The Effect of the American College of Physicians Guideline Change on Screening Pelvic Examinations in an Insured Population

Abstract

Background:

Several advancements in evidence and changes in recommendations for screening pelvic examinations (SPEs) have occurred in the last decade, yet their impact on the use of SPEs is unknown. We assessed the impact of two changes: the September 2013 Quality Enhancement Research Initiative (QUERI) Evidence-based Synthesis Program report finding no evidence that SPEs reduce morbidity or mortality; and the July 2014 American College of Physicians (ACP) guidelines recommending against SPEs in asymptomatic, nonpregnant women.

Methods:

We described monthly rates of SPEs among females ages 15–64 enrolled in a claims database between January 2012 and December 2019. We implemented an interrupted time series analysis to examine the impact of the QUERI evidence report and ACP guidelines on the trend in monthly SPE rates, overall and stratified by age group and geographic region.

Main Findings:

The average monthly rate of SPEs decreased from 357 per 1,000 person-years at risk in 2012 to 302 per 1,000 person-years at risk in 2019. There were no significant immediate changes in rates of SPEs after either the 2013 QUERI white paper or the 2014 ACP guidelines.

Conclusions:

Changes in guidelines had no observable impact on the rate of SPEs. Future work is needed to better understand why recent evidence and guideline developments have not impacted real-world practice to target interventions for providers to decrease SPEs efficiently.

Keywords

gynecology women’s health services gynecologic examination low-value care policy analysis guidelines

Introduction

Bimanual pelvic examinations (henceforth, pelvic exams) have been a usual component of preventive gynecological care in the United States for over a century.¹ A bimanual pelvic exam consists of a provider palpating the internal pelvic organs using two fingers inside the vagina while applying external abdominopelvic pressure. A screening pelvic exam (SPE) is performed for asymptomatic individuals, or those who are not currently experiencing any abnormal or concerning gynecological symptoms, to detect gynecological conditions, such as cancers, infections, and asymptomatic pelvic inflammatory disease.²

A growing evidence base casts doubt on the clinical utility of SPEs.^3–7 Over time, more effective screening mechanisms have been created for several conditions, such as the Papanicolaou test (Pap test) to screen for cervical cancer and self-collected specimens for sexually transmitted infection screening.^2–4,6 SPEs can lead to psychological harms such as fear, embarrassment, or anxiety, and any false-positive signals for ovarian cancer may lead to unnecessary surgery for benign growths.^3,8

Changes to guidelines by major professional societies over the past decade reflect the lack of effectiveness and a growing acknowledgment of the potential harms of SPEs. In 2013, the Quality Enhancement Research Initiative (QUERI) Evidence-based Synthesis Program released a white paper of their systematic review evaluating the benefits and harms of SPEs.⁸ In it, they found “no data [studies] indicating that the performance of the routine pelvic examination in asymptomatic average risk women reduces morbidity or mortality from any condition other than cervical cancer. For cervical cancer the recommended examination is visual inspection of the cervix and cervical swabs for cancer and HPV [Human Papillomavirus].”⁸ The authors published these findings in the Annals of Internal Medicine in 2014, alongside the American College of Physicians (ACP) guidelines recommending against SPEs in asymptomatic, nonpregnant adults.^9,10 These publications were the first systematic review and guidelines recommending against the routine utilization of SPEs. However, other professional societies were slower to change their guidelines or issued less definitive recommendations. In 2017, the American Association of Family Physicians (AAFP) also recommended against SPEs in asymptomatic, nonpregnant adults.¹¹ However, the United States Preventive Services Task Force (USPSTF) concluded that there was insufficient evidence, due to lack of studies, to recommend in favor of or against SPEs,¹² and the American College of Obstetricians and Gynecologists (ACOG) recommended that the decision of whether to conduct an SPE be shared between individuals and their provider.²

To date, the effect of these guideline changes on the occurrence of SPEs in clinical practice over time has not been examined. Prior U.S. studies using data from the National Survey of Family Growth (NSFG) suggest that most pelvic exams in young adults (i.e., ages 15–20) are done for routine or screening purposes^1,13,14 but have not examined trends in utilization of SPEs over time. Specifically, the effect of the seminal QUERI systematic review and corresponding ACP guideline changes is unknown. In addition, there are limited data on the prevalence of SPEs among individuals older than 20 years.

The objectives of this study were to describe rates of SPEs over time and assess the impact of the September 2013 QUERI evidence report and the July 2014 ACP guideline change on SPE rates. We hypothesized that the trend in monthly rates of SPEs would decrease after the 2013 QUERI evidence report and then decrease again after the 2014 ACP guidelines.

Materials and Methods

We examined monthly rates of SPEs as the number of SPEs per 1,000 person-years at risk of SPEs each month from January 2012 to December 2019. We used an interrupted time series (ITS) analysis, a quasi-experimental study design, to assess the population-level impact of policy changes on SPE rates.^15–18 We utilized two interruption time points: (1) October 2013, following the QUERI white paper release in September 2013; and (2) August 2014, following the July 2014 journal publication of the systematic review alongside the ACP recommendation against SPEs.^8–10

This study was reviewed and exempted by the University of North Carolina at Chapel Hill Institutional Review Board (#21-1736).

