Lead/Presenter: Rachel Kimerling,
COIN - Palo Alto
All Authors: Kimerling R (Center for Innovation to Implementation & National Center for PTSD), Zulman DM (Center for Innovation to Implementation, Stanford University School of Medicine), Lewis ET (VA Office of Mental Health and Suicide Prevention), Tamayo, G (Center for Innovation to Implementation), Reise SP (Department of Psychology, University of California, Los Angeles), Schalet BD (Northwestern University Clinical and Translational Sciences Institute, Northwestern School of Medicine)
Objectives:
Healthcare engagement is a key measurement target for value-based healthcare, but a reliable and valid patient-reported measure has not yet been widely adopted. We describe the development and validity of a newly developed brief (8-item), Veteran-centered, patient-reported measure of healthcare engagement, the Patient-Reported Outcome Measure Information System (PROMIS) Healthcare Engagement short form (PHE-8a).
Methods:
Methods: This was a prospective evaluation linking a national survey of Veterans Administration (VA) healthcare users to VA administrative data over the following year. Survey respondents were 9552 Veterans receiving care for diabetes, hypertension, depression or post-traumatic stress disorder. The PHE is written at a 6th grade reading level and was developed with input from VA user and provider stakeholders. VA PHE scores represent Item Response Theory calibrated standardized scores. We examined risk-adjusted associations of PHE scores with CAHPS composite scores for communication and self-management support, outpatient primary care and mental health no-show rates, My HealtheVet use, and HEDIS quality measures for diabetes and hypertension. Survey-adjusted mixed-effects models accounted for facility-level clustering and complex sampling. We reported marginal predicted effects and a Wilcoxin-type test for trends to characterize quality indicators across the continuum of engagement scores.
Results:
The PHE-8a demonstrated good marginal reliability (> .85) from -2.5 SD to + 1.5 SD, (range of scores 25–65) and excellent test re-test reliability (ICC = .89). Engagement scores did not meaningfully differ by gender, race, ethnicity, or education, but were lower among Veterans who were younger, with inadequate health literacy, without a usual source of care, and those reporting financial strain. Higher engagement scores were associated with better healthcare quality across all indicators, with significant trends where each 5-point, or ½ standard deviation increase in PHE scores was associated with statistically significant and clinically meaningful gains in quality. For example, as engagement scores increased, primary care no-show rates decreased from 24% among those with low engagement to 15% among those with high engagement. My HealtheVet use increased from 39% (SE = 1.6%) to 45% (SE = 1.3%) over the continuum of engagement scores; the likelihood of poor diabetes control decreased from 24.7% (SE = 2.8%) to 13.8% (SE = 1.8%); and the likelihood of hypertension control increased from 53% (SE = 2%) to 60% (SE = 1.7%).
Implications:
The PHE-8a is a brief, reliable and valid patient-reported measure of healthcare engagement. The range of reliable scores reflects the emphasis of the PHE on the sensitivity to detect individuals with lower engagement that may benefit from adjunctive services. For all indicators examined, we observed a dose-response relationship between healthcare engagement as measured by the PHE and better healthcare quality.
Impacts:
Our results add support for previously untested hypotheses that patient engagement can promote better healthcare quality. The PHE is broadly applicable and the brevity and reading level support feasible administration at the point of care or in research/quality improvement settings. The measure has good potential for evaluation of Veteran engagement strategies, to enhance population health management, and to improve healthcare delivery.