Change in Provider Beliefs Regarding Cervical Cancer Screening Intervals After an Educational Intervention

Abstract

Background:

Current cervical cancer screening guidelines include the option of lengthening the screening interval to 5 years for average-risk women aged 30–65 years when screened with Pap and human papillomavirus (HPV) test (co-test). Because many providers are reluctant to extend screening intervals, we launched an educational intervention to promote recommended screening practices. The study objective was to assess changes in provider attitudes and beliefs to extending screening intervals among low-income women.

Methods:

The study was conducted in 15 clinics in Federally Qualified Health Centers in Illinois. Providers in the intervention arm received a multicomponent educational intervention. Fifty-six providers (n = 29 intervention and n = 27 control) completed baseline and 12-month follow-up surveys assessing beliefs and intentions about extending screening intervals.

Results:

The 12-month assessment showed providers in the intervention arm were significantly more likely than those in the control arm to recommend a 3-year screening interval (guideline recommendation at time of study) with a normal co-test result. Providers who received the intervention were significantly more likely to agree that routine co-testing is the best way to screen for cervical cancer, that extending the screening interval would be good, easy, and beneficial, and to disagree that the increased screening interval would cause patients to lose contact with the medical system.

Conclusion:

Educating providers on the natural history of HPV infection and cervical cancer and the benefits of extended intervals increased their willingness to follow guidelines. This study provides evidence that an educational intervention delivered with HPV testing materials may be effective in encouraging appropriate cervical screening intervals.

Introduction

Persistent carcinogenic human papillomavirus (HPV) plays a central role in the development of cervical cancer. HPV testing with the Pap test (called co-testing) every 5 years is a recommended option for cervical cancer screening for average-risk women aged 30–65 years.^1
–3 However, both providers and patients may be reluctant to move to the recommended intervals and annual screening remains a common practice throughout the United States.^4

–7 Providers play a critical role in encouraging appropriate screening among their patients. Barriers to longer screening intervals by providers include lack of knowledge of current guidelines, patient preferences for annual Pap tests, fear of patients not returning for other preventive services, and concern about missing early cancers.^6
–8 When assessing adherence to guidelines from the patient perspective, potential barriers to extended screening intervals include lack of knowledge about cervical cancer screening, a desire for frequent care, and perceived risk of cervical cancer.^9,10

To date, no studies have examined provider and patient acceptance of co-testing with longer screening intervals in a clinical setting that serves a low-income population. Federally Qualified Health Centers (FQHCs) provide care regardless of a person's insurance status, income, or ability to pay, and they are a large essential component of the U.S. healthcare safety-net for people who otherwise have limited options to access essential care.¹¹ Conducting research in FQHCs allows for understanding the access and barriers to care among the underserved. In 2009, the Centers for Disease Control and Prevention (CDC) launched the CDC's Cervical Cancer (Cx3) Study—a multicomponent educational intervention to identify facilitators and barriers to guideline-consistent use of the co-test in an underserved population receiving care at FQHCs.¹² The primary objective of the Cx3 Study was to introduce co-testing along with an educational intervention to encourage providers and patients to extend the cervical cancer screening interval for women with normal co-test results. The purpose of this analysis is to describe the baseline and follow-up assessments of providers in the study and determine the effects of the educational intervention on provider beliefs and intentions about extending screening intervals.

