Patterns and correlates of cervical cancer screening initiation in a large integrated health care system.

Background: The latest 2012 US Preventive Services Task Force cervical cancer screening guidelines recommended screening initiation at age 21 years. Little is known about the cervical cancer screening initiation practices in the community and whether there are critical gaps with respect to adherence to current clinical guidelines. Despite an overall decline in cervical cancer incidence across women of all ages, the incidence rate has not declined among 24-25 year olds between 2000 (2.79 per 100,000) and 2013 (2.93 per 100,000). Thus, it is important to understand cervical cancer screening initiation in young women and how woman- and provider-level factors affect the timing of screening initiation to identify areas for improving cervical cancer prevention.

Objective: We examined patterns and correlates of cervical cancer screening initiation among women turning age 21 years in a large community-based practice.

Study Design: Female members of Kaiser Permanente Southern California who turned age 21 years (baseline) during 2013-2015 and had not previously received a Papanicolaou test were included. Cervical cancer screening initiation through October 2016 was captured using electronic health records. Incidence rate and cumulative incidence of screening initiation was calculated. Associations between patient and provider characteristics and screening initiation were evaluated using multivariable Cox models.

Results: A total of 38,257 women were included and the Papanicolaou screening initiation rate was 44 per 100 person-years during the study period. Approximately 40% initiated screening within 1 year after turning age 21 years. In multivariable analyses, Asian/Pacific Islanders (hazard ratio, 0.91; confidence interval, 0.86-0.96 compared with non-Hispanic whites); Medicaid enrollees (hazard ratio, 0.90; confidence interval, 0.83, 0.97); those whose primary language is not English (hazard ratio, 0.71; confidence interval, 0.67, 0.75); those who have a historical inpatient visit, primary care physician in pediatrics, internal medicine, or another specialty compared with family practice; and have a male rather than female primary care physician (hazard ratio, 0.46; confidence interval, 0.36, 0.57) less often initiated screening. On the other hand, those who used other preventive services such as getting a human papilloma virus and influenza vaccination and those with a history of pregnancy, contraception use, and sexually transmitted infections more often had timely screening initiation.

Conclusion: Less than half of the women insured for preventative services initiated screening at age 21 years. Strategies to improve adherence to screening initiation guidelines should consider a tailored approach for at-risk subgroups and addressing initiation challenges associated with male physicians.