Key Points

Question听 Is medication abortion provided using no-test eligibility screening and mailing of medications as effective as in-person care with ultrasonography and safe overall?

Findings听 This prospective, observational study found that medication abortion obtained following no-test telehealth screening and mailing of medications was associated with similar rates of complete abortion compared with in-person care with ultrasonography (94.4% vs 93.3%, respectively) and met the 5% noninferiority margin, with a low prevalence of adverse events.

Meaning听 The findings suggest that models of medication abortion care that rely on no-test eligibility assessment and mailing medications to patients are an option for individuals considering medication abortion.

Importance听 In the US, access to medication abortion using history-based (no-test) eligibility assessment, including through telehealth and mailing of mifepristone, has grown rapidly. Additional evidence on the effectiveness and safety of these models is needed.

Objective听 To evaluate whether medication abortion with no-test eligibility assessment and mailing of medications is as effective as in-person care with ultrasonography and safe overall.

Design, Setting, and Participants听 Prospective, observational study with noninferiority analysis. Sites included 4 abortion-providing organizations in Colorado, Illinois, Maryland, Minnesota, Virginia, and Washington from May 2021 to March 2023. Eligible patients were seeking medication abortion up to and including 70 days鈥� gestation, spoke English or Spanish, and were aged 15 years or older.

Exposure听 Study groups reflected the model of care selected by the patient and clinicians and included: (1) no-test (telehealth) eligibility assessment and mailing of medications (no-test鈥�+鈥塵ail) (n鈥�=鈥�228); (2) no-test eligibility assessment and pickup of medications (no-test鈥�+鈥塸ickup) (n鈥�=鈥�119); or (3) in-person with ultrasonography (n鈥�=鈥�238).

Main Outcomes and Measures听 Effectiveness, defined as a complete abortion without the need for repeating the mifepristone and misoprostol regimen or a follow-up procedure, and safety, defined as an abortion-related serious adverse event, including overnight hospital admission, surgery, or blood transfusion. Outcomes were derived from patient surveys and medical records. Primary analysis focused on the comparison of the no-test鈥�+鈥塵ail group with the in-person with ultrasonography group.

Results听 The mean age of the participants (N鈥�=鈥�585) was 27.3 years; most identified as non-Hispanic White (48.6%) or non-Hispanic Black (28.1%). Median (IQR) gestational duration was 45 days (39-53) and comparable between study groups (P鈥�=鈥�.30). Outcome data were available for 91.8% of participants. Overall effectiveness was 94.4% (95% CI, 90.7%-99.2%) in the no-test鈥�+鈥塵ail group and 93.3% (95% CI, 88.3%-98.2%) in the in-person with ultrasonography group in adjusted models (adjusted risk difference, 1.2 [95% CI, 鈭�4.1 to 6.4]), meeting the prespecified 5% noninferiority margin. Serious adverse events included overnight hospitalization (n鈥�=鈥�4), blood transfusion (n鈥�=鈥�2), and emergency surgery (n鈥�=鈥�1) and were reported by 1.1% (95% CI, 0.4%-2.4%) of participants, with 3 in the no-test鈥�+鈥塵ail group, 3 in the in-person with ultrasonography group, and none in the no-test鈥�+鈥塸ickup group.

Conclusions and Relevance听 This prospective, observational study found that medication abortion obtained following no-test telehealth screening and mailing of medications was associated with similar rates of complete abortion compared with in-person care with ultrasonography and met prespecified criteria for noninferiority, with a low prevalence of adverse events.

Medication abortion with mifepristone and misoprostol was first approved by the US Food and Drug Administration (FDA) in 2000 and is currently approved for use through 70 days鈥� gestation in pregnancy.¹ In 2023, medication abortion accounted for 63% of abortions in the US, an increase from 53% in 2020 and 31% in 2014.^2-4 Multiple studies have firmly established the safety and effectiveness of the mifepristone and misoprostol regimen for medication abortion; the incidence of major complications or serious adverse events (SAEs) is less than 0.5% and overall effectiveness in ending the pregnancy without the need for a follow-up procedure is 95% to 97%.^5-10

