BACKGROUND/RATIONALE:
Prior studies of others and our data demonstrate that a large proportion of Veterans access multiple systems for care. Veterans who use both VA and non-VA services are at increased risk of adverse events, especially during transitions of care when changes in medications and plan of care may occur without the knowledge of VA physicians. Lacking access to adequate information of non-VA care among Veterans may limit ability of VA providers to provide levels of care of a medical home. Given that little is known about how information exchange occurs in Veterans accessing dual care, it is difficult to prescribe solutions to bridge the gap of VA and non-VA care. Understanding how information exchange occurs, the current gaps and reasons behind these gaps, may allow for developing solutions that will help bridge VA and non-VA care.


OBJECTIVE(S):
The objectives of the project were to better understand how information exchange between VA and non-VA settings currently occurs and to assess the quality of information exchange between VA and non-VA providers. The specific aims of the study were (i) to describe how and to what extent information was exchanged with VA primary care teams among Veterans recently discharged from a non-VA hospital or emergency room; (ii) to determine the incidence of medical errors, defined as medication errors, test follow-up errors, and work up errors during the care transition process among these Veterans and (iii) to identify factors that were related to information exchange and with medical errors.

METHODS:
The project enrolled Veterans from James J Peters VA Medical Center and Hudson Valley Health Care System, where up to 40% of Veterans have accessed non-VA care in the previous year. The time period of data collection was from April 2015 to December 2016. We observed in a cohort of Veterans in both urban and rural settings how information exchange occurred after a non-VA hospitalization or emergency room visit, and how and whether VA providers received adequate information for ongoing care for Veterans. We examined if the information exchange occurred according to VA dual care policy using a checklist derived directly from the dual care policy. We then determined the incidence of medical errors, defined as medication errors, tested follow-up errors, and worked up errors, during the care transition process from the review of documentation at the follow up visit at VA. We then identified factors (Veteran, illness episode, and provider factors) that are associated with whether care was consistent with VA policy and with the incidence of medical errors.

FINDINGS/RESULTS:
A total of 167 Veterans, 53.9% from rural areas, were recruited during the time period of April 2015 to December 2016. The average age was 65.0 years with a standard deviation of 13.6. The majority of Veterans (93.4%) were male; 55.2% were White with 11.5% Hispanic; 39.4% were married. Most (>92%) completed high school, GED or above. A total of 37.5% of participants reported having Medicare or Medicare Advantage coverage, 24.5% participants reported having Medicaid coverage, and 28.6% participants reported private insurance coverage. The number of chronic diseases on the modified RAND index was 5.5 on average with SD of 3.0. Average score for Katz independence of activities of daily living (ADL) score was 5.6 (on a scale of 0 to 6 with 6 indicating complete independence). Average health literacy score as measured by the Short-Test of Functional Health Literacy in Adults (S-TOFHLA) was 26.6 (SD 11.3), a breakdown of 73.7% with adequate health literacy, 7.2% with marginal health literacy, and 19.2% with inadequate health literacy. Assessment of patient level of engagement was assessed using the Patient Activation Measure (PAM-13) and the average score was 51.9 (SD 21.2) on a scale of 100 with higher score indicating higher level of patient activation. Of the population examined, 50% of Veterans reported that they had access to the VA patient health information portal-- My HealtheVet. Regarding VA CHOICE Program, only 40% had heard of the CHOICE program, and among those who have heard of it, 18.9% reported having used CHOICE.

Among the checklist items regarding information transfer, reporting of non-VA care to VA providers varied with 63.4% of Veterans reporting that they have informed their VA primary care provider (PCP) of their recent non-VA visit and with 67.7% indicating that they have obtained records for their VA (PCP) about their visit. On chart review, among those who had follow up appointments within 3 months, 66.0% of VA PCP visits documented the non-VA visit and indicated additional review of non-VA documents. Among the 18 VA providers enrolled in the study, 72.2% of providers reported that they "almost always" or "often" receive non-VA information from Veterans, but only 38.9% reported "almost always" or "often" receiving non-VA information from non-VA providers. A total of 33.3% of providers reported almost always using community health information exchange (Regional Health Information Organization) to obtain non-VA patient information. During care episodes, 87% of participating VA providers reported that they received information from patient about non-VA care episode, yet only 43% had discharge summary for review and none had a telephone discussion with non-VA providers. The majority of VA providers (78%) reported that they strongly agree or agree that they received all of the information they needed to be able to take care of the Veteran.

Medication discrepancies were observed in 21.6% of participants, and 22.5% had any errors which include testing, work-up errors, and medication discrepancies during the transition of care episodes from non-VA care to VA care. None of the checklist items regarding Veterans aspect of notification of PCP and obtaining records for PCP, or PCP regarding review of records and documentation were significant factors associated with the incidence of errors. However, non-VA hospital discharge chart information was more likely to be captured in the VA patient medical record when patient accessed non-VA care on fee basis where the referral has a formal process.

Limitations: The lower than expected event rate limited our ability to detect smaller effect sizes.

IMPACT:
The recent transformation of primary care to provide care in a medical home model, PACT, provides increasing focus on providing patient centered care to Veterans. Dually eligible Veterans have the choice to access both VA and non-VA health care and multiple systems use may represent additional risk of errors during transitions of care. We observed discrepancies in medications, tests, and workup in approximately 20% of the episodes when Veterans transition from non-VA acute care to primary care in VA, which was consistent with prior literature.

Although VA providers largely adhered to dual care policy, they were limited by the information that they received at the point of care. VA providers were often notified by patients about the non-VA care episode, only half received discharge summaries and there was also little direct communication among VA providers and non-VA providers in coordinating care. However, information exchange from non-VA sites was better documented when formal processes of referral were utilized. Health information technology advances in health information exchange and patient health information portal may be underutilized and may represent an opportunity to improve information gap.

---

External Links for this Project

NIH Reporter

Grant Number: I01HX001710-01
Link: https://reporter.nih.gov/project-details/8867413

---

PUBLICATIONS:

None at this time.

DRA: Health Systems
DRE: Treatment - Observational
Keywords: none
MeSH Terms: none