Data source

This work was conducted using the Merative MarketScan Commercial Claims and Encounters database (henceforth, MarketScan) from 2011 to 2019. MarketScan is a large, de-identified database of inpatient and outpatient medical claims and outpatient prescription data for employer-sponsored, commercially insured patients younger than 65 years of age.¹⁹

Measures

Time at risk of a SPE

We included individuals in the MarketScan database coded as females, defined by MarketScan to indicate individuals born with pelvic organs, including uterus, cervix, fallopian tubes, and ovaries. We limited to females ages 15–64 years with at least 365 days of continuous enrollment in MarketScan from 2011 to 2019. All females in this age group were considered at risk of experiencing a pelvic exam. During this period, we categorized indications for pelvic exams (described below) to increase the likelihood that we were accurately measuring SPEs only.

To identify SPEs, we considered individuals ineligible for SPEs during days they were admitted as an inpatient, as well as days they had a billed code for a potential indication for a pelvic exam and for a period after the potential indication during which another pelvic exam may be indicated. These periods, reviewed by a board-certified obstetrician-gynecologist (OBGYN), fall into seven categories of the approximate number of days that a pelvic exam could be indicated: 1, 30, 60, 90, 270, 365, and infinite. On these days, individuals were considered ineligible for an SPE (i.e., due to a potential indication for a pelvic exam). Examples of indications in each category and justification for the estimated time frame in which a pelvic exam could be indicated (and therefore a person was not at risk of an SPE) can be found in Table 1 (full details in Supplementary Fig. S1 and Supplementary Table S1).

Table 1.

Number of Days Ineligible for Screening Pelvic Exams with Examples of Corresponding Symptoms and Diagnoses

Approximate number of days ineligible for screening pelvic exam	Example indications	Justification
1	Abdominal pain	These are symptoms that can be addressed the day of a visit, and they are unlikely to indicate a future pelvic exam unless a separate diagnosis is made.
	Urge incontinence
30	Genital herpes, unspecified	Episodic infectious or inflammatory diagnoses should be treated and cured or resolved within one month. A pelvic exam may be indicated to ensure resolution of the problem, but no further pelvic exams are indicated unless a problem persists (coded again).
	Candidiasis of vulva and vagina
	Pelvic inflammatory disease (PID) unspecified
60	Pregnancy deliveries	Pelvic exams may be indicated after a delivery at a postpartum follow-up, recommended to occur in 6–8 weeks after delivery.
90	Dysmenorrhea	A pelvic exam could be indicated for up to 90 days after the diagnosis or IUD insertion, after which a person returns to eligibility for a screening pelvic exam. For example, following an IUD insertion, within 90 days, adverse drug reactions are expected to resolve, and this is the time frame in which, if recommended by the provider, an initial IUD string check would occur.
	Excessive or infrequent menstruation
	Insertion of intrauterine device (IUD)
	Encounter for removal and reinsertion of intrauterine contraceptive device
270	Prenatal care	Prenatal care indicates a pregnancy, but we cannot identify at what stage the pregnancy is in. Therefore, any code for prenatal care could indicate a pelvic exam for up to 9 months or 270 days. If a delivery is coded, indicating the end of the pregnancy, the time ineligible for a screening pelvic exam will follow that time frame (60 days) instead.
365	Moderate cervical dysplasia	These indicate premalignant conditions for which serial exams and treatments are recommended. Pelvic exams could be indicated for approximately 1 year, in which time more tests would indicate another diagnosis or return to baseline risk of a screening pelvic exam.
	Hypertrophy of uterus
Infinite	Malignant neoplasm of the ovary	These diagnoses remove a person indefinitely from the asymptomatic, average-risk population of females who are targeted by screening pelvic exams.
	Carcinoma in situ of cervix, unspecified
	Female pelvic peritoneal adhesions (postinfective)

These estimates are based on the expert opinion of a board-certified obstetrician and gynecologist. All codes and corresponding estimates of eligibility for screening pelvic exam can be found in Supplementary Table S1.

SPE occurrence

We measured pelvic exams during any period an individual was considered at risk of a SPE using the International Classification of Diseases, 9th version, Clinical Modification (ICD-9-CM) diagnosis code V72.31; ICD-10-CM diagnosis codes Z01.411 and Z01.419; and Healthcare Common Procedure Coding System code G0101 (Supplementary Tables S2 and S3).^20,21 Given our study inclusion/exclusion criteria, we assumed that these codes represented SPEs or those done without a medical indication.^20,21

Other measures

We categorized age as 15–17 years, 18–20 years, 21–24 years, 25–34 years, 35–44 years, 45–54 years, and 55–64 years.

Geographic region was measured using a state-level indicator of geographic location of the patient, which we further categorized according to U.S. census region: Northeast, Midwest, South, West, and other/unknown. For individuals with other/unknown values for state of residence, we used a higher-level, region-only variable available in the database to indicate those individuals’ geographic region. We considered the remaining other/unknown geographic region values (1.6% of person-years at risk) to be missing. We used a single (i.e., deterministic) imputation to impute the geographic region for these individuals.