Methods

Study design and participants

Six FQHCs with 15 individual clinics were enrolled in the study by August 2009. Recruitment of FQHCs occurred through the CDC's National Breast and Cervical Cancer Early Detection Program (NBCCEDP).¹³ Each clinic was assigned to one of two study arms: intervention (seven clinics) or control (eight clinics). The intervention and control arms were matched at baseline on clinic attributes, including racial/ethnic characteristics of the patient population, geographical location, and patient volume. Clinics in both groups received free HPV tests and hard copies of clinical practice guidelines current during the time of the study (2009–2010), which specified following a 3-year screening interval for individuals who received normal results on both the Pap and HPV tests (defined as guideline consistent in the analysis).¹⁴ In addition, clinics in the intervention group received a multicomponent educational program, including provider and patient education materials designed for this project.¹² The educational intervention consisted of multiple evidence-based strategies to encourage providers to follow recommendations. These included grand rounds, four separate 1-hour academic detailing sessions, and a password-protected website containing educational podcasts on recommended intervals for cervical cancer screening. Details of the educational intervention are published elsewhere.¹² In total, five educational components were offered to intervention arm providers (one grand round and four academic detailing sessions). This study was approved by the CDC's Institutional Review Board and the Office of Management and Budget (OMB Control #0920–0814).

Providers in the participating FQHCs were eligible to participate in the study if they personally performed Pap testing for routine screening (n = 109). Ninety-eight providers (response rate 89.9%) were enrolled at baseline⁷ with the average age of 41.3 years. Almost 70% of the providers were from the Chicago area, and 75% were female. With respect to race/ethnicity, 48% of the providers were non-Hispanic (NH) white, 22% NH Asian, and 18% NH Black. On average, providers attended to 52 female patients per week. Of the 98 providers with baseline assessments, 19 left the clinic during the study period, 9 became ineligible (stopped performing well-woman exams or routine Pap testing), 14 refused participation in the follow-up survey, and 56 completed the 12-month follow-up survey. Nonresponders to the 12-month survey were more likely to be male and physicians who specialized in obstetrics/gynecology (all chi-square p < 0.05). There were no significant differences in response status by study arm, screening interval practices, or beliefs. Of the 56 providers who completed follow-up surveys, 29 were in the intervention arm. Of those, 25 attended at least one of the teaching educational components; mean sessions attended was 4.20. Providers in the intervention arm who did not attend any training sessions were removed from the analytic sample leaving 84 providers at baseline and 52 providers at follow-up. Among providers in the intervention clinics, there were no significant differences in the measures of attitudes, beliefs, and practices at baseline by attendance at training sessions (Fisher's Exact Test, all p > 0.05).

Measures

Coordinators at participating clinics distributed surveys to providers along with a $50 cash incentive at baseline and 12-month follow-up. The surveys assessed attitudes, beliefs, and practices regarding cervical cancer screening (specifically measures related to co-testing in general and extending the screening intervals).⁷ Measures and clinical vignettes were modeled upon national primary care provider cervical cancer screening practice surveys.^4,15

Statistical analysis

Cross-sectional differences were evaluated with Fisher's Exact Test and the Mantel–Haenszel chi-square test of linear trend. Cross-sectional analyses were conducted on the sample who responded to both the baseline and follow-up surveys (n = 52). Evaluation of longitudinal data was conducted with random effects, ordinal logistic regression models. Models included three terms: main effect of study arm (coded 1 for intervention arm; 0 for control arm, referent category); main effect of time (coded 1 for follow-up; 0 for baseline survey, referent category); and the interaction of study arm with time (coded 1 for intervention arm at follow-up; 0 for other, referent category). The result of the main effect of study arm indicates whether there were significant differences in provider responses between study arms. The result of the main effect of time indicates whether there were significant differences in response between the two surveys. The impact of the intervention is evaluated with the interaction term and is referred to as the “Intervention Effect” in tables and figures. The sample for longitudinal analysis includes baseline data for n = 84 providers with completed baseline surveys and n = 52 providers with completed follow-up surveys. The proportionality assumption was tested with the Brant test; no violations were found. Differences were considered statistically significant with a p-value of less than 5%.

Results

At baseline, in response to a clinical vignette of a 35-year-old woman with normal co-test results, providers suggested the following screening intervals: annual (55%), 2-year (20%), and 3-year (25%) (no significant differences by study arm; data not shown). Limiting the sample to providers who recommended annual screening intervals at baseline (n = 28), the 12-month follow-up assessment showed providers in the intervention arm were significantly more likely than those in the control arm (54% versus 7%, p = 0.008) to recommend a 3-year screening interval (Table 1).

