ÌÇÐÄvlog

Key Points

QuestionÌý Was the US Preventive Services Task Force (USPSTF) 2021 recommendation for average-risk individuals aged 45 to 49 years to receive colorectal cancer (CRC) screening associated with changes in screening uptake?

FindingsÌý In this nationwide cohort study of 10 221 114 privately insured individuals aged 45 to 49 years, mean rates of CRC screening uptake increased significantly in the 20 months after the USPSTF recommendation vs the 20 months before. Those residing in areas with high socioeconomic status and metropolitan areas experienced a higher increase.

MeaningÌý This study found that uptake of CRC screening increased after the USPSTF recommendation, but the increase was uneven based on socioeconomic status and locality.

ImportanceÌý In May 2021, the US Preventive Services Task Force (USPSTF) issued a grade B recommendation encouraging colorectal cancer (CRC) screening among average-risk individuals aged 45 to 49 years. The patterns of screening uptake and possible socioeconomic disparities in screening in this age group remain unknown.

ObjectiveÌý To evaluate changes in CRC screening uptake among average-risk individuals aged 45 to 49 years after the USPSTF recommendation was issued in 2021.

Design, Setting, and ParticipantsÌý This retrospective cohort study used deidentified claims data from commercially insured Blue Cross Blue Shield beneficiaries aged 45 to 49 years across the US between January 1, 2017, and December 31, 2022.

ExposureÌý Publication of the May 2021 USPSTF CRC screening recommendation for adults aged 45 to 49 years.

Main Outcomes and MeasuresÌý Absolute and relative changes in screening uptake were compared between a 20-month period preceding (May 1, 2018, to December 31, 2019) and a 20-month period following (May 1, 2021, to December, 31, 2022) the USPSTF recommendation. Interrupted time-series analysis and autoregressive integrated moving average models were used to evaluate changes in screening rates, adjusting for temporal autocorrelation and seasonality.

ResultsÌý In this cohort study of 10 221 114 distinct beneficiaries aged 45 to 49 years (mean [SD] age, 47.04 [1.41] years; 51.04% female), bimonthly mean (SD) numbers of average-risk beneficiaries were 3 213 935 (31 508) and 2 923 327 (105 716) in the prerecommendation and postrecommendation periods, respectively. Mean (SD) screening uptake increased from 0.50% (0.02%) to 1.51% (0.59%) between the 2 periods (P < .001), representing an absolute change of 1.01 percentage points (95% CI, 0.62-1.40 percentage points) but no significant relative change (202.51%; 95% CI, −30.59% to 436.87%). Compared with average-risk beneficiaries residing in areas with the lowest socioeconomic status (SES), those residing in areas with the highest SES experienced the largest absolute change in screening (1.25 [95% CI, 0.77-1.74] percentage points vs 0.75 [95% CI, 0.47-1.02] percentage points), but relative changes were not significant (214.01% [95% CI, −30.91% to 461.15%] vs 167.73% [95% CI, −16.30% to 352.62%]). After the recommendation was issued, the screening uptake rate also increased fastest among average-risk beneficiaries residing in the areas with highest SES (0.24 [95% CI, 0.23-0.25] percentage points every 2 months) and metropolitan areas (0.20 [95% CI, 0.19-0.21] percentage points every 2 months).

Conclusions and RelevanceÌý This study found that among privately insured beneficiaries aged 45 to 49 years, CRC screening uptake increased after the USPSTF recommendation, with potential disparities based on SES and locality.

