Do age-based discontinuation recommendations influence cervical cancer screening rates? Evidence from the United States’ Behavioral Risk Factor Surveillance System, 2016 and 2018

Abstract

The United States Preventive Services Task Force (USPSTF) recommends that women with adequate prior screening and average cervical cancer risk discontinue routine cervical cancer screening after age 65. This study’s objective was to estimate how the USPSTF recommendation affects Papanicolaou (Pap) test rates among women at age 66 in the United States. We used nationally representative 2016 and 2018 Behavioral Risk Factor Surveillance System data for women ages 56–76 (n = 226,031) and a regression discontinuity design to estimate changes in annual Pap test rates at age 66. Among women age 66–76, 22.5% reported receiving a Pap test within the past one year. At age 66, annual Pap rates declined by 5.9 percentage points (p.p.) (95% Confidence Interval [CI]: −7.7, −4.1) off a pre-66 rate of 39%. The change varied by race/ethnicity, education, and marital status. Pap rates did not change discretely for non-Hispanic Black women (0.8 p.p.; 95% CI: −5.4, 7.1) but did for women from other racial/ethnic groups. The decline was larger for women who graduated college (−9.0 p.p.; 95% CI: −12.0, −5.9) than for women without a college degree and for women who were never married (−9.4 p. p., 95% CI: −17.3, −1.5) than for women who were married/partnered or divorced/separated. The USPSTF recommendation to discontinue cervical cancer screening after age 65 leads to a sharp decline in Pap test rates at age 66 but disparately affects women based on race, education and marital status. Further study is needed to assess the consequences of these changes.

1. Introduction

The United States Preventive Services Task Force (USPSTF), an independent expert panel that issues clinical recommendations for preventive care in the US (Perkins et al., 2021), currently recommends that women ages 21 to 65 years old screen for cervical cancer regularly, commonly through Papanicolaou (Pap) testing (Curry et al., 2018). Since 2003, the USPSTF has recommended against cervical cancer screening in women over age 65 unless they have had inadequate prior screening or are at high risk for cervical cancer (Curry et al., 2018; Moyer, 2012; United States Preventive Services Task Force, 2003). The American College of Obstetricians and Gynecologists and American Cancer Society have published USPSTF-aligned age discontinuation guidelines since 2012 (American College of Obstetricians and Gynecologists, 2021; Centers for Disease Control and Prevention, 2013; Fontham et al., 2020). This age-based threshold for discontinuing cervical cancer screening is somewhat controversial (Bus-Kwofie et al., 2019; Dilley et al., 2021; Dinkelspiel et al., 2012). Recent reviews identify shortcomings in the evidence on age-based screening discontinuation, including recommendations based primarily on theoretical models, inadequate measures of harm, and a lack of qualitative data on patients’ perceptions of screening past age 65 (Bus-Kwofie et al., 2019; Dilley et al., 2021; Gravitt et al., 2018).

In addition, although overall cervical cancer incidence has declined since the 1950s, rates have plateaued recently (Quick et al., 2020; Safaeian et al., 2007). Women over age 65 now account for about 20% of cervical cancer cases in the US and are more likely to be diagnosed late and with poorer prognosis (Fox et al., 2008). Cervical cancer incidence among older women, when adjusted for hysterectomy status, does not decline until after age 85 (White et al., 2017). The net benefit of cervical cancer screening after age 65 may increase as the potential for disease-free life increases (Pankakoski et al., 2019), leading some to call for reevaluating the age-based discontinuation recommendation (Dilley et al., 2021).

Women of color, women with limited educational attainment, and women without health insurance coverage may be disproportionately harmed by ending regular screening after age 65, particularly if prior screening history is not considered. At least one in five women ages 30–65 in the US are not up to date with screening (Suk et al., 2022), and a greater proportion are inadequately screened between ages 56–65 (Cejtin and Schmidt, 2020; Chao et al., 2019; Mills et al., 2021). Screening uptake varies by race and ethnicity. Certain racial/ethnic minorities are at risk of underscreening (McDaniel et al., 2021; Sabatino et al., 2021; Suk et al., 2022). For instance, Hispanic women ages 18–69 are 27% less likely to have ever received Pap testing than Non-Hispanic White women (McDaniel et al., 2021). Disparities by education and insurance status also persist, with women with less than a college degree and uninsured women less likely to receive regular screening (Harper et al., 2020; Sabatino et al., 2021; Watson et al., 2017). In addition, cervical cancer risk is higher among Hispanic/Latina and non-Hispanic Black women. The 2019 incidence of cervical cancer among non-Hispanic Black and Hispanic/Latina women was 8.2 and 9.7 per 100,000 women, compared to 7.3 per 100,000 among non-Hispanic White women (United States Cancer Statistics Working Group, 2022).