Data analysis

All analyses were performed using SAS, v9.3 (SAS Institute Inc., Cary, NC). We calculated the monthly rate of SPEs from January 2012 to December 2019. The denominator of the SPE rate represents all days at risk each month and was converted to 1,000 person-years at risk (where 365 days at risk = 1 person-year at risk); and the numerator is the number of SPEs that occurred during those days at risk. To control for anticipated month-to-month variation in the number of days in a month and the geographic and age distribution in our sample, we standardized the rates for females by age group and U.S. census region to a standard 30-day month.^15,22,23 We used a standard population of females with employer-sponsored health insurance in the United States in the most recently available data (year 2017) from the Current Population Survey from the U.S. Census Bureau (Supplementary Table S4).²⁴

We fit a segmented linear regression model for the standardized monthly rates, allowing for a change in intercepts and slopes after each interruption. The changes in slopes represent the changes in monthly trend (change each month) of SPEs. We allowed for a change in intercepts in the model to allow better model fit and examined whether there were significant immediate changes following the guideline changes. We visually inspected the trend in rates and did not identify any new interruption points to evaluate. We defined statistical significance as alpha < 0.05.

We conducted two sensitivity analyses. First, we limited the analysis to individuals with no missing data (i.e., complete cases) to examine the robustness of the results to different methods of handling missingness. Second, we conducted a sensitivity analysis, which required only 90 days of continuous enrollment before contributing person-time at risk or the occurrence of an SPE, instead of 365 days, to examine the robustness of results to different continuous enrollment requirements.

Stratified analyses

We conducted several stratified analyses to examine whether the guidelines impacted age groups or regions differently. Prior literature examining the occurrence of SPEs has focused on individuals ages 15–20, likely because these studies also examined potentially unnecessary Pap tests (which are rarely recommended for individuals younger than 21 years).^1,13 We theorized that the changes in guidelines may have differential effects across each age group. Each age group was standardized to the distribution of geographic region in its specific stratum (i.e., the distribution of region was specific to each stratum of age group).²⁴

Results

SPE rates

The study sample included over 31 million unique individuals with more than 63 million person-years at risk of an SPE (Table 2). Over 21 million SPEs occurred from 2012 to 2019 (Table 2). The average monthly rate of SPEs decreased from 357 SPEs per 1,000 person-years at risk in 2012 to 302 SPEs per 1,000 person-years at risk in 2019, and these results were similar in sensitivity analyses (Table 3).

Table 2.

Sample Characteristics

	Person-years at risk of a screening pelvic exam		Screening pelvic exams		Rate of screening pelvic exams per 1,000 person-years at risk (2012–2019)
	No.	%	No.	%
Overall	63,973,500	100	21,459,667	100	335
U.S. census region
Northeast	13,854,299	21.7	4,293,297	20.0	310
Midwest	11,824,041	18.5	4,474,555	20.9	378
South	26,382,942	41.2	9,524,643	44.4	361
West	11,912,218	18.6	3,167,172	14.8	266
Age group
15–17	1,827,546	2.9	108,924	0.5	60
18–20	3,231,091	5.1	600,814	2.8	186
21–24	4,802,136	7.5	1,552,475	7.2	323
25–34	9,344,214	14.6	3,648,834	17.0	391
35–44	12,631,376	19.7	4,759,332	22.2	377
45–54	15,833,552	24.8	5,684,928	26.5	359
55–64	16,303,585	25.5	5,104,360	23.8	313
Years
2012	12,485,961	19.5	4,382,489	20.4	351
2013	9,601,738	15.0	3,319,031	15.5	346
2014	10,086,935	15.8	3,435,604	16.0	341
2015	7,026,962	11.0	2,368,988	11.0	337
2016	6,844,985	10.7	2,229,929	10.4	326
2017	6,092,953	9.5	1,996,294	9.3	328
2018	5,796,201	9.1	1,801,507	8.4	311
2019	6,037,764	9.4	1,925,825	89.0	319

Person-years were measured as 365 person-days at risk of a screening pelvic exam. Rates are standardized by age group and U.S. census region to a standard 30-day month. Individuals with missing geographic region measurement (∼1%) were imputed using single imputation methods. The Northeast region includes states Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont. The Midwest region includes states Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, North Dakota, Ohio, South Dakota, and Wisconsin. The South region includes Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Texas, Tennessee, Virginia, Washington D.C., and West Virginia. The West region includes Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, Puerto Rico, Utah, Washington, and Wyoming.

Table 3.

