quality of Life among Congestive Heart Patients
Hayden B. Bosworth, Durham VAMC, Durham, NC, K Steinhauser, Durham VAMC, M Orr, Durham VAMC, J Hoff, Durham VAMC, SC Grambow, Durham VAMC, EZ Oddone, Durham VAMC
Objectives: Patients with congestive heart failure (CHF) experience significant physical and psychosocial decrements that impair their overall quality of life (QOL). This life threatening, chronic disease is punctuated by acute and often unexpected symptom episodes. As a result, patients are confronted with a complex mix of biomedical and psychosocial issues that impact their well-being. In order to improve care and judge efficacy of interventions, clinicians and researchers must have instruments that accurately assess the multiple dimensions of patients' QOL. Unfortunately, most QOL instruments do not capture the range of critical dimensions important to CHF patients. Existing measures focus primarily on physical function, and do not consider individual differences in importance patients attribute to particular domains. This study set out to identify what QOL represents for a sample of CHF patients.
Methods: Patients diagnosed with CHF coded on an outpatient visit to the Durham VAMC over the last year were eligible. Ejection fraction information was obtained from medical records; 60% of the sample had an ejection fraction measure in the last two years. After the subject pool was identified, patients who signed informed consent were categorized by NYHA classifications. Three focus groups with 15 older male veterans were convened. The patients were 47-82 years of age, eight were classified as having NYHA stage 3 and 4 CHF and seven were classified as having NYHA stage 1 and 2 CHF. Discussions were led by a psychologist and lasted an average of 90 minutes. The focus groups were conducted according to a structured, open-ended protocol. Patients were asked to define what QOL meant for them and report factors perceived to influence their QOL as consequence of their disease. All focus groups were audiotaped, transcribed, and coded for common and recurrent themes. The analysis focused on the identification of content categories that described the qualities of the patients' experiences.
Results: Subjects identified over 30 attributes of QOL. While prior domains identified as being important components of CHF patients' QOL were confirmed, many other new factors were also found. Examples of the new domains include spirituality, the consideration and preparation for death, the unpredictability of the disease trajectory, frustration with changes in roles and limitations, and loss of independence. Changes in patients' lives that could be attributed to their disease were not always considered deficiencies; methods of coping to CHF were identified as important attributes. Participants also emphasized the individual nature of definitions of what QOL meant for them; attributes were not perceived to be equally important across patients.
Conclusions: The domains identified contain underlying physical, psychological, and social well-being components. Qualitative data that confirm the current literature and add new QOL domains among CHF patients was obtained. These results will be combined into an instrument designed to assess QOL among CHF patients.
Impact: Identifying the domains important to CHF patients as well as determining the importance patients attribute to these domains is necessary to improve care.
*230. Quality of Life, Smoking, Alcohol, and Depression in Veterans with Head/Neck Cancer: A Pilot Study
Sonia A. Duffy, PhD, RN, VA Ann Arbor Health Care System and University of Michigan Medical Center, Ann Arbor, MI, J Terrell, MD, VA Ann Arbor Health Care System and University of Michigan Medical Center, M Valenstein, MD, VA Ann Arbor Health Care System, D Ronis, PhD, VA Ann Arbor Health Care System and University of Michigan Medical Center, M Connors, MS, RN, VA Ann Arbor Health Care System
Objectives: While rates for head/neck cancer have been declining in the general population, the rates have steadily increased for veterans. Head/neck cancer patients are at increased risk for poor quality of life, smoking, alcohol consumption, and depression, and these behaviors/disorders are often interrelated. Research has shown that despite their past history, head/neck cancer patients who screen negative for smoking, alcohol consumption, and depression have greater rates of survival. The objectives of this pilot study are to: 1. Determine baseline levels of quality of life, smoking, alcohol consumption, and depression among veterans with head/neck cancer; 2. Determine if there is a relationship between quality of life, smoking, alcohol consumption, and depression among veterans with head/neck cancer; and 3. Determine whether veterans with head/neck cancer would be interested in a combined intervention for smoking, alcohol consumption, and depression.
Methods: A self-administered survey was distributed to veterans with head/neck cancer in the ENT clinic at the Ann Arbor VAMC. The survey consists of screening questions for quality of life (SF-36V and Head/Neck Quality of Life Instrument-HNQoL), smoking (Fagerstrom Test for Nicotine Dependence), alcohol (Alcohol Use Disorder Identification Test), and depression (Geriatric Depression Scale). Data analysis remains in progress. The response rate is 78% (N=49).
Results: Both the physical and emotional functioning scores on the SF-36V are lower (34 and 42, respectively) in our pilot sample than the population means for veterans (50.0) reported by Kazis (1997). Furthermore, the mean scores on the emotional domain of the HNQoL questionnaire are lower in our pilot data (42) than those reported by Terrell (1998) for other head/neck cancer patients (85.7 for organ preservation group and 79.5 for surgical group). Of the 49 respondents, 36% currently smoke and 51.1% smoked within the last 6 months, 44% currently consume alcohol while 51.1% consumed alcohol within the last 6 months, and 45% screen positive for depression. In addition, 82% screen positive for one or more of these behaviors/disorders, 46% screen positive for two ore more of these behaviors/disorders, and 17% screen positive for all three of these behaviors/disorders. Furthermore, 31% are interested in smoking interventions, 15% are interested in alcohol interventions, and 29% are interested in depression interventions. Finally, 41% are interested in interventions for one of these behaviors/disorders, 16% are interested in interventions for two of these behaviors/disorders, and 14% are interested in interventions for all three behaviors/disorders.
Conclusions: Veterans with head/neck cancer have decreased quality of life compared to the general population of veterans. In addition, this population remains at high risk for smoking, alcohol consumption, and depression, the very risk factors that contributed to their disease. This pilot data suggests that current intervention strategies for these behaviors/disorders are not reaching a significant number of veterans with head/neck cancer.
Impact: Numerous studies have documented that smoking, alcohol consumption, and depression is interrelated suggesting the need for a multi-faceted intervention. We expect this pilot study to lay the foundation for a larger study designed to improve the quality of life of veterans with head/neck cancer by implementing a combined smoking, alcohol, and depression intervention.
*231. Racial Differences in Quality of Life among Veterans with Localized Prostate Cancer
Sara J. Knight, VA Cancer of the Prostate Outcomes Study (VA CaPOS) Group, Chicago, IL
Objectives: Previous studies have documented racial differences in prostate cancer outcomes with African Americans (AA), compared to whites (W), showing greater incidence, stage at diagnosis, and mortality. While one recent study reported that non-white men have poorer return to baseline quality of life (QOL) after prostate cancer treatment, only 11 AA men were included in these analyses. Our objective was to prospectively evaluate the QOL of AA and W Veterans during the first year after diagnosis.
Methods: Participants were a total of 31 W and 35 AA Veterans diagnosed with localized prostate cancer who completed interviewer-administered European Organization for the Research and Treatment of Cancer Quality of Life Questionnaires. Baseline QOL interviews were conducted before treatment commenced and at three and twelve months after study entry. Only half of the participants were married (46.0%) and almost a third had less than a high school education (31.7%). While AAs and Ws were similar in marital status, a larger proportion of AAs had less than a high school education (42.1% for AAs vs. 20.0% for Ws). Prostate cancer management included radical prostatectomy (RP), hormonal and radiation therapy (HT/RT) and watchful waiting, with fewer AAs than Ws receiving RP (38.5% for AAs vs. 61.5% for Ws) and more receiving HT/RT (58.3% for AAs vs. 41.7% for Ws). Mean differences in QOL scores are reported here with negative values indicating poorer function and positive values indicating better functioning.
Results: While baseline QOL was similar for AA and W Veterans, at three months, AAs reported decrements compared to Ws in social status (mean difference=-9.5 for AAs vs. 0.8 for Ws) and, at twelve months, in role and social status (mean differences=-16.7 and -8.3 for AAs, respectively, vs. -7.7 and 6.4 for Ws, respectively). In contrast, at twelve months, AAs had no change in finances compared to Ws who showed a decrement (mean difference= 0.0 for AAs vs. -15.4 for Ws). On prostate cancer specific items, at twelve months, AAs compared to Ws reported improvement in urinary frequency and control (mean difference=2.1 and 18.8 for AAs, respectively vs.-14.8 and -7.4 for Ws, respectively). At twelve months, AAs and Ws showed comparable improvement in sexual interest, function, activity level, and satisfaction.
Conclusions: Consistent with previous work, our data show differences among races in QOL at one year post-diagnosis of localized prostate cancer with AAs experiencing greater decrements and lesser recovery than Ws on role and social domains. In contrast, AAs had fewer decrements in finances and greater recovery in urinary function compared to Ws. Treatment choice differences (RP vs. HT/RT) among AAs and Ws may have contributed to long-term impact on role and social functioning for AAs. Also, lower levels of education among AAs, compared to Ws, may have influenced QOL on these domains.
Impact: AAs with localized prostate cancer may experience increasing difficulties with daily responsibilities and interpersonal relationships. Additional longitudinal data will help to clarify the impact of education level and patterns of care on the QOL of AA and W Veterans.
HSR&D Funded: IRR-95-120.1.
*232. Rehabilitation Issues of Concern to Women with Spinal Cord Injuries
Carole Ann Bach, Nashville VAMC, Nashville, TN
Objectives: Research Questions: 1. How do female veterans with SCI describe their rehabilitation experiences? 2. Are there common themes that arise in female veterans' descriptions of their rehabilitation experiences? 3. What strategies do female veterans with SCI use to manage their lives after injury? 4. What recommendations do female veterans with SCI have for interventions designed to meet their post-injury rehabilitation needs?
Methods: A grounded theory qualitative research design using three focused interviews was used to answer the research questions, with the goal of developing a theory about females undergoing rehabilitation for SCI and what need to be changed to better equip them for being as productive and independent as possible. The constant comparative technique is being used to analyze data. Thirty-three participants have completed the initial interviews. The mean age of participants was 40 (range 23-75) and a mean income of $3349/month. The levels of injury were classified as cervical (11), thoracic (17), and lumbar (5) with 11 classified as complete and 22 as incomplete.
Results: Preliminary Findings Preliminary discussion indicated that the women: 1). Need information about what to expect when they get home. 2). Need a network of other women with SCI to talk to about what is happening to them. 3). Need help with various issues surrounding management of urinary system. 4). Need to be well-organized in order to be as independent as possible. 5). Felt that exercise was important to them as was participation in the VA Games
Conclusions: The tapes of the interviews have been transcribed and checked for accuracy. Data analysis of each of the transcripts is continuing. Major activities in the last 12 months have included interviewing participants and analysis of data from initial interviews
Impact: This study will provide the foundation for an intervention study designed to individualize care for women veterans undergoing rehabilitation for Spinal Cord Injuries.
HSR&D Funded: 95-145.
*233. Relationship of Organizational Culture, Absenteeism, Turnover and Selected Social Characteristics of Hospital-Based Nursing Personnel
Ruth White Davis, Veterans Affairs Medical Center, New Orelans, LA, New Orleans, Louisiana
Objectives: The objective of the study was to: (1)describe the organizational culture of a federal hospital as perceived by nursing personnel i.e, nursing assistive personnel,licensed practical nurses (LPN), staff RN and nurse managers (2)detemine differences in the perceived organizational culture among hierarchical levels of nursing personnel,(3)determine the relationship among organizational culture, absenteeism and selected social characteristics and,(4)determine the relationship between turnover and selected social characteristics of nursing personnel.
Methods: A descriptive, correlational design using 206 nursing personnel from a southeast federal hospital completed the Organizational Culture Inventory (OCI) to measure culture and selected social characteristics. The OCI measures three types of culture: (1) Constructive culture including four satisfaction-oriented cultural styles: self actualizing, achievement, humanistic-encouraging and affiliative (2) Passive Defensive culture including four security-oriented styles regarding relationships in the workplace: dependent, approval, conventional, avoidance,and (3)Aggressive-Defensive culture including four security-oriented cultural styles about the how work is approached:power,competitive, oppostional, perfectionist. Twelve months of absenteeism and turnover data were collected and entered into formulas to determine rates. Data analysis using descriptive statistics, multiple regression, ANOVA, Chi-squared, Fisher's Exact and Fisher-Freeman-Halton statistical tests were conducted.
Results: Nursing personnel described a Passive Defensive culture. Multiple regression technique revealed a significant relationhship between an avoidance cultural style and absenteeism (p=.01). There was no difference in the perception of the organizational culture among the nursing hierarchy or by selected social characteristics such as age, ethnicity, educational background, tenure or gender. ANOVA revealed a significant positive relationship between absenteeism and turnover (p=0.0001) and educational level (p=.05), and negative correlation to tenure (p=.04). On average, the highest turnover was found among females, 20-29 years old with a tenure of one - ten years.
Conclusions: The findings this study revealed a Passive Defensive culture described by nursing personnel with behavioral norms of risk aversion, acquiescence and antagonism. The avoidance cultural style was directly related to higher absenteeism levels by nursing personnel. Moreover, high absenteeism levels resulted in disruptive turnover by nursing personnel, particular young, female registered nurses.