Measurable decreases in colorectal cancer (CRC) incidence and mortality have been attributed to increased screening.^1,2 In June 2016, the US Preventive Services Task Force (USPSTF) concluded with high certainty that screening for CRC of average-risk, asymptomatic adults aged 50 to 75 years is of substantial net benefit.³ Since then, due to the rising incidence of CRC among younger individuals, CRC screening guidelines have encouraged individuals aged 45 to 49 years to receive screening.^4,5 In May 2018, the American Cancer Society (ACS) issued a qualified recommendation that adults with average risk who are aged 45 years or older undergo regular CRC screening.⁶ Following this recommendation, CRC screening among individuals aged 45 to 49 years more than doubled according to an ACS study of approximately 5800 individuals included in the 2018 National Health Interview Survey.⁷ Subsequently, in May 2021, the USPSTF issued a grade B recommendation for CRC screening in adults aged 45 to 49 years, mandating insurance coverage for this preventive service.^5,8 Despite these efforts, screening remains below the national target of 80% among adults aged 45 years or older, with particularly low screening observed among individuals aged 45 to 49 years; those identifying as American Indian or Alaska Native, Asian, or Hispanic; individuals born outside the US; and individuals with income below 100% of the federal poverty level.^4,9

The association between evolving recommendations and uptake of CRC screening among individuals aged 45 to 49 years is not well understood. Although access to health insurance and a usual source of health care have been identified as key promoters of discussion, recommendation, and delivery of screening, little is known about disparities in screening access in the insured population and the potential contribution of socioeconomic status (SES) in screening disparities in this newly included cohort of individuals.¹⁰ As observed in other age groups, screening uptake may have been lower than expected among those aged 45 to 49 years due to differences in screening access and uptake by SES.^11,12 There are also concerns that detection of earlier-stage disease in younger individuals unnecessarily exposes this group of individuals to medical complications such as perforation, bleeding, and risks associated with sedation and leads to overdetection of adenomas.^7,13,14 In addition, it is predicted that lowering the screening age from 50 to 45 years requires 810 additional colonoscopies per 1000 persons screened using a colonoscopy-based screening strategy, substantially increasing financial costs.^15,16 Furthermore, different recommendations over time may have led to public uncertainty about CRC screening, potentially reducing the impact of future guideline changes. These public health implications warrant the evaluation of CRC screening uptake among individuals aged 45 to 49 years relative to the recently revised guidelines.

In the current study, we assessed the association between the May 2021 USPSTF recommendation and CRC screening uptake in a large cohort of privately insured individuals aged 45 to 49 years across the US. We conducted one of the first and largest studies to identify possible disparities in screening in this age group as well as recent rates in screening uptake after the recommendation was issued.

We conducted a retrospective cohort study using deidentified commercial claims data from the Blue Cross Blue Shield (BCBS) Axis. BCBS is the largest provider of commercial insurance in the US, covering approximately one-third of the US population. The BCBS Axis includes administrative claims data from more than 92% of physicians and 96% of hospitals nationwide, and it is the only data resource in the health care industry that includes information from every zip code in the US.¹⁷ As these data are deidentified, the Yale University institutional review board deemed our study to be non–human participant research, and no informed consent was required. We followed the Strengthening the Reporting of Observational Studies in Epidemiology () reporting guideline.

We categorized the 72 months from January 1, 2017, through December 31, 2022, into 36 consecutive 2-month periods to capture potential temporal changes in screening patterns and analyze critical short-term epidemiologic phenomena (eg, several waves of the COVID-19 pandemic) that may have been associated with CRC screening uptake. For each period, the denominator in our calculation of CRC screening uptake included beneficiaries aged 45 to 49 years who had (1) BCBS as the primary insurance for at least 12 months prior to the start of the period and remained enrolled through the end of the 2-month period and (2) did not receive CRC screening or related procedures in the 12 months preceding the 2-month period. Our numerator included the subset of the denominator who received CRC screening (eTable 1 in Supplement 1). Some individuals were screened more than once in the same period. Those who were screened more than once in the same period were counted as only 1 individual in the numerator when we evaluated overall screening uptake (ie, CRC screening [yes or no]). With use of previously published methods, CRC screening was identified using claims for the following procedures: fecal occult blood test, fecal immunochemical test (FIT), stool DNA test, flexible sigmoidoscopy, double-contrast barium enema, colonoscopy, and computed tomography colonography (eTable 2 in Supplement 1).^18-21 To distinguish screening-related procedures from diagnostic procedures, only outpatient procedures were included, and individuals with claims for gastrointestinal tract symptoms, including abdominal pain, altered bowel habits, weight loss, iron-deficiency anemia, positive fecal occult blood test result, and gastrointestinal bleeding, within the 3 months preceding the period were excluded (eTable 2 in Supplement 1).²² Individuals with a history of CRC were censored before the month of the first claim with a CRC diagnosis code.