Average Monthly Rate of Screening Pelvic Exams (per 1,000 Person-Years) Each Year, 2012–2019

	Average monthly rate screening pelvic exams per 1,000 person-years
	2012	2013	2014	2015	2016	2017	2018	2019
Primary analysis	357	351	338	326	314	311	297	302
Age group strata
15–17	63	66	63	58	53	50	49	52
18–20	207	192	185	182	176	173	160	158
21–24	354	341	327	321	313	310	299	301
25–34	435	422	402	388	370	361	342	339
35–44	408	402	387	371	357	352	337	349
45–54	380	379	366	353	342	341	328	341
55–64	333	333	320	310	300	301	289	294
Region strata
Northeast region	401	396	385	375	362	357	351	367
Midwest region	351	336	318	303	293	291	275	273
South region	382	380	370	366	353	350	330	334
West region	293	290	271	249	238	236	225	233
Sensitivity analyses
Complete case analysis	357	351	337	326	314	311	297	303
90-Day continuous enrollment	341	340	328	319	309	305	291	297

Person-years were measured as 365 person-days at risk of a screening pelvic exam. The rates shown represent the average of all months within a given year. Rates are standardized by age group and U.S. census region to a standard 30-day month. Individuals with missing geographic region measurement (∼1%) were imputed using single imputation methods except in complete case analysis (sensitivity analysis 1). Northeast region includes states Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont. The Midwest region includes states Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, North Dakota, Ohio, South Dakota, and Wisconsin. The South region includes Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Texas, Tennessee, Virginia, Washington D.C., and West Virginia. The West region includes Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, Puerto Rico, Utah, Washington, and Wyoming.

The average monthly rate of SPEs decreased between 2012 and 2019 across all age groups, with heterogeneity by age group (Table 3, Fig. 2). Across all years, the average monthly rate of SPEs was lowest among ages 15–17 (63 and 52 SPEs per 1,000 person-years at risk for years 2012 and 2019, respectively) and only slightly higher among ages 18–20 (207 and 158 SPEs per 1,000 person-years at risk for years 2012 and 2019, respectively; Table 3). For all other age groups, the monthly average rate of SPEs ranged from 333 to 435 and 294 to 349 SPEs per 1,000 person-years at risk for years 2012 and 2019, respectively (Table 3).

ITS—SPE rates

The model results indicate a slight downward, nonsignificant trend in the rate of SPEs per 1,000 person-years at risk for each month of time, and the trend of monthly change in rate of SPEs did not significantly change in level or slope following either interruption (Table 4, Fig. 1). The slopes indicate a consistent monthly decrease in the rate of SPEs over time, although this was only statistically significant for Segment 3, from August 2014 to December 2019, when the rate decreased 0.5 SPEs per 1,000 person-months at risk (95% confidence interval [CI] −0.9, −0.1; Table 4). Across sensitivity analyses, we consistently found no statistically significant changes in the trend of monthly change in rate of SPEs following the 2013 QUERI publication nor the July 2014 guideline changes (Table 4). The findings of the two sensitivity analyses were consistent with the primary results (Figs. S2 and S3).

Table 4.

Interrupted Time Series Analysis Model Results, Segment Slopes, and Absolute Change in Slopes After Interruption

	Segment 1	Segment 2			Segment 3
	Jan 2012–Sept 2013	Oct 2013–Jul 2014			Aug 2014–Dec 2019
	Slope (95% CI)	Immediate change in rate (95% CI)	Slope (95% CI)	Change in slope from segment 1 (95% CI)	Immediate change in rate (95% CI)	Slope (95% CI)	Change in slope from segment 2 (95% CI)
Primary analysis	−0.7 (−1.9, 0.5)	−5.1 (−20.7, 10.5)	−1.4 (−3.9, 1.0)	−0.7 (−3.5, 2.1)	3.6 (−10.5, 17.7)	−0.5 (−0.9, −0.1)	1.0 (−1.6, 3.5)
Age group
15–17	0.4 (0.0, 0.7)	−7.5 (−14.3, −0.7)	0.7 (−0.3, 1.7)	0.4 (−0.7, 1.4)	−7.8 (−13.6, −2.0)	−0.1 (−0.3, −0.0)	−0.8 (−1.8, 0.1)
18–20	−1.3 (−1.7, −0.8)	−3.0 (−3.2, −2.5)	−0.2 (−1.8, 1.3)	1.1 (−0.5, 2.6)	3.9 (−4.7, 12.6)	−0.5 (−0.6, −0.4)	−0.3 (−1.8, 1.2)
21–24	−1.0 (−1.9, −0.2)	−10.9 (−29.0, 7.2)	−0.2 (−2.7, 2.3)	0.9 (−1.7, 3.4)	0.8 (−14.2, 15.9)	−0.4 (−0.8, −0.1)	−0.2 (−2.8, 2.3)
25–34	−1.2 (−1.8, −0.5)	−11.2 (−28.4, 6.0)	−0.8 (−3.2, 1.6)	0.4 (−2.0, 2.7)	−1.9 (−15.5, 11.6)	−1.1 (−1.2, −0.9)	−0.3 (−2.7, 2.1)
35–44	−0.3 (−1.3, 0.6)	−10.1 (−33.0, 12.9)	−1.1 (−4.3, 2.1)	−0.7 (−3.9, 2.5)	−2.9 (−21.2, 15.5)	−0.6 (−0.9, −0.4)	0.4 (−2.7, 3.6)
45–54	−0.1 (−1.1, 1.0)	−6.4 (−29.4, 16.7)	−1.2 (−4.4, 2.1)	−1.1 (−4.4, 2.1)	1.6 (−17.6, 20.7)	−0.4 (−0.7, −0.1)	0.8 (−2.4, 4.0)
55–64	−0.0 (−0.8, 0.7)	−3.3 (−20.7, 14.2)	−1.3 (−3.8, 1.1)	−1.3 (−3.7, 1.1)	0.5 (−13.6, 14.5)	−0.4 (−0.6, −0.2)	0.9 (−1.5, 1.1)
U.S. census region
Northeast	−1.1 (−2.0, −0.1)	−8.7 (−23.9, 6.6)	−0.7 (−3.0, 1.5)	0.3 (−2.3, 2.9)	−6.4 (−19.1, 6.4)	−0.7 (−0.9, −0.5)	0.1 (−2.2, 2.3)
Midwest	−0.8 (−2.2, 0.6)	−0.4 (−23.0, 22.3)	−1.5 (−4.9, 1.8)	−0.7 (−4.4, 3.0)	2.7 (−17.2, 22.6)	−0.2 (−0.7, 0.3)	1.3 (−2.1, 4.8)
South	−0.4 (−1.7, 0.9)	−2.6 (−23.0, 17.9)	−1.3 (−4.4, 1.8)	−0.9 (−4.4, 2.5)	6.7 (−11.2, 24.7)	−0.6 (−1.0, −0.2)	0.7 (−2.4, 3.8)
West	−0.5 (−1.7, 0.77)	−9.7 (−24.6, 5.1)	−2.4 (−4.7, 0.0)	−1.9 (−4.7, 0.9)	5.8 (−7.7, 19.2)	−0.5 (−0.9, −0.1)	1.9 (−0.6, 4.3)
Sensitivity analyses
Complete case analysis	−0.7 (−1.9, 0.5)	−4.5 (−20.0, 11.1)	−1.6 (−4.0, 0.8)	−0.9 (−3.6, 1.9)	3.9 (−10.2, 17.9)	−0.5 (−0.9, −0.0)	1.1 (−1.4, 3.6)
90-Day continuous enrollment	−0.1 (−1.3, 1.0)	−7.2 (−22.1, 7.7)	−1.2 (−3.5, 1.2)	−1.0 (−3.7, 1.6)	3.4 (−10.0, 16.8)	−0.4 (−0.8, −0.0)	0.7 (−1.7, 3.1)

