HSR&D Home » Research » IIR 11-058 – HSR&D Study
Evaluation of VLER_Indiana Health Information Exchange Demonstration Project
David A Haggstrom, MD
Richard L. Roudebush VA Medical Center, Indianapolis, IN
Funding Period: July 2012 - December 2015
Most Veterans receive some portion of their care outside the VA. Health care providers seeing Veterans inside and outside the VA do not have timely access to important health care information about their patients that is documented in other health care systems. To date, such information can only be obtained by a manual release of information process confronted by potential delays and uneven reliability.
In 2009, the President directed the Department of Veterans Affairs and the Department of Defense to create the Virtual Lifetime Electronic Record (VLER). As part of this initiative, in August 2010, the VA began a partnership with Regenstrief Institute to create capacity for the Indianapolis VA to exchange health information with community partners. We have referred to this partnership as the VA-IHIE program. The Indiana Health Information Exchange (IHIE) is arguably the largest and most comprehensive information exchange in the country. Health information exchange (HIE) provides data from a "network of health care networks" that provides clinical data in real-time at the point of care. HIE may enable providers to deliver high-quality care by facilitating continuity of care among multiple providers.
Yet while the potential for HIE to reduce costs and improve the quality of health care is well-recognized, there is a small amount of empirical data regarding the value of HIE in the peer-reviewed literature. Nor has research been designed to examine the amount of additional information provided by the HIE that may be needed to have an effect upon health care quality. There is sparse empiric knowledge of the effect of HIE upon ambulatory care quality or its impact in the primary care setting. Also, little is known about the effect of HIE upon admission rates for ambulatory care sensitive conditions. Another potential of HIE is to reduce unnecessary tests, but there is no evidence base with which to evaluate this claim in the VA.
The objectives and hypotheses of this proposal are as follows:
Objective 1: to assess the proportion and predictors of health care received by Veterans outside the VA
Objective 2: to assess the impact of HIE upon the quality of care received by Veterans
Objective 3: to explore whether the implementation of HIE will reduce the economic expenditures of health care for Veterans
Hypothesis 1: implementation of HIE will improve the quality of ambulatory care among Veterans
Hypothesis 2: implementation of HIE will reduce admission rates for ambulatory care sensitive conditions among Veterans
The VLER health information exchange (HIE) program began signing up Veterans in March 2011. The study employs retrospective, observational data to assess the effect of being exposed to health information exchange.
The study population was drawn from the population of patients who have been seen at the Indianapolis VA (2 outpatient or 1 inpatient visit). From this population availability of patient data in the INPC (Indiana Network for Patient Care) was determined. The recruitment goal for the health information exchange arm of the study was 2,000 patients.
We performed a pre-post evaluation of the HIE implementation among Veterans seen at the Indianapolis VA, with a concurrent control group. HIE users were compared to a control group of non-HIE users with propensity score methods. For purposes of evaluation, we analyzed the HIE implementation on a patient-level given that additional information available from the exchange is available only on a patient-by-patient basis, i.e., a provider cannot use exchange information available for one patient in the care of another.
For subjects in this study, the data managers retrospectively extracted the following data elements, from the Indianapolis VA and the Indiana Network for Patient Care: demographics, medications, problems, immunizations, laboratory, procedures, outpatient clinic, outpatient clinician encounters (visits), and administrative procedure and diagnostic codes.
The study design consisted of a retrospective cohort of Veterans who participated in the VA HIE, compared with a concurrent control group, both before and after the HIE was implemented in March 2012. We collected data from two primary sources, VA administrative data and federated electronic health records from the Indiana Network for Patient Care (the Indianapolis VA's community HIE partner). We have analyzed the data in concordance with participation, utilization, quality of care, and cost metrics.
A significant number of Veterans (6,104) were recruited to participate in the VLER health information exchange (HIE) program at the Indianapolis VA, exceeding our recruitment goal by more than 4,000 individuals. Patients were enrolled in the HIE program onsite at the Indianapolis VHA in outpatient clinics or through the release-of-information office. Our analysis ascertained the characteristics of Veterans who participated in HIE. Increased utilization of medical services, being ages 50-64, possessing higher levels of comorbidity, and being female all increased the likelihood of participation in the HIE program. Residency in a rural area, and being single decreased the likelihood of participation, and there were no differences in participation by race.
In terms of health care utilization, 15% of all laboratory tests, 21% of imaging tests, and 14% of procedures received by Veterans were identified using data from the community HIE (vs. VA data). At the patient-level, 26% of Veterans had at least some lab data, 26% had imaging data, and 9% had procedure data from the community HIE partner (i.e., this proportion of Veterans had either some or all of their utilization data identified from the community HIE partner, in each of these categories). The community HIE partner was more likely to provide additional data about coagulation labs, glucose tests, flexible sigmoidoscopies, cardiac catheterizations, and CT scans. Veterans who have utilization data from the community HIE partner are more likely to be older, white, married, live in urban areas, and have less disability.
