Significant racial disparities exist in cervical cancer incidence and mortality in the U.S., with Black women experiencing disproportionately higher rates of both. While timely cervical cancer screening (Pap tests, HPV tests, or co-testing) is a powerful preventative tool, its utilization is suboptimal. Previous studies examining racial/ethnic differences in self-reported screening yielded mixed results, with self-reporting known to overestimate screening and introduce differential misclassification by race. Single-site studies may also miss crucial patterns across different healthcare settings. To address this, this multi-site study uses electronic health record (EHR) and administrative data to evaluate racial and ethnic disparities in cervical cancer screening and timely follow-up of abnormal results across three diverse U.S. healthcare settings. Understanding these disparities is critical to improving equity in cancer prevention.

The existing literature on racial and ethnic disparities in cervical cancer screening shows inconsistent findings, partly due to reliance on self-reported data, which is susceptible to recall bias and social desirability bias. Studies using health record data offer greater accuracy but may not capture systemic factors contributing to disparities if limited to a single healthcare setting. The lack of multi-site studies analyzing health record data on screening and follow-up across diverse settings highlights a gap in the literature that this research aims to fill. Studies focusing on disparities in follow-up after abnormal screening results are also limited, underscoring the importance of this aspect of the research.

This retrospective study utilized data from the Multi-level Optimization of the Cervical Cancer Screening Process in Diverse Settings & Populations (METRICS) Research Center, part of the Population-based Research to Optimize the Screening Process (PROSPR) consortium. METRICS comprises three sites: a safety-net system in the Southwest, a mixed-model system in the Northwest, and an integrated system in the Northeast. The study population included female patients aged 18-89 years, with inclusion criteria based on visit history (Sites A and C) or health plan enrollment (Site B). Data were collected and harmonized from EHR and administrative databases from January 1, 2010 to December 31, 2019, with analysis focusing on a three-year period (June 1, 2016 to May 31, 2019). The primary outcomes were receipt of any cervical cancer screening test within the three-year window and timely follow-up (within six months) for abnormal findings. Patients were grouped by race and ethnicity based on EHR data. Covariates included age, insurance, smoking status, comorbidity score, number of primary care visits, and pregnancy status. Statistical analyses included chi-square tests and modified Poisson regression models (three models: unadjusted, clinically adjusted, and fully adjusted) to assess the association between race/ethnicity and the outcomes, accounting for confounding and mediating variables. The study was approved by the IRBs of each participating site.

The study included 188,415 eligible patients. Overall, 62.8% received cervical cancer screening during the three-year study period. Screening uptake varied significantly by race and ethnicity. Non-Hispanic Black patients had the lowest screening rate (53.2%), while Hispanic (65.4%) and Asian/Pacific Islander (66.5%) patients had higher rates than non-Hispanic White patients (63.5%). Bivariate analyses revealed that site of care was a strong predictor of screening rates, with the safety-net system (Site A) having the lowest rate (58.6%). Black patients had lower screening rates than non-Hispanic White patients overall, despite having similar or higher rates within individual sites. This disparity was largely attributable to the disproportionate representation of Black patients at Site A. Hispanic patients showed consistently higher screening rates than non-Hispanic White patients across all sites. Among those screened, 3.5% had abnormal results requiring follow-up. Black and Hispanic patients had a higher prevalence of abnormalities than non-Hispanic White patients. Timely follow-up (within six months of an abnormal result) was low overall (72.5%), with Hispanic patients showing the highest rate (78.8%) and Black patients the lowest (65%). Multivariable regression analyses showed that after adjusting for clinical and socioeconomic factors, including insurance and site of care, the lower screening rate for Black patients was attenuated, suggesting the role of systemic inequities. In contrast, the higher screening rate for Hispanic patients persisted and even increased after adjustment, indicating a complex interplay of factors beyond individual-level characteristics. Black and Hispanic patients were more likely to have screen-detected abnormalities even after adjustment for covariates. The difference in follow-up rates between Hispanic and non-Hispanic White patients was no longer significant after adjusting for insurance and site.

This multi-site study demonstrates persistent racial and ethnic disparities in cervical cancer screening and follow-up. The lower screening rate for Black patients is largely explained by the distribution of patients across healthcare settings and insurance coverage, highlighting the impact of systemic inequities. The higher screening rate for Hispanic patients, even after adjustment, suggests the importance of considering contextual factors influencing healthcare access and utilization within this population. The consistently low follow-up rates across all racial and ethnic groups underscore the need for interventions targeting improvements in this area. The study's findings reinforce the need to address systemic racism, which contributes to unequal access to quality healthcare and resources. Further research should focus on understanding the factors driving the higher screening rates observed in Hispanic patients and exploring the barriers to co-testing among Hispanic and Black individuals.

This study reveals persistent disparities in cervical cancer screening and follow-up, with Black patients experiencing lower screening rates and Hispanic patients higher rates than non-Hispanic White patients, largely explained by healthcare setting and insurance. Addressing systemic inequities impacting access to quality care is crucial. Future research should focus on understanding the positive factors driving higher screening in Hispanic populations and implementing interventions to improve timely follow-up for all patients with abnormal findings.

The study may have misclassified some patients as unscreened if they received care outside the study's healthcare systems. The three-year study period may introduce selection bias. The study did not assess all potential mediators of disparities, such as structural and interpersonal racism. The site effect likely reflects both within and outside healthcare system characteristics. The analysis relied on EHR data for race and ethnicity, leading to potential misclassification.