Where slope is the change in rate of screening pelvic exams per 1,000 person-years at risk each month; no interruptions significantly changed the trend in rate of screening pelvic exams at alpha = 0.05; immediate change in rate refers to the change in the rate of SPE from the month prior to the policy change to the month after; change in slope from segment 1 and change in slope from segment 2 refer to the relative change in the slope from the previous segment or timeframe. Rates are standardized by age group and U.S. census region to a standard 30-day month. Individuals with missing geographic region measurement (∼1%) were imputed using single imputation methods except in complete case analysis (sensitivity analysis 1). Northeast region includes states Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont. The Midwest region includes states Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, North Dakota, Ohio, South Dakota, and Wisconsin. The South region includes Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Texas, Tennessee, Virginia, Washington D.C., and West Virginia. The West region includes Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, Puerto Rico, Utah, Washington, and Wyoming.

CI, confidence interval.

FIG. 1.

The impact of evidence and guideline changes on screening pelvic exam rates estimated using interrupted time series, 2012–2019. Abbreviations: CI, Confidence Interval; QUERI, Quality Enhancement Research Initiative; ACP, American College of Physicians; Rates are standardized by age group and U.S. census region to a standard 30-day month.

Stratified analyses: age group

From January 2012 to September 2013, the rate of SPEs trended down for all age groups except 15–17 years, but only age groups 18–20 years, 21–24 years, and 25–34 years trended down significantly (Table 4). The monthly rate of SPEs from January 2012 to September 2013 for the age group 15–17 years significantly increased over time, albeit only slightly, at 0.4 SPEs per 1,000 person-years at risk each month (95% CI 0.002, 0.7; Table 4, Fig. 2).

FIG. 2.

Age group strata, the impact of evidence and guideline changes on screening pelvic exam rates estimated using interrupted time series, 2012–2019. Abbreviations: QUERI, Quality Enhancement Research Initiative; ACP, American College of Physicians.

Across all age groups, the change in trend of monthly rate of SPEs after the September 2013 QUERI evidence report and the July 2014 ACP guideline change was not statistically significant (Table 4, Segment 2; Fig. 2). However, there were statistically significant decreases in the immediate rate of SPEs immediately after the 2013 QUERI evidence report for age groups 15–17 (−7.5 [95% CI −14.3, 0.0.7]) and 18–20 (−3.0 [95% CI −3.2, −2.5]; Table 4, Segment 2; Fig. 2) and immediately after the 2014 ACP guidelines for age group 15–17 (−7.8 [95% CI −13.6, −2.0]; Table 4, Segment 3; Fig. 2).

Stratified analyses: region

The overall monthly rates of SPEs were slightly lower in the West and Midwest regions than in the South and Northeast. The monthly rate of SPEs significantly decreased from January 2012 to September 2013 in the Northeast by 1.1 SPEs per 1,000 person-years at risk (95% CI −2.0, −0.1) but did not significantly change in the Midwest, South, and West regions (Table 4, Segment 1; Fig. 3). There were no significant immediate changes in the monthly rate of SPEs in any region after the September 2013 QUERI evidence report. Following the 2014 ACP guideline update, there were no significant changes in slope in any region (Table 4, Segment 3; Fig. 3).