Quiz Ref IDIn recent years, access to various medication abortion models has expanded rapidly, including telehealth or virtual eligibility assessment and counseling, mailing medications directly to patients, and, in some instances, forgoing the requirement for an in-person ultrasonography or physical examination to establish location and duration of pregnancy prior to provision of medication abortion.¹¹ The FDA鈥檚 removal of the in-person dispensing requirement for mifepristone, which had long been part of the drug鈥檚 Risk Evaluation and Mitigation Strategy, facilitated this expansion. Emerging research evaluating the safety and effectiveness of telehealth and in-person, no-test models^12-14 builds on prior evidence demonstrating that provision of medication abortion by telemedicine (with ultrasonography and/or physical examination) has a comparable safety and effectiveness profile to in-person provision.^15-19 A 2022 retrospective cohort study of 2379 pregnant people accessing medication abortion using history-based (no-test) eligibility assessment in the US reported 94.5% effectiveness in ending the pregnancy and rate of SAEs of 0.42%.²⁰ Similarly, a large, retrospective cohort study of over 50鈥�000 patients receiving medication abortion in the UK reported effectiveness rates of 98% to 99%, with no difference between those accessing care with and without routine pretreatment ultrasonography.²¹

Further evidence on the effectiveness and safety of telehealth and in-person, no-test (without ultrasonography or physical examination) medication abortion informs expansion of these models to all pregnant people seeking abortion, particularly those now living in the 21 states where abortion is banned or heavily restricted²² or the 6 states where abortion remains legal but telehealth or mailing of medications is banned.²³ The present study compares the effectiveness and safety of medication abortion provided using 2 models of no-test, telehealth abortion care (mailed or pickup medications) with in-person care with ultrasonography.

This was an observational, prospective study comparing medication abortion effectiveness and safety using different methods of eligibility assessment (ultrasonography vs no-test) and dispensing of medications (mail vs in-person). The study protocol, including the statistical analysis plan, was published at Open Science Framework in April 2022 prior to initiating any analyses.²⁴

Participants were recruited from 4 abortion-providing organizations in 6 states (Colorado, Illinois, Maryland, Minnesota, Virginia, and Washington) between May 2021 and March 2023. We selected organizations that offered direct-to-patient telehealth or other no-test, history-based eligibility screening models that could dispense mifepristone using a mail-order pharmacy and, to allow for an appropriate comparison group, that also offered in-person provision at a facility within the same clinic network and state. One organization had facilities in 3 states and offered telehealth services to patients in those same states through a centralized call center. Thus, there were a total of 6 facilities, 1 in each state (eFigure 1 in Supplement 1). Eligibility was restricted to patients aged 15 years or older (3 facilities) or aged 18 years or older (3 facilities), English or Spanish speakers, seeking and eligible for medication abortion at that facility, and intending to take the medication up to and including 70 days鈥� gestation. Participants were typically approached by facility staff during initial counseling; those interested in participating received a longer description of the study and then completed electronic informed consent and signed a Health Insurance Portability and Accountability Act waiver allowing clinic staff to review the patients鈥� medical records with the study team. Patients then shared contact information for survey follow-up and received a $15 gift card for enrolling.

Study data were obtained from 2 sources. First, participants were asked to complete electronic surveys 1 day, 2 weeks, and 1 month after receiving the medications. Survey links were sent via text or email and included up to 5 reminders. Surveys were designed such that participants with an unclear abortion outcome (eg, who responded 鈥淚鈥檓 not sure鈥� if the medications worked to end the pregnancy) were sent repeat surveys until an outcome was ascertained or the patient was considered lost to follow-up. Participants were remunerated with gift cards worth $25 for completing the first and third surveys and $15 for completing the second survey.

Clinics abstracted medical record data using a standardized tool²⁵ 6 weeks after the initial appointment. If care was ongoing, clinics continued to abstract clinical data until the outcome of the abortion was known or adverse events (AEs) had resolved. In-depth, follow-up chart review was conducted for all cases with an unclear abortion outcome or potential AE.

The primary exposure was model of medication abortion care received. There were 3 groups: (1) no-test (telehealth) eligibility assessment by history and mailing of medications (no-test鈥�+鈥塵ail); (2) no-test eligibility assessment (via telehealth or in-person) by history and pickup of medications (no-test鈥�+鈥塸ickup); or (3) in-person assessment with ultrasonography and in-person dispensing of medications (in-person).

