2019 HSR&D/QUERI National Conference
1139 — Development of a Veteran-Centered Measure of Healthcare Engagement
Lead/Presenter: Rachel Kimerling, COIN - Palo Alto
All Authors: Kimerling R (Center for Innovation to Implementation, Palo Alto), Schalet, BD (Feinberg School of Medicine, Northwestern University), Reise, SP (Department of Psychology, University of California, Los Angeles) Zulman, DM (Center for Innovation to Implementation, Palo Alto) Lewis, ET (Center for Innovation to Implementation, Palo Alto)
The National Academy of Medicine has identified measurement of healthcare engagement as a core metric for health care, yet valid measures applicable to a wide range of patient populations are not available. We describe the first quantitative results from a study to develop a Veteran-centered measure of healthcare engagement.
A prior study derived items and an engagement framework comprised of 4 domains: collaborative communication, health information use, healthcare navigation, and self-management. Items were administered to 9,538 Veterans diagnosed with chronic conditions and/or mental health conditions via mail survey fielded October, 2018 through January, 2019 (RR = 38%). Current results describe data available at submission date (n = 7,948; 83% of sample). We used hierarchical item clustering and exploratory Item Response Theory derived from ordinal factor analysis to evaluate measurement structure and item characteristics. We used Spearman correlations and ordinal logit models to evaluate concurrent validity.
A pool of 41 candidate items formed one general dimension of engagement, with clusters around navigation (alpha = .89), self-management (alpha = .89), and communication (alpha = .88). Items demonstrated good reliability and measurement precision (alpha = .95; reliability > .80 and SEM < .35 from 3 standard deviations (SD) below the mean to 2 SD above the mean). Several items emerged as particularly informative, such as "I can get the care I need without getting discouraged"; "I know I can get the information I need about the pros and cons of treatments"; and "I know I can express my doubts, even if my provider might disagree." Higher item scores were significantly associated with higher scores for related outcomes, including CAHPS composite scores for communication with providers (r = .46-.66), health literacy (r = .22-.34), and ratings for how well VA care met their needs (r = .47-.67).
Analyses suggest very good feasibility for these items to establish a valid and reliable measure of healthcare engagement. Associations with related outcomes suggests potential for a predictive measure.
Measurement of healthcare engagement can enhance population health management by identifying Veterans in need of adjunctive services (e.g., coaching, care coordination, high-touch care) or optimizing outcomes through tailored engagement strategies. Research also suggests applications for risk adjustment and performance benchmarking.