Estimated Quality of Life and Economic Outcomes Associated With 12 Cervical Cancer Screening Strategies: A Cost-effectiveness Analysis

Key Points

Question  After incorporating women’s preferences into a cost-effectiveness analysis, what are the estimated quality of life and economic outcomes associated with cervical cancer screening strategies currently recommended in the United States?

Findings  Of 12 strategies evaluated in a cost-effectiveness model, cytologic testing every 3 years for women aged 21 to 29 years with either continued triennial cytologic testing or switching to a low-cost high-risk human papillomavirus test every 5 years from age 30 to 65 years conferred a reasonable balance of benefits, harms, and costs from both a societal and health care sector perspective.

Meaning  Cytologic testing every 3 years and low-cost high-risk human papillomavirus testing every 5 years both may be considered reasonable cervical cancer screening options for women aged 30 to 65 years.

Importance  Many cervical cancer screening strategies are now recommended in the United States, but the benefits, harms, and costs of each option are unclear.

Objective  To estimate the cost-effectiveness of 12 cervical cancer screening strategies.

Design, Setting, and Participants  The cross-sectional portion of this study enrolled a convenience sample of 451 English-speaking or Spanish-speaking women aged 21 to 65 years from September 22, 2014, to June 16, 2016, identified at women's health clinics in San Francisco. In this group, utilities (preferences) were measured for 23 cervical cancer screening–associated health states and were applied to a decision model of type-specific high-risk human papillomavirus (hrHPV)–induced cervical carcinogenesis. Test accuracy estimates were abstracted from systematic reviews. The evaluated strategies were cytologic testing every 3 years for women aged 21 to 65 years with either repeat cytologic testing in 1 year or immediate hrHPV triage for atypical squamous cells of undetermined significance (ASC-US), cytologic testing every 3 years for women age 21 to 29 years followed by cytologic testing plus hrHPV testing (cotesting), or primary hrHPV testing alone for women aged 30 to 65 years. Screening frequency, abnormal test result management, and the age to switch from cytologic testing to hrHPV testing (25 or 30 years) were varied. Analyses were conducted from both the societal and health care sector perspectives.

Main Outcomes and Measures  Utilities for 23 cervical cancer screening–associated health states (cross-sectional study) and quality-adjusted life-years (QALYs) and total costs for each strategy.

Results  Utilities were measured in a sociodemographically diverse group of 451 women (mean [SD] age, 38.2 [10.7] years; 258 nonwhite [57.2%]). Cytologic testing every 3 years with repeat cytologic testing for ASC-US yielded the most lifetime QALYs and conferred more QALYs at higher costs ($2166 per QALY) than the lowest-cost strategy (cytologic testing every 3 years with hrHPV triage of ASC-US). All cytologic testing plus hrHPV testing (cotesting) and primary hrHPV testing strategies provided fewer QALYs at higher costs. Adding indirect costs did not change the conclusions. In sensitivity analyses, hrHPV testing every 5 years with genotyping triage beginning at age 30 years was the lowest-cost strategy when hrHPV test sensitivity was markedly higher than cytologic test sensitivity or when hrHPV test cost was equated to the lowest reported cytologic test cost ($14).

Conclusions and Relevance  Cytologic testing every 3 years for women aged 21 to 29 years with either continued cytologic testing every 3 years or switching to a low-cost hrHPV test every 5 years confers a reasonable balance of benefits, harms, and costs. Comparative modeling is needed to confirm the association of these novel utilities with cost-effectiveness.

An estimated 13 240 women in the United States received a diagnosis of cervical cancer in 2018, and 4170 died from the disease.¹ Although large declines in cervical cancer incidence and mortality in the United States have accompanied widespread screening with cervical cytologic testing, screening options have greatly expanded beyond cytologic testing alone. In 2012, major guideline groups in the United States endorsed 4 cervical cancer screening strategies for women aged 21 to 65 years: cervical cytologic testing every 3 years with 2 options for managing atypical squamous cells of undetermined significance (ASC-US), and cervical cytologic testing every 3 years for women aged 21 to 29 years followed by cytologic testing plus testing for high-risk human papillomavirus (hrHPV) every 5 years for women aged 30 to 65 years with 2 options for managing those with normal cytologic test results and positive results of hrHPV testing.^2-4

In 2014, the US Food and Drug Administration approved a hrHPV test for primary cervical cancer screening in women aged 25 years or older. The American College of Obstetricians and Gynecologists stated that this strategy could be considered as an alternative to current cytologic test–based screening methods⁵ and recommended that interim guidance published by the Society of Gynecologic Oncology be followed regarding its use; these guidelines recommended that screening begin at age 25 years, that rescreening occur “no sooner than every 3 years”^6(p334) and that women with positive results of hrHPV tests be managed based on HPV genotyping results and, in some cases, cytologic test results.