Impact: A dissonant hospital culture of risk aversion, dependency, antagonism and ineffective communication impact the cost of during business by increasing the levels of absenteeism among nursing personnel. Increased levels of absenteeism is further aggravated by increased turnover among nursing personnel. Excessive absenteeism and turnover negatively impact cost, the quality of care, continuity of care, and staff morale.
*234. Smart Search: A Quick Easy Method for Searching Medical Evidence
Cynthia D. Mulrow, MD, MSc, South Texas Veterans Health Care System, Audie L. Murphy Division, San Antonio, TX, Robert Badgett, MD, South Texas Veterans Health Care System, Audie L. Murphy Division, Cynthia D. Mulrow, MD, MSc, South TX Veterans Health Care System, Audie L. Murphy Division, Linda Levy, Briscoe Library, UTHSCSA
Objectives: Because few clinicians have skills or time to search for up-to-date medical evidence, we developed and tested an automated method of metasearching the Internet for medical evidence.
Methods: SmartSearch (http://SmartSearch.UTHSCSA.edu)is an automated metasearch engine that simultaneously identifies medical evidence from multiple sources such as MEDLINE and the National Guideline Clearinghouse. It examines clinician’s simple queries then automatically chooses best resources to search, formats the search for each resource, and executes the searches. Based on the search results, SmartSearch may search alternative resources when too few articles are retrieved, or SmartSearch may reformat a more restrictive search when too many articles are retrieved. In addition, SmartSearch analyzes searches that find no articles and suggests how to correct them. SmartSearch also can manage passwords to journal web sites. To use SmartSearch, the clinician only needs to know how to construct a search query using common terms and Boolean logic (Boolean logic is the proper use of 'ands' and 'ors'). Although SmartSearch has not yet been formally announced outside of our institution, our clinicians are encouraged to use SmartSearch and are instructed regarding its use. We used two Internet server logs to measure the impact of SmartSearch. First, we analyzed the frequency of selection of SmartSearch and other resources by clinicians who use our clinical web site. Second, we analyzed the server log of SmartSearch to assess the frequency and characteristics of "successful" searches. "Successful" searches are defined as retrieving at least one article from MEDLINE.
Results: In the last one month, SmartSearch was queried 1767 times. Users of the clinical web site requested SmartSearch 449 times and requested OVID, the next most frequent resource, 69 times. In the last month, 81% of queries were successful while 19% of searches failed. The was no difference in the number of search terms between successful and failed searches. Successful searches, as compared to failed searches, were more likely to use two or less words within all search terms (83% versus 51%, 2=158, p<0.001). When clinicians used more than two words within a search term, 41% of searches failed. Forty-four percent of searches for etiology questions and 33% of searches for adverse effects failed, whereas only 14% of searches for therapeutic questions failed.
Conclusions: Clinicians who use our clinical web site use SmartSearch more often than any other medical resource. Eighty-one percent of searches find at least one article. "Successful" searches are more likely to use search terms that are two words or less in length and more likely to address therapeutic than other questions. Although our definition of "successful" searches is certainly an overestimate, it allows practical identification of characteristics of better search strategies. SmartSearch can then be programmed to warn clinicians when they are about to submit poor strategies such as using three words in a search term.
Impact: As SmartSearch is quick and requires little training, it has great potential for improving health care professionals' abilities to seek, find and use medical evidence. It can be used in any Veterans Affairs clinical settings that have access to Internet.
*235. Social Isolation Increases Risk of Re-Hospitalization
Joel Rosansky, Greater Los Angeles Healthcare System - West LA Healthcare Center, Los Angeles, California, R Mistry, Greater Los Angeles Healthcare System - West LA Healthcare Center, J McGuire, Greater Los Angeles Healthcare System - West LA Healthcare Center, C McDermott, Greater Los Angeles Healthcare System - West LA Healthcare Center, L Jarvik, Greater Los Angeles Healthcare System - West LA Healthcare Center (Retired), UCLA Department of Psychiatry and Biobehavioral Sciences, National UPBEAT Collaborative Group, Greater Los Angeles Healthcare System - West LA Healthcare Center
Objectives: The relevance of social networks to health, well-being and healthcare utilization has been demonstrated previously. The role of social support in recovery and management of chronic disease in the elderly is important since this population is at increased risk for both isolation and chronic illness. Current literature lacks data on social isolation risk and likelihood of re-hospitalization. The present study examines social isolation risk and likelihood of re-hospitalization in a group of elderly male veterans followed for one year at a VA Healthcare Center. It is postulated that greater risk for social isolation is associated with higher rates of re-hospitalization.
Methods: Participants in this 12-month longitudinal study are part of UPBEAT (Unified Psychogeriatric Biopsychosocial Evaluation and Treatment), a mental health care coordination demonstration project at nine VA Healthcare Centers nationwide. 144 UPBEAT patients at the VA West Los Angeles Healthcare Center were evaluated using the Lubben Social Network Scale (LSNS) and Rand 36-Item Health Survey Short Form, and inpatient utilization was tracked over the year following enrollment. Patients were categorized as isolated, at high risk for isolation, at moderate risk for isolation, or at low risk for isolation based on LSNS scores. These variables and other covariates (years of education, age, annual household income and ethnicity) were used to develop a logistic regression model. To test what aspects of social networks were related to re-hospitalization, bivariate analysis (Chi-squared; t-test) was conducted between re-hospitalization and items on the Lubben scales.
Results: The regression model was significant (Chi-squared = 26.9; p = .0007). Variables significant in the model were physical health status (OR = .92; p = .000), mental health status (R=.93; p=.04), years of education (OR=1.2; p= .04) and social isolation risk (OR = 3.2; p = .01). The group of patients who were isolated or at moderate to high risk for social isolation was 3 times more likely to be re-hospitalized within the study period than were low-risk patients. Ethnicity, age and income were not significant predictors. Bivariate analyses indicated that patients not re-hospitalized, when compared to re-hospitalized patients, had more close relatives ( t = 2.54; p= .01) and people to talk to about important decisions ( t = 2.07; p= .04) .
Conclusions: Social isolation is an important risk factor for increased re-hospitalization in this population. Patients with no re-hospitalizations were characterized by having close relatives whom they can call for help and someone to talk to about important decisions.
Impact: Due to the high costs of inpatient care and social isolation of the elderly, these findings have important implications for policy as well as care coordination. Healthcare interventions which assist elderly veterans in building or maintaining social networks may reduce re-hospitalizations and improve overall quality of life. This analysis augments the scant data on social networks of hospitalized geriatric veterans.
*236. Spiritual Experience and Advanced Directives
Oscar C. Stine, MD, VAMC Martinsburg, Shepherdstown, WV, JT Cunningham, VAMC Martinsburg, CE Drebing, VAMC Bedford, OC Stine, VAMC Martinsburg, KM Lyttle, VAMC Martinsburg, DE Lafaber, VAMC Martinsburg, SJ Jenks, VAMC Martinsburg
Objectives: We propose to improve the outcomes of advance directive counseling by correlating spiritual experience reported by Berg’s COMPUTER ASSESSMENT PROGRAM (CAP) with repeated rejection of advance directives (AD). (Berg, Gary E. Department of Veterans Affairs. 1992.) Veterans admitted to the addiction and homeless rehabilitation services at the Martinsburg Center have completed this assessment for four years. The chaplains have used the results to initiate and guide their counseling services. The CAP has been shown by Lawson et al. to be valid and capable of showing the late effects of childhood abuse that correlate with spiritual injury score. (Lawson, R. et al. The Long Term Impact of Child Abuse on Religious Behavior of Men. Child Abuse and Neglect. Vol. 22. No.5. 369-380. 1998.) Since the majority of veterans answer no past and present offers of AD’s, and the Self Determination Act of 1992 requires an AD, it is important to find the positive and negative spiritual experiences associated with use or disuse of AD.
Methods: The chaplain sets up the computer and leaves to avoid biasing the veteran’s "y" or "n" answers. The answers are printed out and scores calculated to guide the chaplain’s counseling. There were 1542 assessments performed. Frequency Distributions, discriminant analyses, and cluster analyses were performed.
Results: The 825 veterans that answered no to both AD questions compared with the 717 others. There is a distinct excess of disusers answering "0" on the spiritual dimension score: 18.5% compared with 13.2 % of the others. In contrast the disusers exceeded others in report of daily use of prayer, 31.1 % to 28.1 %. On the life change / loss score, 53.0 % of the disusers scored below 600 compared with 42.9 % of the others. The curve for the disusers in the spiritual injury score exceeds the curve for the others between scores of 8 and 15, 41.9 % compared with 31.2 %, and between scores of 19 and 21, 19.1% compared with 15.1% of the others.. The disusers reported worrying about or fearing death "sometimes" or "never" 91.4 while 87.6 % of the others used those answers.
Conclusions: Veterans who deny use or interest in AD’s may include those that avoid or deny organized religion, but seek support through personal prayer. They may report less life change and loss score. They are more likely than others to report minimal fear of death and to minimize the experiences causing spiritual injury.
Impact: Chaplains' confirmation of the computerized report to identify denials, need for counseling or referral for resolution of stress and spiritual injury may need to precede counseling for AD’s.
*237. Spiritual Well-Being in Older Individuals with Chronic Illness: Development of a Spiritual Function Scale
Kenneth E. Rosenfeld, MD, VA Greater Los Angeles Healthcare System, Los Angeles, CA, R Hays, Rand, MJ Steckart, VA Greater Los Angeles Healthcare System
Objectives: There have recently been increased interests in the relationship between spirituality and health, including empirical evidence that spiritual well-being represents a unique dimension of health-related quality of life (HRQOL). Nevertheless, efforts to integrate spirituality into HRQOL assessment have been limited by difficulties in conceptualizing spiritual health and in translating conceptual models into reliable and valid measurement instruments. The purpose of this study was to develop, pilot test, and psychometrically evaluate an instrument to measure spiritual well-being in an older, chronically ill population.
Methods: We conducted eight focus groups with chronically ill individuals over 65 years of age, stratified by race, illness category, and gender to identify patients’ perspectives on the meaning of spiritual health. Analysis of focus group transcripts was integrated with a literature review and input from an expert advisory panel in order to develop a preliminary spiritual well-being instrument. The instrument was then pilot tested in cognitive interviews with 45 older individuals from three ethnic groups. Pilot testing involved instrument completion followed by concurrent think-aloud and retrospective interviews, which identified problematic items and explored subjects’ perception of item responses. Detailed field notes were taken and reviewed to ensure that items successfully represented the intended spirituality dimension. The instrument was revised and will now be administered to a convenience sample of 300 individuals 65 years of age or older (at least 200 of whom will have major chronic illness) from local senior centers in West Los Angeles and from outpatient medical clinics at our VA and at UCLA. Respondents also will complete the MOS SF-36 and Paloutzian and Ellison’s Spiritual Well-Being Scale, and provide demographic information.
Results: The revised spiritual function scale contains – items representing six dimensions of spiritual health – general spiritual orientation, existential perspective, morality, interpersonal relationships, positive outlook, and overall affect. Data collection is ongoing at the time of this publication, however, final analyses will include extensive evaluation of the psychometric properties of the pilot tested instrument. Item and scale distribution , scale internal consistency, reliability, the factorial structure of spirituality items, and the relationships between spirituality and other quality of life constructs will be examined. These analyses will identify ways of shortening the instrument and asses whether the spirituality items enhance information provided in standard HRQOL tools.
Conclusions: We are able to rigorously develop an instrument to measure spiritual well-being in older, chronically ill adults. Pilot data currently being collected and analyzed will display the instrument’s psychometric and factor properties, as well as its criterion validity. Comparisons with the SF-36 will provide preliminary information on the legitimacy of including spiritual functioning as a unique dimension of HRQOL in older, chronically ill individuals.
Impact: If the spiritual function instrument developed in this study adds important information to that obtained in the SF-36, then the instrument would be appropriate to include as a standard part of HRQOL assessment.
*238. Staff Member Integration within Psychiatric Interdisciplinary Treatment Teams
Rebecca Wells, Penn State, University Park, PA, J Alexander, U. of Michigan, F Blow, Serious Mental Illness Treatment Research and Evaluation Center, R Lichtenstein, U. of Michigan, F Lee, U. of Michigan
Objectives: Quality improvement depends upon the active engagement of staff from a range of disciplines, yet previous research has shown that individuals tend to withdraw from diverse groups. In this study we examined the effects of demographic differences from coworkers on individual staff member integration within psychiatric interdisciplinary treatment teams. In addition, we tested whether staff members who were more integrated within their treatment teams provided better patient care than less integrated staff members.
Methods: We tested both the demographic antecedents and the behavioral consequences of integration within treatment teams using data collected through the US Department of Veterans Affairs Long Term Mental Health Enhancement Program (LTMHEP). We used a lagged structure with demographic and integration measures from 1994 (n=992 staff members in 97 teams) and patient care data from 1995 (n=447 staff members in 79 teams). We tested the overall effects of demographic difference in the full sample, and also tested for varying effects depending on social status by splitting the sample into high and low status subgroups for each attribute (e.g. high and low status disciplines). Multilevel modeling accommodated the nested nature of the data.