FIG. 3.

Region group strata, the impact of evidence and guideline changes on screening pelvic exam rates estimated using interrupted time series, 2012–2019. Abbreviations: QUERI, Quality Enhancement Research Initiative; ACP, American College of Physicians.

Discussion

We observed a gradual decrease in the monthly rate of SPEs from 2012 to 2019, with a significant downward trend from August 2014 to December 2019. The monthly rate of SPEs was notably lower among 15- to 20-year-olds compared with older age groups, and the monthly rates of SPEs were lowest among 15- to 17-year-olds. We found no significant overall change in trend of change in the monthly rate of SPEs following the publication of the September 2013 white paper nor the July 2014 peer-reviewed paper and ACP recommendations, which recommended against SPEs for average-risk, nonpregnant women. However, our results showed a statistically significant immediate decrease in SPE rates immediately after the 2013 QUERI evidence report for age groups 15–17 and 18–20 and immediately after the 2014 ACP guidelines for age group 15–17, although the magnitude of these decreases is modest. Our findings indicate that changes in the pelvic exam guidelines had limited immediate impact on the change in rate of SPEs. To our knowledge, this is the first study examining the impact of these guidelines on the change in SPE rates.

Implications for practice and policy

The lack of immediate change in trend of SPE rates after the QUERI publication and ACP guideline change may reflect the lack of consensus among medical providers about how to include SPEs in their practice. Although ACP and AAFP agree that routine SPEs should not be performed,^10,11 the USPSTF and the ACOG have avoided issuing such recommendations, respectively claiming insufficient evidence about benefits and harm for asymptomatic women¹² and deferring decisions to individual patients and their providers.²

SPEs are not without potential harm, but the USPSTF has noted the lack of studies that quantify the anxiety and psychological harms associated with SPEs as a particular limitation in being able to weigh the harms and benefits.^12,25 The potential for physical and psychological harm due to unnecessary SPEs must be examined and quantified to support women’s well-being and receipt of evidence-based care. In addition, consensus among stakeholder organizations is necessary to provide clarity in best practices regarding pelvic exams.

It is unclear to what extent providers are aware of these guideline changes, which could be a barrier to practice change.²⁶ Clinical inertia, or the habits and routines of past years of clinical practice that are a barrier to change for both clinicians and patients, could also contribute to the slow and gradual rate of change in SPE rates.²⁶ Studies examining the adoption of longer cervical cancer screening intervals also show a substantial lag between guideline issuance and clinical practice change.^27,28 Even if providers are aware of the guideline changes, whether they are ready to change their clinical practice accordingly is unknown.²⁶ Although the immediate impact of the policies did not result in significantly different rates of screening, we were able to describe an overall downward trend starting in 2014. This trend may reflect a clinically significant but much more gradual acceptance and implementation of the accumulated guideline recommendations over time by both clinicians and patients.

SPEs in specific populations

We found notably lower monthly rates of SPEs among individuals younger than 21 years, perhaps due to more awareness of the lack of benefits for SPEs in younger individuals. The higher rates of SPEs among other age groups may also be related to Pap test guidelines and the historic connection recommending annual Pap tests alongside SPEs. A 1987 consensus statement regarding cervical cancer screening recommended that all women have an annual pelvic examination, including a Pap test, beginning with the onset of sexual activity or by age 18.^29,30 Since 2012, guidelines from ACOG, USPSTF, and ACS agree that before age 21, no cervical cancer screening is recommended, regardless of sexual history,^5–7 which may result in lower provision of SPEs without the prompt of an accompanying Pap test.

Moreover, our findings suggest that SPEs continue in clinical practice today, aligning with prior literature. Mason et al. analyzed the 2013–2014 NSFG data to identify factors associated with receiving a pelvic exam (not necessarily for screening purposes) before age 21.¹³ Among 1,170 respondents ages 15–20 years, 31% had received a pelvic exam before age 21.¹³ Similarly, Qin et al. estimated that from 2011 to 2017, 23% of women ages 15–20 reported having a pelvic exam in the previous year¹; an estimated 54% (95% CI 49–60%) of those pelvic exams were potentially unnecessary screening exams, affecting an estimated 1.4 million women.¹

Ellison et al. found that 24.2% of females ages 15–49 enrolled in MarketScan experienced SPEs at a contraceptive encounter between 2007 and 2017.²⁰ In contrast to our results, they found that SPEs increased from 4.3% in 2007 to 13.3% in 2017.²⁰ Notably, this study was limited to only visits that were billed as contraceptive encounters, which may not be representative of all preventive care visits that females attend and may experience an SPE. Future work should examine subgroups that may have a higher rate of SPEs, such as individuals who use self-administered hormonal contraception.

