Lead/Presenter: Susan Frayne, COIN - Palo Alto
All Authors: Frayne SM (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System; Stanford University School of Medicine)
Phibbs CS (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System; Stanford University School of Medicine)
Berg E (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System)
Saechao F (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System)
Hoggatt KJ (HSR&D Center for the Study of Healthcare Innovation, Implementation & Policy (CSHIIP), VA Greater Los Angeles; UCLA Fielding School of Public Health)
Yano EM (HSR&D Center for the Study of Healthcare Innovation, Implementation & Policy (CSHIIP), VA Greater Los Angeles; UCLA Fielding School of Public Health)
Finlay AK (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System)
Breland JY (HSR&D Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System)
Washington DL (HSR&D Center for the Study of Healthcare Innovation, Implementation & Policy (CSHIIP), VA Greater Los Angeles; UCLA Geffen School of Medicine)
Hamilton AB (HSR&D Center for the Study of Healthcare Innovation, Implementation & Policy (CSHIIP), VA Greater Los Angeles; UCLA Geffen School of Medicine)
Objectives:
In 2008, VA launched a far-reaching policy designed to ensure new women Veterans' (WVs') access to comprehensive women's health care. Knowing that WVs new to VA have high rates of attrition from VA, we examined whether attrition rates decreased for new WV patients in the years after this policy's rollout, benchmarked against attrition rates among new men Veteran (MV) patients over the same period.
Methods:
For each Fiscal Year (FY) examined (FY07-FY13), we used national VA administrative data to create a cross-sectional national cohort of new WVs (no VA or fee basis use in prior 8 years), and, for benchmarking, new MVs. We created person-specific timelines starting with the first VA outpatient visit. Attrition was defined as no subsequent VA care in person-specific years 2-3 (which could be as late as FY16). Within each annual cohort, we examined rate of attrition among WVs (and separately, among MVs). We fit gender-stratified logistic regression models for attrition (yes/no) as a function of cohort year, first unadjusted and then adjusting for age, race-ethnicity, and service-connected status.
Results:
Between FY07 to FY13, number of new WVs increased 24% from 15,963 to 19,772, while new MVs declined 4% from 226,784 to 218,146. Among new WVs, the attrition rate was 23.0% in FY07 and declined yearly to 17.2% by FY13; for new MVs, the attrition rate was 22.9% in FY07 and 20.6% by FY13. For WVs, there was a decline over time in unadjusted odds of attrition (OR 0.89 [0.84-0.94] for FY08 vs FY07; OR 0.69 [0.66-0.73] for FY13 vs FY07) and adjusted odds (0.92 [0.86-0.97] and 0.72 [0.68-0.77], respectively). Improvements in attrition over time were less pronounced for new MVs, both unadjusted (OR 0.94 [0.92-0.95] and 0.88 [0.86-0.89], respectively) and adjusted (OR 0.94 [0.93-0.95] and 0.90 [0.86-0.91], respectively).
Implications:
Though attrition was similar for women and men in FY07, from FY07-FY13 the reduction in attrition was greater for new WV patients compared with MVs.
Impacts:
While causal conclusions cannot be drawn, the substantial improvement in WVs' attrition rate over time is consistent with a potential beneficial effect of VA's comprehensive women's health care policy.