Impact of COVID-19 on HIV Preexposure Prophylaxis Prescriptions in the United States – A Time Series Analysis

Abstract Background Uptake of HIV preexposure prophylaxis (PrEP) has been increasing in the United States since its FDA approval in 2012; however, the COVID-19 pandemic may have affected this trend. Our objective was to assess the impact of the COVID-19 pandemic on PrEP prescriptions in the United States. Methods We analyzed data from a national pharmacy database from January 2017 through March 2021 to fit an interrupted time-series model that predicted PrEP prescriptions and new PrEP users had the pandemic not occurred. Observed PrEP prescriptions and new users were compared with those predicted by the model. Main outcomes were weekly numbers of PrEP prescriptions and new PrEP users based on a previously developed algorithm. The impact of the COVID-19 pandemic was quantified by computing rate ratios and percent decreases between the observed and predicted counts during 3/15/2020 – 3/31/2021. Results In the absence of the pandemic, our model predicted that there would have been 1,058,162 PrEP prescriptions during 3/15/2020 – 3/31/2021. We observed 825,239 PrEP prescriptions, a 22.0% reduction (95% CI: 19.1%-24.8%) after the emergency declaration. The model predicted 167,720 new PrEP users during the same period; we observed 125,793 new PrEP users, a 25.0% reduction (95% CI: 20.9%-28.9%). The COVID-19 impact was greater among younger persons and those with commercial insurance. The impact of the pandemic varied markedly across states. Conclusion The COVID-19 pandemic disrupted an increasing trend in PrEP prescriptions in the United States, highlighting the need for innovative interventions to maintain access to HIV prevention services during similar emergencies.

A c c e p t e d M a n u s c r i p t 3 BACKGROUND On March 13, 2020, the President of the United States declared a national emergency in response to the outbreak of coronavirus disease 2019  in the United States caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. Many states and localities issued mandatory "stay-at-home" or "shelter-in-place" orders and other protective measures in an effort to reduce the spread of SARS-CoV-2 [2][3][4]. The Centers for Disease Control and Prevention (CDC) also recommended individuals and groups practice social distancing to reduce exposure to SARS-CoV-2 [5]. These policies, as well as individuals' fear of COVID-19 exposure, resulted in decreased use of health services, especially preventive and elective health care [6,7].
HIV preexposure prophylaxis (PrEP) with daily oral antiretroviral medications is a safe and effective intervention that reduces the risk of HIV acquisition among men who have sex with men (MSM), heterosexual men and women, and persons who inject drugs [8,9]. In 2012, the U.S. Food and Drug Administration (FDA) approved tenofovir disoproxil fumarate combined with emtricitabine (FTC/TDF) as PrEP [10]. The CDC published clinical PrEP practice guidelines in 2014 and updated guidelines in 2017 and in 2021 [11][12][13]. Nondaily event-driven PrEP (also called "2-1-1" PrEP), while not an FDA-approved regimen, has been prescribed and used among selected patients, as two clinical trials have demonstrated its HIV prevention efficacy among MSM [13]. The number of persons prescribed PrEP had been increasing since its approval. Compared with the estimated 1.1 million persons with indications for PrEP in the United States, approximately 280,000 (23%) were prescribed PrEP in 2019 [14]. In October 2019, the FDA approved a second drug for PrEP -tenofovir alafenamide combined with emtricitabine (FTC/TAF) [15]. About one third of existing PrEP users switched to the newer formulation within 12 months of its approval [16]. Several generic formulations of FTC/TDF were also approved by FDA in the fall of 2020.
In the absence of the COVID-19 pandemic, the trend in national PrEP prescriptions was expected to A c c e p t e d M a n u s c r i p t 4 continue to increase. However, when shelter-in-place and social distancing orders were issued, many healthcare providers temporarily closed their practice, limited it to providing urgent care, or provided telemedicine services [17,18]. The COVID-related closures as well as individuals' concerns about potential exposure to SARS CoV-2 likely affected the use of PrEP services. Persons who choose to initiate PrEP need to be assessed by a clinician for existing HIV, sexually transmitted infections (STIs), hepatitis B, hepatitis C, and their renal function. Persons who have been taking PrEP are recommended to have monitoring health care visits every 3 months for assessment of ongoing risk of HIV acquisition, PrEP adherence and persistence counseling, and laboratory testing for HIV, STIs, and renal function [12]. The pandemic also likely affected adherence to these recommended clinical guidelines. The objective of this study was to quantify the impact of the COVID-19 pandemic on PrEP prescriptions and new PrEP users in the United States by analyzing a national pharmacy database.

Data Source
We analyzed data from the IQVIA Real World Data-Longitudinal Prescriptions Database (hereafter, IQVIA database) from January 2017 to March 2021. The IQVIA database captures prescriptions from all payers and represents approximately 92% of all prescriptions dispensed from retail pharmacies and 60-86% from mail order outlets in the United States [19]. The database does not include prescriptions from closed healthcare systems such as health maintenance organizations or the Veterans Administration. Prescriptions in the IQVIA database are linked to medical claims to identify associated diagnoses, and to the Experian consumer database to identify patient demographic characteristics. Race/ethnicity data were available for <40% of persons prescribed PrEP.