The primary outcome was effectiveness of medication abortion defined as a complete abortion without need for procedural abortion or repeating the mifepristone and misoprostol regimen.²⁶ We relied on a combination of patient self-report (survey) and clinic (chart-abstracted) data to determine effectiveness. Survey questions included patients鈥� self-assessment of whether they 鈥渂elieve[d] that the abortion pills worked to end the pregnancy,鈥� had taken a pregnancy test and its result, 鈥渉ad a suction procedure (vacuum aspiration or dilation and curettage [D&C]) to complete the abortion,鈥� or had been 鈥済iven another dose of the first abortion pill (mifepristone)鈥� as part of follow-up care. Chart-abstracted data included the clinician鈥檚 assessment of the medication abortion outcome at last contact with the patient (complete, incomplete, ongoing pregnancy, unknown), any tests performed to document pregnancy outcome (ie, ultrasonography, physical examination, review of symptoms, urine pregnancy test, serum human chorionic gonadotropin testing), and details on any follow-up care received or reported to the clinic. A flowchart was developed a priori to resolve discordant patient self-report and chart-abstracted outcomes (eFigure 2 in Supplement 1). Clinic staff completing abstraction were not blind to the study group because this information was in the patient鈥檚 medical record.

A secondary outcome was safety, defined as reporting of any abortion-related SAE. This included overnight hospital admission, surgery, blood transfusion, or death. Surveys assessed whether patients had 鈥済o[ne] to a hospital, emergency department, doctor鈥檚 office, or clinic (other than your regularly scheduled follow-up visit)鈥� since taking the mifepristone and, if so, what type of care they received (including 鈥淚 was given a blood transfusion鈥� or 鈥淚 stayed overnight in the hospital鈥�). Chart-abstracted data included any documented care received at an outside facility as well as details on the care received. We searched across all chart-abstracted data fields for a list of terms related to the defined SAEs to ensure all potential cases were identified. All open-ended chart data fields were also reviewed by a physician (D.G.) to determine whether an SAE was reported.

We collected sociodemographic and pregnancy characteristics prespecified as potential confounders, including participants鈥� age, race and ethnicity, gestational duration, parity, and prior abortion. Race and ethnicity data were based on 2 prespecified questions regarding Hispanic, Latine, or Spanish ethnicity (yes or no) and self-reported race (Asian, Black or African American, Middle Eastern or North African, Native American or Other Pacific Islander, White, or other), with the option to select all that apply and/or write in a response. Gestational duration at initial appointment collected by facility staff was estimated using date of last menstrual period for the telehealth groups and ultrasonography for the in-person with ultrasonography group. For race and ethnicity, prior abortion, and parity, we relied on participants鈥� survey responses; when possible, missing responses were completed using medical record data.

We described the distribution of sociodemographic and pregnancy covariates in the study sample overall. We then tested for differences in the distribution of these covariates as well as the proportion with abortion outcome data by study group using bivariable mixed effects linear or multinomial regression models adjusted for clustering of observations by facility as well as standardized mean differences. Although our prespecified statistical analysis plan describes adjusting for clustering by site, our subsequent inclusion of facilities from the same organization but located in different states led to a preference for adjustment for clustering at the state and facility level, in the spirit of the prespecified plan. However, to ensure robustness of findings, we also included sensitivity analyses with random effects for organization (n鈥�=鈥�4).

Among enrolled participants who took mifepristone or misoprostol and had abortion outcome data available, the unadjusted proportion (with 95% CI) with a complete abortion by study group was calculated. For the primary, prespecified analysis, we examined noninferiority of the telehealth (no-test) with mailed medications care group compared with the in-person care with ultrasonography group. From a logistic mixed effects model with a random effect for facility, we calculated the unadjusted risk difference (and 95% CI) in effectiveness between telehealth (no-test) with mailed medications and in-person care. We then reran the same model with the addition of prespecified confounders, including age and gestational duration, as well as any variables found to differ by study group in bivariable analysis. Telehealth (no-test) with mailed medications care was considered noninferior if the lower side of the 2-sided CI was above 鈭�5%, chosen to be consistent with prior research.^27-29 As secondary, prespecified analyses, we then repeated these steps to examine noninferiority of no-test鈥�+鈥塸ickup (vs in-person) and either no-test (vs in-person with ultrasonography) care. We also descriptively examined the proportion (with 95% CI) in each study group that experienced an SAE and, when possible, differences in proportions by study group from adjusted models.

