Use Trends and Recent Expenditures for Cervical Cancer Screening–Associated Services in Medicare Fee-for-Service Beneficiaries Older Than 65 Years

Key Points

Question

What are the use trends and expenditures for cervical cancer screening–associated services in different age groups in women older than 65 years?

Findings

In this population-based, observational cross-sectional study using Medicare fee-for-service claims data from 1999 to 2019, annual use rates of cytology and/or human papillomavirus testing, colposcopy, and cervical procedures decreased by 55.3%, 43.2%, and 64.4%, respectively. In 2019, 8% (1.3 million women) received at least 1 service at an expenditure of $83 527 181, including $7 391 101 spent among women older than 80 years.

Meaning

The study results suggest that more than 1.3 million women in the Medicare fee-for-service program recently received cervical cancer screening–associated services after age 65 years at substantial cost but of unclear clinical appropriateness.

Abstract

Importance

Since 1996, the US Preventive Services Task Force has recommended against cervical cancer screening in average-risk women 65 years or older with adequate prior screening. Little is known about the use of cervical cancer screening–associated services in this age group.

Objective

To examine annual use trends in cervical cancer screening–associated services, specifically cytology and human papillomavirus (HPV) tests, colposcopy, and cervical procedures (loop electrosurgical excision procedure, cone biopsy, and ablation) in Medicare fee-for-service beneficiaries during January 1, 1999, to December 31, 2019, and estimate expenditures for services performed in 2019.

Design, Setting, and Participants

This population-based, cross-sectional analysis included health service use data across 21 years for women aged 65 to 114 years with Medicare fee-for-service coverage (15-16 million women per year). Data analysis was conducted between July 2021 and April 2022.

Main Outcomes and Measures

Proportion of testing modalities (cytology alone, cytology plus HPV testing [cotesting], HPV testing alone); annual use rate per 100 000 women of cytology and HPV testing, colposcopy, and cervical procedures from 1999 to 2019; Medicare expenditure for these services in 2019.

Results

There were 15 323 635 women 65 years and older with Medicare fee-for-service coverage in 1999 and 15 298 656 in 2019. In 2019, the mean (SD) age of study population was 76.2 (8.1) years, 5.1% were Hispanic, 0.5% were non-Hispanic American Indian/Alaska Native, 3.0% were non-Hispanic Asian/Pacific Islander, 7.4% were non-Hispanic Black, and 82.0% were non-Hispanic White. From 1999 to 2019, the percentage of women who received at least 1 cytology or HPV test decreased from 18.9% (2.9 million women) in 1999 to 8.5% (1.3 million women) in 2019, a reduction of 55.3%; use rates of colposcopy and cervical procedures decreased 43.2% and 64.4%, respectively. Trend analyses showed a 4.6% average annual reduction in use of cytology or HPV testing during 1999 to 2019 (P < .001). Use rates of colposcopy and cervical procedures decreased before 2015 then plateaued during 2015 to 2019. The total Medicare expenditure for all services rendered in 2019 was about $83.5 million. About 3% of women older than 80 years received at least 1 service at a cost of $7.4 million in 2019.

Conclusions and Relevance

The results of this cross-sectional study suggest that while annual use of cervical cancer screening–associated services in the Medicare fee-for-service population older than 65 years has decreased during the last 2 decades, more than 1.3 million women received these services in 2019 at substantial costs.

This cross-sectional study examines annual use trends in cervical cancer screening–associated services in Medicare fee-for-service beneficiaries and estimates expenditures for services.