In August 2018, the US Preventive Services Task Force endorsed primary hrHPV testing alone performed at 5-year intervals beginning at age 30 years as a preferred screening strategy⁷ along with cytologic testing at 3-year intervals for women aged 21 to 65 years. The group continued to recommend cotesting (cytologic testing plus hrHPV testing) for women aged 30 to 65 years as an alternative strategy.

With various possible test combinations, screening frequencies, and ages to switch from one screening strategy to another, many different cervical cancer screening strategies are now being recommended in the United States. To help identify which strategies might constitute high-value care, some groups have recommended that cost-effectiveness analyses be performed.² Prior analyses have been hampered by the lack of a comprehensive, population-derived set of process utilities (preferences) that capture important quality of life outcomes anticipated throughout the contemporary screening process.^8,9 In an effort to contribute to policy discussions regarding high-value cervical cancer screening, we estimated quality of life and economic outcomes associated with 12 strategies by measuring women’s preferences and incorporating them into a cost-effectiveness analysis.

We conducted 4 focus groups to aid in constructing scenarios for health states associated with cervical cancer screening (eTable 1 in the Supplement). A sociodemographically diverse group of English-speaking or Spanish-speaking women aged 21 to 65 years was then recruited from 2 women’s health clinics in San Francisco, California, between September 22, 2014, and June 16, 2016. The sociodemographic and clinical characteristics of the sample are in eTable 2 in the Supplement. These women were enrolled in a cross-sectional study consisting of a 50-minute, face-to-face interview during which they completed an interviewer-administered questionnaire and viewed a 7-minute educational video. Using a computerized tool, preferences were elicited from participants using the time tradeoff method^10,11 and were used to generate utilities for 23 health states. To prevent fatigue, health states were grouped into 3 sets of 7 or 8; each participant was randomly assigned by computer to assess 2 of the 3 sets. To minimize possible effects of the order in which scenarios were presented, we also randomized the set presentation order. Prior to viewing the educational materials and performing the preference elicitation exercise, the first 262 participants were asked to select a preferred screening strategy; after the exercise was completed, they were again asked to select a preferred strategy (eTable 3 in the Supplement). The University of California, San Francisco and Zuckerberg San Francisco General Hospital Institutional Review Boards approved this study, and written informed consent was obtained from the participants. Details are in eAppendices 1, 2, and 3 in the Supplement.

We constructed an HPV type-specific Markov decision model using data on the natural history of HPV and cervical neoplasia. The model was constructed using TreeAge Pro 2017 (TreeAge Software Inc), and R, version 3.5.0 (R Foundation for Statistical Computing).¹² The natural history model simulated a birth cohort of women at average risk of developing cervical cancer (not immune-compromised and not vaccinated against HPV). In the model, the transition probabilities between health states were HPV type–specific (eFigure 1 in the Supplement). To fully capture changes in health states associated with screening, we used a 1-year Markov cycle length.

The cohort started at age 10 years with no existing HPV infections and was followed over a lifetime (until death or age 100 years). Every year, women were at risk of becoming infected with HPV 16 or 18 or other hrHPV types (eTable 4 in the Supplement). Women could clear their infections, stay infected, or progress to cervical intraepithelial neoplasia, a precancerous lesion. Cervical intraepithelial neoplasia could regress, persist, or progress to higher grades or cancer. Cervical intraepithelial neoplasia grade 1 was included as a health state because current management guidelines endorse treatment if these lesions persist.¹³ Women with cancer could have their cancer detected by symptoms as the stage progressed and were at risk of cancer death (eTable 5 in the Supplement). Women were also at risk of age-specific causes of death and of undergoing hysterectomy for noncancerous conditions.

