3083 — Use of Provider Feedback and Clinical Champions Associated with Depression Screening, Findings from EPRP and National Survey of PC Directors (2007)
Rose DE (Center for the Study of Healthcare Provider Behavior) , Klap RS
(Center for the Study of Healthcare Provider Behavior), Farmer MM
(Center for the Study of Healthcare Provider Behavior), Mitchell MN
(Center for the Study of Healthcare Provider Behavior), Canelo I
(Center for the Study of Healthcare Provider Behavior), Yano EM
(Center for the Study of Healthcare Provider Behavior)
Depression is a common, costly and chronic condition. According to the US Preventive Service Task Force, in primary care settings, prevalence of major depression ranges from 5% - 9%, and up to 50% of depressed patients are not recognized. Here we seek to assess the association of quality improvement methods (e.g., computerized reminders, specialized CPRS templates, performance profiling, incentives, designated local clinical champion, delegated RN for disease specific management, and provider education) with depression screening rates.
We linked data from the External Peer Review Program (2007), National Survey of Primary Care Directors (2007), and Area Resource File to create an analytic sample of 34,964 patients at 220 VA Medical Centers and Community Based Outpatient Clinics. Organization-level controls included: primary care clinic size (number of patients), academic affiliation, and scales based on primary care directors’ report of authority over and adequate resources for primary care programs. Patient-level controls include age, gender, race/ethnicity, marital status, income, and extent of service connection. We performed multi-level logistic regression to predict receipt of depression screening, controlling for clustering of patients within facilities, and testing for statistically significant interactions between quality improvement methods. We used Wald tests to test simple effects associated with statistically significant interactions terms.
When we tested each quality improvement method separately, no single strategy was associated with increased depression screening rates. However, sites using both performance profiling and clinical champions were twice as likely to report depression screening (OR: 2.3, 95%CI: 1.1, 4.5, p < 0.05). Among sites with performance profiling, the use of clinical champions statistically significantly increases the probability of depression screening (86% v. 79%, Wald test with one degree of freedom = -2.00, p = 0.046). Only 14% or 30 sites in our analytic sample reported using both performance profiling and clinical champions.
When used together, performance profiling and local clinical champions were associated with significantly greater likelihood of receipt of depression screening.
If all sites in our sample had adopted both these practices, an additional 5% (or 17,000 veterans) would have received depression screening.