More than a third of Veterans' hospital admissions were identified using non-VA data. At the patient level, 30% of Veterans only had hospitalizations identified at non-VA hospitals, while 7% had hospitalizations identified at both VA and non-VA hospitals in the community HIE. Hypothetically, HIE may reduce hospitalization rates for ambulatory care sensitive (ACS) conditions through the sharing of health care information. The difference-in-differences in the rates of ACS hospitalization (per 100,000 population) was not statistically significant among patients who were enrolled in the HIE program (4,694 before vs. 4,633 after implementation) compared to those who were not enrolled in the HIE program (3,090 before vs. 3,184 after implementation). Thus, the impact of HIE use in primary care upon ambulatory-care sensitive conditions was not detectable, at least within the first year of program implementation.
Findings regarding ambulatory care quality indicate that the VA HIE program had no measurable effect upon the quality of diabetes care, in terms of the proportion of patients with HgA1c < 9% or the proportion with either LDL cholesterol < 130 or < 100. Potential reasons for the lack of effect may be that health information exchange (HIE) does not influence these types of care management processes, the amount of data shared with the VA was not sufficient to affect care quality, or the evaluation was performed too early in the course of the HIE program implementation.
We did find that participation in the VA HIE program had significant effects upon the overuse of health care tests. Specifically, patients who participated in the VA HIE were less likely to receive lab tests at more frequent intervals, including complete blood counts and renal profiles (absolute difference in differences:-1.98% (95% CI, -2.93%, - 1.05%) and lipid and liver tests (absolute difference in differences: -3.19% (95% CI, -4.30%, - 2.18%) Furthermore, participants in VA HIE were less likely to receive high-cost imaging tests (CT, MRI, and PET) at more frequent intervals (absolute difference in differences: -1.31% (95% CI, -2.56%, - 0.12%).
With regards to costs, we assessed whether HIE reduced VHA medical costs 1 year post-implementation. A cohort evaluation with a concurrent control group compared VHA healthcare costs using propensity score adjustment. The annual adjusted total cost difference per patient was associated with a higher cost for HIE participants $1152 (95% CI: $807-1433) (P < 0.01) (in 2014 dollars) than HIE participants. Short-term evaluation of this demonstration project did not show immediate reductions in healthcare costs, although cost reductions from shared health information may be realized with longer time horizons.
The Virtual Lifetime Electronic Record (VLER) Health program provides the Veterans Health Administration (VHA) a framework whereby VHA providers can access the veterans' electronic health record information to coordinate healthcare across multiple sites of care. As an early adopter of VLER, the Indianapolis VHA and Regenstrief Institute implemented a regional demonstration program involving bi-directional health information exchange (HIE) between VHA and non-VHA providers.
Many stakeholders in health care, clinical informatics, and management intuitively believe HIE will increase the quality and reduce costs of care received. This project implemented a framework to assess, and deliver empirical data about, the value of health information exchange. The project also assessed the characteristics of Veterans who had data available through the HIE, as well as what type and volume of health care services they received outside the VA healthcare system. Because health care reform provides Veterans with more options for seeking care, it is important to understand their current patterns of utilization across health care systems. Our analysis demonstrates the potential for information sharing, as well as some early effects of health information exchange.
One of the implications of the study, given that large amounts of non-VA clinical information were identified in the health information exchange (HIE), is that HIE offers great potential to coordinate care. Although HIE did not demonstrably increase the receipt of evidence-base care, reductions were found in the frequency of routine testing. These findings suggest that, through the increased sharing of information, HIE may have the most immediate impact in reducing unnecessary care. In addition, our research team believes that future HIE interventions will benefit from an active, tailored human component in order to maximize the impact of newly available information upon systems of care.
As of March 2012, the HIE demonstration program was operational at our site. We subsequently completed a training video for providers to orient them to the VLER program. Since then, the VLER program has been disseminated more widely among more than a dozen VA facilities nationwide. We completed a training video for providers to orient them to the VLER program.
In terms of next research investigations, one of our team's co-investigators Dr. Brian Dixon is participating as a site PI for another VA HSR&D study (IIR-14-049) led by Dr. Ken Boockvar (PI): "Regional data exchange to improve care for veterans after non-VA hospitalization". Investigators at the Indianapolis VA HSR&D Center for Health Information and Communication plan to submit a couple VA HSR&D proposals on HIE as next step of IIR 11-058.
External Links for this Project
NIH ReporterGrant Number: I01HX000722-01
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
DRA: Health Systems
DRE: Research Infrastructure
MeSH Terms: none