Prespecified analyses focused on a complete-case analysis. However, to ensure robustness of study findings, we repeated our primary, prespecified analysis using multiple imputation by chained equations. We imputed missing data for all covariates and the primary outcome (effectiveness) using all covariates in the original adjusted models as well as receipt of public assistance, educational attainment, and whether they completed a survey. We then used Stata鈥檚 MI estimate command and a logistic regression model using imputed values and the mimrgns command to estimate the risk difference between study groups (StataCorp). We utilized an inverse probability of treatment-weighted regression adjustment approach to balance covariates between study groups and re-estimate the risk difference. Specifically, logistic regression was used to calculate the propensity score or probability of treatment (telehealth鈥�+鈥塵ail). We then calculated weights as 1/propensity score for the treated group and 1/1-propensity score for the untreated (in-person with ultrasonography) and reran a logistic regression model applying this weight. Finally, to allow for interpretation of marginal, population-level effects, we repeated the primary analysis using a logistic regression model with robust standard errors for facility.

We determined the minimum sample size required to establish noninferiority of telehealth (no-test) with mailed medications care to in-person care for the primary effectiveness outcome. Assuming 20% loss to follow-up, moderate correlation of observations by site (increasing sample size by an additional 15%), and effectiveness of 95%,^7,30 a sample size of 170 participants per study group, or 510 participants overall, gave 80% power with a 1-sided 伪 of .025 to establish noninferiority of telehealth (no-test) with mailed medications care. The study was not powered to detect noninferiority of different models with respect to safety.

All study activities were approved by the University of California San Francisco institutional review board in January 2021 (#20-32514). This manuscript was prepared in accordance with guidance.³¹

A total of 1107 patients were approached, of whom 878 were eligible; 588 consented and enrolled (Figure). There were no differences in mean age (P鈥�=鈥�.38), appointment type (P鈥�=鈥�.72), or gestational duration (P鈥�=鈥�.51) between those eligible who enrolled vs did not enroll.

Among 588 enrolled patients, 3 were excluded from the analysis because they did not receive (n鈥�=鈥�1) or take (n鈥�=鈥�1) pills or because they sought emergency care for an ectopic pregnancy before picking up the pills (n鈥�=鈥�1), resulting in a final sample of 585. Abortion outcome data were available for 537 of 585 (91.8%) of the participants; the proportion with abortion outcome data did not differ by study group (Table 1). Further, there were no differences by sociodemographic and pregnancy characteristics in the proportion missing abortion outcome data (eTable 1 in Supplement 1).

Median (IQR) participant age was 26 years (23-31). Nearly one-half (48.6%) were non-Hispanic White, with fewer non-Hispanic Black (28.1%), non-Hispanic other, or multiple races (12.1%) or Hispanic (9.6%). Approximately 4 in 10 participants reported a prior abortion (42.6%). Median (IQR) gestational duration at screening was 45 days (39-53) and did not differ by study group (P鈥�=鈥�.39). In the no-test鈥�+鈥塸ickup medications study group, two-thirds (66%) completed eligibility assessment virtually and one-third (33%) completed no-test eligibility assessment in-person at a clinic.

Overall, 510 of 537 participants had a complete abortion without a procedure or repeating the mifepristone and misoprostol regimen, for an unadjusted effectiveness rate of 95.0% (95% CI, 92.8%-96.7%). Among the 27 participants classified as having an incomplete abortion, 18 had a follow-up procedure to complete the abortion, 6 reported a positive pregnancy test or uncertainty about whether the pills worked and were then lost to follow-up, 2 had a confirmed or suspected ectopic pregnancy, and 1 reported receiving outside care to complete the abortion to the study clinic. One ectopic pregnancy was confirmed by the patient, who reported to the clinic that they had sought care at an emergency department after taking the medications and had an operation to remove a fallopian tube. The other participant was referred to a hospital for evaluation and treatment of a suspected ectopic pregnancy after reporting no bleeding for more than 24 hours after taking misoprostol.