As the USPSTF 2021 recommendation for starting CRC screening at the age of 45 years applied to individuals at average risk for CRC, we assessed an individual’s CRC risk (eMethods and eTable 2 in Supplement 1). We obtained demographic data of the beneficiaries from BCBS; age and sex were known for all, while imputed data on race and ethnicity were only available for about one-third of beneficiaries. Race and ethnicity information in BCBS is imputed using multilevel modeling among beneficiaries with sufficient identifying information at the individual or household level.²³ We included race and ethnicity in the analysis to assess potential disparities in screening by these factors; categories were Hispanic, non-Hispanic Asian (hereafter, Asian), non-Hispanic Black (hereafter, Black), non-Hispanic Native American or Pacific Islander (hereafter, Native American or Pacific Islander), non-Hispanic White (hereafter, White), and unknown. Additionally, the Social Deprivation Index (SDI), a composite measure of area-level deprivation based on 7 demographic characteristics (proportion living in poverty, with less than 12 years of education, of single-parent households, living in rented housing units, living in overcrowded housing units, of households without a car, and of unemployed adults younger than 65 years) collected in the American Community Survey, was calculated using beneficiaries’ zip code of residence.²⁴ We further categorized beneficiaries’ locality of residence based on metropolitan and nonmetropolitan area (including micropolitan, small town, and rural areas) using rural-urban commuting area codes that classify US census zip codes using measures of population density, urbanization, and daily commuting.²⁵

We categorized May 1, 2018, to December 31, 2019, the 20-month period preceding the USPSTF 2021 recommendation and the COVID-19 pandemic, as the prerecommendation period and the 20-month period from May 1, 2021, to December 31, 2022, following the USPSTF 2021 recommendation and the first 2 waves of the COVID-19 pandemic, as the postrecommendation period (Figure 1).^26,27 We calculated the absolute and relative changes in screening uptake comparing the prerecommendation and postrecommendation periods.²⁸ In addition, we used interrupted time-series analysis to assess the association between the USPSTF 2021 recommendation and CRC screening uptake.²⁹ This method has been used to study the outcomes of large-scale policy interventions, including changes in USPSTF screening guidelines.^30-32 Based on prior observation that changes in CRC screening are rapidly implemented following changes in guidelines,⁷ we assumed an immediate effect in our analysis.

To calculate rate of uptake since the recommendation was issued, we fit segmented regression to the series of bimonthly screening uptake, with parameters for intercept, baseline trend, and changes in level and trend comparing the prerecommendation and postrecommendation periods. The Durbin-Watson statistic was used to check for autocorrelation of residuals.²⁹ If autocorrelation was observed, we used autoregressive integrated moving average (ARIMA) models that adjusted for autocorrelation and seasonality of the data.^29,33,34 All analyses were performed using R, version 4.2.4 (R Project for Statistical Computing). All P values were from 2-sided tests, and results were deemed statistically significant at P &±ô³Ù; .05.