Measures
We identified PrEP prescriptions in the IQVIA database from January 1, 2017 to March 31, 2021 using a previously developed and validated algorithm [20][21][22]. We Identified all FTC/TDF, FTC/TAF  The outcome measures were reported by patient sex, age group, geographic region, payer type, and race/ethnicity. Payer type at the person-level was calculated based on a hierarchical variable, constructed using a payer hierarchy of public insurance (Medicaid/CHIP and Medicare), commercial insurance, cash, and other. The other payer type category included coupon/voucher programs, discount card programs, and state or manufacturer medication assistance programs (MAPs). We also stratified the outcomes by state. A state was identified using 3-digit ZIP codes of patients' residential location in the IQVIA database.

Analysis
Weekly PrEP prescription data before and after the start of the COVID-19 pandemic were modeled as an interrupted time series using a generalized linear quasi-Poisson model adjusted for seasonality.
This approach models the sequence of repeated weekly observations which is interrupted by an event, in this case the start of the COVID-19 pandemic, occurring at a known timepoint. The impact

National Trends
The  Table 1).
The observed and expected numbers and percent decreases are broken down by month ( Table 1).
The monthly percent reduction in the number of PrEP prescriptions was mostly around 20% throughout the observation period, compared to the expected number. We observed a 17.4% A c c e p t e d M a n u s c r i p t 7 reduction in the number of PrEP prescriptions in June 2020, and decreases greater than 25% in December 2020, and February and March 2021. We observed a 39.5% and 34.2% reduction in the number of new PrEP users in April and May 2020, compared to the expected number of new PrEP users. New PrEP users rebounded in June 2020 with only a 16.5% reduction, then we observed wider gaps until October 2020. In December 2020, we observed a 30.9% reduction in the number of new PrEP users, but after then the gaps were around 20%.

DISCUSSION
We found a 22% decrease in the total number of PrEP prescriptions and a 25% decrease in the total number of new PrEP users between March 2020 and March 2021 compared to predicted numbers assuming the COVID-19 pandemic shutdown had never occurred. We observed a partial rebound in the number of new users in June 2020, but then followed by declining numbers towards the end of 2020. We observed another rebound in the number of new users after December 2020, when the COVID-19 vaccines became available.
Our finding of reductions in PrEP prescriptions was consistent with other studies that found declines in the use of preventive and elective healthcare services [6,7]. PrEP requires adherent and persistent use for its effectiveness as a biomedical tool for HIV prevention. Persons who stopped taking PrEP but had ongoing risk behaviors during the pandemic might have acquired and subsequently transmitted HIV infection. At least one study found that HIV testing rates decreased substantially during the COVID-19 pandemic [23], which may be partially due to decreases in PrEP A c c e p t e d M a n u s c r i p t 9 prescriptions. HIV testing is an important part of integrated PrEP services, that is, PrEP users are required to have a negative HIV test result prior to initiating PrEP, and testing is recommended every 3 months at follow-up visits before a new prescription is provided for PrEP continuation. Decreases in PrEP initiation and ongoing PrEP prescriptions resulted in fewer HIV tests, as well as fewer opportunities to diagnose HIV.
The pandemic caused more disruption in new PrEP prescriptions among younger persons. PrEP coverage was lower among persons in younger age groups prior to the COVID pandemic, and it decreased even more during the pandemic shutdown [14]. Young persons are typically less likely to adhere to and persist with daily medications [24][25][26]. In addition, young persons might have had less access to care during the pandemic compared to older persons [27,28], likely because older persons had established relationships with health care providers prior to the shutdown. Decreased PrEP uptake and persistence due to lack of access to care, along with lack of perceived HIV risk, might have resulted in increased HIV transmission risk among persons in younger populations [29].
Innovative interventions such as risk assessment tools, educational messages, PrEP provider locator tools, and other resources linked to social media apps could help reach this population to improve their PrEP initiation, adherence, and persistence.
We also observed larger reductions in PrEP use among persons with commercial health insurance. A study that reported on an analysis of the IQVIA database found that out-of-pocket payments for PrEP were lower among persons with Medicaid or Medicare than among those with commercial insurance [30]. With loss of employment and health insurance coverage during the COVID-19 shutdown, high copayments might have been a barrier to PrEP use among persons with commercial insurance or those who paid with cash. Starting January 2021, most health plans were required to offer PrEP to their beneficiaries without copays under the Affordable Care Act (ACA), which can increase access to PrEP by removing financial barriers. Our study has some limitations. We did not capture PrEP prescriptions from closed health systems such as health maintenance organizations. We might have over-or under-estimated the COVID-19 impact on PrEP prescriptions in some states due to low precision in the measurement of some statelevel numbers. Decreased PrEP prescriptions could be due to lack of access to care or decreased risk behavior during the pandemic [33]; we were unable to distinguish between these factors.
Additionally, some users might take event-driven (or 2-1-1) PrEP regimens, even if their refill patterns changed, which could not be observed in this analysis. Race/ethnicity data were available for fewer than 40% of persons prescribed PrEP. The race/ethnicity data that were available in the IQVIA database were from the linked Experian consumer database that likely included larger proportions of persons who were white, older, and with higher incomes. Because of the lack of A c c e p t e d M a n u s c r i p t 11 race/ethnicity data in the IQVIA database for most PrEP users, our finding of no significant differences by race/ethnicity should be interpreted with caution.
In conclusion, our analysis of a national pharmacy database found that the COVID-19 pandemic disrupted an increasing trend in PrEP use in the United States, highlighting the need for innovative interventions to maintain access to HIV prevention services during similar emergencies. Strategies like the expansion of telemedicine and HIV self-testing or self-sample collection can provide access to PrEP care during such emergencies or as a convenient health service option for some PrEP users.   M a n u s c r i p t 16 M a n u s c r i p t 17