Unadjusted effectiveness was 94.1% (95% CI, 90.1%-96.8%) in the in-person group, 95.1% (95% CI, 91.2%-97.6%) in the no-test鈥�+鈥塵ail group, and 96.5% (95% CI, 91.2%-99.0%) in the no-test鈥�+鈥塸ickup group. Quiz Ref IDAfter adjustment for covariates including age, race and ethnicity, and gestational duration, the risk difference between the in-person and no-test鈥�+鈥塵ail groups was 1.2% (95% CI, 鈭�4.1% to 6.4%), above the limit of 鈭�5% and indicating noninferiority of telehealth (no-test) with mailed medications care. In this adjusted model, effectiveness was 94.4% (95% CI, 89.6%-99.2%) in the no-test鈥�+鈥塵ail group and 93.3% (95% CI, 88.3%-98.2%) in the in-person group. Similarly, the adjusted risk difference between the in-person and no-test鈥�+鈥塸ickup care groups was 1.7% (95% CI, 鈭�4.6% to 8.1%), indicating noninferiority of no-test鈥�+鈥塵edications pickup care. In this adjusted model, effectiveness was 95.0% (95% CI, 88.8%-100.0%) in the no-test鈥�+鈥塸ickup care group and 93.3% (95% CI, 88.8%-97.8%) in the in-person group (Table 2).

In sensitivity analyses, rerunning models with imputed missing observations, with application of inverse probability of treatment weights, and with adjustment for clustering of observations by organization (vs facility), results were consistent and demonstrated noninferiority of the no-test model (eTable 2 in Supplement 1).

Overall, 6 of 537 participants experienced an SAE, for a rate of 1.1% (95% CI, 0.4%-2.4%). Events included 4 overnight hospitalizations, 2 blood transfusions, and 1 surgery (for treatment of ectopic pregnancy). Quiz Ref IDThe unadjusted rate of SAEs was 1.5% (95% CI, 0.3%-4.2%) in the telehealth鈥�+鈥塵ail group (n鈥�=鈥�3) and 1.4% (95% CI, 0.3%-3.9%) in the in-person group (n鈥�=鈥�3); no SAEs were reported in the no-test鈥�+鈥塸ickup care group. In adjusted logistic models with robust standard errors for state, the adjusted risk difference between the in-person with ultrasonography and no-test鈥�+鈥塵ail care group was 0.4 (95% CI, 鈭�1.0 to 1.7), indicating noninferiority of the no-test鈥�+鈥塵ail model (eTable 3 in Supplement 1).

In this prospective, observational study of 585 people obtaining medication abortion, it was found that medication abortion following no-test telehealth screening and mail-order pharmacy dispensing of medications was associated with similar rates of complete abortion as in-person care with ultrasonography, met the prespecified threshold for noninferiority, and had a low rate of AEs overall. Further, medication abortion following no-test screening with pickup of medications was also associated with similar outcomes as in-person care with ultrasonography and had low rates of AEs overall. Levels of effectiveness with models of care incorporating telehealth and eligibility assessment in this study were comparable to large 2022 and 2024 studies of no-test telehealth patients.^20,32 This study contributes new evidence pertaining to no-test telehealth models in direct comparison with a concurrent control group of people receiving care from similar facilities through in-person care with ultrasonography. Taken together, these findings indicate that no-test and telehealth models with mailing of medications may be offered as an option to people seeking abortion. Although geographic and logistical barriers have always delayed and complicated pregnant people鈥檚 access to abortion,^33,34 these barriers have become increasingly problematic following the Supreme Court鈥檚 June 2022 decision revoking federal protections on abortion.^35,36 Thus, expansion of telehealth and no-test or history-based models offers a safe and effective, as well as urgently needed, way to overcome at least some of the logistical and geographic barriers to accessing abortion.³⁷

Study findings may also be used to refute ongoing efforts to further regulate and restrict access to medication abortion, most notably in the Alliance for Hippocratic Medicine v FDA case currently before the US Supreme Court. One of the central issues being considered in this case is the quality of the evidence the FDA relied on when it removed the in-person dispensing requirement for mifepristone in 2021.³⁸ Although the data were robust in 2021 when the FDA performed its review, this current analysis, as well as other research published since then,^20,32 further strengthens the evidence documenting the safety and effectiveness of medication abortion provided using no-test telehealth screening for eligibility and dispensing the medications by mail.