A total of 10 221 114 distinct beneficiaries aged 45 to 49 years were included in the study sample (mean [SD] age, 47.04 [1.41] years; 51.04% female; 48.96% male), with a mean (SD) of 3 304 879 (108 378) individuals eligible for CRC screening in every 2-month period. Of this cohort of beneficiaries, 1.15% were Asian, 2.65% were Black, 3.86% were Hispanic, 0.15% were Native American or Pacific Islander, 25.43% were White, and 66.76% had unknown race and ethnicity. As shown in Figure 1, mean (SD) CRC screening uptake among individuals aged 45 to 49 years was comparable before (January 1, 2017, to April 30, 2018; 0.57% [0.03%]) and after (May 1, 2018, to December 31, 2019; 0.58% [0.02%]) the ACS recommendation (P = .30). Screening uptake was only 0.24% in March and April 2020 (ie, onset of COVID-19) but increased steadily after May 2021 (ie, when the USPSTF recommendation was issued) (range, 0.71%-2.41%). Between November 1, 2021, and February 8, 2022, uptake increased slowly, possibly coinciding with the onset of the Omicron COVID-19 variant. In comparison, screening uptake among individuals aged 50 to 75 years remained relatively stable except during the period from March 1 to July 31, 2020. Furthermore, by December 31, 2022, screening uptake among individuals aged 45 to 49 years reached a level similar to that among individuals aged 50 to 75 years (2.37% and 2.40%, respectively).

The bimonthly mean (SD) number of beneficiaries aged 45 to 49 years who were eligible for CRC screening in the prerecommendation period was 3 385 126 (32 878) (Table). This number decreased to a mean (SD) of 3 194 246 (91 357) beneficiaries during the postrecommendation period. The bimonthly mean (SD) uptake of CRC screening increased significantly from 0.58% (0.02%) in the prerecommendation period to 1.58% (0.57%) in the postrecommendation period (P < .001). Overall, this represented an absolute change of 1.00 percentage points (95% CI, 0.62-1.38 percentage points) (Figure 2) but no significant relative change (171.67%; 95% CI, −20.90% to 365.71%) (Figure 3).

Stratified analysis by risk status indicated that both average- and high-risk beneficiaries experienced significant increases in screening between the prerecommendation and postrecommendation periods. A bimonthly mean (SD) of 3 213 935 (31 508) and 2 923 327 (105 716) beneficiaries were classified as average risk in the prerecommendation and postrecommendation periods, respectively (Table). Among average-risk beneficiaries, mean (SD) screening uptake increased from 0.50% (0.02%) to 1.51% (0.59%) between the 2 periods (P < .001). There were a mean (SD) of 171 191 (25 201) and 270 919 (14 948) beneficiaries classified as having high risk for CRC during the prerecommendation and postrecommendation periods, respectively. In this high-risk group, mean (SD) screening uptake changed from 2.12% (0.11%) to 2.33% (0.27%) (P < .001). Although high-risk beneficiaries had significantly higher screening uptake than average-risk beneficiaries in both periods, average-risk beneficiaries experienced a higher absolute change in screening (1.01 percentage points [95% CI, 0.62-1.40 percentage points] vs 0.21 percentage points [95% CI, 0.02-0.40 percentage points]); however, average-risk beneficiaries had no significant relative change in screening (202.51%; 95% CI, −30.59% to 436.87%) (Figure 2 and Figure 3).

In the prerecommendation period, colonoscopy accounted for 41.3% of all screening tests followed by FIT (32.2%) and fecal occult blood test (24.6%) (eFigure and eTable 3 in Supplement 1). Stool DNA tests, flexible sigmoidoscopy, double-contrast barium enemas, and computed tomography colonography comprised the remaining modalities. In the postrecommendation period, colonoscopy remained the most common screening modality (52.7%) followed by stool DNA test (25.0%) and FIT (16.2%).

Females with average risk had significantly higher uptake of overall screening in the postrecommendation period compared with the prerecommendation period (1.56% vs 0.55%; P < .001). Similarly, males with average risk had significantly higher uptake of overall screening in the postrecommendation period compared with the prerecommendation period (1.46% vs 0.45%; P < .001). Postrecommendation screening uptake did not differ significantly between females and males (1.56% vs 0.45%; P = .69). In the postrecommendation period, use of colonoscopy was 0.65% among males compared with 0.60% among females (P = .74), and use of stool DNA test was 0.31% among females compared with 0.26% among males (P = .57) (Table and eTable 3 in Supplement 1).