We used Bayesian calibration methods to estimate the parameters of the natural history by matching model-predicted outcomes with calibration targets (eFigures 2-4 in the Supplement). Validation was by comparison with 2 sources: outcomes from a recent randomized trial of hrHPV testing compared with cytologic testing¹⁴ and Surveillance, Epidemiology, and End Results (SEER) data¹⁵ (eFigures 5-7 in the Supplement). Details are in eAppendices 4, 5, and 6 and eTable 6 in the Supplement.

We evaluated 12 strategies: cytologic testing every 3 years for women aged 21 to 65 years with either repeat cytologic testing in 1 year or immediate hrHPV triage for women with ASC-US; cytologic testing every 3 years for women aged 21 to 29 years followed by cotesting for women aged 30 to 65 years with either repeat cotesting in 1 year or immediate genotyping triage for women with normal cytologic test results and positive hrHPV test results; cytologic testing every 3 years for women aged 21 to 29 years followed by primary hrHPV testing alone every 3 years or every 5 years for women aged 30 to 65 years with either immediate cytologic testing triage for women with positive hrHPV test results or immediate genotyping triage for women with positive hrHPV test results with additional cytologic testing triage of women with positive hrHPV test results and negative genotyping results. In strategies switching from cytologic testing to hrHPV testing, we also evaluated switching at aged 25 years instead of 30 years (eAppendix 7 and eTable 7 in the Supplement).

For all cotesting and primary hrHPV testing strategies, women underwent cytologic screening prior to beginning hrHPV testing. We included annual cytologic testing as an additional comparator because it is a screening strategy still preferred by many US women.^16,17 Management of abnormal test results and cervical intraepithelial neoplasia treatment were programmed in the model to reflect the complexity of current American Society of Colposcopy and Cervical Pathology guidelines¹³ (eTables 7 and 8, eAppendix 8, and eFigures 8-14 in the Supplement). We assumed that all women adhered with screening, follow-up and treatment.

In primary analyses, estimates of screening test accuracy (sensitivity and specificity) were abstracted from recent systematic reviews, including a 2017 Cochrane review (Table 1).^18-25 We used cervical intraepithelial neoplasia grade 2 as defining disease because treatment is recommended for most women in the United States with this lesion.¹³ In sensitivity analyses, we used accuracy estimates from a multicentered, United States–based study that enrolled more than 47 000 women (the Addressing the Need for Advanced HPV Diagnostics study [ATHENA])²⁶; we used unadjusted accuracy estimates and those adjusted for verification bias. We used separate summary accuracy estimates for tests applied in surveillance^21,22 and posttreatment follow-up.²³ Clinical algorithms and utility maps are in eFigures 8-14 in the Supplement.

We incorporated direct medical costs into our model, reported in 2016 US dollars and accounting for medical inflation. Direct costs associated with screening, diagnosis, and treatment were based on Medicare reimbursement rates (Table 1).^18-25 Costs for cancer, including cancer death, were based on SEER-Medicare claims data.²⁵ In sensitivity analyses, we included indirect nonmedical costs, including time for patient travel, waiting, and examination.²⁷ In the absence of published studies on the indirect costs for cervical cancer, we used costs for uterine cancer.²⁸

The primary outcomes were lifetime costs in 2016 US dollars and quality-adjusted life-years (QALYs) for each strategy, both discounted at an annual rate of 3%. Each utility was mapped to an outcome for a specific strategy that would occur during a 1-year period (eFigures 8-14 in the Supplement); only 1 utility per unique outcome was applied. As recommended,²⁹ cost-effectiveness analyses were conducted from both a societal and health care sector perspective. Incremental cost-effectiveness ratios were calculated by dividing the additional cost by the additional health benefit of a specific strategy compared with the next less costly, nondominated strategy.

We performed deterministic sensitivity analyses to examine the independent effect of the following on total costs, QALYs, and incremental cost-effectiveness ratios: adding indirect costs, substituting ATHENA test accuracy estimates, and equating hrHPV test cost with the reported lower bound for the cost of a cytologic test ($14). We explored the result of substituting accuracy estimates in the primary screening setting with the highest summary specificity estimates reported for hrHPV testing (Table 1)^18-25; because of a lack of direct evidence regarding how this change in specificity affects accuracy estimates for cotesting and primary hrHPV testing with genotyping triage, these strategies were excluded from this model. To explore the independent effect of utilities on our results, we removed utilities stepwise by the categories defined in Table 1^18-25 (screening, then surveillance, then false-positive testing, then treatment and cancers).