Introduction

More than 20% of 65 061 cervical cancer cases during 2014 to 2018 were diagnosed in women older than 65 years in the US, and more than 36% of 20 823 cervical cancer–related deaths during this period occurred in this group.¹ Many cervical cancer cases occur in women who were never screened or not adequately screened.² In 1996, the US Preventive Services Task Force (USPSTF) concluded that there was insufficient evidence to recommend for or against an age to end cervical cancer screening but suggested that age 65 years was reasonable in women with consistently normal prior screening.³ Although the American Cancer Society (ACS) recommended ending screening in average-risk women aged 70 years and older in 2002,⁴ the American College of Obstetricians and Gynecologists (ACOG) did not recommend an age to end screening until 2009, endorsing screening cessation between ages 65 and 70 years in those with recent prior normal testing.⁵ In 2012, all major organizations (ie, ACOG, ACS, and USPSTF) endorsed ending screening in average-risk women at age 65 years with adequate prior screening. Average risk was defined as having no diagnosis of a cervical precancer or cancer within the last 20 years, no immunocompromising medical condition (such as HIV), and no in utero exposure to diethylstilbestrol.^6,7 The current unified definition of adequate prior screening is having 3 consecutive negative cytology test results or 2 or more consecutive prior human papillomavirus (HPV) test results (with or without cytology) within the prior 10 years with the last normal test result documented within the prior 5 years (if an HPV test) or 3 years (if a cytology test).^{6,7,8,9,10,11}

Use of cervical cancer screening–associated services in women 65 years and older at the population level is largely unknown. It is also unknown how changing guidelines and recommendations have been associated with use of these services in women older than 65 years. While the National Health Interview Survey and the Behavioral Risk Factor Surveillance System have been used to estimate self-reported cervical cancer screening prevalence, to our knowledge, few studies have included analyses in populations older than 65 years.^{12,13,14,15,16} In addition, about 20% of women in these surveys reported not knowing whether they had an HPV test in the past.^17,18 While other studies have investigated cervical cancer screening test use using alternative data sources, they have excluded women older than 65 years^19,20,21 or offered limited analysis of use in this population.^{22,23,24,25,26} Furthermore, limited data have been published regarding use of colposcopy and cervical procedures (loop electrosurgical excision procedure, cone biopsy, and ablation).^25,27 Little is known about expenditures for these services in populations older than 65.

Medicare claims data provide an opportunity to understand use and the cost of cervical cancer screening–associated services in different age groups after age 65 years. The objectives of this study were to investigate trends in use of cytology/HPV testing, colposcopy, and cervical procedures during 1999 to 2019 and estimate total expenditures by Medicare fee-for-service beneficiaries for these services in 2019. Because age has implications for the benefits and harms of cervical cancer screening–associated service use in older women, we also analyzed use by age group.

Methods

Data Source and Study Population

We used Medicare claims data for 100% of the Medicare fee-for-service beneficiaries from January 1, 1999, through December 31, 2019 (the most recent year available). Specifically, we obtained 100% carrier claims (professional services), outpatient claims (hospital outpatient facility use), and master beneficiary summary files (enrollment and demographic characteristics) from the US Centers for Medicare & Medicaid Services Chronic Conditions Data Warehouse.²⁸ The study population included women 65 years and older who had 11 or 12 months of Medicare Part A and B fee-for-service coverage (or coverage until time of death) and 1 month or less of health maintenance organization coverage in each calendar year. We analyzed 4 age groups: 65 to 69 years, 70 to 74 years, 75 to 79 years, and 80 years o older. Women 115 years or older were excluded because the recording of their mortality information in the database may not have been timely. In addition, we analyzed use by race and ethnicity, which included Hispanic, non-Hispanic American Indian/Alaska Native, non-Hispanic Asian/Pacific Islander, non-Hispanic Black, and non-Hispanic White. We used the Research Triangle Institute race code variable available in the US Centers for Medicare & Medicaid Services Master Beneficiary Summary File for race and ethnicity data.

Because most Medicare beneficiaries entered the program at age 65 years and their prior medical information was not available in the database, we were not able to determine if women had a prior hysterectomy, adequate screening before age 65 years, surveillance because of a previous abnormal cervical cancer screening result, or cervical precancer or cancer diagnosis. Thus, we refer to outcomes of this study as cervical cancer screening–associated tests and procedures. Institutional review board approval was waived due to the use of deidentified data.

