Declining Attrition of New Women Patients After Veterans Health Administration Rollout of Comprehensive Women's Health Care Policy.

INTRODUCTION: Women veterans represent a numeric minority group in the Veterans Health Administration (VA). Responding to documented gaps in their care, the VA Office of Women's Health in 2010 launched the Comprehensive Women's Health Care Policy, designed to transform VA health care for women nationally. Before this policy, there was a high rate of attrition (i.e., discontinuation of VA use) among women Veterans new to VA care. We examined whether attrition rates declined among women veterans new to VA across the 7-year period before and after this 2010 policy. Women were benchmarked against men for these analyses, because the policy was women-specific. MATERIALS AND METHODS: For this retrospective longitudinal analysis of existing national VA databases, we identified date of first VA outpatient use (index date) for all veterans new to VA in fiscal year (FY) 2007-2013. We considered a patient to have attrited if they had no VA-based or outsourced outpatient or inpatient care in years 2 and 3 following the index date (e.g., through FY2016 for those with an index date in FY2013). In sex-stratified analyses, we conducted logistic regression analyses to predict the odds of attrition as a function of fiscal year of the index date, with adjusted regression analyses controlled for age, race/ethnicity, and service-connected disability status. RESULTS: The final cohort of new veteran VA outpatients included n = 131,313 women and n = 1,753,240 men. Among new women patients, the attrition rate was 23% in FY2007 and 17% in FY2013; among new men it was 23% in FY2007 and 21% in FY2013. Among women, the odds of attrition were lower for those with an FY2013 index date than for those with an FY2007 index date (adjusted OR 0.72 [0.68-0.76]). Reduction of attrition was less pronounced for men over this same time period (adjusted OR 0.89 [0.88-0.91]). Focusing on changes that occurred after the rollout of the women's health policy, attrition decreased post-FY2010 among women but not among men. Specifically, women with an FY2013 index date had lower odds of attrition than those with an FY2010 index date (adjusted OR 0.92 [0.88-0.98]), but attrition increased over that time period for men (adjusted OR 1.05 [1.04-1.07]). CONCLUSIONS: Although causal conclusions are not possible, findings from this national longitudinal study suggest a beneficial impact of the 2010 Comprehensive Women's Health Policy upon attrition from VA care among new women veteran VA patients. The effects of more recent legislation expanding veterans' health care access deserve attention, to confirm whether in this changed landscape, women veterans continue to remain in VA for their health care.