Cervical Cancer Screening and Prevalence Among Older US Medicare Beneficiaries with and Without HIV

Abstract

Females with HIV (FWH) are recommended to receive cervical cancer screening annually, with the interval extended to every three years after three sequential normal results. Lifelong screening is highly recommended due to their increased risk of HPV infection and cervical cancer. We assessed the trends in cervical cancer screening rates and cervical cancer/precancer prevalence among older FWH and females without HIV (FWOH) using 2007–2019 U.S. Medicare data. We found that age-adjusted cervical cancer screening rates decreased similarly in both FWH and FWOH (Average annual percentage change [AAPC]: −4.4 [95% CI: −5.2, −3.6] vs. −5.7 [95% CI: −6.8, −4.7], p=0.11). However, the age-adjusted cervical cancer/precancer prevalence showed increasing rates among FWH (5.4, [2.9, 7.9]) while stable in FWOH (−0.6, [−1.4, 0.1]). These findings underscore the need for strict adherence to clinical practice guidelines for cervical cancer screening in older FWH.

Introduction

Cervical cancer was once the leading cause of cancer-related deaths in America, with current estimates indicating that more than 20% cases are found in women aged 65 and older^1–5. Human papillomavirus (HPV), the precursor event to cervical cancer, is the most prevalent sexually transmitted infection in the United States, with high-risk (HR) types detected in 99% of cervical precancers ⁶.

Human immunodeficiency virus (HIV) compromises the immune system, decreasing the body’s ability to clear HPV and thereby increasing the risk of developing precancerous lesions and cervical cancer¹, particularly with advancing age. Females with HIV (FWH) are about six times more susceptible to cervical cancer than females without HIV (FWOH)⁷. However, to our knowledge, no studies have compared risk differences between older FWH and FWOH across calendar years.

Cervical cancer screenings have declined over the years, with Pap testing use dropping from 10.7% in 2016 to 8% in 2020 among over 20 million U.S. females aged 18 and older⁸. Recent data from the National Health Interview Survey showed that cervical cancer screening in 2023 remained 14% lower than in 2019⁹, indicating a continued and persistent decline in recent years¹⁰. Additionally, the number of women aged 65 and older with Medicare fee-for-service (FFS) coverage screened for cervical cancer (cytology or HPV testing) fell from 2.9 million in 1999 to 1.3 million in 2019¹¹, likely as a result of changes in screening guideline in 2012, in which screening cessation was recommended for average-risk women over 65 with adequate prior cervical cancer screening and no history of high grade cervical intraepithelial neoplasia¹¹.

Screening options and periods now vary based on risk level and age. Regular cervical cancer screening is recommended for individuals aged 21 through 65 years, but it is no longer recommended for the general population beyond age 65, after adequate negative prior screenings^12–14. For FWH, annual screening is recommended and may begin before age 21, with the interval extended to every three years after three sequential normal results¹⁴. Lifelong screening is highly recommended for FWH due to their increased risk of HPV infection and cervical cancer^13–14,17. FWH aged 65 or older should continue regular screening, as they remain at higher risk for persistent HPV infections and cervical intraepithelial neoplasia due to reduced CD4+ T-lymphocytes^15–16. Persistent HR-HPV infection promotes progression from high-grade squamous intraepithelial lesions to invasive cancer¹⁸. Despite these risks, cervical cancer screening trends in FWH have been understudied since implementation of the guidelines. We therefore used population-level data to assess the prevalence of HPV-related cervical cancers and precancers as well as cervical cancer screening rates and their temporal trends among U.S. Medicare beneficiaries ages 65 and older with and without HIV.

Materials and Methods

We used U.S. Medicare data from 100% of female beneficiaries aged 65 and older with an HIV diagnosis (flagged as a chronic condition in the database) and a 5% national sample of older female enrollees without an HIV diagnosis in any year between 2007–2019. The details on the U.S. Medicare databases have been described in a previous study ¹⁹. We constructed successive cross-sectional cohorts of FWH and FWOH for each year that they had continuous FFS coverage (Medicare parts A and B with no Health Maintenance Organization enrollment) (Table S1). These cohorts served as the denominators. We did not limit the cohort to women with a history of cervical intraepithelial neoplasia II (CIN II) or above due to the lack of accurate or complete information on prior diagnoses in the available data. Master Beneficiary Summary Files were used to extract the demographic characteristics including age, race/ethnicity, Medicaid Dual Eligibility, original entitlement for Medicare and region. Cervical cancers and precancers were identified via ICD-9-CM/ICD-10-CM diagnosis codes (with 1 inpatient or 2 outpatient diagnoses, Table S2). The precancer diagnosis was determined either based on cytologic Pap tests or cervical histology. We defined high grade squamous intraepithelial lesions, moderate dysplasia, and severe dysplasia as precancer. The prevalences of precancer, cancer and combined cases (precancer or cancer) were calculated for each calendar year. Cytology and HPV testing for cervical cancer screening were identified using the ICD-9-CM/ICD-10-CM, CPT or HCPCS codes (Table S3). Screening rate was then calculated using the same denominators. Crude rates were adjusted by direct standardization using the reference age distribution from FWOH in 2007. Average annual percentage changes (AAPC) and their 95% confidence intervals (CI) in cervical cancers/precancers prevalence and cervical cancer screening for FWH and FWOH were determined via Joinpoint regression analysis²⁰.

