Pap-HPV co-testing adoption trends for cervical cancer screening in a multi-state Practice Research Network (PBRN) 2012-2017

Context: In 2012, the United States cervical cancer screening (CCS) guidelines changed to add co-testing (Papanicolaou [Pap] and human papillomavirus [HPV] test) to Pap-only. Little is known about variation in adoption of CCS modalities in federally qualified health centers (FQHCs) in communities where cervical cancer disparities persist.

Objective: Describe adoption of co-testing as a CCS modality in FQHCs serving socioeconomically disadvantaged populations.

Study design and analysis: In this observational study, we compared the distribution of screening modalities by 1) FQHC and year and 2) patient characteristics by screening modalities using chi-squared and t-tests.

Setting or dataset: Electronic health record data from 25 primary care clinics in 3 FQHCs in Washington and Idaho PBRN from 2012-2017.

Population Studied: Average-risk females ages 21-64 with ≥1 medical encounter and ≥1 routine CCS during 2012-2017.

Intervention/instruments: N/A Outcome measures: Proportion of CCS with Pap-only, co-testing (Pap and HPV testing occurring within 30 days), and HPV-only.

Results: From 2012-2017, 19,555 CCS were performed across 3 FQHCs, and the proportion of CCS with Pap-only decreased from 99.6% to 52.5%, co-testing increased from 0.4% to 46.7%, and HPV-only from 0.1% to 0.8%. Though co-testing increased steadily across 3 FQHCs, the proportion varied across organizations, ranging from 19.7% to 60.8% in 2017. Among 12,506 screened average-risk individuals, the mean age was 39.0 and 86.3% were White, 16.0% Hispanic or Latino, 30% rural and 27% not insured. Compared to individuals screened with Pap-only, those screened with co-testing were significantly older and more likely to be American Indian or Alaska Native (1.5% vs 0.6%, p <0.001), Black or African American (3.4% vs 2.5%, p <0.001), multiple race (0.3% vs 0.1%, p <0.001) and Hispanic or Latino (28.5% vs 11.6%, p <0.001), live in urban areas (87.3% vs 59.4%, p <0.001), and be insured (83.5% vs 69.3%, p <0.001).

Conclusions: Significant variation in adoption of co-testing across multi-state PBRN may reflect disparate access to CCS and HPV testing among subpopulations and FQHCs. Failure to adopt HPV testing, which is more sensitive than Pap-only for detecting cervical precancers, may exacerbate existing disparities. As new screening modalities emerge (e.g., HPV self-sampling), multilevel interventions that are tailored for both clinic and patients’ needs will be critical to advance equitable access to CCS.