After the recommendation was issued in May 2021, uptake of screening was significantly higher among average-risk beneficiaries residing in areas with SDI in the first quintile (ie, highest SES) compared with beneficiaries in the fifth quintile (ie, lowest SES) (1.84% vs 1.19%; P = .02) (Table). Beneficiaries residing in areas in the first SDI quintile also experienced the largest absolute change in screening between the prerecommendation and postrecommendation periods (1.25 percentage points; 95% CI, 0.77-1.74 percentage points), although there was no relative change in screening (214.01%; 95% CI, −30.91% to 461.15%) (Figure 2 and Figure 3). In comparison, beneficiaries residing in areas in the fifth quintile of SDI experienced an absolute change in screening uptake of 0.75 percentage points (95% CI, 0.47-1.02 percentage points) but no relative change (167.73%; 95% CI, −16.30% to 352.62%) between the 2 periods. Stratified analysis by screening modality similarly showed highest and lowest screening in individuals residing in areas in the first and fifth SDI quintiles, respectively (eTable 3 in Supplement 1).

Overall, average-risk beneficiaries residing in metropolitan areas had significantly higher uptake of screening during both periods compared with beneficiaries residing in nonmetropolitan areas (0.53% vs 0.38% prerecommendation [P < .001] and 1.59% vs 1.11% postrecommendation [P = .06]) (Table). However, beneficiaries in both metropolitan and nonmetropolitan areas experienced a significant increase in screening between the 2 periods. Among average-risk beneficiaries residing in metropolitan areas, mean (SD) uptake of screening increased from 0.53% (0.02%) to 1.59% (0.62%), but there was no significant relative change (203.56%; 95% CI, −27.56% to 436.04%) (Figure 2 and Figure 3). Among beneficiaries residing in nonmetropolitan areas, mean (SD) uptake of screening increased from 0.38% (0.02%) to 1.11% (0.48%) but with no significant relative change (194.79%; 95% CI, −54.57% to 446.36%). Stratified analysis by screening modality showed higher, though not significantly different, use of colonoscopy between metropolitan and nonmetropolitan areas in the postrecommendation period (0.66% vs 0.45%; P = .10). Similarly, use of stool DNA tests did not vary significantly between metropolitan and nonmetropolitan areas in the postrecommendation period (0.29% vs 0.25%; P = .59) (eTable 3 in Supplement 1).

Results from stratified ARIMA models assessing the rate of screening uptake since the recommendation was issued indicated an increasing rate of screening uptake among beneficiaries aged 45 to 49 years overall and across subgroups. After the recommendation was issued in May 2021, uptake of CRC screening in the overall cohort increased 0.19 percentage points (95% CI, 0.18-0.20 percentage points) every 2 months (Figure 4). Among average-risk beneficiaries, screening uptake overall increased 0.19 percentage points (95% CI, 0.18-0.20 percentage points) every 2 months. Beneficiaries residing in the areas in the first SDI quintile experienced the highest increase in trend followed by beneficiaries in the second to fifth quintiles. Screening uptake increased 0.24 percentage points (95% CI, 0.23-0.25 percentage points) every 2 months among those residing in the first SDI quintile areas and 0.14 percentage points (95% CI, 0.12-0.15 percentage points) every 2 months among those residing in the fifth SDI quintile areas. Screening uptake among metropolitan area residents increased 0.20 percentage points (95% CI, 0.19-0.21 percentage points) every 2 months, whereas nonmetropolitan area residents experienced an increase of 0.16 percentage points (95% CI, 0.15-0.17 percentage points) in screening uptake every 2 months.