To account for model input parameter uncertainty, we randomly sampled values from parameter distributions. For the calibrated parameters, we sampled from the posterior distribution obtained from the Bayesian calibration approach (eAppendix 5 in the Supplement). In total, we performed 10 000 iterations of Monte Carlo simulations to evaluate the effect of varying all model inputs simultaneously for each strategy (ranges in Table 1^18-25) on cost-effectiveness results.

We measured utilities (Table 1) in 451 women. Their mean (SD) age was 38.2 (10.7) years, 258 (57.2%) were nonwhite, and 151 (35.7%) had less than a college degree (eTable 2 in the Supplement). Estimated lifetime screening outcomes per 1000 women demonstrated more false-positive test results associated with hrHPV test–based strategies with concurrent lower cancer incidence (Table 2). Cancer mortality ranged from 0.9 to 1.4 per 1000 women who underwent screening compared with 19.0 per 1000 women who did not undergo screening.

Screening was cost-saving compared with no screening ($1267-$2577 per woman vs $2891 per woman). Cytologic testing every 3 years with repeat cytologic testing for ASC-US yielded the most lifetime QALYs (28.91174) (Table 3). Cytologic testing every 3 years with hrHPV triage of ASC-US was the lowest-cost strategy ($1267 per woman), and cytologic testing every 3 years with repeat cytologic testing for ASC-US conferred more QALYs at higher costs ($2166 per QALY). Cotesting and primary hrHPV testing provided fewer QALYs at higher costs (ie, were dominated). Annual cytologic testing was the most costly strategy but provided fewer QALYs than did cytologic testing every 3 years (lifetime costs, $2577; QALYs, 28.80491).

Adding indirect costs did not change our conclusions about cost-effectiveness (Table 3). When the costs of hrHPV testing were equated to the lower bound of cytologic testing ($14), hrHPV testing every 5 years with genotyping triage for women with positive hrHPV test results and additional cytologic triage of women with positive hrHPV test results and negative genotyping- results beginning at age 30 years was the lowest-cost strategy ($1183). Additional QALYs at higher costs were conferred by cytologic testing every 3 years with hrHPV triage of ASC-US ($708 per QALY) and further with cytologic testing every 3 years with repeat cytologic testing in 1 year for ASC-US ($2590 per QALY).

When we substituted ATHENA test accuracy estimates adjusted for verification bias in which hrHPV test sensitivity was markedly greater than that of cytologic testing (0.649 vs 0.313), hrHPV testing every 5 years with genotyping triage for women with positive hrHPV test results and additional cytologic triage of women with positive hrHPV test results and negative genotyping results beginning at age 30 years was the lowest-cost strategy; additional QALYs at higher costs were conferred by cytologic testing every 3 years with hrHPV triage of ASC-US ($715 per QALY) and further with cytologic testing every 3 years with repeat cytologic testing for ASC-US ($2446 per QALY). Conclusions were similar when we used unadjusted ATHENA accuracy estimates (eAppendix 9 and eTable 9 in the Supplement). When we substituted a high-specificity estimate for hrHPV testing (0.933),¹⁸ hrHPV testing every 5 years with cytologic triage beginning at age 30 years was the least costly strategy ($1230 per woman); additional QALYs at higher costs could be achieved with cytologic testing every 3 years with hrHPV triage of ASC-US ($488 per QALY) and further with cytologic testing every 3 years with repeat cytologic testing in 1 year for ASC-US ($2166 per QALY; eTable 9 in the Supplement).

Although utilities for health states involving hrHPV testing for screening and surveillance were generally lower than those for cytologic testing (Table 1), removal of these utilities did not change our conclusions about cost-effectiveness; the incremental cost-effectiveness ratio of cytologic testing every 3 years with repeat cytologic testing in 1 year for ASC-US compared with cytologic testing every 3 years with hrHPV triage of ASC-US increased from $2166 per QALY to $21 795 per QALY. Additionally removing utilities associated with false-positive test results suggested that cytologic testing every 3 years with hrHPV triage of ASC-US dominated all other strategies. When all utilities were removed, hrHPV testing every 3 years with genotyping triage for women with positive hrHPV test results and additional cytologic triage of women with positive hrHPV test results and negative genotyping results beginning at age 25 years conferred more life-years than did cytologic testing with hrHPV triage of ASC-US (29.54523 vs 29.54286 life-years; $135 865 per life-year; Figure 1).