Results

We identified 32,586 FWH and 1,548,980 FWOH, with a mean follow-up of 5.0 person-years for FWH and 6.7 person-years for FWOH (Table S1). Table 1 shows the demographic characteristics of FWH and FWOH in 2007 and 2019. In 2007, FWH were about 3 years younger on average than FWOH. A higher proportion of Black and Hispanic beneficiaries was observed among FWH compared with FWOH (34.8% vs. 8.8%). More FWH had Medicare-Medicaid dual coverage than FWOH (44.5% vs. 18.4%), and fewer FWH than FWOH qualified for Medicare due to age (81.9% vs. 93.2%). Most FWH resided in the South region (53.4% vs. 38.5%). Over time, demographic shifts were observed, particularly among FWH, whose racial/ethnic composition changed from 62.6% White and 23.5% Black in 2007 to 49.8% White and 40.2% Black in 2019. In contrast, the racial/ethnic distribution of FWOH remained relatively stable. Patterns in Medicare–Medicaid dual coverage also changed during the study period. Among FWOH, the proportion with dual coverage declined from 18.4% in 2007 to 13.8% in 2019, whereas among FWH the proportion decreased from 44.5% in 2007 to 39.9% in 2019.

Table 1.

Demographics by HIV status and period

	FWOH		FWH
Characteristics N (%)	2007 (N=872,580)	2019 (N=766,153)	2007 (N=10,415)	2019 (N=14,850)
Age (Mean, SD)	78.5 (7.8)	77.5 (8.2)	75.6 (7.0)	75.5 (8.0)
Age group
66–69	120,307 (13.8)	141,135 (18.4)	2,427 (23.3)	4,452 (30.0)
70–74	193,141 (22.1)	204,608 (26.7)	2,815 (27.0)	3,653 (24.6)
75–79	186,550 (21.4)	143,791 (18.8)	2,243 (21.5)	2,477 (16.7)
80+	372,582 (42.7)	276,619 (36.1)	2,930 (28.1)	4,268 (28.8)
Race/Ethnicity
White	771,205 (88.4)	660,035 (86.2)	6,518 (62.6)	7,401 (49.8)
Black	64,871 (7.4)	54,121 (7.1)	2,449 (23.5)	5,974 (40.2)
Hispanic	12,452 (1.4)	10,871 (1.4)	1,175 (11.3)	829 (5.6)
Other/Unknown	24,052 (2.8)	41,126 (5.4)	273 (2.6)	646 (4.4)
Medicare-Medicaid Dual coverage	160,529 (18.4)	105,880 (13.8)	4,638 (44.5)	7,410 (49.9)
Original Medicare entitlement (old age)	813,316 (93.2)	699,306 (91.3)	8,527 (81.9)	9,682 (65.2)
Region
Northeast	224,826 (25.8)	171,402 (22.4)	1,492 (14.3)	2,051 (13.8)
South	336,077 (38.5)	303,642 (39.6)	5,566 (53.4)	6,953 (46.8)
Midwest	173,444 (19.9)	140,211 (18.3)	2,015 (19.4)	3,746 (25.2)
West	134,077 (15.4)	148,060 (19.3)	1,279 (12.3)	2,002 (13.5)

The adjusted prevalence of cervical cancers/precancers among FWH increased from 275 per 100,000 women in 2007 to 406 per 100,000 women in 2019, while the rate among FWOH remained stable (103 to 97 per 100,000 women) (Figure 1A, solid lines). The AAPC of cervical cancers/precancers increased significantly among FWH (5.4 [95% CI: 2.9, 7.9]) but showed no significant changes among FWOH (−0.6 [−1.4, 0.1]). Separate analyses of precancer and cancer (Figure 1B and 1C, solid lines) showed similar trends between FWH and FWOH. In FWH, the AAPC was slightly higher for cancer (5.2 [95% CI: 2.1, 8.4]) than for precancer (4.0 [95% CI:1.0, 7.2]). In contrast, in FWOH, the precancer rate showed a significant decline (AAPC: −2.5 [95% CI: −3.7, −1.3]), while the change in cancer rate was not statistically significant (AAPC: −0.4 [95% CI: −1.3, 0.4]). Age-adjusted cervical cancer screening rates decreased over time in both groups. Among FWH the rate decreased from 14,591 per 100,000 women in 2007 to 8,475 per 100,000 women in 2019, while among FWOH the rate decreased from 13,986 per 100,000 women in 2007 to 7,006 per 100,000 women in 2019 (Figure 1A, 1B and 1C, dashed lines). The AAPC of cervical cancer screening showed decreasing rates in both groups (−4.4 [95% CI: −5.2, −3.6]) for FWH vs (−5.7 [95% CI: −6.8, −4.7]) for FWOH; (p=0.11).