Overall, 13.9% of beneficiaries aged 45 to 49 years during the postrecommendation period received CRC screening before reaching the age of 50 years (eTable 4 in Supplement 1). A total of 11.5% and 37.1% of average- and high-risk beneficiaries aged 45 to 49 years during the postrecommendation period, respectively, received CRC screening.

In the subgroup of average-risk beneficiaries with imputed data on race and ethnicity (35.2% and 32.1% of beneficiaries during the prerecommendation and postrecommendation periods, respectively), each racial and ethnic group experienced a significant increase in screening between the 2 periods. Among White beneficiaries, mean (SD) uptake of screening increased from 0.50% (0.02%) to 1.50% (0.59%) (P < .001) (Figure 2). Among Black beneficiaries, mean (SD) uptake increased similarly from 0.51% (0.04%) to 1.50% (0.60%) (P < .001). Mean (SD) screening uptake among Asian, Native American or Pacific Islander, and Hispanic beneficiaries ranged from 0.50% (0.03%) to 0.52% (0.02%) in the prerecommendation period and from 1.50% (0.61%) to 1.55% (0.62%) in the postrecommendation period (P < .001 for both). No significant difference in screening uptake was observed between White beneficiaries and other beneficiaries, including those with unknown race and ethnicity, during the postrecommendation period (Table). Stratified analysis by screening modality showed no significant difference in colonoscopy and stool DNA test use based on race and ethnicity (eTable 3 in Supplement 1). Furthermore, stratified analysis evaluating differences in screening uptake by SDI within each racial and ethnic group showed that individuals residing in areas in the first and fifth SDI quintile had the highest and lowest uptake of CRC screening, respectively (eTable 5 in Supplement 1).

Our results indicated that CRC screening uptake among individuals aged 45 to 49 years significantly increased following the May 2021 USPSTF recommendation. Although CRC screening was expectedly higher among high-risk individuals, the 2-fold increase in screening among average-risk beneficiaries suggests a shift in the screening paradigm of this newly included age group. Although statistically significantly different, screening uptake was relatively low in all subgroups in the prerecommendation period. However, in the postrecommendation period, widening disparities based on SDI and locality were observed. These disparities persisted among individuals with known or unknown race and ethnicity information.

We observed a similar uptake of CRC screening in beneficiaries 45 to 49 years of age preceding and following the ACS recommendation in 2018, which may be attributable to the ACS recommendation not stipulating any insurance mandate. Although a prior study identified significant increases in screening following the 2018 ACS recommendation,⁷ our study in a large cohort of privately insured individuals did not find evidence of increased screening following this recommendation. However, the shift in insurance coverage due to the USPSTF 2021 recommendation was significantly associated with changes in screening practices among beneficiaries aged 45 to 49 years. Additionally, during the first wave of the COVID-19 pandemic (March and April 2020), almost all screening was eliminated in this age group; however, screening recovered to prepandemic levels in September 2020, and screening uptake remained relatively unchanged during the second and third waves of the pandemic. Overall, mean frequency of eligible beneficiaries was lower in the postrecommendation period compared with the prerecommendation period (3.2 million vs 3.4 million). This may be due to a decrease in job-based insurance coverage following the COVID-19 pandemic.³⁵

Insurance status is a key factor contributing to disparities in screening.⁴ All individuals in the study population had private insurance through BCBS, but the observed disparity in CRC screening based on neighborhood SDI and locality highlight the persistent role of socioeconomic and community-level factors that contribute to gaps in screening among subgroups of beneficiaries. The mechanisms behind these persistent disparities have not been fully elucidated, although factors such as perception of low health care quality, no routine checkup within 2 years, and perceived discrimination may have exacerbated these disparities.^11,36

The 3-fold increase in screening uptake among average-risk individuals aged 45 to 49 years reflects an accomplishment, yet evidence of widening disparities based on SDI and locality indicate that population subgroups may not be benefiting equally from this change in CRC screening recommendation. Furthermore, given that only 11.5% of average-risk individuals aged 45 to 49 years during the postrecommendation period received CRC screening before the age of 50 years, targeted initiatives to improve screening in this age group are warranted to reach the national goal of screening 80% of the population in every community.