Outcomes

We analyzed annual use rates of cervical cancer screening–associated services in the study population: testing (cytology and/or HPV testing), colposcopy, and cervical procedures (loop electrosurgical excision procedure, cervical conization, or ablation). These tests and procedures were identified using Current Procedural Terminology, Healthcare Common Procedure Coding System, and International Classification of Diseases, Ninth Revision (ICD-9) and ICD-10 diagnosis and procedure codes (eAppendix in the Supplement). Use was analyzed at the person level by querying the databases and identifying women in the study population who received these tests and procedures. For each calendar year, we created a binary (yes/no) variable indicating whether a woman had at least 1 cytology or HPV test. Similarly, we created binary (yes/no) outcome variables indicating receipt of colposcopy and cervical procedures for each woman in a year. We analyzed each outcome (testing, colposcopy, and cervical procedures) separately regardless of whether the woman received other services.

During the past 2 decades, cervical cancer screening guidelines and recommendations evolved to include 3 screening modality options for women 30 years and older: cytology alone every 3 years, HPV testing alone every 5 years, or cytology plus HPV testing (known as cotesting) every 5 years.²⁴ Thus, we analyzed 4 mutually exclusive modalities of cytology and HPV test use: (1) cytology alone: at least 1 cytology test but no HPV test in a year; (2) HPV alone: at least 1 HPV test but no cytology test in a year; (3) cotesting: defined as a cytology plus HPV test within 3 days before or 30 days after the date of the cytology test; and (4) other: had a cytology and HPV test in a year, but not in the cotesting time frame defined previously. For women who had multiple cytology or HPV tests, we used the date of the first test in the year. To assess changes in the screening approach during the study period, we examined the proportion of these 4 modalities used each year.

Statistical Analysis

Annual utilization rates per 100 000 women were calculated for each calendar year from 1999 through 2019, overall, by age group, and race and ethnicity. Age-adjusted rates (overall and by race and ethnicity) were calculated using a direct standardization method, with age standardized to the 2010 US standard female population.²⁹ Data preparation and analyses were conducted using SAS Enterprise Guide (version 7.1; SAS Institute). We used Joinpoint software (version 4.9.0.0)³⁰ to characterize piecewise log-linear time calendar trends in the age-adjusted rates and calculated annual percentage change (APC) and average APC. We used the weighted Bayesian Information Criteria model selection method. The relative percentage change in utilization rates between 1999 and 2019 was calculated. All P values were 2-sided, and we considered P < .05 as statistically significant.

We also examined actual Medicare payment amounts for cytology and HPV testing, colposcopy, and cervical procedures in 2019. We used 100% carrier line records and outpatient revenue center records to identify individual payment for eligible procedure codes (listed in the eAppendix in the Supplement), and we presented Medicare payment for each outcome in total (the sum of all the line item–level payment in 2019) and by age group.

Results

There were 15 323 635 women 65 years and older with Medicare fee-for-service coverage in 1999 and 15 298 656 in 2019. In 2019, the mean (SD) age of study population was 76.2 (8.1) years, 5.1% were Hispanic, 0.5% were non-Hispanic American Indian/Alaska Native, 3.0% were non-Hispanic Asian/Pacific Islander, 7.4% were non-Hispanic Black, and 82.0% were non-Hispanic White. About 15 to 16 million women in each year were included in the study population. Cytology alone was the most used test modality during 1999 to 2019, and its proportion among the total tests decreased from nearly 100% in 1999 to 83.4% in 2019 (Table 1). Concurrently, the cotesting proportion of the total tests increased over time from 0% in 1999 to 15.4% in 2019. An HPV test alone was rarely used, but there was a noticeable increase in its use from 2018 (0.5%) to 2019 (1.0%).

