Although diabetes outcomes are improving within Veterans Affairs (VA), 16% of Veterans with diabetes have a hemoglobin A1c (HbA1c) >9.0%. Veterans with persistent poorly-controlled diabetes mellitus (PPDM) - pragmatically defined as maintenance of HbA1c >9.0% for >1 year despite ongoing VA diabetes care - are at high risk for complications and costs. There is an urgent need for novel approaches to PPDM that: 1) can address refractory diabetes; and 2) can be implemented using existing resources (without relying upon research-funded interventionists/devices).
We developed Advanced Comprehensive Diabetes Care (ACDC) to address PPDM within VA. The ACDC intervention targets common barriers to improvement in PPDM by combining telemonitoring, medication management, self-management behavioral support, and depression management. Delivery of ACDC requires only existing VA clinical staff and equipment, maximizing its potential for implementation and scaling.
This pilot aimed to: 1) evaluate ACDC's effectiveness; 2) utilize mixed-method formative analysis to assess ACDC's implementation strategy - use of existing VA Care Coordination Home Telehealth (CCHT) services; and 3) generate data to refine ACDC and inform broader study of its implementation. We hypothesized that ACDC would improve HbA1c, diabetes self-care, medication adherence, and depressive symptoms versus usual care.
We evaluated ACDC using a type I effectiveness-implementation hybrid design. Eligible patients had PPDM as defined above; determination of PPDM required 2 HbA1c readings over the prior year and an established Primary Care/Endocrinology provider at Durham VA. Exclusion criteria included inability to communicate by phone, prior hypoglycemia complications, refusal to self-monitor blood glucose, and others.
Fifty Veterans were randomized in an unblinded fashion to receive ACDC or usual care or for 6 months. All patients continued to follow with their existing providers. ACDC patients were enrolled in the Durham VA CCHT program; all patients with poorly-controlled diabetes qualify for CCHT at the Durham VA. Each ACDC patient was assigned to one of two CCHT nurses working with the study. These nurses provided patients a standard CCHT orientation and VA-approved telemonitoring equipment.
ACDC comprised 4 phone-delivered components, administered concurrently. First, patients performed telemonitoring by collecting self-monitored blood glucose (SMBG) four times daily and transmitting data to their assigned CCHT nurse. Second, patients received medication management. At 2-week intervals, CCHT nurses called patients to reconcile medications and review SMBG data and medication adherence. This information was compiled into an electronic health record-based report and forwarded to the study Endocrinologist, who recommended any needed medication changes via a report addendum (primary provider copied). CCHT nurses enacted any recommended changes with patients. Third, patients received self-management behavioral support. During each 2-week phone encounter, CCHT nurses delivered modules addressing PPDM-relevant content (8 modules, with topics including SMBG, hypoglycemia recognition/management, proper insulin use, etc.). Fourth, if applicable, patients received depression management. Individuals endorsing depressive symptoms received a phone call from the study Psychiatrist (or their existing VA Psychiatrist) to arrange an expedited evaluation. Our treatment approach emphasized antidepressant pharmacotherapy proven effective in diabetes populations.
All patients attended baseline, 3-month, and 6-month outcome visits; a survey was completed and lab measures obtained at each outcome visit. We used linear mixed models to analyze our primary effectiveness outcome, HbA1c. Using validated scales, we also assessed diabetes self-care, self-reported medication adherence, and depressive symptoms. Adverse events were ascertained in both randomization groups by structured self-report, and CCHT data were carefully analyzed in the ACDC group for evidence of hypoglycemia.
We conducted a mixed-method formative analysis to assess ACDC and its CCHT-based implementation strategy. Intervention patients completed a quantitative acceptability measure and a phone-based semi-structured qualitative interview. We also completed qualitative interviews with both CCHT nurses to assess ACDC's implementation. Interview guides were informed by a conceptual model, and qualitative analyses followed established processes.