Results: In the pooled sample, controlling for simple demographics (respondent discipline, age, sex, and position tenure) we found scattered effects of demographic difference from coworkers on some aspects of staff member integration within teams. Difference in age from other team members was negatively related to role clarity, and difference in position tenure was negatively associated with both participation and role clarity. In the split samples, we found that disciplinary difference from colleagues increased participation and perceived use of skills for members of higher status disciplines. The strongest pattern of results, however, was from the variables intended to serve as controls. Members of higher status disciplines were more integrated within treatment teams, whereas members of lower status disciplines experienced lower integration than other staff. Also, contrary to our expectations, integration within treatment teams was not related to therapeutically appropriate patterns of patient care behavior.
Conclusions: This study indicated that members of higher status disciplines were significantly more integrated within treatment teams than were members of lower status disciplines. Contrary to our expectations, these effects were generally not relational. That is, what appeared to matter was what someone was (e.g. psychiatrist or psychologist versus LPN or aide), not who else was in the room. We were also surprised by the fact that more integrated team members did not appear to provide more therapeutically appropriate patient care than less integrated individuals. This latter result may have occurred because we could not capture the actual outcomes of integration within interdisciplinary treatment teams with available data.
Impact: Members of higher status disciplines are more integrated within interdisciplinary treatment teams than are members of lower status disciplines. A leader who wants every discipline to be well represented in team discussions may need to take extra steps to both encourage individuals from lower status disciplines to talk and ensure that their perspectives are valued.
HSR&D Funded: Long Term Mental Health Enhancement Program (LTMHEP) 03.
*239. Systematizing Intensive Care Unit Surveillance within a VA Healthcare Network
Linda L. McKinley, Philadelphia VA Medical Center, Philadelphia, PA, CC Johnson, Philadelphia VA Medical Center, HJ Moriarty, Philadelphia VA Medical Center, TH Short, Villanova University
Objectives: This study was designed to standardize infection control surveillance of nosocomial infection rates (NIR) within a VA healthcare network in which interhospital comparisons are available. The specific objectives are: 1) provide network hospitals with a systematized data management system for intensive care unit (ICU) surveillance and 2) to compare ICU NIR in hospitals that receive comparative data feedback to those that do not receive the comparative data.
Methods: Since infection control surveillance is not performed using standardized methodology within the VA health system, interhospital comparisons are of limited value. In addition many smaller VA hospitals do not meet entry criteria nor have the resources to participate in national surveillance projects, such as the Centers for Disease Control National Nosocomial Surveillance System (CDC NNIS). To facilitate standardization of surveillance within a VA healthcare network, this study used one central coordination site to collect and analyze data for other network hospitals. VA Stars and Stripes (VISN 4) is a health care network of ten hospitals. Among VISN 4 hospitals only the Philadelphia VA Medical Center (PVAMC) has been a CDC NNIS participant. In this study, PVAMC serves as the central coordination site where surveillance data is analyzed according to CDC NNIS criteria and reported back to the VISN hospitals. The eight participating hospitals in the study were randomized into one of two groups. The experimental group (N=4 hospitals) receives risk adjusted infection rates with national comparative data and the control group (N=4 hospitals) receives only the risk adjusted infection rates without the comparative data. Data has been submitted and analyzed monthly since January 1, 1999.
Results: The healthcare network in VISN 4 has successfully implemented a standardized infection control surveillance system for NIR in the ICU. Analysis from the fist year of data collection compares the effect of NIR in hospitals that received comparative data to those hospitals that did not receive comparative data.
Conclusions: The use of comparative data for interhospital comparisons aims to reduce NIR, decrease morbidity and mortality, and reduce cost.
Impact: The study may serve as a model for other infection control surveillance indicators within VISN 4, as well as an infection control surveillance model for other VA healthcare networks.
*240. The Association of Illness Trajectory with Hospital Mortality: An Example of Longitudinal Logistic Regression
Michael L. Johnson, PhD, Houston Center of Excellence, Houston, TX, W Chan, University of Texas School of Public Health, AS Kapadia, University of Texas School of Public Health, CS Slater, University of Texas School of Public Health, CM Ashton, Houston Center of Excellence
Objectives: To demonstrate the association of illness trajectory measured by daily laboratory tests with hospital mortality using longitudinal logistic regression, a method to analyze repeated measures of risk factors and a binary outcome.
Methods: Laboratory data was collected using the Health Summary component of VISTA for 476 consecutive patients admitted with non-psychiatric principal diagnoses to the general medicine section of the Houston VAMC. Laboratory test results were scored to create APACHE-L severity of illness measures for each day of the hospital stay that laboratory tests were conducted. The APACHE-L is based on the laboratory component of the original APACHE (Acute Physiology and Chronic Health Evaluation) severity of illness measure. Repeated measures of severity of illness form a pattern, or illness trajectory, leading to discharge or hospital death, defined as died in the hospital or within 30 days of discharge. The association of illness trajectory with mortality was analyzed with a longitudinal logistic regression model using SAS PROC GENMOD. The regression framework allowed control of fixed and time-dependent covariates. Fixed covariates were demographic factors, including age, race, and marital status, and comorbidities. The APACHE-L severity of illness measure was the time-dependent risk factor tested in the model. Model performance was assessed by calculating a "time-dependent" c-index, or area under the ROC curve.
Results: A total of 44 patients died, for a crude mortality rate of 9.2% (44/476). In unadjusted analyses, married patients were at decreased risk for mortality, while patients with cancer comorbidity were at increased risk. The average APACHE-L score at admission was 7.9 among patients who died, compared to 3.6 among survivors, a statistically significant difference. In the longitudinal logistic regression model, patients of white race had an odds ratio (OR) of 0.50, and 95% confidence interval (CI) of (0.28, 0.91), indicating a decreased risk for mortality. Married patients had an OR=0.54 and 95% CI of (0.28, 1.06), a decreased risk for mortality at the p <0.10 level. The OR for cancer comorbidity was 2.3 (1.3, 4.0). APACHE-L was highly statistically significant (p < 0.001), with an OR of 1.2 (1.15, 1.3) for each increasing point in the score. The c-index ranged from a low of 0.773 at 21 days of follow-up to a high of 0.856 at 4 days of follow-up.
Conclusions: Longitudinal logistic regression can be used to model the association of illness trajectory with hospital mortality. Because of the longitudinal nature of the APACHE-L severity of illness measure, the positive odds ratio indicates that higher and increasing APACHE-L scores, i.e. worsening trajectory, is associated with hospital mortality. Lower and decreasing APACHE-L scores, i.e., improving trajectory, is associated with survival at discharge.
Impact: Regression techniques to handle longitudinal data are becoming more commonly used in health services research, but longitudinal logistic regression for binary outcomes is not well-understood. Quality assessment efforts could expand from focus on condition at admission to include monitoring of changes in illness trajectory leading to hospital discharge or death. Further research is needed to explore methods to use longitudinal data in risk-adjustment of health care outcomes.
*241. The Chronic Disease Score and Pharmacy-Based Risk Adjusters – Recommendations for Replicability
Kevin L. Sloan, MD, VA Puget Sound Heath Care System, Seattle, WA, AE Sales, VA Puget Sound Heath Care System, PA Fishman, Center for Health Studies, Group Health Cooperative of Puget Sound, JD Malkin, Covance Health Economics and Outcomes Services Inc
Objectives: As efforts to compare clinical care across providers and settings continue to increase, the need for adequate risk adjustment methodologies continues to become more acute. Unfortunately, most currently available risk-adjustment models are proprietary, require burdensome clinical data collection or both. One exception is the Chronic Disease Score (CDS), a prescription-data-based risk adjuster developed at Group Health Cooperative (GHC) in Seattle and subsequently replicated at a number of sites in the US and abroad. We report our experiences in translating the CDS algorithm to the VISN 20 site formularies and make recommendations for the description of this and other open-source pharmacy-based risk adjusters.
Methods: We obtained the SAS source code for the current GHC CDS, which is coded via National Drug Code (NDC) identifiers, along with a publicly available database from Multum Corp. that translates NDC codes to drug names and classes. (This is the same database used by the original developers of the CDS.) We also used the CHIPS relational database to access the VA National Drug File and formularies at all VISN 20 sites. In a sequential manner, we attempted to match on a variety of fields and combinations including NCD code, drug name (VA generic name), route of administration and VA-specific fields including VA Product Name and National Drug File ID.
Results: A number of difficulties were identified. Problems with content definitions included mapping multiple oral potassium preparations to the Asthma category and inclusion of antinausea agents (e.g., promethazine) and exclusion of the atypical antipsychotics (save clozapine) from the Psychotic Illness category. There were also a number of single-drug mappings that were clearly erroneous on clinical grounds (e.g., one ibuprofen preparation mapping to the Diabetes category). Furthermore, the lack of a clear and explicit rationale for drug inclusion or exclusion for CDS categories made it difficult, without content area specific clinical knowledge, to judge whether drugs not found on the current version of the CDS should be added to the algorithm.
Conclusions: Although the CDS, and by extension other pharmacy based methodologies, show some promise in providing an open-source alternative to current diagnosis based risk adjusters, a more structured approach to category definition and specification would be useful. We propose a schema that defines categories at a minimum of three levels of abstraction: clinical disease or organ system (e.g., ischemic heart disease), drug class (e.g., nitrates) with route, if applicable, and drug (e.g., isosorbide dinitrate). Each description should contain a sufficient level of detail to provide a clear rationale for decision making with regard to novel agents and drug classes. We particularly recognize that, in the absence of any drug categorization schema analogous to ICD-9 for diagnostic data, a clear multi-level algorithm description is necessary to allow optimal portability as well as cross-study comparison.
Impact: Establishment of a more systematic method for describing pharmacy-based risk adjusters would allow for improved portability and better cross-study comparison. This, in turn, will potentially aid investigators in developing improved risk-adjustment methodologies.
*242. The Cost of Family Caregiving for the Elderly with Dementia: New Estimates from a Nationally Representative Sample
Kenneth M. Langa, MD, PhD, VA Health Services Research and Development Service, VA Ann Arbor Healthcare System, Ann Arbor, MI, and the University of Michigan, Ann Arbor, MI, AR Herzog, University of Michigan, MB Ofstedal, University of Michigan, RJ Willis, University of Michigan, MU Kabeto, University of Michigan, ME Chernew, AM Fendrick, University of Michigan
Objectives: As the US population ages, the increased prevalence of dementia will create a substantial societal burden. Studies conclude that the cost associated with family caregiving--not direct medical costs--accounts for most of the total cost of dementia. However, existing estimates of caregiving costs derived from small, non-representative samples are likely biased. Accordingly, our objective was to obtain better estimates of the cost of family caregiving in a nationally representative sample.
Methods: We used data from the first two waves (1993 and 1995) of the Asset and Health Dynamics (AHEAD) Study, a national longitudinal survey of people aged 70 or older (N=7,443). Using the validated AHEAD cognitive status measure and place of residence at Wave 2, we identified four dementia stages of increasing severity: EARLY STAGE (no dementia at Wave 1 / dementia and living at home at Wave 2); MIDDLE STAGE (dementia at Wave 1 / living at home at Wave 2); LATE STAGE (dementia at Wave 1 / in a nursing home at Wave 2); and END STAGE (dementia at Wave 1 / dead at Wave 2).
Results: Of the 834 (11%) respondents with dementia, 65% received informal care from family members while living in the community at Wave 1, compared to only 21% for those with normal cognition (p<.01). Among those receiving family care, there was a significant association between dementia stage and intensity (hours per week) of caregiving (p<.01). Those with NORMAL cognition received an average of 23 hours per week of family care; EARLY STAGE dementia cases received 35 hours; MIDDLE STAGE received 42 hours; LATE STAGE received 48 hours; and END STAGE received 71 hours. Using the average hourly wage of a home health aide in 1993 ($11) as the value of a family caregiver's time, the expected yearly caregiving cost per case ranged from $20,000 for EARLY STAGE cases to $40,600 for END STAGE cases.
Conclusions: Family caregiving costs for those with dementia are substantial and increase sharply with the severity of dementia. Accurate estimates of caregiving costs are, therefore, essential to the validity of future evaluations of interventions that decrease the incidence or slow the progression of dementia.
Impact: Family caregiving costs for those with dementia are substantial and increase sharply with the severity of dementia.
*243. The Cost-utility of Screening for Depression in Primary Care
John Zeber, MHA, VA Center for Practice Management and Outcomes Research, Ann Arbor, MI, S Vijan, VA Center for Practice Management and Outcomes Research, M Valenstein, VA Center for Practice Management and Outcomes Research
Objectives: Depression is a common disorder in primary care that causes substantial disability but often remains undiagnosed or untreated. Screening for depression is one strategy for increasing detection and has been considered or implemented by several healthcare organizations, including the Department of Veterans' Affairs. We conducted a formal cost-utility analysis of screening for depression compared to "usual care" in order to examine the cost-utility of screening in typical primary care settings and to evaluate the conditions under which screening might be most valuable.