Table 1. Number and Proportion of Medicare Fee-for-Service Beneficiaries^a Who Received Cytology and/or HPV Testing by Modality From 1999 to 2019.

Year	No. (%)
	Cytology alone	Cotestingb	HPV testing alone	Otherc
1999	2 899 355 (99.98)	352 (0.01)	63 (0.00)	254 (0.01)
2000	2 962 448 (99.94)	994 (0.03)	122 (0.00)	654 (0.02)
2001	2 978 789 (99.80)	4087 (0.14)	233 (0.01)	1525 (0.05)
2002	3 050 115 (99.58)	9436 (0.31)	456 (0.01)	2897 (0.09)
2003	2 796 896 (99.38)	14 360 (0.51)	507 (0.02)	2694 (0.10)
2004	2 667 501 (98.71)	30 841 (1.14)	753 (0.03)	3314 (0.12)
2005	2 508 230 (97.39)	61 806 (2.40)	1279 (0.05)	4125 (0.16)
2006	2 342 249 (95.65)	100 000 (4.08)	1487 (0.06)	5089 (0.21)
2007	2 185 085 (94.29)	125 066 (5.40)	2359 (0.10)	4989 (0.22)
2008	2 049 444 (92.94)	147 633 (6.69)	3020 (0.14)	5138 (0.23)
2009	1 945 172 (91.94)	161 598 (7.64)	3801 (0.18)	5090 (0.24)
2010	1 783 298 (91.45)	158 345 (8.12)	3738 (0.19)	4618 (0.24)
2011	1 770 762 (91.11)	164 578 (8.47)	3688 (0.19)	4611 (0.24)
2012	1 631 253 (90.53)	161 803 (8.98)	4446 (0.25)	4322 (0.24)
2013	1 513 573 (89.13)	175 275 (10.32)	5031 (0.30)	4273 (0.25)
2014	1 360 464 (87.74)	181 211 (11.69)	4840 (0.31)	4066 (0.26)
2015	1 279 938 (86.31)	194 138 (13.09)	5332 (0.36)	3630 (0.24)
2016	1 236 231 (85.27)	204 117 (14.08)	5894 (0.41)	3512 (0.24)
2017	1 169 284 (84.58)	203 347 (14.71)	6541 (0.47)	3345 (0.24)
2018	1 115 827 (84.14)	200 348 (15.11)	7007 (0.53)	3016 (0.23)
2019	1 080 341 (83.43)	198 750 (15.35)	12 497 (0.97)	3271 (0.25)

Abbreviation: HPV, human papillomavirus.

^a Women aged 65 to 114 years who were enrolled in Medicare fee-for-service plans for at least 11 months each calendar year.

^b Cotesting was defined as cytology plus HPV testing within 3 days before or 30 days after the date of cytology test.

^c Received cytology and HPV test in a year, but not in the cotesting time frame defined in the table.

The total number of women who received at least 1 cytology and/or HPV test in a calendar year ranged from about 3 million 2 decades ago to 1.3 million in 2019, and the annual use rate of tests decreased from 18 925 per 100 000 women in 1999 to 8464 per 100 000 women in 2019, a reduction of 55.3% (Table 2). The annual use rate was the highest in 1999 for colposcopy (338 per 100 000 women) and cervical procedures (55 per 100 000 women), then decreased 43.2% and 64.4%, respectively, over time.

Table 2. Number and Annual Use Rate of Cytology and/or HPV Testing, Colposcopy, and Cervical Procedures in Medicare Fee-for-Service Beneficiaries^a From 1999 to 2019.