Table 1.

Screening Interval Recommended at 12-Month Follow-Up for a 35-Year-Old Woman with Normal Co-Test, Providers Who Would Screen Annually at Baseline (n = 28)

		12-month follow-up screening interval, %
	n	1 year	2 years	3 years	p^a
Baseline, providers who report screening annually
Control	15	73	20	7	0.008
Intervention	13	31	15	54

Mantel–Haenszel chi-square test of linear trend.

Provider attitudes about extending the cervical cancer screening interval to three or more years for a woman over age 30 with normal co-test results are shown in Table 2. There were no significant differences in attitudes about extending the screening interval by study arm at baseline. However, at follow-up, there was evidence of an intervention effect with providers in the intervention arm more likely to report that extending the screening interval would be good (OR 6.45, p = 0.038), easy (OR = 5.18, p = 0.032), and beneficial (OR = 8.53, p = 0.034), and less likely to report that extending the screening interval would cause patients to lose contact with the medical care system (OR 9.80, p = 0.044).

Table 2.

Provider Attitudes Reported for Extending the Cervical Cancer Screening Interval for Women with Normal Co-Tests by Study Arm and Survey Period (Baseline—2009/10 and Follow-Up 2010/11)

	Cross-sectional analysis ^a						Longitudinal analysis random effects model ^b
	Baseline			Follow-up
Extending screening interval with normal co-test would…	C %	I %	p	C %	I %	p		Study arm	Time effect	Interaction arm × time
Good-Bad
Bad	37	18	0.616	23	0	0.004	OR	1.25	1.29	6.45
Neither	11	36		27	14		p	0.638	0.608	0.038
Good	52	45		50	86
Difficult-Easy
Difficult	22	29	0.358	42	5	0.033	OR	0.68	0.53	5.18
Neither	26	33		19	43		p	0.400	0.205	0.032
Easy	52	38		38	52
Beneficial-Harmful
Harmful	30	18	0.631	23	0	0.007	OR	1.36	2.55	8.53
Neither	33	45		23	14		p	0.596	0.103	0.034
Beneficial	37	36		54	86
Cause patients to lose contact with medical care system
Agree	70	67	0.541	81	43	0.007	OR	2.20	0.32	9.80
Neither	15	8		7	17		p	0.337	0.139	0.044
Disagree	15	25		11	39
Help reduce healthcare costs
Disagree	11	13	0.812	7	0	0.719	OR	1.04	3.82	0.94
Neither	15	17		4	14		p	0.951	0.081	0.952
Agree	74	71		89	86
Put patient at increased risk for cervical cancer
Agree	33	38	0.840	15	9	0.449	OR	0.95	2.68	1.77
Neither	26	13		33	30		p	0.934	0.093	0.516
Disagree	41	50		52	61
Result in higher rates of cervical precancer
Agree	33	38	0.912	30	30	0.556	OR	0.92	0.68	2.00
Neither	22	17		33	17		p	0.885	0.470	0.405
Disagree	44	46		37	52
Increase patient concerns about missing cervical cancer
Agree	70	67	0.791	70	65	0.477	OR	1.86	0.72	1.36
Neither	11	13		15	9		p	0.424	0.636	0.757
Disagree	19	21		15	26
Put me and practice at risk for liability if patient's next result is abnormal
Agree	56	50	0.956	48	30	0.113	OR	1.41	1.17	3.33
Neither	11	21		26	22		p	0.650	0.802	0.201
Disagree	33	29		26	48
Take too much time to explain to the patient
Agree	0	13	0.153	4	13	0.453	OR	0.68	1.15	1.16
Neither	22	21		19	13		p	0.616	0.842	0.891
Disagree	78	67		78	74

Cross-sectional analysis presents percent distributions by study arm at each survey period. For comparability across time points, the sample was limited to the n = 52 providers who completed both baseline and follow-up surveys. Significant differences were tested with Mantel–Haenszel chi-square test of linear trend.