There are limitations to our study. Data on race and ethnicity were only available for a subset of the population, undermining our ability to thoroughly evaluate the potential influence of these demographic factors. Although we analyzed SDI and locality as factors that represent multiple domains of the social and physical environment, prior studies showed that SES alone does not capture the pervasive association of race with health outcomes.^37,38 As BCBS is currently updating race and ethnicity information for a larger proportion of the study cohort, future studies may continue to evaluate changes in screening patterns among diverse population subgroups. We did not have specific information on the type of coverage provided by each beneficiary’s insurance plan, which may have impacted their screening uptake in the prerecommendation period. Since the USPSTF recommendation mandates all insurance programs to cover the cost of CRC screening among individuals aged 45 to 49 years, potential differences in uptake based on type of insurance may have been alleviated in the postrecommendation period. Although BCBS is the largest provider of commercial health insurance, the study cohort may not be fully representative of the general US population because BCBS beneficiaries tend to be younger and more socioeconomically advantaged with employer-based insurance. Nevertheless, it has been demonstrated that patterns of cancer care in the BCBS population are generalizable to the Medicare population.³⁹ As the focus of our analysis was CRC screening among individuals 45 to 49 years of age, BCBS is a great resource. Additionally, there are substantial SES variations among BCBS beneficiaries, which allowed for stratified analysis to assess disparity in CRC screening.

This cohort study of 10 221 114 individuals found that CRC screening uptake increased significantly among those aged 45 to 49 years after the USPSTF issued its qualified recommendation in May 2021 that encouraged screening in this age group. A significant increasing rate was observed overall and in racial and ethnic, SES, and metropolitan and nonmetropolitan locality subgroups, but disparities in screening were observed based on socioeconomic status and locality, underlining an urgent need to improve CRC screening uptake for all.

Accepted for Publication: July 29, 2024.

Published: October 3, 2024. doi:10.1001/jamanetworkopen.2024.36358

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2024 Siddique S et al. ÌÇÐÄvlog Open.

Corresponding Author: Sunny Siddique, MPH, Department of Chronic Disease Epidemiology, Yale School of Public Health, 60 College St, New Haven, CT 06510 (sunny.siddique@yale.edu).

Author Contributions: Mr Siddique and Dr Wang had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Siddique, Wang, Gaddy, Gross, Ma.

Acquisition, analysis, or interpretation of data: Siddique, Wang, Yasin, Zhang, Ma.

Drafting of the manuscript: Siddique, Yasin, Gaddy.

Critical revision of the manuscript for important intellectual content: Wang, Yasin, Gaddy, Zhang, Gross, Ma.

Statistical analysis: Siddique, Wang, Ma.

Obtained funding: Siddique.

Administrative, technical, or material support: Gaddy.

Supervision: Wang, Yasin, Gross, Ma.

Conflict of Interest Disclosures: Dr Gross reported receiving grants from Johnson & Johnson to help develop new approaches to clinical trial data sharing, personal fees from Genentech for research, and grants from the National Comprehensive Cancer Network (NCCN) Foundation (with funding to the NCCN provided by AstraZeneca) to fund research outside the submitted work. Dr Ma reported receiving personal fees from Bristol Myers Squibb outside the submitted work. No other disclosures were reported.

Funding/Support: This work was supported by award 1F31CA281335-01 from the National Cancer Institute of the National Institutes of Health (NIH) under the Ruth L. Kirschstein National Research Service Award Individual Predoctoral Fellowship (Mr Siddique).

Role of the Funder/Sponsor: The National Cancer Institute 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.

Disclaimer: The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the NIH.

Data Sharing Statement: See Supplement 2.