Discontinuation of government subsidized HIV pre-exposure prophylaxis in Australia: a whole-of-population analysis of dispensing records

Introduction: HIV pre-exposure prophylaxis (PrEP) has been government subsidized in Australia since April 2018 and while uptake is high among men who have sex with men, rates of discontinuation are also high. The aims of this study were to examine the impact of discontinuation on overall PrEP usage, the proportion of PrEP users who discontinue and the predictors of discontinuation. Methods: We used linked de-identified dispensing records of all government subsidized PrEP in Australia between April 2018 and September 2021: a whole-of-population data set. Defining discontinuation as 180 days or more without PrEP after the final dispensed supply, we calculated the number of people who discontinued at each 6-month interval during the study period, the proportion who had discontinued 2 years after the first supply and, using Cox regression, predictors of discontinuation. Results: Of 49,164 people dispensed PrEP (98.5% male, median age 34 years), 40.3% (19,815) had discontinued by September 2021. Within 2 years of their first supply, 11,150 (37.7%) of 29,549 PrEP users had discontinued, including 10.0% after a single dispensed supply. Large variations were observed, particularly according to prescriber characteristics: discontinuation was higher among people prescribed PrEP by low caseload (≤10 patients) prescribers (61.2%) than by high caseload (>100 patients) prescribers (31.1%, p<0.001), and by prescribers practising in areas with low estimated prevalence (<1.0%) of gay men (64.1%) than high (>5%) prevalence (36.7%, p<0.001). Women and younger people were more likely to discontinue, while patients receiving a higher level of government subsidy were less likely. The independent predictors of discontinuation with the greatest effect size were female sex (adjusted hazards ratio [aHR] 2.99, p<0.001), low estimated gay prevalence of prescriber location (aHR 1.98, p<0.001) and low prescriber PrEP caseload (aHR 1.79, p<0.001). Conclusions: There are high rates of PrEP discontinuation in Australia and some populations are at increased risk of discontinuation. Strategies are needed to support persistence on PrEP and the re-starting of PrEP during periods of risk.