Year	Study population, No.	Cytology/HPV testingb		Colposcopy		Cervical proceduresc
		No.	Rated	No.	Rated	No.	Rated
1999	15 323 635	2 900 024	18 925.2	51 849	338.4	8494	55.4
2000	15 370 214	2 964 218	19 285.5	50 464	328.3	7846	51.0
2001	15 732 188	2 984 634	18 971.5	49 997	317.8	7331	46.6
2002	16 121 155	3 062 904	18 999.3	48 621	301.6	6894	42.8
2003	16 334 516	2 814 457	17 230.1	41 142	251.9	5773	35.3
2004	16 378 192	2 702 409	16 500.0	37 054	226.2	5009	30.6
2005	16 285 638	2 575 440	15 814.2	34 522	212.0	4547	27.9
2006	15 732 372	2 448 825	15 565.5	33 241	211.3	4377	27.8
2007	15 334 177	2 317 499	15 113.3	31 911	208.1	3954	25.8
2008	15 032 705	2 205 235	14 669.6	30 964	206.0	3881	25.8
2009	14 874 937	2 115 661	14 223.0	31 502	211.8	3734	25.1
2010	14 950 510	1 949 999	13 043.0	29 829	199.5	3635	24.3
2011	15 003 050	1 943 639	12 955.0	30 345	202.3	3291	21.9
2012	15 072 549	1 801 824	11 954.3	29 602	196.4	3143	20.9
2013	15 129 397	1 698 152	11 224.2	28 685	189.6	2880	19.0
2014	15 052 888	1 550 581	10 300.9	27 536	182.9	2823	18.8
2015	15 136 074	1 483 038	9798.0	27 072	178.9	2863	18.9
2016	15 382 007	1 449 754	9425.0	27 836	181.0	2921	19.0
2017	15 374 493	1 382 517	8992.3	28 523	185.5	2969	19.3
2018	15 327 860	1 326 198	8652.2	28 830	188.1	2950	19.2
2019	15 298 656	1 294 859	8463.9	29 392	192.1	3017	19.7
Percentage change 1999-2019e	NA	NA	–55.3	NA	–43.2	NA	–64.4

Abbreviations: HPV, human papillomavirus; NA, not applicable.

^a Women aged 65 to 114 years enrolled in Medicare fee-for-service plans for at least 11 months each calendar year.

^b One or more cytology tests or HPV tests in a calendar year, which combines all modalities presented in Table 1.

^c Loop electrosurgical excision procedures, conizations, and ablations.

^d Annual use rate = (number of women received test or procedure/number of total study population) multiplied by 100 000.

^e Calculated by subtracting the rate in 2019 from that in 1999 and dividing the result by the rate in 1999, and multiplying by 100.

Annual use rates of cytology/HPV testing, colposcopy, and cervical procedures overall and by age group decreased over time. Of the 2.9 million cytology/HPV tests performed in 1999, 28.5% were performed in women aged 65 to 69 years and 16.7% in women 80 years or older. In contrast, of the 1.3 million cytology/HPV tests performed in 2019, 41.2% were performed in women aged 65 to 69 years and 10.0% in women 80 years or older.

A Joinpoint analysis showed that the overall age-adjusted annual use rate of cytology/HPV testing decreased 4.6% each year on average from 1999 to 2019 (P < .001; Figure 1A). We identified 2 distinct trend segments of cytology/HPV testing, with a 3.1% annual decrease in the first segment (1999-2009; P < .001) and a 6.1% annual decrease in the second segment (2009-2019; P < .001).

Figure 1. Trends in Annual Use Rates of Cervical Cancer Screening–Associated Services in Medicare Fee-for-Service Beneficiaries^a Overall^b and by Age Group From 1999 to 2019.

AAPC indicates average annual percentage change; ACS, American Cancer Society; USPSTF, US Preventive Services Task Force; ACOG, American College of Obstetricians and Gynecologists, ASCCP, American Society of Colposcopy and Cervical Pathology; APC, annual percentage change; HPV, human papillomavirus.

^aWomen aged 65 to 114 years enrolled in Medicare fee-for-service plans for at least 11 months each calendar year.

^bStandardized to the 2010 US female population age distribution.

^cP < .001.

^dP < .01.