Methods: We developed a semi-Markov model to examine the lifetime costs and benefits of screening for depression for hypothetical cohorts of primary care patients, aged 20, 40, and 60 years. Estimates of prevalence of depression, detection and treatment in usual care, the impact of screening, patient response to treatment, and treatment costs were drawn from a comprehensive review of the published literature. Cost-utility analyses were conducted from the perspective of a single payer system, and costs and benefits were discounted at an annual rate of 3%.
Results: Using baseline assumptions, annual depression screening had a cost-utility (C/U) ratio of over $200,000/QALY gained. Opportunistic screening had a higher C/U ratio, while screening only at entry into the system had a lower ratio, but all screening strategies exceeded C/U ratios of $75,000/QALY gained. C/U ratios did not decrease to ranges often considered to support program implementation (below $50,000/QALY) in one-way or two-way sensitivity analyses. Only when 3 or more model parameters were changed to feasible but "optimistic" values that favored screening did C/U ratios fall below $50,000/QALY. Estimates for C/U ratios were most sensitive to potentially changeable factors of costs of screening, initiation of treatment by PCPs once depression is diagnosed, and remission rates in patients who receive treatment.
Conclusions: Screening for depression is unlikely to be a cost-effective intervention in most primary care settings. In order for screening to become cost-effective, the costs of the screening program must be minimized and substantial improvements must be made in the quality and effectiveness of treatment that is provided within primary care.
Impact: These findings have clear relevance to policy makers and administrators within the VA and other healthcare organizations who are considering initiating or continuing depression screening programs. In most primary care settings, benefits that ensue from "stand alone" screening programs are prohibitively expensive. Improvements in the quality of depression treatment in primary care must precede or accompany screening efforts.
HSR&D Funded: Career Development Award RCD 98-350.
*244. The D-E-N-T-A-L Identifies Veterans in need of Dental Care
Judith A. Jones, DDS, MPH, VA Center for Health Quality, Outcomes and Economic Research, Bedford, MA and Boston University, Bedford, MA, NR Kressin, VA Center for Health Quality, Outcomes and Economic Research, Bedford, MA and Boston University, RI Garcia, VA New England Health Care System (MAVERIC), and Boston University, CW Randall, VA Center for Health Quality, Outcomes and Economic Research, Bedford, MA and Boston University, DR Miller, VA Center for Health Quality, Outcomes and Economic Research, Bedford, MA, MAVERIC, and Boston University, A Spiro III, VA New England Health Care System (MAVERIC), VA Center for Health Quality, Outcomes and Economic Research, Bedford, MA and Boston University
Objectives: The Veterans Health Administration's 36 million annual outpatient visits include 1.2 million dental visits. In such a large organization, a screening methodology that would correctly identify persons needing dental care would be useful. We examined a six item self-administered questionnaire for its ability to correctly identify veterans in need of dental care. This instrument, the D-E-N-T-A-L (Bush et al., JAGS 1996;44:979), queries whether participants have Dry mouth, Eating or swallowing problems, Not had a dental exam in the last 2 years, Tooth or mouth problems, Altered eating habits because of teeth or mouth, and Lesions or sores in the mouth.
Methods: Clinical and self-reported oral health (questionnaire) data were collected in two samples of veterans: (1) 233 men who used VA outpatient medical care from the Veterans Health Study (VHS: age=62+12); and (2) 206 relatively healthy men from the VA Dental Longitudinal Study (DLS: age=71+7) who do not use VA health care. The gold standard for need for care was determined by clinical exam conducted by a trained and calibrated examiner. Specific criteria were established for 4 categories of clinically determined need including diagnostic, restorative, periodontal, or denture-related need. Persons with need in at least one category were considered to have clinically determined need for care. We examined multiple cutpoints out of 8 possible (2 items were given weights of 2) on the D-E-N-T-A-L as indicating need.
Results: Results: Overall 97% of VHS and 64% of the DLS participants had at least 1 clinically determined need while 44% in the VHS and 13% had D-E-N-T-A-L scores of 2 or higher. In the VHS, sensitivities ranged from 43% for a cutpoint of 2 to 14% for a cutpoint 4 and specificities ranged from 33-83%. In the DLS, sensitivities were very low, (e.g. 14% for a cutpoint of 2) while specificities ranged from 88% for a cutpoint of 2 to 99% for a cutpoint of 4. Efficiency (sensitivity + specificity)/2 was maximized in the VHS in predicting persons needing periodontal care using a cutpoint of 2.
Conclusions: Need for care was high in the VA patients and moderate in the healthier, community residing men. Specificities improved with higher cutpoints in both populations. In populations where need for care is high (e.g. the VHS) and where need for care is modest (e.g. DLS) the D-E-N-T-A-L may be of value in identifying persons who need care.
Impact: Declining resources require that VHA facilities target services to persons in greatest need for care. Use of the D-E-N-T-A-L may assist VHA in targeting oral health care to veterans with specific oral health needs.
HSR&D Funded: IIR93-025.B
*245. The Effect of a Narcotics Management System on Narcotics Prescribing and Unscheduled Visits in a Large Veterans Primary Care Clinic
Kenneth Charles Goldberg, HSR&D, Durham VAMC, Durham, NC, EZ Oddone, HSR&D, Durham VAMC, DL Simel, HSR&D, Durham VAMC
Objectives: Several national guidelines advocate specific recommendations for prescribing narcotics for chronic pain; these include establishing a written pain contract, limiting the number of providers, and providing consistent, scheduled supplies of narcotics. It is not known how adoption of these interventions affects provider or patient behavior. This study measured the effect of an explicit pain management program on unscheduled patient visits, narcotics prescribing behavior, and oxycodone use in a large ambulatory care clinic at the a university-affiliated VA hospital.
Methods: In 1997, the Durham VA Ambulatory Care Clinic initiated a comprehensive pain management program to help primary care providers manage more effectively patients whose treatment involved chronic narcotics. This program involved a specific pain contract signed by the provider and the patient and a clerical mechanism that assisted with scheduled prescribing of controlled substances. All patients with evidence of such a pain contract were identified retrospectively. A matched control group of patients followed in the ambulatory care clinics during the same period that received comparable amounts of narcotics, but did not have a pain management plan on their chart, was also identified. Demographic information, all outpatient visits, and narcotics prescriptions were abstracted and summarized for both groups. Visits and narcotics prescriptions from other nearby VA hospitals were collected for intervention patients. Emergency department visits, the number of separate providers issuing narcotics prescriptions, the number of unique narcotics prescriptions, and the amount of oxycodone consumed were summarized by calendar quarter for intervention patients before and after the execution of their pain management contract, and for control patients throughout the study period.
Results: 91 patients with a pain management contract were identified. Compared to the time period prior to the contract, visits to the emergency department, the number of providers issuing prescriptions, the number of separate prescriptions for narcotics, and the number of dispensed oxycodone tablets decreased significantly (p<0.001 for each measure) after implementation of the contract. Visits to other nearby VA hospitals did not increase. In a matched group of 224 patients receiving narcotics, emergency department visits decreased during the observation period, but to a lesser degree. The number of separate providers issuing narcotics and the number of unique prescriptions did not change over time, although the number of oxycodone tablets consumed increased steadily. Patients in both groups frequently visited specialty clinics that would be expected to assist in a multidisciplinary approach toward the management of chronic pain.
Conclusions: A pain management program made explicit by a provider-patient contract can result in fewer unscheduled visits and consolidation of narcotics prescribing.
Impact: For patients that receive narcotics for chronic pain, an intervention that incorporates frank and explicit communication between patients and their providers, and communication between providers and a larger health-care system in the form of a shared medical record, can reduce unscheduled visits, reduce fragmentation in narcotics prescribing, and possibly decrease the amount of controlled substances that patients consume. The intervention itself is inexpensive, could be implemented in any large, integrated health-care system, and leads to changes recommended by national treatment guidelines.
*246. The Effects of Individualized vs. Classical "Relaxation" Music on the Frequency of Agitation in Elderly Persons with Alzheimer's Disease and Related Disorders
Linda A. Gerdner PhD, RN, HSR&D Center for Mental Healthcare and Outcomes Research Little Rock VA Medical Center, North Little Rock, Arkansas
Objectives: Agitation has been identified as a major problem in persons with dementia such as Alzheimer's disease and related disorders (ADRD). Limitations and concerns related to the current managment of agitation suggest the need to investigate alternative interventions. The purpose of this study was to compare the immediate and 30-minute residual effects of individualized to classical "relaxation" music relative to baseline on the frequency of agitation in persons with ADRD.
Methods: An experimental repeated measures pretest-postest cross-over design was used to study 39 subjects recruited from six long-term care facilities in Iowa (5 community and 1 VA). The sample was comprised of 30 females and 9 males (mean age 82 years)with severe cognitive impairment as evidenced by a Global Deterioration Scale score of 6. Each subject was studied over a 17 week period. Baseline data were collected during the initial three weeks. Group A received individualized music for 6 weeks followed by a 2-week "washout" period and six weeks of classical "relaxation" music. Group B received the same protocol but in reverse order. The selection of individualized music was based on findings from the Modified Hartsock Music Preference Questionnaire (Family Version). The time of intervention was individualized based on application of the Progressively Lowered Stress Threshold Model. Music interventions were presented "free field" for 30 minutes, 2 times per week. The frequency of agitation was measured using the Modified Cohen-Mansfield Agitation Inventory.
Results: A repeated measures ANOVA (Mixed Procedure) with Bonferroni post hoc test revealed a significant reduction in agitation during the 30 minute presentation of individualized music (p<.0001) and the 30 minutes immediately following (p<.0001) compared to classical "relaxation" music relative to baseline.
Conclusions: The findings of this study support the use of individualized music as an effective intervention to reduce agitation in persons with ADRD. This intervention is relatively inexpensive and requires minimal time expenditure.
Impact: Currently there are an estimated 600,000 veterans diagnosed with severe ADRD. The majority of these individuals are cared for in long-term care facilities within the Veteran's Administration. Studies have shown that as many as 64% of persons with severe ADRD exhibit agitation. The prevalence and severity of agitation in this population, adversely effects the quality of care for these persons. Ineffective management of agitation also results in an increase in staff stress and provider cost. The use of individualized music as an alternative intervention for agitation is a potential cost-effective approach to the management of agitation in veteran's with dementia. This would thereby improve the quality of life for veterans and would also serve to reduce staff stress associated with their care.
*247. The Relationship Between Self-rated Health and Health Care Services Among Women Veterans in a Primary Care Clinic
Hayden B. Bosworth, Durham VAMC, Durham, NC, MI Butterfield, Durham VAMC, KM Stechuchak, Durham VAMC, LA Bastian, Durham VAMC
Objectives: Self-rated health (usually measured by asking respondents whether they would rate their health as excellent, very good, good, fair, or poor) has been found to be a predictor of mortality and functional ability even after controlling for a variety of physical health status indices. However, few studies have examined the relationship between self-rated health and health care service use, particularly, among women veterans. This study examined whether self-rated health predicted health service use among women in an equal access primary care clinic setting after adjusting for demographics, mental disorders and physical symptoms.
Methods: Women veterans (n=139), between the ages of 23-76 years were given the Primary Care Evaluation of Mental Disorders questionnaire (PRIME-MD) at their primary care outpatient clinic visit. Health service utilization was monitored for 12 months using the Decentralized Hospital Computer Program. Specific symptoms examined in this study included depressed mood, headaches, joint and back aches, gastrointestinal problems, and cardiac and pulmonary problems. These categories of symptoms were selected because they were the most commonly reported.
Results: The median age of the women was 41.7 years; 48.6% of the women were minorities. Women who reported their health as poor/fair had a median 18.5 outpatient visits to any VAMC outpatient clinic over 12 months as compared to 7 visits for women who reported having excellent/very good health. Women who reported their health as poor/fair reported more depressive symptoms, gastrointestinal problems, joint and back ache, headaches, and cardiac and pulmonary problems as compared to those women who reported their health was excellent/very good. In logistic regression models, women who reported their health as poor/fair were 3.25 times more likely to have a higher number (>12) outpatient visits over 12 months than women who reported their health as being excellent/very good adjusting for age, race, marital status, depressed mood, headaches, joint and back aches, gastrointestinal problems, and cardiac and pulmonary problems.
Conclusions: Women veterans who perceived their health as fair/poor were more than three times likely to have a higher number of outpatient visits than those who perceived their health as excellent/very good after considering demographic factors, mental disorders and physical symptoms.
Impact: Among women veterans using outpatient services, poor self-rated health may be an important predictor of outpatient health care utilization. Poor self-rated health may indicate psychosocial problems that clinicians may not readily detect. Poor health perception among women veterans may indicate greater health risks that warrants further follow-up.