Limitations

The results of this study should be interpreted in light of several limitations. There is currently no gold standard for identifying SPEs in insurance claims data. We used prior literature to guide our approach to measure SPEs, but no approach in measuring SPEs has been validated yet.^15,20 We do not know to what extent we may be over- or underestimating SPE occurrence. The exclusion criteria used to isolate SPEs (versus pelvic exams indicated due to a problem or complaint) seem robust but have not been formally examined. A validation study is needed to understand how providers bill for pelvic exams in general and how to distinguish between SPEs and indicated pelvic exams in insurance claims data.

We used a 365-day lookback period, during which we required continuous enrollment, before an individual was included in measuring the rates of SPEs. However, some exclusion criteria could indicate a lifetime of potential pelvic exams (e.g., gynecological cancers), which we may not have observed in the lookback period. Although these diagnosis codes would be expected to be associated with all future encounters, if they were not, this could lead to an overestimating of person-time at risk of an SPE. In the sensitivity analysis, we found similar results to the primary analysis, although we cannot rule out potential mismeasurement due to enrollment criteria in either approach.

In this analysis, we used “months” as our unit of time. Although this is a commonly used measure in ITS,³¹ other units of time, such as quarters or years, may be helpful in policy analysis to capture the intended and unintended impacts. Future analyses on SPE guideline changes may benefit from incorporating varying units of time.

While we found no significant changes in trend of the monthly rate of SPEs after the 2013 QUERI report nor the 2014 ACP guideline change overall, these results may not be generalizable across all groups of provider specialty (e.g., obstetrician-gynecologists, primary care providers). It is possible that providers in some specialties implemented changes. The MarketScan database includes only individuals insured by employer-sponsored health insurance, and as such, the results may have limited generalizability to other populations.

Conclusions

Our findings suggest that neither the 2013 QUERI evidence report nor the 2014 ACP guideline change, recommending against SPEs, altered the overall trend in monthly rates of SPEs from 2012 to 2019. SPEs have questionable benefits in routine care, and false positives for diseases such as ovarian cancer can cause significant undue harm. Future work is needed to understand these harms as well as concrete morbidity and mortality benefits of SPEs for conditions other than cervical cancer, ovarian cancer, gonorrhea, and chlamydia and to better understand why recent evidence and guideline developments have not impacted real-world practice.

Authors' Contributions

R.A.P. conceptualized the study idea and methods, curated the data, performed the analyses, acquired funding, created the tables and figures, and wrote the original draft. C.T.T. and C.R.T. assisted with the study conceptualization, supervised the work, and assisted with funding acquisition. A.K.K. assisted with the study conceptualization and validated study approach. A.C.K. provided methodological expertise. D.M.C. validated study approach. V.P. assisted with software implementation. All authors were involved with reviewing and editing the manuscript and approved of the final draft.

Footnotes

Author Disclosure Statement

No interests to disclose.

Funding Information

The project described was supported by the Agency for Health Research Quality (project number 1R36HS028539-01) and the National Center for Advancing Translational Sciences (NCATS), National Institutes of Health, through Grant Award Number UL1TR002489. The first author’s time was, in part, supported by a Bristol-Myers Squibb—University of North Carolina Worldwide Health Economics and Outcomes Research Fellowship. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Department of Veterans Affairs. The authors are responsible for the contents of this article.

Supplemental Material

References

Qin

, Saraiya

, Martinez

, et al. Prevalence of Potentially Unnecessary Bimanual Pelvic Examinations and Papanicolaou Tests. 2020; 180(2):274–280.

754 ACON. The Utility of and Indications for Routine Pelvic Examination. Obstet Gynecol, 2018; doi: 10.1097/AOG.0000000000002895

Grossman

, Curry

, Owens

, et al.; US Preventive Services Task Force. Screening for ovarian cancer US preventive services task force recommendation statement. JAMA, 2018; 319(6):588–594; doi: 10.1001/jama.2017.21926

Lefevre

, U.S. Preventive Services Task Force. Screening for chlamydia and gonorrhea: U.S. Preventive services task force recommendation statement. Ann Intern Med, 2014; 161(12):902–910; doi: 10.7326/M14-1981

Moyer

, U.S. Preventive Services Task Force. Screening for Cervical Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med, 2012; 156(12):880–891, W312; doi: 10.7326/0003-4819-156-12-201206190-00424

168 PBN. Cervical Cancer Screening and Prevention. Obstet Gynecol, 2016:128; doi: 10.1097/AOG.0000000000001699

Saslow

, Solomon

, Lawson

, et al.; ACS-ASCCP-ASCP Cervical Cancer Guideline Committee. American Cancer Society, American Society for Colposcopy and Cervical Pathology, and American Society for Clinical Pathology screening guidelines for the prevention and early detection of cervical cancer. CA Cancer J Clin, 2012; 62(3):147–172; doi: 10.3322/caac.21139

Bloomfield

, Olson

, Cantor

, et al. Screening Pelvic Examinations in Asymptomatic Average Risk Adult Women. 2013.

Bloomfield

, Olson

, Greer

, et al. Screening pelvic examinations in asymptomatic, average-risk adult women: An evidence report for a clinical practice guideline from the American College of Physicians. Ann Intern Med, 2014; 161(1):46–53; doi: 10.7326/M13-2881

10.