Furthermore, cervical cancer screening rates have declined over the past two decades in the US. This trend has been demonstrated for women ages 21–65 as well as women ages 65 and over with traditional Medicare coverage (National Cancer Institute, 2022; Qin et al., 2023; Suk et al., 2022). Whether the USPSTF recommendation to discontinue screening after age 65 has contributed to this trend for older women remains unclear. In this study, we analyze the causal effect of the screening recommendation on Pap testing rates at age 66. In addition to estimating changes overall, we characterize changes by race/ethnicity, education and marital status to understand who is most affected by the recommendation.

2. Methods

We used public-use data from the 2016 and 2018 Behavioral Risk Factor Surveillance System (BRFSS) (Centers for Disease Control and Prevention, 2019a). Since 1984, BRFSS has collected data on the risk behaviors and preventive health practices of non-institutionalized US residents ages 18 and over through monthly telephone surveys. The 2016 and 2018 datasets are the most recent survey years that capture cervical cancer screening and were not impacted by the COVID-19 pandemic (Wentzensen et al., 2021). Our study was deemed human subjects exempt by the University of Southern California Institutional Review Board.

The 2016 and 2018 BRFSS data include about 925,000 respondents. We restrict to the approximately 25% of respondents ages 56 to 76 whose sex is designated as female (n = 226,031). We use the term women when describing our sample to align with the BRFSS and USPSTF language for cervical cancer screening (Centers for Disease Control and Prevention, 2019b; Curry et al., 2018; Moyer, 2012). However, we acknowledge this conflation of sex and gender is imprecise and incomplete since not all individuals with cervixes are women and vice versa (Caughey et al., 2021).

2.1. Measures

Our primary outcome was receipt of a Pap test within the past year. We categorized respondents who answered Pap test questions with “Don’t know/Not sure” or “Refused” as missing but test the sensitivity of our results to this coding (Centers for Disease Control and Prevention, 2019b).

We conducted sub-group analyses based on self-reported race/ethnicity, education, and marital status. We categorized respondents as non-Hispanic White, non-Hispanic Black, Hispanic/Latina, and non-Hispanic other/multi-race. The other/multi-race category includes respondents who identified as Asian, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander, or as more than one race. We characterized respondents based on highest grade of school completed (college graduate, some college, high school or less) and marital status (never married, married/partnered, divorced/separated, widowed). Finally, we included indicators for having had a hysterectomy and for having health insurance as covariates.

2.2. Statistical analysis

We calculated population-weighted means and proportions of demographic characteristics, annual Pap testing, hysterectomy prevalence, and health insurance coverage in our overall sample and compared these characteristics between respondents ages 56–65 (n = 113,242) and ages 66–76 (n = 112,789).

We employed a regression discontinuity design (RDD) to estimate changes in annual Pap testing at age 66, the first age that USPSTF recommends against routine cervical cancer screening for most women. The RDD, which has been used previously to study age-based cancer screening guidelines (Jacobson and Kadiyala, 2017; Kadiyala and Strumpf, 2016; Kadiyala and Strumpf, 2011), uses a discrete change in policy eligibility to estimate the causal effect of that policy on outcomes among individuals just barely eligible relative to just barely ineligible. In the current context, we compared Pap test rates for women just after relative to just before age 66. We analyzed our data graphically, plotting the regression-adjusted average rate of annual Pap testing at each age in years along with linear age-profiles on either side of age 66. Operationally, our linear estimates are just least-squares-fitted values from our RDD model:

$${\mathrm{Y}}_{\mathrm{i}}=\mathrm{\alpha }+\beta {\mathrm{D}}_{\mathrm{i}}+\mathrm{\gamma }\left({\mathrm{A}}_{\mathrm{i}}-66\right)+\mathrm{\delta }\left[\left({\mathrm{A}}_{\mathrm{i}}\mathrm{i}-66\right)\mathrm{D}\mathrm{i}\right]+{\mathrm{e}}_{\mathrm{i}}$$ (1)