Patients were predominantly male and majority African-American. Mean age was 60, and 48% of the population had depression by ICD-9 codes or positive depression screen at baseline. Baseline diabetes control was poor (estimated mean HbA1c 10.5%). Of the 50 Veterans enrolled in the study, 46 completed 6-month follow-up (23 in each group).
At 6 months, estimated mean HbA1c was 9.2% for ACDC patients vs. 10.2% for usual care (estimated between-group difference -1.0%, 95% CI -2.0 to -0.0, p=0.05). Within the ACDC group, 'engaged' patients (i.e., those completing 6 phone encounters, n=16) had an observed mean HbA1c of 8.3% at 6 months vs. 10.8% for 'non-engaged' patients ( 3 encounters, n=7). ACDC patients had a significant improvement in diabetes self-care versus usual care, but self-reported medication adherence and depressive symptoms did not differ between groups. Of note, virtually every intervention patient with a positive depression screen (n=7) had existing VA Psychiatry care; only one patient received expedited evaluation by the study Psychiatrist (and was felt not to be depressed).
Adverse events did not differ between study arms. No patients reported hypoglycemia requiring assistance. Among the 20 ACDC patients transmitting blood glucose data to CCHT, hypoglycemia was most common before the first phone encounter. Rates of hypoglycemia following encounter 1 were low (median blood glucose values <70 =0, IQR 0-4; median blood glucose values <60 =0, IQR 0-1).
We conducted qualitative interviews with 20 ACDC patients and the two study CCHT nurses. Formative analysis is ongoing.
ACDC produced a statistically and clinically significant improvement in HbA1c among Veterans with PPDM; this improvement was particularly impressive for engaged patients. This improved diabetes control occurred without excessive rates of hypoglycemia or other adverse events; hypoglycemia was actually most common at study start and rarer thereafter. Veterans' improved diabetes self-care may have contributed to these observed gains. While depression was prevalent in the study population, ACDC did not impact depressive symptoms, perhaps due to high rates of preexisting Psychiatry care.
By definition, Veterans with PPDM remain poorly-controlled despite engagement with available diabetes care, so are among the VA's highest-risk diabetes patients. ACDC effectively improves control in this refractory population, so represents a high-impact tool in VA's diabetes management arsenal. Critically, ACDC is designed for telemedicine-based delivery using only existing VA clinical infrastructure, so is amenable to implementation and scaling. After honing ACDC through this study's formative analysis, we will study broader implementation of the intervention to maximize its impact throughout VA.
- Andrews SM, Sperber NR, Gierisch JM, Danus S, Macy SL, Bosworth HB, Edelman D, Crowley MJ. Patient perceptions of a comprehensive telemedicine intervention to address persistent poorly controlled diabetes. Patient preference and adherence. 2017 Mar 3; 11:469-478.
- Crowley MJ, Edelman D, McAndrew AT, Kistler S, Danus S, Webb JA, Zanga J, Sanders LL, Coffman CJ, Jackson GL, Bosworth HB. Practical Telemedicine for Veterans with Persistently Poor Diabetes Control: A Randomized Pilot Trial. Telemedicine journal and e-health : the official journal of the American Telemedicine Association. 2016 May 1; 22(5):376-84.
- Crowley M, Edelman D, McAndrew AT, Kistler SN, Danus SM, Webb JA, Zanga J, Sanders L, Coffman CJ, Jackson GL, Bosworth HB. Effectiveness of a Scalable Telemedicine Intervention for Veterans with Persistent Poor Diabetes Control. Paper presented at: American Diabetes Association Annual Scientific Session; 2015 Jul 9; Boston, MA.
- Crowley M, Sperber N, Gierisch JM, Edelman D, McAndrew AT, Kistler S, Danus SM, Jackson GL, Bosworth HB. Patient Perceptions of Effective, Scalable Telemedicine Care for Veterans with Persistent Poor Diabetes Control. Paper presented at: VA HSR&D / QUERI National Meeting; 2015 Jul 8; Philadelphia, PA.