The overall age-adjusted annual use rate of colposcopy decreased 3.0% each year on average from 1999 to 2019 (P < .001; Figure 1B). There was a steep 11.6% annual decline in colposcopy use during 2001 to 2004 (P = .01), which was followed by a gradual annual decrease of 2.1% during 2004 to 2016 (P < .001). However, receipt of colposcopy among women aged 65 to 69 years was associated with an increase after 2015 (APC, 2.8%; P < .001). The trends in use of cervical procedures were similar to trends in colposcopy use. The overall age-adjusted annual use rate of cervical procedures decreased 5.3% each year on average from 1999 to 2019 (P < .001; Figure 1C).

Similar trends were noted in the annual use rates of cytology/HPV testing, colposcopy, and cervical procedure analyses stratified by race and ethnicity (Figure 2). The age-adjusted annual use rate of cytology/HPV testing was the highest in non-Hispanic White women and the lowest in non-Hispanic American Indian/Alaska Native women during 1999 to 2019 (Figure 2A). However, the age-adjusted annual use rate of colposcopy was consistently higher in non-Hispanic Black compared with non-Hispanic White women (Figure 2B), and Hispanic and non-Hispanic Black women had a higher age-adjusted annual use rate of cervical procedures compared with non-Hispanic White women, although the differences decreased during the study period (Figure 2C).

Figure 2. Trends in Annual Use Rates of Cervical Cancer Screening–Associated Services in Medicare Fee-for-Service Beneficiaries^a by Race and Ethnicity^b From 1999 to 2019.

Data for women with unknown or other race and ethnicity (≤2% in any given year) are not displayed.

^aWomen aged 65 to 114 years enrolled in Medicare fee-for-service plans for at least 11 months each calendar year.

^bStandardized to the 2010 US female population age distribution.

In 2019, the total amount paid by Medicare for all cervical cancer screening–associated services in women 65 years or older was $83 527 181 (Table 3). Most expenditures were for cytology and/or HPV testing ($76 083 414), and 2.8% (130,415/ 4 719 771) of beneficiaries 80 years or older received 1 or more services at a cost of $7 391 101.

Table 3. Expenditures by the Medicare Fee-for-Service Program for Beneficiaries Older Than 65 Years Who Received 1 or More Cervical Cancer Screening–Related Services in 2019 by Age Group and Service.

Service, No. and cost	Age groups, y				Total
	65-69 (n = 3 724 099)	70-74 (n = 3 972 323)	75-79 (n = 2 882 463)	≥80 (n = 4 719 771)	≥65 (n = 15 298 656)
Cytology and/or HPV testing	533 152	420 756	211 846	129 105	1 294 859
	$33 378 319	$24 256 510	$11 703 029	$6 745 556	$76 083 414
Colposcopy	13 649	8776	4124	2843	29 392
	$2 013 165	$1 286 575	$599 767	$420 922	$4 320 429
Cervical procedures	1530	889	358	240	3017
	$1 608 024	$941 683	$349 008	$224 623	$3 123 338
Women receiving ≥1 servicea	535 793	422 794	213 080	130 415	1 302 082
% Receiving ≥1 serviceb	14.4	10.6	7.4	2.8	8.5
Total expenditures	$35 999 508	$26 484 768	$12 651 804	$7 391 101	$83 527 181

Abbreviation: HPV, human papillomavirus.

^a Number of women who received 1 or more services, ie, cytology, HPV testing, colposcopy, or any of the cervical procedures (loop electrosurgical excision procedure, cone biopsy, and ablation). A woman may have multiple services.

^b Number of women who received 1 or more services/total number of beneficiaries meeting inclusion criteria.