This study has limitations. First, approximately 8% of the participants were lost to follow-up and did not complete a survey or have any follow-up contact with the clinic. Although this follow-up rate represents an improvement over prior research that relied only on chart-abstracted outcomes,^12,13,19,20 the possibility that those who were lost to follow-up had worse outcomes remains. Second, only patient-reported outcomes up to 4 weeks after initiation of treatment were collected; it is possible that additional treatment could have been accessed after that point. Third, as an observational study in which patients could select which model of medication abortion care they received, some differences in the sociodemographic and pregnancy characteristics by study group were expected and therefore the study relied on adjustment for these characteristics to account for them in analyses. Reassuringly, differences were not observed in gestational duration by study group, which is likely most strongly correlated with outcomes of safety and effectiveness,⁷ and models with imputed data produced comparable findings. However, confounding on unobserved covariates always remains a possibility in observational data. Fourth, this study sample size was not powered to detect noninferiority of, or differences in, safety of different models of care. Although few SAEs (n鈥�=鈥�6 [1.1%]), including ectopic pregnancy, were observed, the estimate is qualitatively higher than a large 2022 retrospective study of patients receiving medication abortion through history-based screening, which found that 0.5% experienced an SAE.²⁰ Inclusion of patient self-reported outcomes, including overnight hospitalization, in defining SAEs likely contributed to this. Notably, the concurrent control group allowed direct comparison of the rate of SAEs between those receiving no-test telehealth medication abortion with mailing of medications vs in-person care with ultrasonography.

This prospective, observational study found that medication abortion following no-test telehealth screening and mailing of medications was associated with similar rates of abortion completion compared with in-person care with ultrasonography, meeting the prespecified threshold for noninferiority and with a low prevalence of adverse events.

Accepted for Publication: May 17, 2024.

Published Online: June 24, 2024. doi:10.1001/jama.2024.10680

Corresponding Author: Lauren J. Ralph, PhD, MPH, Advancing New Standards in Reproductive Health (ANSIRH), University of California San Francisco, 1330 Broadway, Ste 1100, Oakland, CA 94612 (Lauren.Ralph@ucsf.edu).

Author Contributions: Dr Ralph had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Ralph, Baba, Biggs, Grossman.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Ralph, Baba.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Ralph, Baba, Biggs.

Obtained funding: Grossman.

Administrative, technical, or material support: Ralph, Baba, McNicholas, Hagstrom Miller, Grossman.

Supervision: Grossman.

Conflict of Interest Disclosures: Dr McNicholas reported receiving grants from University of California San Francisco (UCSF) as co-investigator during the conduct of the study and being the Chief Medical Officer for Planned Parenthood of the St. Louis Region. Ms Hagstrom Miller reported receiving grants from UCSF during the conduct of the study. Dr Grossman reported receiving personal fees from The Lawyering Project and Planned Parenthood Federation of America for serving as an expert witness in cases challenging abortion restrictions, including restrictions on telemedicine. No other disclosures were reported.

Funding/Support: This study was supported by grants from the Demartini Family Foundation and The Susan Thompson Buffett Foundation.

Role of the Funder/Sponsor: The Demartini Family Foundation and The Susan Thompson Buffett Foundation had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: The authors express deep gratitude to members of this study鈥檚 Community Advisory Board for their input on survey questions and study design. Members include Claudie Kiti Bustamante, BA (Black Women for Wellness); Sung Yeon Choimorrow, MDiv (National Asian Pacific American Women's Forum); Debra Hauser, MPH (Advocates for Youth); Yamani Hernandez, MArch (National Network of Abortion Funds); Tammi Kromenaker, BSW (Red River Women's Clinic); Ghazaleh Moayedi, DO (Pegasus Health Justice Center); Jamila Perritt, MD (Physicians for Reproductive Health); and Ena Valladares, MPH (California Latinas for Reproductive Justice). We are also grateful to John Boscardin, PhD, for statistical guidance; Rosalyn Shroeder, MSc, for analysis support; and Rana Barar, MPH, Sabrina Serrano, BA, Tanvi Gurazada, MS, Gillian Borges, BA, and Jessica Navarro, MPH, for study support activities (University of California San Francisco).