where ${\mathrm{Y}}_{\mathrm{i}}$ is an indicator for women who received Pap testing during the prior 12 months, ${\mathrm{D}}_{\mathrm{i}}$ is an indicator for women age 66 and over, and ${\mathrm{A}}_{\mathrm{i}}$ is age in years. We specified linear age-specific trends that vary on either side of age 66. We centered these trends at age 66, which allows us to interpret $\mathrm{\beta }$ as the change in the share of women screened due to the recommendation at age 66. We applied the final weight variable from BRFSS to all statistical results to produce nationally representative estimates. We performed all analyses using Stata/MP 16.1 (StataCorp, 2019).

Our main analysis included women between ages 56–76, a bandwidth of 10 years before and after the age 66 discontinuation threshold. Women between ages 66 and 76 are “treated” since they are subject to the recommendation against Pap testing, and women between ages 56–65 are “control.” We stratified our analysis by race/ethnicity, education, and marital status to better understand the impact of the USPSTF age guidelines across different demographic and socioeconomic subgroups. We also stratified our sample by hysterectomy status to account for the effect of hysterectomy prevalence on Pap test rates.

The RDD implicitly assumes that people around the age 66 threshold have similar characteristics but for the cervical cancer screening recommendation. While characteristics can vary, they should not change discretely at the threshold. We tested this assumption by analyzing observable characteristics – including race/ethnicity, educational attainment, hysterectomy, marital status, and health insurance status – around age 66 using the model specified in (1).

We conducted several sensitivity checks. First, to ensure that our results isolated the impact of discontinuation rather than other factors, we estimated RDD models that controlled for health insurance and race/ethnicity (as measured by the share identifying as Non-Hispanic White) - the two covariates in our data that change discontinuously at age 66. Second, to ensure that the RDD estimates were not driven by trends distant from the age 66 threshold, we estimated models using narrower age bandwidths (61–71, 63–69, 65–67). An additional aim of analyzing only respondents ages 65–67 was to test the potential confounding impact of Medicare entitlement beginning at age 65. Third, we modified how we defined our Pap test outcome. We estimated changes in Pap test receipt in the last two years to align with Medicare coverage frequency for beneficiaries with low or average risk of cervical cancer (Centers for Medicare and Medicaid Services, 2006). We shifted the age threshold to age 67 to account for the two-year retrospective period and examine if Medicare benefits, which may be more restrictive than private health insurance benefits, contribute to the drop in respondents reporting an annual Pap test annually between ages 65 and 66. We also estimated Pap test receipt within the last three years, as reported at age 68, to align with the USPSTF’s recommended Pap testing timeline for women ages 21–65. We estimated changes in receipt of annual human papillomavirus (HPV) test and/or a Pap testing to determine if including HPV testing affected our findings, since the USPSTF alternately recommends HPV/Pap co-testing or, as of 2018, only HPV testing as cervical cancer screening options for women ages 30–65 (Appendix Table A). As placebo checks of our Pap test findings, we also assessed changes at age 66 in rates of annual mammography and routine medical checkups, preventive services that the USPSTF does not recommend discontinuing after age 65 (Siu, 2016). Finally, we tested the sensitivity of our results to missing data - estimating RDD models with the missingness of each variable as the outcome to determine if, in alignment with RDD assumptions, the change in missingness of each variable was smooth across the age 66 threshold. We also performed two RDD analyses - one that coded respondents with missing Pap test data (approximately 8% of our sample) as having received an annual Pap test and the other that coded them as not having received an annual Pap test.

3. Results

The demographic characteristics of women ages 56 to 65 and 66 to 76 were comparable (Table 1), although the treated group had a higher percentage who identified as Non-Hispanic White (75.1% compared to 69.5%) and a higher percentage with a high school or less education (43.4% compared to 39.4%) relative to the control group. The control group, conversely, had a higher percentage who were married/partnered (61.5% compared to 52.8%). As expected, the treated group had lower rates of receiving a Pap test within one year (22.5% compared to 39.0%) compared to the control group, despite higher rates of health insurance.

Table 1.

Weighted Characteristics of Female BRFSS Respondents Age 56–76 Years: United States, 2016 and 2018.