*248. The Roles of the Family and Physician in Patient Decision Making about Serious Illness
Dorcas Mansell, MD, MPH, University of Alabama Birmingham and Birmingham VAMC, Birmingham, AL, R Poses, Memorial Hospital of Rhode Island and Brown University, Providence Rhode Island, C Duefield, CHQOER, Bedford VAMC, L Kazis, CHQOER, Bedford VAMC and Boston University School of Public Health
Objectives: Much effort has been made to promote patient autonomy in clinical decsisions. Most studies about patient decision making focus on the doctor and patient. However, many patients also involve their family in decision making. We examined the roles of demographic characteristics, comorbidities, quality of life, and prior experience with serious illness on the patient's desire to have family and physician involvement in decision making about serious illness; these variables have predicted patient desire for involvement in decision making in previous studies.
Methods: We performed a cross-sectional survey of VA outpatients; 255 of 591(43%) completed the survey. Independent variables included demographics and prior serious illness (in patient, spouse, family, friends) measured by self report, quality of life with the PCS and MCS from the SF-36, and case mix via chart abstracts. We asked patients a three part global question about the proportion of the responsibility (in %) that they wanted for themselves, their family, and physician for decisions about serious illness. The % input from family and patient were chosen as dependent variables in a MANCOVA (% physician input equals 100 - %family - %patient, and the correlation between desire for input from patient and input from physician was -0.84).
Results: Patients had a mean age of 63.2 and 4.0 +/- 2.0 chronic illnesses; 95% were male, 62% were married, 55% had a high school education or less, 99% were white. 235 completed the global question; on average patients wanted 39% of the input for decisions from their physician, 18% from their family, and 43% from themselves. The highest proportion, 11% (26/235), wanted roughly equal input from themselves, their physician, and family. MANCOVA showed that age, marital status, and prior serious illness in family were the only multivariate predictors in the overall model. Patients who desired more input from family were more likely to be married (p<0.05) and have a prior serious illness in family (p<0.05). Patients who desired more decision making for themselves were younger (p=0.0001).
Conclusions: Patients wanted to share decision making about serious illness with their physicians; many wanted some family involvement. Prior studies of patient decision making in less serious illness have found that patients wanted to be less involved in the decisions than their physician, but our study, which focused on decisions about serious illness, found that patients wanted involvement equal to that of the physician.
Impact: Clinical practice and future research on patient decision making in serious illnesses should consider not only the physician-patient relationship but also the roles of families in order to promote patient preferences.
*249. The Unreliability of Single-Item Scales: Implications for Health Services Research
Kimberly Wristers, Houston Center for Quality of Care and Utilization Studies: A VA HSR&D Center of Excellence; Baylor College of Medicine, Houston, Texas, L Rabeneck, Houston Center for Quality of Care and Utilization Studies: A VA HSR&D Center of Excellence; Baylor College of Medicine, KF Cook, VA Medical Center; Baylor College of Medicine, J Souchek, Houston Center for Quality of Care and Utilization Studies: A VA HSR&D Center of Excellence, TJ Menke, Houston Center for Quality of Care and Utilization Studies: A VA HSR&D Center of Excellence; Baylor College of Medicine, NP Wray, Houston Center for Quality of Care and Utilization Studies: A VA HSR&D Center of Excellence; Baylor College of Medicine
Objectives: Reliability of self-report instruments has been the subject of previous research, yet few investigators have examined the reliability of single-item self-report scales. The unreliability of a measure impacts the accuracy of the patients' scores and consequently the accuracy of outcome assessment based on the scores. Given the importance of self-report instruments and the role that reliability plays in the accuracy of such instruments, the present research investigated the impact of a typical single-item self-report scale's reliability on outcome assessment. The goals of this paper were twofold. First, we evaluated the reliability of a single-item scale. Second, after demonstrating its lack of reliability, we offered a theoretical explanation for the limitations of basing outcome assessment on single-item scales.
Methods: Our research focused on dyspepsia which is upper abdominal discomfort or pain thought to arise in the upper gut. We analyzed data from a double-blind, randomized clinical trial of 128 dyspepsia patients. We investigated reliability and its impact on outcome assessment using a single-item dyspepsia pain scale and the 6-item SODA (Severity of Dyspepsia Assessment) Pain Intensity Scale. The reliability of both scales was estimated within a factor analytic framework. The reliability estimates were then used to simulate data to evaluate the effect of reliability on outcome assessment.
Results: The reliability of the single-item pain scale was .48 and the multi-item SODA Pain Intensity Scale was .95. We then compared the simulated scores of one patient to a predefined cut-score, or a score that indicated the need for treatment. The comparison indicated that the probability of making incorrect treatment decisions was 30% greater with the single-item scale than with the multi-item scale.
Conclusions: Single-item scales are not reliable tools for health services researchers. The dyspepsia example and simulated data demonstrated how poor reliability of a single-item scale can cause significant errors in scores and how these errors lead to incorrect clinical judgements.
Impact: The increased recognition of the importance and need for patient-centered outcome measures in clinical research has not been accompanied by a proportionate recognition of the need for empirical testing of the measurement properties of these outcome measures. The development of clinically relevant instruments with sound measurement properties requires collaboration between health care researchers and psychometricians. By bringing the two disciplines together, we have the greatest chance of developing sound measures that will lead to accurate patient information and improved patient outcomes.
*250. The Use and Effects of Consumer Health Information Following a City-wide Informational Intervention
Todd Wagner, VA Palo Alto Health Care System, Menlo Park, CA, JH Hibbard, University of Oregon, MR Greenlick, Oregon Health Sciences University
Objectives: This study sought to determine whether the provision consumer health information had any effect on the use of self-care information and subsequent medical utilization.
Methods: The intervention was a social experiment in which self-care information, including reference books, a telephone advice line, and computerized information, were provided free of charge to the entire city of Boise. For the evaluation, random household surveys were conducted in Boise, ID (experimental site), Billings, MT and Eugene, OR (control sites). A total of 5,909 surveys were returned for a response rate of 42%. The dependent variables were self-reported use of self-care resources in the past few months and self-reported medical utilization in the past year. The community-wide intervention's effect on self-care resource and medical utilization was analyzed with difference-in-differences (DD) estimators.
Results: The intervention, which the consumer saw as a price decrease, was associated with statistically significant increases in the use of medical reference books, the telephone advice nurse and computers for health information. Results also show that Boise residents had a higher odds of accessing care and had a lower average number of doctor visits compared to residents in the control cities. These differences were not statistically significant for the average resident, but they were statistically significant for persons with diabetes. In addition, variation in the number of doctor visits decreased in Boise compared the control cities over time, suggesting a shift towards more appropriate utilization.
Conclusions: The overall results suggest that the intervention was associated with dramatic increases in the use of self-care information. Presence of a chronic illness, age, education and travel time to the physician were associated with using self-care information resources, although few variables were associated with using all three types of self-care information (books, telephone advice nurses and computers). These results provide direction for future informational interventions within the VA and they provide recommendations on how to tailor such interventions to the patient population. Although the intervention had a very small effect on overall self-reported utilization, larger reductions in utilization may be achieved by targeting chronically ill persons (e.g., persons with diabetes). Future research is needed to determine if such interventions improve appropriate utilization and decrease inappropriate utilization. Also, more economic research is appropriate as scale and scope economies will likely impact the cost of providing such services.
Impact: The use of self-care resources is price sensitive. The provision of self-care resources was not associated with appreciably better or worse changes in medical utilization. Decreases in utilization were detected in persons with diabetes.
*251. The Validity of the Clinical Signs and Symptoms Used to Identify Localized Chronic Wound Infection
Sue Ellen Gardner, Iowa City VAMC, Iowa City, IA
Objectives: The purpose of this study is to examine the validity of the clinical signs and symptoms used to identify localized chronic wound infection among a sample of subjects with chronic wounds. Research Questions 1. What is the sensitivity and specificity of each clinical sign and symptom used to identify localized chronic wound infection among a sample of infected and non-infected, non-arterial, chronic wounds? 2. What is the discriminatory power of each sign and symptom? 3. What is the positive predictive value for each sign and symptom?
Methods: A cross-sectional, descriptive design was used to examine the validity of 5 "classical" signs and symptoms (i.e., pain, erythema, edema, heat, and purulence) and 7 signs specific to secondary wounds (i.e., serous exudate with concurrent inflammation, delayed healing, discoloration of granulation tissue, friable granulation tissue, pocketing at the base of the wound, foul odor, and wound breakdown) in identifying chronic wound infection. A convenience sample of 36 chronic wounds was assessed for the presence of these signs and symptoms of infection. The wound sample was subsequently classified as infected or non-infected based on quantitative and qualitative cultures of viable wound tissue. Wounds with 100,000 or more organisms per gram of tissue or that contained beta-hemolytic streptococcus were defined as infected. A 2X2 contingency table was constructed for each sign and symptom that reflected the presence/absence of the sign or symptom and the "true" infection status of the wounds. The sensitivity (SENS), specificity (SPEC), discriminatory power (LR), and positive predictive value (PV+) of each sign and symptom were calculated based on the information in these tables.
Results: Of the 36 study wounds, 11 were infected and 25 were non-infected.Each sign or symptom had a sensitivity and specificity value greater than 0.00, except pocketing. However, the mean sensitivity for the classical signs and symptoms was 0.38, while the mean for the signs specific to secondary wounds was 0.62. Of all the signs and symptoms, only increasing pain, friable granulation tissue, foul odor, and wound breakdown had acceptable LR, and PV+ values.
Conclusions: The signs specific to chronic wounds appear to be better indicators of chronic wound infection than the classical signs and symptoms. Inclusion of theses signs in the routine monitoring of chronic wounds needs to be addressed and encouraged. To successfully integrate these signs into clinical practice, nurses responsible for wound monitoring must be made aware that their presence may be just as significant as the presence of classical signs and symptoms. Additional study is needed to identify patient and wound characteristics that alter expression of signs and symptoms of infection.
Impact: Less well-known signs and symptoms of infection appear to be better indicators of infection in chronic wounds than the "classical" signs of infection. Clinicians responsible for monioring chronic wound healing need to be aware of these signs and symptoms. Further research is needed in the area of chronic wound infection.
*252. Third-Party Insurance and Length of Acute Care Inpatient Stays At VA Hospitals
John A. Gardner, Center for Health Quality, Outcomes and Economic Research, Bedford, MA, Ann M. Hendricks, Center for Health Quality, Outcomes, and Economic Research
Objectives: To determine whether VA patients who have billable, third-party insurance have longer inpatient stays than other patients, correcting for primary diagnosis and patient’s age. Important because accuracy of forecasts of potential alternative revenues for the VA from third-party insurance is affected by differences in care for insured and uninsured patients.
Methods: Data: 1) billing and collection data for FY1998 downloaded by CHQOER staff from 24 VA hospitals that account for about 1/5 of all acute care discharges; 2) all admissions during FY1998 (except to observation bedsections) in acute care inpatient data from the VA National Patient Care Database maintained at the Austin Automation Center,. We considered patients to have third-party insurance if the VA billed their insurer for either inpatient or outpatient services performed during FY1998. Hospitals were chosen to provide a broad geographic representation and a range of past success in collections on bills to third-party insurers. The billing data allow for at least a three month lag for the submission of a bill. We calculated length of stay (LOS) from administrative data through the end of FY1999, with only 34 admissions out of more than 126,000 censored. We compared LOS within groupings defined by a combination of age (eight ranges, including three for those 65 and older) and diagnosis (3-digit ICD-9-CM codes).
Results: Of 2,534 age/diagnosis groupings in our data, 111 groupings (with 47% of all inpatient stays) had enough insured patients (30) to allow calculation of a reliable LOS for insured patient stays. Of the 111 groupings, only 24 had significantly different (at the 0.10 level) LOS for insured and non-insured patients. Seven of those were in diagnoses involving psychological disorders or substance abuse: senile/presenile organic psychoses (ages 75-84), affective psychoses (45-54,75-84), alcohol dependence syndrome (65-74), drug dependence (35-44, 45-54), adjustment reaction (45-54). For 22 of the 24 groupings, insured patients had shorter average LOS than uninsured patients: 2 to 4 days out of an average LOS near 10 days. The smallest significant difference in LOS was 4.8 versus 4.0 days for other acute ischemic heart disease (45-54); the largest was 51.9 days versus 32.1 days for other paralytic syndrome (65-74). In the two cases with significantly longer LOS for insured patients, the differences were 7 and 9 days out of the nearly 20 day average for insured patients: diabetes mellitus (75-84) and other cellulitis and abscess (65-74).
Conclusions: For most combinations of diagnosis and age group, we found no systematic differences in inpatient acute care length of stay for patients with third-party insurance. Where we did find significant differences, they consistently show shorter average stays for patients with third-party insurance.
Impact: The current VA forecasting practice of assuming that stays for insured and uninsured patients are the same average length leads to a slight overestimate of potential alternative revenues.
*253. Timing of Alcohol and Smoking Cessation Study: Screenees’ Attitudes Toward Participating in a Clinical Trial
Anne M. Joseph, MD, MPH, Minneapolis VAMC, CCDOR, University of MN, Minneapolis, MN, ML Willenbring, Minneapolis VAMC, CCDOR, University of MN, Ben Lexau, Minneapolis VAMC, CCDOR, University of MN, Herb Stockley, Minneapolis VAMC, CCDOR, Sean Nugent, Minneapolis VAMC, CCDOR
Objectives: The Timing of Alcohol and Smoking Cessation Study (TASC) is a multicenter, randomized controlled trial of the effectiveness of smoking intervention delivered concurrently with intensive substance abuse treatment vs. smoking intervention that is delayed 6 months. It will compare 1) short and long-term smoking cessation rates and 2) substance abuse treatment outcomes.