Discussion

The results of this cross-sectional study suggest that cytology test use in the Medicare fee-for-service female population 65 years or older has been declining steadily since 1999. We found a distinct change in overall age-adjusted screening test use trends at 2009, which may be associated with finalized recommendations from ACOG that screening could end at age 65 to 70 years.⁵ Furthermore, a significant decrease in cytology/HPV test use can be seen starting in 2002 that may be associated with ACS guidelines and USPTSF recommendations regarding screening cessation.^4,31

Despite decreasing use of testing, in 2019, more than 8% of the female Medicare fee-for-service population older than 65 years received 1 or both of cytology and HPV tests. More than 41% of the cytology/HPV tests conducted in 2019 were performed for women aged 65 to 69 years, suggesting that many of these tests may have been performed to meet the criteria for screening cessation after age 65 years. Screening may be clinically indicated in women older than 65 years with an inadequate or unknown screening history and those who are otherwise at high risk (ie, women with a history of high-grade precancerous lesions or cervical cancer, in utero exposure to diethylstilbestrol, or a compromised immune system).⁷ Previous studies have found that 24% to 65% of women failed to meet criteria to end cervical cancer screening by age 65 years in various populations, including women with private health insurance,³² women in an integrated health care system,³³ and those who received services in safety net hospitals.^32,34 Furthermore, few of these underscreened women received screening after age 65 years.^33,34 More in-depth investigation of service use after age 65 years could better elucidate whether they are indicated by women’s screening and medical history or whether they represent overscreening.³⁵

Continued testing in women 80 years or older may represent overuse. As part of the Choosing Wisely initiative, the Society for General Internal Medicine does not recommend cancer screening be performed for individuals with a life expectancy of fewer than 10 years,³⁶ which corresponds to age 80 years or older.³⁷ The Society for General Internal Medicine cites a higher likelihood of experiencing harms, including complications of testing and treatment. While cervical procedures are considered safe when performed in most people aged 21 to 65 years, to our knowledge, the safety of these procedures has not been well established in older individuals.^38,39

A 2018 study found that many health care clinicians do not agree with guidelines to end cervical cancer screening.⁴⁰ The study reported that up to 25% of health care clinicians surveyed (including obstetrician-gynecologists, family medicine physicians, internists, nurse practitioners, and physician assistants) would continue to perform screening, largely because of concerns about missing cervical cancer cases; 32% of survey respondents would continue screening in populations older than 65 years despite believing that the risk of cancer is low.⁴⁰ This is consistent with other studies that found that many clinicians would recommend additional cancer screening in older populations, even with substantial comorbidities and despite guidelines and recommendations against screening.⁴¹ Overuse and factors driving overuse of cervical cancer screening tests in older patients are worth investigating, as many clinicians may be providing services that are not warranted and could be exposing patients to iatrogenic harms, such as false-positive testing and unnecessary procedures.⁴² In this context, overuse can be defined as providing cervical cancer screening in populations that no longer require these services per national guidelines/recommendations.

While rates of colposcopy and cervical procedures have decreased substantially since 1999, we observed a plateau in the overall age-adjusted rates and an increase in colposcopy among women aged 65 to 69 years since 2015. Potential explanations could be major organizations recommending the use of cotesting in women 30 years or older. While cotesting is more sensitive than cytology testing alone, its lower specificity is associated with more colposcopies and cervical procedures.⁴³ More women may continue screening beyond age 65 years because they were unable to meet the criteria to end screening because of positive HPV test results. Furthermore, when stratified by race and ethnicity, rates of colposcopy and cervical procedures followed similar decreases during the study period. Despite having relatively lower age-adjusted rates of cytology/HPV testing, Hispanic and non-Hispanic Black populations have maintained higher overall age-adjusted rates of colposcopy and cervical procedures. This could represent increased use of health care services after becoming Medicare eligible and an associated detection of previously undetected abnormalities, especially as older Hispanic and non-Hispanic Black populations are much more likely to be overdue for cervical cancer screening.^15,44,45 Future studies investigating these populations on a more granular level could provide insight into the cervical cancer incidence disparities seen in older individuals by racial and ethnic groups.^46,47

This study found that the Medicare fee-for-service program spent more than $83 million for cervical cancer screening–associated services in 2019. A significant proportion, more than $76 million (or 91% of 2019’s total expenditures) was spent on cytology/HPV testing. This is more than a third higher than previous estimates using older National Ambulatory Medical Care Survey data.²⁶ Future studies that evaluate the need for these services (eg, benefits, harms, and cost-effectiveness) being performed in older populations can provide valuable information given the high costs of these services found in the present study.

