Introduction: Pre-exposure prophylaxis (PrEP) is highly effective, although PrEP adherence and persistence has been variable during real world implementation. Little is known about missed opportunities to enhance PrEP adherence among individuals who later HIV seroconverted after using PrEP. The goal of this analysis was to identify all HIV infections among individuals who had accessed PrEP in an integrated health system in San Francisco, and to identify potentially intervenable factors that could have prevented HIV infection through in-depth interviews with people who HIV seroconverted after using PrEP. / Methods: We identified individuals who initiated PrEP in an integrated safety-net public health system and performed indepth chart review to determine person-time on and after stopping PrEP over six years. We identified all PrEP seroconversions using the Centers for Disease Control and Prevention's Enhanced HIV/AIDS Reporting System and then calculated HIV incidence while using PrEP and during gaps in use. We then performed in-depth interviews with those who seroconverted. / Results: Overall, 986 initiated PrEP across the San Francisco Department of Public Health from July 2012 to November 2018. Data were gathered from 895 person-years on PrEP and 953 after stopping PrEP. The HIV incidence was 7.5-fold higher after stopping PrEP compared to while on PrEP (95% CI 1 to 336). Of the eight individuals who HIV seroconverted; only one was taking PrEP at the time of seroconversion but was using on-demand PrEP inconsistently. All eight agreed to qualitative interviews. Major barriers to PrEP persistence included substance use, mental health and housing loss; difficulty accessing PrEP due to cost, insurance, and the cost and time of medical visits; difficulty weighing PrEP's benefit versus selfperceived risk; and entering a primary partnership. The individual who developed HIV using on-demand PrEP reported confusion about the dosing regimen and which sexual encounters required accompanying PrEP dosing. / Conclusions: HIV incidence during gaps in PrEP use was nearly eight-fold higher than while on PrEP in this large cohort in San Francisco. Many individuals who stop PrEP remain at risk of HIV, and participants reported that proactive outreach could potentially have prevented HIV infections. Individuals using non-daily PrEP may require additional education and support in the United States.