Methods: Subjects are not required to quit smoking. To date 1270 patients have been screened from three substance use treatment programs; 308 subjects have been randomized. The majority of subjects are dependent on multiple substances. This report describes attitudes of smokers toward addressing nicotine dependence and participating in a research study.
Results: Forty-six percent scored >7 on a contemplation ladder (range 0-10). Fifty-four percent agreed or strongly agreed that they were "too concerned about alcohol" to quit smoking, and 50% agreed or strongly agreed that they were "too stressed by alcohol treatment" to quit smoking. The majority (67%) disagreed or strongly disagreed with the statement, "I don’t see smoking as a big problem right now." A small proportion agreed or strongly agreed that "quitting smoking will make me drink," (12%), or that "I don’t think I will ever quit smoking" (14%). Seventy-one percent (n=886) reported that they would consider participation in a research protocol. Of those who were not interested in participation (n=371), 60% did not like the idea of being in a research study, 53% thought that there would be too much pressure to quit smoking, and 36% were concerned about giving out contact information.
Conclusions: These data suggest that patients in intensive substance abuse treatment have considerable interest in smoking cessation, but that recruitment to clinical trials may be adversely affected by two factors: perceived pressure to quit smoking and the need to provide contact information.
Impact: These results have implications for inclusion of nicotine dependence treatment in intensive alcohol interventions. They also describe barriers that patients in alcohol treatment perceive to participation in research studies that could be addressed with alternative study designs.
*254. Toward Gender-Aware VA Health Care: Year 2 Findings
Dawne S. Vogt, PhD, National Center for Posttraumatic Stress Disorder, Brookline, MA, VW Savarese, National Center for Posttraumatic Stress Disorder, LA King, National Center for Posttraumatic Stress Disorder, J Wolfe, National Center for Posttraumatic Stress Disorder
Objectives: Despite VHA efforts to better accommodate the health care needs of a growing number of women veterans, an historical orientation towards treating male patients may have resulted in a less than optimal care environment for female patients. Our guiding hypothesis is that this environment is influenced by three overlapping personal characteristics of VHA employees: (1) the tendency to work without gender stereotypes (ideology), (2) sensitivity to female patients’ special needs, and (3) knowledge of women veterans and their health care. Together these comprise what we have termed gender awareness. For this three-year project, we seek to develop a reliable and valid self-report measure of gender awareness for use with all VHA personnel. The project has four objectives: (1) to refine the definitions of the gender awareness components and create the instrument, (2) to establish the instrument’s psychometric properties, (3) to obtain normative data, and (4) to make the instrument available for large-scale use.
Methods: The design is observational and cross-sectional, with multiple waves of data for reliability and validity analyses. Participants are VHA employees selected using proportional stratified random sampling from the VA New England Healthcare System. In the first two years, 891 employees received the instrument, and 488 (55%) were completed. Analyses for both years included computation of item-total correlations, coefficients of internal consistency reliability (alpha), interscale correlations, and correlations with other measures.
Results: The subscales of the Gender Awareness (GA) instrument have been refined, resulting in a Cronbach’s alpha = .85 for both the 20-item ideology subscale and 29-item sensitivity subscale (N = 114). Providing evidence that the subscales tapped distinct facets of GA, we found that ideology and sensitivity were moderately correlated (r = .36, p < .001), as were ideology and knowledge (r = .41, p < .001), and that sensitivity and knowledge were not correlated (r = .05, p = .62). The GA instrument demonstrated sufficient convergent and discriminant validity: as expected, the three subscales were moderately related to measures of attitudes towards women and sex-role egalitarianism (p < .05), but were unrelated to a measure of social desirability.
Conclusions: The project is ongoing. The item set has been refined and finalized, and convergent and discriminant validity have been tested. Year 3 is well under way, as we are currently collecting normative data on the scales from medical centers across VISN 1.
Impact: The GA assessment instrument will provide a means to identify and measure factors that contribute to the quality of VA healthcare for women veterans, and will provide a mechanism to pinpoint training needs for the education and remediation of employees. Moreover, the instrument will afford a means to monitor organizational improvements in gender awareness over time. Our long-term goal is to have a product that has justifiable transportability on a large scale. The GA assessment package can fit nicely within a program to monitor patient satisfaction among the growing number of women veterans, as that satisfaction relates to features of organizational climate.
HSR&D Funded: GEN 97-014.
*255. Uncovering the Stress Response to Bathing in Veterans with Dementia.
Ann C. Hurley, RN, DNSc, FAAN, Bedford VAMC; Northeastern University, Boston, MA, Bedford, MA, L Volicer, Bedford VAMC; Boston University, Boston, MA, EK Mahoney, Bedford VAMC; Boston College, Chestnut Hill, MA
Objectives: Uncomfortable behavioral symptoms during bathing, an activity of daily living, are common for persons who suffer from dementia. These behaviors are linked theoretically to a stress response; thus, reducing stress associated with bathing should alleviate or minimize resistiveness to care and discomfort.
Methods: This pilot work used the Antecedent Behavior Consequence model of behavior analysis and an interrupted time series design with multiple switching replications to examine the effect of 4 protocols during bathing using the techniques of the gentle towel bath: 1) relaxation, 2) reminiscence, 3) relaxation and reminiscence, and 4) control. Thirteen veterans identified by nursing staff as resistive to care during bathing participated during 208 baths (16/veteran), randomly ordered and videotaped for rating by 2 blinded observers. Independent variables were nurse and protocol. Noninvasive physiological indicators to approximate the stress response were continuous heart rate monitoring by telemetry and salivary cortisol levels. Behavioral outcomes were observed resistiveness to care (8 item tool, alpha = .71) and discomfort(9 item tool, alpha = .68).
Results: All veterans had severe dementia evidenced by cognitive scores and were nearly mute, (mean MMSE = .9 + 1.2 SD, mean Speech Assessment = 7.6 + 3.8). Using the area under the curve for percent heart rate change from baseline through the bath, relaxation alone showed the least stress while the control protocol showed the most stress response. We were able to obtain 678 of the desired 832 saliva samples (4/bath), at baseline, immediately after bath, after dressing, and one hour after bath. Cortisol levels were similar in all protocols. Behavioral outcomes (effect size by eta squared) comparing intervention protocols with the control condition for resistiveness and discomforts respectively were: relaxation .01, .20; reminiscence .38, .48, both relaxation and reminiscence .13, .24.
Conclusions: The repeated measures analysis of variance for the sample as a whole did not show a statistically significant main effect of experimental protocol on either physiological or behavioral outcomes. We were specifically funded to do pilot work and did not have a sample size large enough to detect the small to moderate, but clinically meaningful differences observed. We also used the gentle towel bath given by expert nurses with advanced degree preparation as interventionists, a protocol which in itself is a low stress method of bathing. Single-subject analysis demonstrated significant within-subject variation in behavior and heart rate response to the four protocols. Individual patterns emerged with differential responses.
Impact: This study supports the validity of the single-subject design for identifying persons who benefit from specific nursing interventions and will facilitate the conduct of further research using these methods. Future research is needed to uncover brain-behavior relationships as a basis for designing interventions and to test protocols with a single family caregiver providing home based care for a veteran.
HSR&D Funded: NRI-96-023.1.
*256. UPBEAT: Utilization, Cost and Consumer Satisfaction
Maria D. Llorente MD, Miami VAMedical Center; GRECC; University of Miami Department of Psychiatry, Geriatric Division, Miami, FL, EJ Olsen, Miami VA Medical Center; GRECC; University of Miami Department of Psychiatry, Geriatric Division, Miami, FL, E Antonio, Miami VA Medical Center; GRECC; University of Miami Department of Psychiatry, Geriatric Division, Miami, FL; D Duffy, Miami VAMC; GRECC, A MONTALVO, Miami VAMC; GRECC, LF SAMOS, Miami VAMC; GRECC; Univ of Miami, Department of Medicine, Division of Endocrinology
Objectives: 1.To determine whether early identification and treatment of psychiatric disorders in medically ill older veterans can prevent unnecessary medical utilization and reduce medical costs. 2.To examine the effectiveness of clinical care coordination models on veteran satisfaction.
Methods: Target population: Veterans aged 60 or older, admitted to acute medical/surgical services and meet symptom criteria for anxiety, depression or substance abuse.
Study Procedure: Baseline structured symptom and functional assessments are administered with randomization by patient to either active intervention (UPBEAT) or usual care control). The UPBEAT model includes a structured and clinical psychiatric interview, a medical evaluation by a geriatrician and the development of a multidisciplinary treatmen plan that addresses medical and psychosocial problems. Each patient in this arm is assigned a clinical care coordinator with whom he/she will have monthly contact at a minimum. Prospective follow-up assessments are conducted at 6, 12,and 24 months post-enrollment. Outcome measures: Patient physical/mental functioning, satisfaction (when compared to national VA data). Medical costs are based on acute inpatient lengths of stay multiplied by the VA national weighted average cost per day for FY 1998 of $1048. Statistical analyses were conducted were appropriate using chi-square with statistical significance set at p<0.05.
Results: Between June 1995 and September 1998, 262 veterans were enrolled in the program in Miami: 133 in UPBEAT and 129 controls. There were no differences between the groups in demographic features and severity of psychiatric symptoms at baseline. While both the UPBEAT and the control patients hadsignificant improvement in baseline scores of psychological symptoms, there were no between-group differences. 80% of UPBEAT patients reported complete satisfaction with care versus 45% of nationally surveyed veterans. Quality of care was rated as excellent by 51% of UPBEAT patients versus 30% of nationally surveyed veterans. During the first 12 months of the program, on average, Miami UPBEAT patients were hospitalized post-enrollment four fewer days than controls. This resulted in inpatient cost savings of $4,192 per patient.
Conclusions: 1.UPBEAT successfully reduced patient symptoms of anxiety, depression and substance abuse. 2.UPBEAT patients demonstrated higher levels of satisfaction with care and quality of care than nationallly surveyed veterans. 3. UPBEAT was able to reduce medical services costs by a reduction in inpatient lengths of stay. 4. Veteran consumer levels of satisfaction were very high in the presence of cost savings.
Impact: Had all 262 patients that agreed to participate in the program been enrolled in UPBEAT active intervention, the total potential cost savings would have approached 1.1 million dollars. Conservatively, it has been estimated that 10% of primary care patients meet symptom criteria for anxiety, depression or substance abuse. At the Miami VAMC, it is estimated that almost 1,000 patients could be eligible for this program. Generalizing the above results, the UPBEAT program has the potential to reduce inpatient medical costs at the Miami VAMC by 4.2 million dollars in a 12-month period.
*257. Use of Clustering and Weighting Techniques in a Complex Survey of HIV Testing in Homeless Women
Martin L. Lee, PhD, VA HSR&D Center of Excellence; UCLA School of Public Health, Sepulveda, CA, B Herndon, VA HSR&D Center of Excellence; VA Greater LA Healthcare System, L Gelberg, UCLA Department of Family Medicine, G Sumner, UCLA School of Public Policy and Social Research
Objectives: In a study to determine, among other issues, the predictors of HIV testing in homeless women in Los Angeles County, a complex probability survey was designed of 974 subjects. This involved the selection of participants from various shelters and meal programs and also raised the question of potential bias arising from visit-based sampling. The aim of the analysis was to develop a statistical model that could utilize a set of demographic and other subject characteristics to predict the probability of HIV testing, but taking into account the nature of the sample.
Methods: Bivariate analyses of various patient characteristics were initially utilized to reduce the number of potential predictor variables. Additionally, a forward stepwise algorithm was used to develop a parsimonious logistic regression model that incorporated the use of a cross-validation technique based on repeated random sampling of the data. The clustering of subjects within site was handled by the use of a Huber adjustment model within the context of the logistic model, while the issue of frequent users of the facilities was dealt with using covariate adjustment within the model.
Results: Initial significant predictors of HIV testing included a history of drug dependence, depression, having a regular source of care, pregnancy, multiple partners and a history of pap or TB testing. The adjustment for the potential clustering in the data was important in that the significance of many of the potential predictors was changed when this factor was taken into account. However, the frequency of attendance at the facilities did not appear to be a significant covariate in the analysis. An alternative model based on the survey scheme and a weighting adjustment for subject utilization of the facilities resulted in similar findings.
Conclusions: In the particular population under consideration here, the development of a simple logistic model for the prediction of the probability of HIV testing under a complex survey scheme was possible using straightforward statistical techniques of cluster and covariate adjustment for frequent users of the sites where the sampling took place.
Impact: The need for more involved techniques using specialized statistical programs may be obviated by thoughtful consideration of more available approaches.