Probabilistic sensitivity analyses showed that at cost-effectiveness thresholds of $50 000 per QALY, $100 000 per QALY and $150 000 per QALY, cytologic testing every 3 years with repeat cytologic testing for ASC-US was cost-effective in 95% to 96% of iterations (Figure 2). Cytologic testing with hrHPV triage of ASC-US every 3 years was cost-effective in 4% to 5% of iterations and primary hrHPV testing every 5 years was cost-effective in 0.01% to 0.04% of iterations. Beginning hrHPV testing prior to age 30 years, performing hrHPV testing every 3 years, and cotesting were cost-effective in 0% of iterations (ie, not cost-effective).

Of 12 strategies evaluated, our findings suggest that cytologic testing every 3 years for women aged 21 to 29 years with either continued cytologic testing every 3 years or switching to a low-cost hrHPV test every 5 years from age 30 to 65 years confers a reasonable balance of benefits, harms, and costs from both a societal and health care sector perspective. Cytologic testing plus hrHPV testing (cotesting) did not appear to be cost-effective under any condition we evaluated. Both the American College of Obstetricians and Gynecologists and the American Cancer Society consider cotesting the preferred cervical cancer screening strategy, and the US Preventive Services Task Force considers it an alternative strategy. Our findings challenge these endorsements. Furthermore, our analyses suggest that it is not cost-effective to begin primary hrHPV testing prior to age 30 years, to perform hrHPV testing every 3 years, or to perform cytologic testing annually. Comparative modeling is needed to confirm these findings.

Our sensitivity analyses identified 3 factors that are potentially associated with the cost-effectiveness of hrHPV testing: test sensitivity, test specificity, and test cost. When we used ATHENA estimates in which hrHPV test sensitivity was markedly greater than that of cytologic testing (0.649 vs 0.313), hrHPV testing every 5 years with genotyping triage for women with positive hrHPV test results and additional cytologic triage of women with positive hrHPV test results and negative genotyping results beginning at age 30 years was the lowest-cost strategy, suggesting that full-cost hrHPV testing may be a reasonable approach to screening in some US settings. The ATHENA estimate for cytologic test sensitivity, however, was lower than all 15 estimates summarized in the 2017 Cochrane review¹⁸ (range, 0.52-0.94); the generalizability of ATHENA findings to other screening settings has been questioned.³⁰

Our finding that primary hrHPV testing every 5 years with cytologic triage is the lowest-cost strategy when test specificity is relatively high demonstrates a tangible way in which the screening process might be improved; using tests with a relatively low specificity often leads to surveillance, a costly health state to which our participants assigned relatively low utilities. Our high-specificity estimate was derived from a hrHPV test targeting E6 and E7 mRNA.¹⁸ This analysis, however, was limited by our inability to compare primary hrHPV testing with all other 11 strategies. Our sensitivity analysis of hrHPV test costs demonstrated that the attractiveness of hrHPV testing could be improved directly by lowering the cost of the test.

We chose the QALY as our primary outcome because it is a widely recognized measure that combines length of life and population-derived health-related quality of life in a single measure.^31,32 Other recent cost-effectiveness analyses that have found hrHPV test–based strategies to be more favorable to cytologic testing–based strategies in US settings have used outcomes other than the QALY.^33,34 Other analyses have been limited by the use of utilities derived from expert panels.^35,36

Investigators of a recent randomized clinical trial³⁷ comparing 2 of the 12 strategies we evaluated (cytologic testing with hrHPV triage of ASC-US vs hrHPV testing with cytologic triage of positive hrHPV tests) stated that the cost-effectiveness of these strategies would need to be understood. Our cost-effectiveness analysis suggests that over a lifetime of screening, cytologic testing every 3 years with hrHPV triage of ASC-US dominates hrHPV testing with cytologic triage of positive hrHPV tests performed either every 3 years or every 5 years (Table 3).