Evaluation of longitudinal data was conducted with random effects, ordinal logistic regression. Models included three terms: study arm (coded 1 for intervention arm; 0 for control arm, referent category); time (coded 1 for follow-up; 0 for baseline survey, referent category); and, the interaction of study arm with time (coded 1 for intervention arm at follow-up; 0 for other, referent category). The interaction of study arm with time is the estimate that tests the effect of the educational intervention. Sample includes baseline data for n = 84 providers with completed baseline surveys and n = 52 providers with completed follow-up surveys.

Bold values indicate p-value was significant at p < 0.05.

C, control arm; I, intervention arm; OR, odds ratio.

In contrast to attitudes about extending the screening interval, the vast majority of providers had positive beliefs about co-testing women over 30 years of age at study inception. At baseline, most providers believed the co-test to be beneficial (90%), good (80%), and easy (73%) (Fig. 1). The majority of providers agreed that co-testing would help determine a plan for follow-up if Pap result is abnormal (94%), help explain cervical cancer risk to patients (88%), help determine the appropriate future screening intervals (85%), and would give a more complete understanding of patients' current state of health and risk for disease (79%). In addition, the majority of providers disagreed that co-testing would take too much time (94%), would be an extra burden for office staff (92%), would not provide any more useful information than the Pap test alone (87%), would be a test patients would not want (79%), was not needed (75%), and would only be needed for high-risk patients (71%). There were no significant differences in these beliefs by study arm at baseline. There was a significant interaction effect for one measure of beliefs about co-testing (i.e., “co-testing is the best way to screen for cervical cancer”) (Fig. 2; OR 25.17, p = 0.035).

FIG. 1.

Provider baseline beliefs about using the co-test for routine cervical cancer screening (%). Testing with random effects binary and ordered logistic regression determined no significant differences by study arm or time (p < 0.10), therefore only study arm combined, baseline percent distributions are presented.

FIG. 2.

Provider beliefs about co-testing for routine cervical cancer screening by survey time period and intervention group. Co-testing is the best way to screen for cervical cancer… Significant differences were tested with random effects ordinal logistic regression. Model included three terms: study arm (control arm, referent category); time (baseline survey, referent category); and the interaction of study arm with time (for study arm, OR 0.29, p = 0.260; for time, OR 0.61; p = 0.532; for interaction, OR 25.17, p = 0.035).

Discussion

Our findings showed the impact of an enhanced educational intervention on the attitudes and intentions of providers who reported screening annually at the initiation of the study. Overall, providers in this study had positive beliefs at baseline about using the co-test and that positivity continued through to the 12-month follow-up. However, providers varied in their beliefs about extending the screening interval according to guidelines. Providers in the intervention arm gained more favorable attitudes toward extended screening intervals. In addition, more intervention providers reported they would increase the screening interval according to recommended guidelines.

Before the 2012 revisions to the screening guidelines, the recommendations for screening were not consistent across all organizations and included recommendations to screen every 2 to 3 years with a history of negative Pap tests.¹⁴ When the Cx3 Study began in 2009, annual screening was the most common screening practice, even though not recommended by national organizations.^15,16 Our baseline data found that only 25% of providers would recommend a 3-year screening interval with a normal co-test result.⁷ When we examined the same baseline data as part of a separate analysis to determine whether providers were more comfortable with extending the screening interval with a history of negative Pap tests, the same percent (27%) agreed to a 3-year interval (data not shown). Adding the HPV test to the Pap test as a screening strategy meant the history of negative tests was no longer a requirement to establish longer intervals between screenings. Current guidelines, revised in 2012, now recommend a 3-year interval with a Pap test alone and do not require the previous criteria involving a history of normal results. Studies have shown a small risk of diagnosis of cervical cancer within the 3-year time period.³