*258. Using Teleform® Software Application to Design and Conduct Automated Data Entry of Self-administered Questionnaires for Health Services Research in the VA
Britta I. Neugaard, James A. Haley Veterans' Hospital, Tampa, Florida, SL Luther, James A. Haley Veterans' Hospital, GM Powell-Cope, James A. Haley Veterans' Hospital
Objectives: To describe our experience with using Teleform® software to design, distribute and evaluate both simple and complex self-administered questionnaires. To provide insights and recommendations concerning the use of Teleform® to other VA researchers.
Methods: The use of self-administered survey questionnaires is an essential strategy for health services research. One important component of successful research involving self-administered questionnaires is the design and production of clear, professional looking forms. The best forms are designed in a way that facilitates data coding and analysis. Teleform® is an inexpensive and easy to use questionnaire design software product that allows for direct scanning of questionnaires into a variety of database formats. While Teleform® was primarily designed for use in market research, we have adopted it for health services research projects including both short (two pages) and long (35 pages) questionnaires.
Results: Our research team has identified the following advantages to using Teleform®: survey questionnaires can be professionally designed while a database is simulatenously created, data is automatically entered into a database saving time in manual data entry, and the forms can be easily modified to use as a data collection tool for other research projects. Problems we have experienced with the software include a weak interface between the scanner and the computer, delays in receiving technical support from the manufacturer, and the inability to make changes in the form once it has been activated.
Conclusions: Teleform® can be used to create a professional looking questionnaire that can be easily scanned into a data set ready for statistical analysis. The learning curve to use Teleform® is modest and can be accelerated by training sessions provided by the software vendor. Our experience suggests that there are fewer technical problems with shorter forms. We believe Teleform® is efficient to operate in research departments with a high volume of self-administered surveys, but not in those with a low volume.
Impact: With a relatively modest investment VA health services researchers can establish the capability to design professional looking self-administered questionnaires, the results of which can be transformed into a data format for analysis.
*259. Utilization of VISTA Databases to Assess Outcomes of Care Provided Patients with Diabetes
John Douglas Burke, Department of Veterans Affairs Medical Center, Louisville, KY
Objectives: To determine if the data stored in VISTA databases can be used to assess the impact of VA guidelines for the treatment of Type 2 diabetes.
Methods: Data were extracted from patient files in VISTA using coded reports, locally written programmed reports, and V21.0 File Man routines. The data extracted were sent to a spooler, downloaded from the created VISTA files and imported into Excel 2000 spreadsheets and Access 2000 databases. Data were organized into a useable format using Excel and Access. Relational databases were used to combine data from separate databases and spreadsheets. Data from Fiscal Years 1995 through 1999 were compared. Diabetic patients were identified by comparing the list of diabetic patients in the FY 1999 patients database with diabetic patients in the previous four fiscal years using relational databases in Access 2000. This method identified 806 diabetic patients with at least one Hemoglobin A1c (HbA1c) in each year who continuously received treatment from the VA medical center in Louisville.
Results: HbA1c data were obtained on the same 806 diabetic patients treated each year over a five-year period. The average HbA1c for this population was 8.31% in FY 1995, 7.54% in FY 96, 7.35% in FY 97, 6.81% in FY 98 and 7.12% in FY 99. HbA1c labs were obtained in 80.1% of the total diabetic population in FY 99; HbA1c < 8%: 74.5%, HbA1c > 9.5%: 8.8%. Blood pressure data were available for three fiscal years. Those average measures were 146/77 in FY 97, 145/76 in FY 98, and 145/75 in FY 99. Lipid data review in FY 99 demonstrated the following: Cholesterol measured in the diabetic population: 77.3%, measured under 200: 66.3%, LDL < 130: 75.2%, LDL < 100: 40.9%.
Conclusions: This measure assessed the care provided 806 diabetic patients over five years as judged by their average HbA1c. It also assessed the percentage of diabetic patients who have had at least one HbA1c test in the past year. An HbA1c less than 8% is considered to be good control of a patient's diabetes. HbA1c testing is fundamental to assessing the underlying control of the disease. Measurement of HbA1c quantifies glucose control over the previous three to four months and is the preferable measure of long-term glycemic control. In this patient population, the HbA1c measures fell consistently from FY 95 through FY 98. The average rose approximately 4.6% in FY 99 from the previous fiscal year's low. Measures of lipid and blood pressure control were obtained for part of the review period. Blood pressure and lipid measures do not indicate improvement and offer opportunities to use the data obtained to improve patient care in these important areas.
Impact: Data from VISTA databases can be used to assess care provided in a diabetic patient population. One cannot say conclusively that care provided the VA diabetic was influenced by the VA diabetes guidelines but, if one uses the average HbA1c for this population, then the care for these patients appears to have improved over time.
*260. VA Chicago Health Care System Integration Evaluation
Marylou Guihan, PhD, MCHSPR, Hines, IL, JD Thomas, MCHSPR, S LaVela, MCHSPR, DM Hynes, MCHSPR, LM Manheim, MCHSPR and Northwestern University
Objectives: In June, 1996, VA approved the Integration of Lakeside and West Side VA Hospitals into VA Chicago Health Care System (VACHCS), as part of a larger restructuring plan for VHA. Among the 14 other VA integrations that were approved or in planning by this time, this integration was unique because they were two urban, tertiary care hospitals, and a common patient population. Study objectives included identifying and describing key players and organizational structures and comparing VA and private sector academic medical centers (AMCs) in terms of resources, facilitators and obstacles to integrating.
Methods: Descriptive analyses using qualitative data collected during in-person, semi-structured interviews were conducted with managers and service chiefs at VACHCS, and two private sector AMCs and their medical school affiliates.
Results: Respondents perceived that VA leadership never clearly identified or communicated the rationale for and the goals of the Integration. Respondents unanimously and consistently indicated that the major obstacle to this Integration was poor communication, for which VA leadership was held accountable. Respondents perceived that the goals were not explicit enough, not prioritized and not action oriented. A general perception held by participants was that, despite official assurances to the contrary, cost savings, not patient care, were the major motivation of the Integration. Most respondents described academic affiliations as being extremely important to VA and expressed concern that the Integration would adversely affect the relationship with the affiliates.
Conclusions: VA leadership should anticipate resistance and fear as likely first reactions to integration. Therefore, communication with and among integration participants should be early and often. Development and deployment of methods for communicating information about the integration across management, staff and divisions is key. The sooner new leadership is appointed the better. Careful consideration should given to whether the new director should come from within or outside VA. Academic affiliates should be included in early discussions regarding integrating VA facilities. Our results suggest that experience in previous integrations or specific training in organizational change and consolidation helped staff create and implement more detailed integration plans. Even before announcing a merger, public relations personnel should be deployed. VA should consider exploring a number of financial and non-financial incentives to help motivate and reward those individuals actively pursuing integration. Major facility integration committees should be organized and convened as soon as possible after the integration announcement. We will discuss what the composition of such committees should include.
Impact: With the ongoing changes in VA, as outlined in "Vision for Change," our results provide valuable information about the structures and processes of merging two urban, tertiary care facilities. Since integrating similar VA facilities is anticipated for the future, identifying and understanding the facilitators and obstacles, and describing the lessons learned in this context will benefit the strategic planning of future complex integrations.
*261. VA Cooperative Study #430: "Reducing the Efficacy-Effectiveness Gap in Bipolar Disorder"
Erinn E. Dawson, Providence VA Medical Center, Providence, RI, WO Williford, Perry Point Cooperative Studies Coordinating Center, MS Bauer, Providence VA Medical Center; Brown University
Objectives: The effectiveness of efficacious treatment for bipolar disorder in general clinical practice is greatly attenuated, resulting in what has been called an "efficacy-effectiveness gap". The VA Cooperative Study has funded a 12-site randomized controlled trial of an easy-access program for bipolar disorder from 1997-2003 that is an intervention designed to reduce this gap. The Bipolar Disorders Program (BDP) is a clinic-based treatment delivery system that emphasizes (1) algorithm-driven somatotherapy, (2) standardized patient education, and (3) easy access to a single primary mental health care provider to maximize continuity of care. This study proposes that increased access and provider and patient education will reduce the "efficacy-effectiveness gap" for bipolar disorder. We specifically hypothesize that compared to usual VA care (UVAC) the easy-access BDP, will improve clinical, functional, and economic outcome.
Methods: Patients who are admitted to a acute psychiatric ward with a primary or co-primary admission diagnosis of bipolar disorder type I or II are screened for the study. The study randomly assigns patients with bipolar disorder that are currently hospitalized in a VA acute psychiatric unit for a major affective episode to one of the two treatment arms (BDP and UVAC). Randomized patients can have any comorbid diagnosis and must have had at least two hospitalizations on psychiatric wards in the past five years. Additional exclusion criteria are minimal. One hundred and ninety-one patients are being randomized and all enrolled patients are followed up for three years using a bimonthly follow-up schedule.
Results: Randomization began on January 1, 1998. All 12 sites have been trained to criterion and are fully functioning. As of October 31, 1999, 2,523 patients have been screened and 227 patients have been randomized (113 BDP and 112 UVAC). Systematic and regular monitoring of the 12 participating sites has ensured adherence to protocol and maintained a high level of fidelity to the experimental treatment. Characteristics of the sample are consistent with its design as an effectiveness trial.
Conclusions: CSP #430 provides an example of a long-term, multi-site randomized controlled trial of a multi-modal intervention that was developed form the outset as an effectiveness, rather that efficacy, trial. Current status of CSP #430 indicates that it is on track in recruiting and retaining appropriate subjects and collecting required data. Coupled with ongoing measures developed to maintain internal validity, this provides confidence that the study will be successful in addressing its major goals.
Impact: The findings of CSP #430 will be of relevance both to the VHA and to other healthcare sectors. It is designed to evaluate the basic principle that augmenting ambulatory access for major mental illness will improve outcome and reduce overall treatment costs. If results are positive, this study will provide reason to reconsider the prevailing trend toward limitation of ambulatory services that is characteristic of many managed care systems today. Further, the study will provide specific direction with regard to how to structure such ambulatory services.
HSR&D Funded: CSP #430; DEV-97-015
*262. VA Hospital-Hotel (Hoptel) Equalizes Inpatient Length of Stay Among Homeless and Domiciled Patients
James F. McGuire, LCSW, PhD, VA Greater Los Angeles Healthcare System, Los Angeles, CA, AS Mares, VA Greater Los Angeles Healthcare System
Objectives: Homelessness has been associated with longer inpatient stays and higher medical costs among veteran patients. Nationwide in 1996, homeless veteran inpatients were hospitalized 68-75% longer than domiciled inpatients. One large VA medical center with a large homeless population developed a hospital-hotel (hoptel) in 1996 to reduce inpatient lengths of stay among homeless patients. The hoptel is a 30-bed facility located on the grounds of the hospital and operated by the Salvation Army. The hoptel provides short-term recuperation and discharge planning/housing placement services at a daily rate of $22. The purpose of this study was to evaluate the effectiveness of the hoptel in equalizing inpatient stays among domiciled and homeless patients. We hypothesized that the presence of a hoptel would equalize inpatient lengths of stay (adjusted for severity of illness and demographic differences) between homeless and domiciled patients by allowing homeless patients to be discharged from inpatients wards/units when medically stable.
Methods: In this cross-sectional study of patients at a large urban VA medical center, inpatient lengths of stay among domiciled (N=6,586) and homeless hoptel users (N=441) were compared. Racial/ethnic, income, substance abuse, and severity of illness differences between domiciled and homeless hoptel groups were adjusted for using analysis of covariance (ANCOVA) statistical analyses. Risk-adjusted mean lengths of inpatient stays were compared for these two groups during the first three years of hoptel operation (1996-98). Severity of illness adjustments were made by combining major diagnosis-related group (DRG) and principal diagnosis into a single variable, and then stratifying by DRG sub-groups containing two or more domiciled and homeless hoptel subjects. Demographic, diagnosis, and length of stay data were obtained from medical center computer files (Patient File and Patient Treatment File) for all patients discharged from an inpatient ward/unit during the three year study period.
Results: No significant differences were found in risk-adjusted mean inpatient lengths of stay between domiciled and homeless hoptel patient groups. Overall, homeless hoptel patients stayed 0.7 days longer than domiciled patients (p=.50). Homeless patients stayed an average 0.3 (p=.63), 0.4 (p=.84), 5.2 (p=.67), and 3.3 (p=.29) days longer than domiciled patients for general medicine, psychiatric, oncology, and surgical hospitalizations, respectively.
Conclusions: No significant differences in risk-adjusted inpatient lengths of stay were observed during the first three years of hoptel operation. These data support the hypothesis that risk-adjusted inpatient lengths of stay among domiciled and homeless patients would be equalized in the presence of a hospital-hotel (hoptel).
Impact: VA medical centers and healthcare systems serving large numbers of homeless patients ought to consider developing hospital hotels (hoptels) as one way of equalizing inpatient lengths of stay among domiciled and homeless patients.
*263. Visions of Integration: Multiple Approaches to Structuring VISNs
Victoria A. Parker, DBA, CHQOER, MDRC, Bedford, MA, LG Pucci, MDRC, MP Charns, MDRC
Objectives: To describe the extent to which network-level service lines or other strategies were adopted by VISN leaders to achieve integration of services within their networks.