Characteristic	56–76	56–65	66–76
Unweighted n	226,031	113,242	112,789
Age (mean)	65.0	60.6	70.1
Race / ethnicity (%)
Non-Hispanic White	72.0	69.5	75.1
Non-Hispanic black	10.6	11.5	9.5
Hispanic/Latina	10.1	11.4	8.6
Non-Hispanic other or multiracial	5.5	5.9	5.0
Missing	1.9	1.8	1.9
Education (%)
High school or less	41.1	39.4	43.4
Some college	32.7	32.8	32.7
College graduate	25.8	27.5	23.5
Missing	0.4	0.3	0.4
Marital status (%)
Never married	6.3	7.8	4.6
Married/partnered	57.7	61.5	52.8
Divorced/separated	19.2	20.3	17.9
Widowed	16.0	9.8	24.1
Missing	0.7	0.7	0.6
Had pap test within 1 year (%)
Unweighted n	208,100	105,041	103,059
(%)	31.8	39.0	22.5
Had HPV test within 1 year
Unweighted n	150,658	75,210	75,448
(%)	9.8	13.2	5.3
Had hysterectomy
Unweighted n	193,723	96,756	96,967
(%)	37.0	32.4	42.9
Have health insurance
Unweighted n	225,462	112,920	112,542
(%)	95.1	92.5	98.4

All covariates except for health insurance status and race/ethnicity were smooth across the age threshold (Appendix Fig. B). The share insured increased by 3.7 percentage points between ages 65 and 66 (95% Confidence Interval [CI]: 2.8, 4.6; Appendix Table B), likely reflecting age-65 Medicare entitlement. Although this change violates an identifying assumption of the RDD, it should lead us to underestimate the relationship between the discontinuation guidance and the decline in Pap test rates since prior work has linked Medicare eligibility to increased cancer screening, particularly among previously uninsured individuals (McWilliams et al., 2003; Sun et al., 2022). The share who identified as Non-Hispanic White increased by 2.5 percentage points between ages 65 and 66 (95% CI: 0.5, 4.5; Appendix Table B). Combined with lower annual Pap test rates among Non-Hispanic White women in this age band (Table 2), the increase could overstate the estimated impact of the USPSTF guidelines on Pap test discontinuation at age 66. In sensitivity analyses, we show that our results are insensitive to controlling for insurance status and Non-Hispanic White race (Appendix Table C).

Table 2.

Change in Annual Pap Test Rate for Women at Age 66 by Race and Ethnicity and Education in BRFSS: United States, 2016 and 2018.

	All	Non-Hispanic White	Non-Hispanic black	Hispanic/Latina	Other or multiracial	High school or less	Some college	College Graduate
Change at age 66a	−5.9	−6.0	0.8	−6.7	−12.2	−5.8	−3.4	−9.0
(95% CIb)	(−7.7, −4.1)	(−7.8, −4.2)	(−5.4, 7.1)	(−14.4, 1.1)	(−25.2, 0.9)	(−8.8, −2.9)	(−6.6, −0.2)	(−12.0, −5.9)
Pap test rate, ages 56–65 (weighted)	39.0%	37.2%	46.2%	44.0%	36.9%	36.8%	37.3%	43.9%
Observations	208,100	168,157	17,074	10,123	9722	71,025	61,193	75,412
R-squared	0.037	0.039	0.031	0.020	0.031	0.032	0.034	0.045

^aEstimates show percentage point change.

^b95% CI refers to 95% confidence interval.

Our main results are shown in Fig. 1, which plots regression-adjusted rates of annual Pap testing by age in years (open circles) and the linear fit from the RDD model in Eq. (1). Panel A shows a sharp change at age 66 in the proportion of women who received a Pap test within the last year. Based on the RDD estimate, the decline was 5.9 percentage points (95% CI: −7.7, −4.1) at age 66 (Table 2). This sharp drop contrasts with the otherwise smooth and more gradual decline in Pap test rates with age reported before 65 and after 66.

Fig. 1.

Annual Pap Test Rate for Women Age 56–76 by (A) Age and (B) Age and Race/Ethnicity in BRFSS: United States, 2016 and 2018.