Qaseem

, Humphrey

, Harris

, et al.; Clinical Guidelines Committee of the American College of Physicians. Screening pelvic examination in adult women: A clinical practice guideline from the American College of Physicians. Ann Intern Med, 2014; 161(1):67–72; doi: 10.7326/M14-0701

11.

American Academy of Family Physicians. Clinical Prevantative Service Recommendation: Screening Pelvic Exam. Available from: https://www.aafp.org/family-physician/patient-care/clinical-recommendations/all-clinical-recommendations/screening-pelvic-exam.html [Last accessed: September 15, 2020].

12.

Bibbins-Domingo

, Grossman

, Curry

, et al.; US Preventive Services Task Force. Screening for gynecologic conditions with pelvic examination US preventive services task force recommendation statement. JAMA, 2017; 317(9):947–953; doi: 10.1001/jama.2017.0807

13.

Mason

, Sappenfield

, Turyk

, et al. Prevalence of and factors associated with receiving a pelvic examination or pap testing under the age of 21 Years. Journal of Adolescent Health, 2020; 67(4):562–568; doi: 10.1016/j.jadohealth.2020.03.002

14.

Martinez

, Qin

, Saraiya

, et al. Receipt of pelvic examinations among women aged 15-44 in the United States, 1988-2017. NCHS Data Brief, 2019(339):1–8; doi: 10.1001/jamainternmed.2019.5727

15.

Kinlaw

, Jonsson Funk

, Conover

, et al. Impact of new medications and $4 generic programs on overactive bladder treatment among older adults in the united states, 2000–2015. Med Care, 2018; 56(2):162–170; doi: 10.1097/MLR.0000000000000858

16.

Penfold

, Zhang

. Use of interrupted time series analysis in evaluating health care quality improvements. Acad Pediatr, 2013; 13(6 Suppl):S38–S44; doi: 10.1016/j.acap.2013.08.002

17.

Shadish

, Cook

, Campbell

. Experimental and Quasi-experimental Designs for Generalized Causal Inference. Houghton Mifflin; 2002.

18.

Wagner

, Soumerai

, Zhang

, et al. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther, 2002; 27(4):299–309; doi: 10.1046/j.1365-2710.2002.00430.x

19.

Kulaylat

, Schaefer

, Messaris

, et al. Truven Health Analytics MarketScan Databases for Clinical Research in Colon and Rectal Surgery. Clin Colon Rectal Surg, 2019; 32(1):54–60; doi: 10.1055/s-0038-1673354

20.

Ellison

, Cole

, Hanchate

, et al. Nonindicated pelvic examinations during contraceptive encounters: Prevalence and provider variations. Contraception, 2021; 103(4):239–245; doi: 10.1016/j.contraception.2020.12.011

21.

Rosenlund

, Leivseth

, Nilsen

, et al. Extent, regional variation and impact of gynecologist payment models in routine pelvic examinations: A nationwide cross-sectional study. BMC Womens Health, 2017; 17(1):114; doi: 10.1186/s12905-017-0471-2

22.

Freeman

, Holford

. Summary Rates. Biometrics, 1980; 36(2):195–205; doi: 10.2307/2529971

23.

Greenland

, Rothman

. Introduction to stratified analysis. In: Rothman

, Greenland

, Lash

., eds. Modern Epidemiology. Lippincott Williams & Wilkins; 2008:258–282.

24.

U.S. Census Bureau. Current Population Survey. Available from: https://www.census.gov/cps/data/cpstablecreator.html [Last accessed: June 9, 2021].

25.

Guirguis-Blake

, Henderson

, Perdue

, et al. Screening for gynecologic conditions with pelvic examination: A systematic review for the U.S. preventive services task force. 2017; 317(9):947–953.

26.

Cabana

, Rand

, Powe

, et al. Why Don’t Physicians Follow Clinical Practice Guidelines? JAMA, 1999; 282(15):1458–1465; doi: 10.1001/jama.282.15.1458

27.

Meissner

, Tiro

, Yabroff

, et al. Too much of a good thing? Physician practices and patient willingness for less frequent pap test screening intervals. Med Care, 2010; 48(3):249–259; doi: 10.1097/MLR.0b013e3181ca4015

28.

Lee

, Morgan

, Fiascone

, et al. Underscreening, overscreening, and guideline-adherent cervical cancer screening in a national cohort. Gynecol Oncol, 2022; 167(2):181–188; doi: 10.1016/j.ygyno.2022.09.012

29.

Fink

. Change in american cancer society checkup guidelines for detection of cervical cancer. CA Cancer J Clin, 1988; 38(2):127–128; doi: 10.3322/canjclin.38.2.127

30.

Waxman

. Guidelines for cervical cancer screening: History and scientific rationale. Clin Obstet Gynecol, 2005; 48(1):77–97; doi: 10.1097/01.grf.0000151590.08451.26

31.

Hategeka

, Ruton

, Karamouzian

, et al. Use of interrupted time series methods in the evaluation of health system quality improvement interventions: A methodological systematic review. BMJ Glob Health, 2020; 5(10):e003567; doi: 10.1136/bmjgh-2020-003567