Methods: At least one site visit to each of 22 VISNs (regional networks) was conducted as part of the 3-year service line evaluation project. Key informants from the network offices and selected facilities were interviewed, using a semi-structured interview guide. Informants were selected to maximize the number of different perspectives represented within each VISN. Respondents were asked to describe major changes in their networks during the previous year, focusing on service line development and other approaches to integrating services across multiple sites. Interviews were conducted by teams of two researchers; one managed the interview while the other entered responses into an electronic database. The database facilitated the sorting of responses by topic, such as VISN, respondent role or other dimension of interest during the analysis process. Interviews from each VISN were read and analyzed together in order to develop a composite assessment of the change strategy underway in that VISN. Finally, integration strategies were identified for each VISN. During the process, categories emerged as we learned of the varied ways in which the VISNs formulated and implemented their missions.
Results: Three strategies were identified: 1) VISN sub- regions. These were health care systems/services that crossed 2-3 facilities, or alliances, which brought facilities together in line with geographic proximity. 2) VISN-level service line divisions, where specific bundles of services (such as mental health or primary care) are managed at the network level. In this case facility-based service line managers report to their network-level counterparts; 3) VISN-level task forces or councils that have no operational authority, but are charged with developing clinical improvements and/or common clinical practices in specific service areas (e.g. mental health or extended care). In this case there is no change in reporting relationships. These approaches were not necessarily mutually exclusive, and appeared in many cases to be evolutionary in nature, changing as network leadership, facility leadership or perceived barriers to innovation changed. There were often striking differences from one respondent to the next as to their understanding and reactions to the respective strategies within their VISN. Rationales cited for carrying out particular plans also helped in interpreting different strategies. For example, respondents in VISNs with very strong network-level service lines were more likely to report that the service line structure was closely linked to strategies for improving performance measures and other indices of quality.
Conclusions: Little consensus on the best strategies for bringing about operational integration within VISNs appears to exist. Instead, VISNs have pursued strategies that appear to be related to differences in: leadership approach, geography, stakeholder interests, and leadership personnel.
Impact: VISN leaders looking for organizational change models aimed at integration must examine underlying management philosophy and strategy as well as the various change models, to determine what will "work" best in their VISN.
*264. Voice Recognition Dictation System reduces Report Turnaround Time
Lorraine M. Fig, MD, MPH, VA Nuclear Medicine Program Office, Ann Arbor, MI, MD Gross, VA Nuclear Medicine Program Office, RA duFour, VA Nuclear Medicine Program Office
Objectives: Prompt delivery of accurately transcribed Nuclear Medicine (NM) reports is the goal of every NM department. Computer systems that convert dictation directly into text are becoming more widely available. The purpose of this study was to determine the suitability of a commercial voice recognition dictation system (VRDS) for implementation in a VA teaching NM department and to evaluate its effect on efficiency, as measured by report turnaround time.
Methods: 1) The accuracy of the VRDS (MedSpeak, IBM) was evaluated in a simulated clinical setting: 13 residents (8M, 5F) and 3 staff NM physicians without prior exposure to VRDS dictated standardized reports which contained typical wording for a variety of representative nuclear medicine studies, both complex and simple. Dictations were graded for accuracy and expressed as % words incorrectly transcribed. 2) Turnaround time for clinical report completion was assessed for comparable periods before and after implementation of the VRDS.
Results: Overall, mean inaccuracy of 10 subjects with standard US accents was 5.9% and for 5 with non-US accents 11.25%. First day individual training took a mean of 47 min (without personal voice enrollment) plus 60 min for enrollment; additional support was required for 2-3 days to answer specific questions (total estimated time 30 min/resident/day). Report turnaround time was reduced by 51.2% after introduction of VRDS.
Conclusions: VRDS can be successfully integrated into a VA academic multiuser NM setting and significantly reduces report turnaround time. In the implementation phase, serious consideration must be given to the time/personnel effort involved in adequate training and supervision of rotating residents. VRDS represents an opportunity to improve efficiency of services, increase productivity and realize cost savings and is potentially applicable to all clinical diagnostic disciplines.
Impact: A voice recognition dictation system has the following impacts upon clincal services: 1. Quality of services is enhanced due to decreased report turnaround time which allows more timely access of clinical report by referring physicians. 2. Cost savings/cost avoidance occur due to increased efficiency of existing staff and reduction in transcriptionist FTEE. 3. Report accuracy is improved because potential transcription errors are eliminated.
*265. Health-related Quality of Life after Major Abdominal Surgery in Elders
Valerie A. Lawrence, VERDICT Center of Excellence (Health Services Research and Development, Veterans Affairs), San Antonio, Texas, HP Hazuda, University of Texas Health Science Center at San Antonio, CD Mulrow, VERDICT Center of Excellence (Health Services Research and Development, Veterans Affairs)
Objectives: Despite its importance to patients, little is known of the course of functional recovery and health-related quality of life in elders having major abdominal operations. Most functional recovery research has focused on coronary artery bypass operations, hip or knee replacement, and cataract operations.
Methods: In a prospective cohort study of elders undergoing elective major abdominal surgery (ABD), patients 60 years old had assessments of health-related quality of life with the Medical Outcomes Study Short Form-36 (MOS SF-36) preoperatively (preop) and postoperatively (postop) at 3 weeks, 6 weeks, and 3 and 6 months. The MOS SF-36 covers 8 domains of function (detailed in Results below) with higher scores indicating better function. The individual domains can also be summarized into two summary domains (Physical Component and Mental Component).
Results: Physical Functioning and Role Physical scores both declined significantly 3 weeks postoperatively (21 +/- 24 and 25 +/-42, p  0.0001 for both), then recovered to preop baseline by 3 months, and improved statistically signficantly beyond baseline by 6 months. Bodily pain scores worsened modestly at 3 weeks postop (3.1, +/- 34, p  0.08), returned to baseline at 6 weeks, then improved significantly beyond preop baseline at 3 and 6 months. The trajectories of Vitality, Social Functioning echoed each other: significant declines at 3 weeks postop (6.4 +/- 26, 18 +/- 38, p  0.0001 for both), returned to baseline at 6 weeks, and improved significantly beyond preop baseline at 3 months, and 6 months. Role Emotional scores declined nonsignificantly at 3 weeks postop, then also returned to and significantly improved beyond baseline at 6 weeks, 3 and 6 months. General Health and Mental Health improved beyond baseline at all followups. The Physical Component summary score fell by 6.5 +/- 10.0 at 3 weeks postop (p  0.0001), returned to preop baseline by 3 months and was significantly better than the preop baseline at 6 months. The Mental Component Summary.improved beyond baseline at all postop assessments.
Conclusions: In this large, sentinel cohort study of elders undergoing major abdominal surgery, dimensions of HRQOL varied substantially in the early postop course. The largest decrements early postop occurred in physical, role physical, and social functioning dimensions, wiith smaller decrements in vitality and bodily pain. General health and mental health showed no decrements at all. Summary scores showed a small decrement for the Physical Component and no decline for the Mental Coomponent. In the later postoperative period, all dimensions and summary scores were consistent in significant improvement beyond preop status. There are at least two possible explanations for this finding. Patients’ preoperative symptoms related to the need for surgery may affect HRQOL. Alternatively the anticipation of surgery may reduce HRQOL preoperatively, with the postoperative rebound actually reflecting a return to usual HRQOL.
*266. Medical Complications after Hip Fracture Repair: Incidence and Outcomes in 8930 Patients
Valerie A. Lawrence, VERDICT Center of Excellence (Health Services Research and Development, Veterans Affairs) and University of Texas Health Science Center, San Antonio, Texas, H Noveck, University of Medicine and Dentistry of New Jersey, A Duff, University of Medicine and Dentistry of New Jersey, J Carson, University of Medicine and Dentistry of New Jersey
Objectives: Hip Fracture typically occurs in elders, usually with multiple comorbidities. Little evidence on the spectrum, incidence and relative clinical impact of postoperative (postop) medical complications in large populations is available. We studied the incidence and outcomes of medical complications in a large cohort of 8930 patients undergoing hip fracture repair.
Methods: We used a retrospective cohort of patients 60 and older undergoing surgical repair of hip fracture in 20 hospitals in 1982-1993. Participating hospitals included 5 university-affiliated hospitals, two Veterans Affairs hospitals, four community teaching hospitals, and nine community hospitals. Initial exclusion criteria were: metatstatic cancer; multiple myeloma; multiple trauma; amputation; para- or quadriplegia; for a cohort of 9598 patients. Further xclusion criteria for this analysis were: acute preoperative medical complications (e.g., myocardial infarction, congestive heart failure, pulmonary embolism, pneumonia), leaving a cohort of 8930 patients. Data were collected by chart audit, using trained abstractors, standardized pretested abstraction form, and explicit criteria for medical complications, out to death, discharge, or 30-days, whichever occurred first. Cardiac and pulmonary complications were further characterized as serious or nonserious. The index complication was the first complication. Multiple index complications were defined as those occurring within 1 day of the first identified postop medical complication.
Results: 7193 patients (81%) had no postop medical complications; 1737 had one or more complications. Of the 1737, 1531 (88%) had only one index complication; 182, 19, and 5 had 2,3, and 4 index complications, respectively. Numbers of patients with different types of complications (and associated 30-day mortality rates were): serious cardiac – 178 (23%); serious pulmonary – 229 (17%); gastrointestinal bleed (GIB) – 195 (8%); venous thromboembolism – 89 (4%); stroke or transient ischemic attack – 85 (18%); combined cardiac and pulmonary complications – 117 (33%); 3 or more complications – 89 (38%); renal failure – 8 (38%); and septic shock – 5 (0%). Of the most frequent complications, serious cardiac occurred earlier than serious pulmonary (1 vs 4 median days, p<0.0001) and days to inpatient death were fewer for serious cardiac than serious pulmonary (2 vs 8 median days, p<0.0001). Mortality rates at 30 days and one-year did not differ for serious cardiac and serious pulmonary complications (p=.17 and.10, respectively.
Conclusions: From this first large study using patients, rather than complications, as the unit of analysis, we conclude that most patients do not have medical complications after hip fracture surgery. Frequency and mortality rates were similar for both serious pulmonary and cardiac complications. However, serious cardiac complications occur significantly earlier and time to death is significantly longer for serious pulmonary complications.
*267. Functional Recovery after Major Abdominal Surgery in Elders
Valerie A. Lawrence, VERDICT Center of Excellence (Health Services Research and Development, Veterans Affairs), San Antonio, Texas, H Hazuda, University of Texas Health Science Center at San Antono, C Mulrow, VERDICT Center of Excellence (Health Services Research and Development, Veterans Affairs)
Objectives: Despite its importance to patients, little is known of the course of functional recovery in elders having major abdominal operations. Most functional recovery research has focused on coronary artery bypass operations, hip or knee replacement, and cataract operations. The little data on abdominal surgery suggest that 10-20% elders may not be fully functionally recovered as long as six months after surgery.
Methods: In a prospective cohort of elders undergoing elective major abdominal surgery, patients >=60 years old underwent assessment of Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), timed walk, and grip strength preoperatively (preop) and five times postoperatively (postop) out to six months.
Results: Patients were recruited from a Veterans Hospital (19%), a public teaching hospital (32%) and private hospitals (49%). Of 383 enrolled patients, 11 dropped out after the preop assessment and 280 have 6 month followup. Procedure types, confirmed by chart review in 360 patients, are: colectomy – 131; aortic – 88; ventral/umbilical hernia repair – 61; upper abdominal procedures – 21 (Whipple, gastrectomy, hepatectomy, splenectomy, Nissen); open cholecystectomy – 19; hysterectomy – 13; nephrectomy – 8; and other – 10. Patients comprise 44% women, 42% Mexican American and 47% EuroAmerican;. Age ranges are 60-69 years old (57%), 70-79 years old (37%), and >= 80 years old (6%). Mean ADL summary score worsened significantly at one week postop by 2.8 (SD 3.6) (p  0.0001) reamined significantly worse at 3 and 6 weeks (p  0.001, 0.05) and returned to preop baseline by 3 months. Mean IADL summary score worsened at one week postop by 7.3 (SD 4.8) (p  0.0001), remained significantly worse at 3 weeks, 6 weeks, and 3 months (p  0.0001 for all) and returned to baseline by 6 months. Mean timed walk returned to baseline by 3 months postop but mean grip strength (a measure of conditioning) was still signficantly worse than preop baseline at 6 months. 14% and 10% of participants had continued impairments in ADLs at 3 and 6 months, respectively. In IADLs, 37% and 20% had not fully recovered at 3 and 6 months. For the timed walk, 30% and 33% were not back to baseline at 3 and 6 months, respectively. And for grip strength, a measure of overall conditioning, 62% and 53% were not recovered to preop baseline at 3 and 6 months.
Conclusions: This sentinel large cohort study showed approximately one third of patients undergoing major elective abdominal surgery do not recover to baseline IADLs by 3 months and 20% have not recovered at 6 months. Recovery in basic ADLs occurs earlier but 10% continue to have impairments 6 months after surgery.
|| | | | ||