Stratifying our sample by race/ethnicity, self-reported annual Pap test rates were consistently higher for non-Hispanic Black and Hispanic/Latina women, and of varying difference for non-Hispanic other and multi-race women compared to non-Hispanic White women ages 56–76. The change in annual Pap test rate at age 66 also varied by race/ethnicity. Annual Pap test rate change was virtually indiscernible at age 66 for non-Hispanic Black women, similar in size for non-Hispanic White and Hispanic/Latina women, and larger for other and multi-race women (Fig. 1B). Estimates from the RDD models (Table 2) show a decline of 6.0 percentage points (95% CI: −7.8, −4.2) for non-Hispanic White women, a statistically insignificant decline of similar magnitude for Hispanic/Latina women (−6.7 percentage points [p.p.], 95% CI: −14.4, 1.1), an even larger but also imprecisely estimated decline of 12.2 percentage points (95% CI: −25.2, 0.9) for non-Hispanic other and multi-race women, and no change for non-Hispanic Black women (0.8 p.p., 95% CI: −5.4, 7.1).

The change in annual Pap test rate at age 66 also varied by education (Table 2 and Appendix Fig. A) and marital status (Appendix Fig. E and Table H). For women who had a high school education or less, the decline at age 66 was similar to the full sample change (−5.8 p.p., 95% CI: −8.8, −2.9). Women with some college education had a smaller decline (−3.4 p.p., 95% CI: −6.6, −0.2) while women who graduated college had a larger decline (−9.0 p.p., 95% CI: −12.0, −5.9). Although Pap test rates declined for all marital status groups, the decline was largest for those who were never married (−9.4 p.p., 95% CI: −17.3, −1.5 compared to −6.7 p.p., 95% CI: −11.0, −2.3 for divorced/separated and −5.0 p.p., 95% CI: −7.3, −2.6 for married/partnered.

We also examined differential hysterectomy rates around the age 66 threshold. Although hysterectomy prevalence increases with age, the change is not discontinuous at age 66 (Appendix Fig. B.iv). The RDD estimate shows a small, non-significant 1.1 percentage point increase in the probability of hysterectomy at age 66 (Appendix Table B). Stratifying the sample by hysterectomy status, we found that the annual Pap test rate for women ages 56–65 who had had a hysterectomy was significantly lower than for women of the same age who had not had a hysterectomy (27.9% compared to 44.3%) and the estimated decline at age 66 was smaller - a decrease of 3.8 percentage points (95% CI: −6.9, −0.8) compared to 6.7 percentage points (95% CI: −9.1, −4.3; Table 3).

Table 3.

Change in Annual Pap Test Rate for Women at Age 66 by Hysterectomy Status in BRFSS: United States, 2016 and 2018.

	Pap test rate Hysterectomy	Pap test rate No hysterectomy
Change at age 66a	−3.8	−6.7
(95% CIb)	(−6.9, −0.8)	(−9.1, −4.3)
Pap test rate, ages 56–65 (weighted)	27.9%	44.3%
Observations	71,064	116,617
R-squared	0.018	0.040

^aEstimates show percentage point change.

^b95% CI refers to 95% confidence interval.

Our results were robust to alternate model and outcome specifications. Amending our RDD model to control for insurance status and non-Hispanic White race only nominally increased the estimated decline in Pap test rates at age 66 to 6.3 percentage points (95% CI: −8.1, −4.5; Appendix Table C), suggesting these discontinuities minimally impact the estimate. Models using narrower age bands produced similar estimated declines between −6.2 and −4.7 percentage points (Appendix Table D). Substituting combined annual HPV and/or Pap test, two-year Pap test, and three-year Pap test rates as outcomes generated similar 6.1, 6.1, and 6.4 percentage point declines respectively (Appendix Fig. C, Fig. D, and Table E). In placebo checks, we found no discrete drop at age 66 in annual mammogram (1.5 p.p., 95% CI: −0.2, 3.3; Appendix Fig. F and Table I) or routine checkup rates (1.9 p.p., 95% CI: 0.7, 3.0; Appendix Fig. G and Table I), services not subject to age 66 discontinuation recommendation. Finally, analyses of data missingness did not identify concerns. Missingness rates for all variables changed smoothly across the age 66 threshold and alternative coding of missing Pap test data did not meaningfully impact our results (Appendix Tables F and G).