Our study has important limitations. Although our participants were recruited from 2 clinical settings, the preferences we measured may not be generalizable to other populations. We did not identify any demographic or clinical characteristic by which the utilities varied; these analyses, however, may have been underpowered to detect differences (eTables 10 and 11 in the Supplement). Despite providing participants with clinical information in visual, written, and audio formats along with a research assistant to provide clarifications, we could not be assured that participants understood the information or its relevance to the time tradeoff exercise. Our estimates of incident cancers are higher than those reported recently,³⁸ as were our estimates of cancer deaths, perhaps because of differential assumptions regarding natural history. Other differences may be because of the intensity with which we modeled guideline-recommended surveillance and posttreatment follow-up. We did not consider the projected effect of HPV vaccination because of limited evidence concerning screening outcomes among vaccinated women as well as current recommendations by all major guideline groups that vaccinated women be screened no differently than unvaccinated women. To allow comparisons with outcomes from other studies, we assumed 100% adherence. How adherence at the numerous steps in the screening process may differ by strategy is uncertain, but may affect screening effectiveness. Finally, we used age-specific transitions to model natural history in contrast with time-in-state dependent transitions,³⁸ highlighting the need to confirm our findings using comparative modeling.

Our study also has strengths. Our measurement of a comprehensive set of health process utilities specific to cervical cancer screening in a sociodemographically diverse group of women allowed comparisons of quality of life outcomes conferred by each and demonstration of the independent effect of utilities on cost-effectiveness results. By using accuracy estimates drawn from enrollees in a single large US study, we provided more direct comparisons of all strategies in sensitivity analyses. We also followed detailed contemporary clinical algorithms and adjusted both sensitivity and specificity in the surveillance and posttreatment settings to more accurately reflect the benefits, harms, and costs incurred throughout the full trajectory of recommended care in the United States.

Without quality adjustments, conferred life-years varied little among all 12 screening strategies (29.54175-29.54523 life-years; Figure 1), a finding that underscores the importance of quantifying screening harms and costs unique to each strategy. Identifying and promoting strategies that maximize quality of life outcomes and minimize costs at all steps throughout the screening process will provide higher-value cervical cancer screening from the perspectives of society, the health care sector, and women.

Accepted for Publication: March 11, 2019.

Published Online: May 13, 2019. doi:10.1001/jamainternmed.2019.0299

Correction: This article was corrected on June 17, 2019, to fix an error in Table 2.

Corresponding Author: George F. Sawaya, MD, Department of Obstetrics, Gynecology & Reproductive Sciences, University of California, San Francisco, 550 16th St, Floor 7, San Francisco, CA 94143 (george.sawaya@ucsf.edu).

Author Contributions: Drs Sawaya and Kulasingam had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sawaya, Smith-McCune, Gregorich, Silverberg, Kuppermann, Kulasingam.

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

Drafting of the manuscript: Sawaya, Alarid-Escudero, Kuppermann, Kulasingam.

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

Statistical analysis: Alarid-Escudero, Leyden.

Obtained funding: Sawaya, Smith-McCune, Gregorich, Kuppermann.

Administrative, technical, or material support: Sawaya, Smith-McCune, Leyden, Kulasingam.

Supervision: Sawaya, Kuppermann.

Conflict of Interest Disclosures: Dr Sawaya reported receiving grants from the National Cancer Institute during the conduct of the study. Dr Smith-McCune reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Gregorich reported receiving grants from the National Institutes of Health/National Cancer Institute during the conduct of the study. Dr Leyden reported receiving grants from the National Cancer Institute during the conduct of the study. Dr. Huchko reported receiving grants from the University of California, San Francisco, Department of Obstetrics/Gynecology during the conduct of the study. Dr Kulasingam reported receiving grants from the National Cancer Institute during the conduct of the study. No other disclosures were reported.

Funding/Support: This work was funded by grant 1R01CA169093 from the US National Cancer Institute.

Role of the Funder/Sponsor: The funding source 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.

Additional Contributions: Allison O’Leary, MPH, and Melissa Duncanson, BA, University of California, San Francisco, provided research administration assistance. Allison O’Leary, MPH, Mayra Carrillo, BS, Xochilt Borja, BS, and Rachel Freyre, BS, University of California, San Francisco, conducted participant interviews. Michelle Moghadassi, MPH, and Cinthia Blat, PhD, University of California, San Francisco, provided statistical support. Hilary Whitham, PhD, and Ran Zhao, MPH, University of Minnesota, provided modeling assistance. All individuals except Ms Duncanson were compensated for their contributions.