Figure 1. Age adjusted screening and precancer/cancer prevalence over year by HIV status.

Discussion

The increasing age-adjusted prevalence rates of cervical cancer and precancer, particularly cancer among older FWH U.S. Medicare beneficiaries, alongside declining age-adjusted screening rates during 2007–2019, raises a public health concern, particularly given the aging population and the need for lifelong screening in this high-risk group. These findings also present a significant obstacle to achieving the WHO goal of eliminating cervical cancer by 2030^21-22. Despite being a highly preventable disease, cervical cancer screening remains suboptimal, especially among FWH, who face an elevated risk of developing invasive cervical cancer. A study conducted in the United States across three healthcare settings reported similarly lower screening rates in later calendar years among women living with HIV ²³, which is consistent with our findings. Cervical cancer screening should continue after age 65 for FWH, and the decreasing screening rate over time, similar to those observed in FWOH, is worrisome.

These results suggest that providers may not consistently implement the recommended screening practices. As FWH often require multidisciplinary care, preventive screenings may be deprioritized in favor of more urgent health needs. To improve uptake of screening services for FWH, cervical cancer screening should be routinely and proactively integrated into HIV care, coupled with targeted patient education on prevention.

Evidence indicates that cervical cancer is not limited to younger women, with about 20% of cases diagnosed in women over age 65 in the United States^4–5. While our findings also revealed relatively stable age-adjusted prevalence rates of cervical cancer/precancers among older FWOH during the same period, decreased screening rates were observed. Despite the increased risk of late-stage diagnosis of cervical cancer in older women, many remain unscreened due to current guidelines that do not recommend screening beyond age 65. Late-stage cervical cancer has a poorer prognosis and higher treatment costs, potentially increasing the healthcare burden and contributing to future disparities in cervical cancer incidence and mortality. Thus, the exclusion of women over age 65 from screening may need to be reconsidered.

One important limitation of our study is that we were unable to restrict the analysis to women with a documented history of CIN II or higher, who are recommended to continue screening beyond age 65. As a result, our estimates may underestimate screening rates among women for whom continued surveillance is clinically indicated. Another limitation is the inability to assess guideline-concordant cervical cancer screening intervals due to insufficient longitudinal data and lack of prior screening histories. Therefore, we used annual screening rates as a proxy to examine patterns across groups. Although this approach does not reflect current guideline recommendations, it enables relative comparisons of screening behavior between FWH and FWOH. Additional limitations are inherent to claims data, including potential inaccuracies or incomplete capture of cervical cancer/precancer diagnoses. The number of cancer cases among FWH per year was also relatively small. Nevertheless, to our knowledge, this is the largest nationwide study to examine temporal trends over more than a decade in cervical cancer screening and the prevalence of cervical precancers and cancers among older Medicare enrollees with HIV compared to those without HIV. These results are generalizable to the older FFS Medicare population and hold important implications for cervical cancer prevention in women living with HIV.

Future studies using Medicare encounter data for the HMO population are needed. Our findings did not establish a causal relationship between cervical cancer screening and cancer prevalence. Instead, we provide population-based estimates in light of observed trends, and the burden of disease warrants further in-depth research on the impact of screening on outcomes using longitudinal studies and cancer registries. Future work should also explore interventions to improve screening uptake among older women living with HIV and evaluate the cost-effectiveness of extending screening beyond age 65 in high-risk groups.

In conclusion, despite the higher and increasing burden of cervical cancer/precancer over time among older FWH and recommendation of lifetime cervical cancer screening in this population, cervical cancer screening decreased in a similar rate as in FWOH. These findings highlight the need to enhance the cervical cancer screening implementation in the routine primary care of older FWH.

Novelty and Impact.

The novelty of this study lies in its assessment of temporal trends in the prevalence of HPV-related cervical cancers and precancers alongside cervical cancer screening among older U.S. Medicare beneficiaries, both with and without HIV. The finding of increasing age-adjusted prevalence rates of cervical cancer and precancers, coupled with declining age-adjusted screening rates in females with HIV (FWH), underscores the need to adhere to clinical practice guidelines for cervical cancer screening in older FWH.

Abbreviations

AAPC

Average annual percentage change

CI

Confidence interva

CPT

Current Procedural Terminology

FFS

Fee-for-service

FWH

Females with HIV

FWOH

Females without HIV

HCPCS

Healthcare Common Procedure Coding System

HIV

Human immunodeficiency virus

HMO

Health Maintenance Organization

HPV

Human papillomavirus

HR

High-risk

ICD-9-CM

International Classification of Diseases, 9th Revision, Clinical Modification

ICD-10-CM

International Classification of Diseases, 10th Revision, Clinical Modification

IRB

Institutional Review Board

U.S.

United States

WHO

World Health Organization

Data Availability Statement:

Access to Medicare data can be obtained directly through CMS at at https://resdac.org/, subject to standard data use agreements. Further details and other data that support the findings of this study are available from the corresponding authors upon request.