4. Discussion

The USPSTF recommends that most women stop regular cervical cancer screening after age 65. Consistent with this recommendation, we found a sizeable drop in the annual Pap test rate at age 66 – nearly 6 percentage points or about 15% off the pre-66 annual test rate of 39% – that was robust to several sensitivity analyses. We also found differences in both the level and change in Pap test rates by race/ethnicity. Between ages 56–65, Non-Hispanic Black and Hispanic/Latina women had higher annual Pap test rates than non-Hispanic White women (46.2% and 44.0% compared to 37.2%, respectively), gaps that remained through age 76. This pattern is noteworthy given prior evidence that non-Hispanic Black and Hispanic/Latina women have lower lifetime Pap test rates (McDaniel et al., 2021). In contrast to every other racial/ethnic group, non-Hispanic Black women were less affected by the discontinuation recommendation, which mitigates the concern that screening rates among older non-Hispanic Black women, who have heightened cervical cancer risk, might be disproportionately adversely impacted by the recommendation.

We also found differential changes in Pap test rates by education and marital status. Women who graduated from college had higher annual Pap test rates between ages 56–65 and a steeper decline at age 66 compared to women without a college degree. Changes at the age threshold narrowed but did not eliminate disparities in annual Pap test rates by education; despite a large drop, college graduates maintained the highest annual Pap test rates past age 70 (Appendix Fig. A). We found similar annual Pap test rates among married/partnered, divorced/separated, and never married women between ages 56–65 but a steeper decline at age 66 among women who were never married. Real or assumed differences in sexual activity by marital status may influence how patients or providers respond to the discontinuation recommendation. Variation in the size of the decline in Pap test rates at age 66 by education, marital status, and race may also reflect differences in prior screening history and risk, per the USPSTF recommendations, although our data do not allow us to test this directly.

Our study has several limitations. First, the RDD estimates a local average treatment effect and cannot estimate the average impact of the recommendation on all women over age 66. Second, the RDD model fit was limited because BRFSS measures age in years. Thus, some respondents age 66 may have received a Pap test within the last 12 months while still age 65. This imprecision has a smoothing effect, however, minimizing the difference in annual Pap test rates between ages 65 and 66 and limiting our ability to detect a discontinuity at age 66. That we still detect a discontinuity in Pap test rates at age 66 supports our findings that Pap test rates decrease discontinuously at the USPSTF-recommended age threshold. Third, the RDD assumes that no factor other than the USPSTF recommendation changes at age 66 could influence Pap test rates. Since Medicare eligibility and thus insurance coverage changes discretely at age 65 (Appendix Fig. B.i, B.iii, and Table B), just a year prior to the discontinuation age, it risks biasing the result. Given that health service utilization increases with insurance coverage (Card et al., 2008), however, this change actually bolsters our conclusion: Pap testing declines at the discontinuation age threshold. Fourth, we studied annual Pap test receipt to more readily discern a discrete change at the age threshold but the USPSTF recommendation specifies testing every three years; sensitivity analyses considering testing every three years still demonstrated a discontinuous decline. Fifth, our primary analysis excluded HPV screening history and did not capture potential respondents who opted for only HPV testing or Pap/HPV co-testing, received their last Pap test more than three years ago, and still met USPSTF recommendations. However, sensitivity analysis accounting for both annual HPV and Pap test rates yielded similar results (Appendix Fig. C and Table E). Finally, our data are cross-sectional and do not capture medical history or lifetime Pap test receipt, information necessary to fully determine adherence to USPSTF recommendations.

5. Conclusions

While our results demonstrate that the USPSTF discontinuation recommendation influences Pap test rates, our work cannot speak directly to mechanisms. Research focused on healthcare providers’ knowledge, interpretation, and implementation of USPSTF’s cervical cancer screening recommendations and on how patients make screening decisions past age 65 is needed to disentangle patient- and provider-driven aspects of the change in Pap test utilization at age 66, and to identify factors that explain variation in the decline by race, ethnicity, education and marital status. Future work might also examine changes in the Pap test rate further from the age 66 threshold. In addition to the discontinuous drop in Pap test rates at age 66, the slope of Pap test rate decline steepens after age 66 (Fig. 1A). Most importantly, our work cannot speak to whether the decline in screening at age 66 negatively or positively affects older women. Given recent concerns about the applicability and value of the USPSTF recommendation, our findings underscore the importance of re-evaluating the health implications of discontinuing cervical cancer screening on the incidence and timing of cervical cancer diagnosis, morbidity, and mortality (Dilley et al., 2021; Gravitt et al., 2018).

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ypmed.2023.107543.

Data availability

The data that support this study’s findings are publicly available at www.cdc.gov/brfss. Code used for analysis is available on request from the corresponding author.