Investigating cytology/HPV testing, colposcopy, and cervical procedure use in the Medicare population is a novel use of this database. The Medicare fee-for-service population includes millions of women and allows for a population-level exploration of health care utilization and cost for cervical cancer screening–related services. Furthermore, it allows us to evaluate trends in use of these services during a 21-year period in the context of changing guidelines and recommendations. The study results provided a baseline overview by year, age group, and race and ethnicity and provide a potential foundation for future analyses to explore topics, such as appropriateness of cervical cancer screening–associated services in older women overall and in specific subgroups, including socioeconomic status and rural/urban residence.

Limitations

This study had several limitations. First, we were only able to ascertain what we describe as cervical cancer screening–associated tests and procedures as we were unable to discern if cytology/HPV tests were used for screening or surveillance because of unknown medical history before age 65 years, especially as women with previously diagnosed and treated precancer should be followed up for 25 years per national guidelines.^6,48 However, it is believed that most of these tests were for screening, as indicated in the description of the procedure codes used. The specificity of the cervical procedure codes (eg, loop excisions) indicated that these would have been performed exclusively in persons with no prior hysterectomy. Second, because of the nature of the Medicare database, we were unable to discern if a woman had undergone a hysterectomy before becoming Medicare eligible. Since 1996, the USPSTF has recommended that women who have undergone a hysterectomy with removal of the cervix for benign disease do not need cervical cancer screening.³ Estimates from self-reported health surveys indicate that about 40% of women report having undergone a hysterectomy by age 65 years,^49,50 and more than half of women reporting previous hysterectomy for benign reasons also report a Papanicolaou test during the past 3 years.⁴¹ Service use rates are assumed to be higher among women who have not undergone a hysterectomy. Third, we were unable to accurately identify women for whom lifelong screening is recommended (eg, living with HIV⁵¹) and women with in utero exposure to diethylstilbestrol for whom screening may be indicated past age 65 years.^6,7 Fourth, we were only able to use the fee-for-service population, and this may not represent the Medicare Advantage population. Despite this, Medicare fee-for-service represents more than two-thirds of the Medicare eligible population, and previous studies have found little difference in health care utilization,⁵² especially in association with cancer screening in older populations.⁵³ Fifth, because we relied on claims data, we cannot know whether these claims reflect what was actually done in clinical practice⁵⁴; future studies investigating electronic health records may be able to further elucidate actual clinical practice in older populations. Sixth, the estimated Medicare payments may be an underestimate of the total cost to the health care system because coinsurance payments and payments billed to Medicare Part A for institutional services were not included.

Conclusions

While the results of this cross-sectional study suggest that the use of cervical cancer screening–associated services declined significantly since 1999, many women older than 65 years received these services, which cost the Medicare fee-for-service system more than $83 million in 2019. The extent to which these trends represent appropriate screening cessation or overtesting is unknown. The decision to end screening in women at average risk for cervical cancer requires review of up to 10 years of medical history to determine whether continued routine cervical cancer screening is appropriate after age 65 years. Such information is not always obtainable for new Medicare enrollees, and the balance of screening benefits and harms changes with advancing age. Strategies to ameliorate access to longitudinal historical medical information and tools to facilitate decision-making for clinicians and patients could support cervical cancer prevention in older women and prevent harms and costs from unnecessary tests and diagnostic procedures.

Supplement.

eAppendix. Medical Codes Used in Identifying Cytology, HPV Tests, Colposcopy, and Cervical Procedures