The educational materials developed for this study focused heavily on the natural history of cervical cancer and the long latency period (up to decades) from HPV infection to invasive cervical cancer. They also addressed common concerns and barriers to extending the screening intervals.¹² One of the common barriers reported during the early formative work of this study involved legal aspects such as liability of missing cancers, which we addressed during the first academic detailing session. In addition, providers felt that women may not return for other preventive care. Thus, our patient materials also focused on the importance of well-woman exams and the importance of returning for preventive care. The significant increase in the percent of intervention providers willing to extend the screening interval from annual to 3 years compared to the control over the 12-month follow-up (Table 1), indicates that the messages provided in the enhanced educational intervention appropriately addressed some of the concerns about extending the interval of this group of providers.

In 2012, when guidelines from all major organizations consistently recommended when and how often to screen, they also were revised to recommend extending the interval to 5 years (from 3 years) with a co-test. Although our study did show an increased percentage of providers willing to extend the interval to 3 years (from annual at baseline), other studies have found that providers and women were more resistant to longer intervals (up to 5 years).^8,10 Countries with organized screening programs have had later initiation of screening and longer screening intervals with no increase in mortality when compared to the United States.¹⁷ Clinicians in managed care organizations have been the most successful at increasing intervals.¹⁸

The provider beliefs, attitudes, and practice recommendations are self-reported and might not reflect actual behavior in the clinical setting. As a demonstration study in one state, our results may not be generalizable. However, when we examined the racial and ethnic distribution of this study population in Illinois, it was very similar to the NBCCEDP client population nationwide. In addition, it was up to the provider whether he/she wanted to use/attend the offered intervention resources. Thus, while the intervention components were consistently delivered, uptake and use may have varied by the provider. In addition, there was a very high percentage of staff turnover at all clinics and we did not include new providers for the follow-up survey. Both study arms received HPV tests and therefore we did not have access to a true control group (in which they did not receive any intervention).

Conclusion

While many studies are aimed at introducing new technology, very few are focused on introducing new technology in a low-income population coupled with evidence-based strategies that emphasize the best care through less frequent screening. The Cx3 Study's findings on practices and attitudes regarding HPV testing and screening interval extension are very timely, now that national organizations have aligned their recommendations for co-testing and screening intervals. In addition, primary HPV testing has recently been FDA approved with interim guidelines¹⁹ and management guidelines have been updated with a focus on triaging women with specific high-risk HPV types.²⁰ Thus as the science advances and additional choices are given, providers may need adequate materials and tools to facilitate talking to women about the new testing options.²¹ The information on the harms of overtesting, which may lead to overdiagnosis and overtreatment, is only useful to the extent that it is understood and accepted by providers and patients. CDC has produced several materials to educate providers and patients on the appropriate use of the co-test and these materials may be adapted over time to meet the changing needs of the population and the latest science (www.cdc.gov/cancer/cervical/).

Footnotes

Acknowledgments

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC. This article was written in the course of employment by the United States Government or under a contract with Battelle (200–2002-00573, Task Order No. 0006 and No. 0017) and it is not subject to copyright in the United States. The authors gratefully acknowledge the medical directors and administrators from the FQHCs who participated in the study, and the Illinois Breast and Cervical Cancer Early Detection Program.

The CDC Foundation provided in-kind support for testing reagents from Qiagen and Roche Molecular Diagnostics.

This research was supported, in part, by an appointment (LL) to the Research Participation Program at the CDC administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and CDC.

Author Disclosure Statement

The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

ACOG. ACOG Practice Bulletin Number 131: Screening for cervical cancer. Obstet Gynecol, 2012; 120:1222–1238.

U.S. Preventive Services Task Force Recommendation Statement: Screening for Cervical Cancer. 2012. Available at: www.uspreventiveservicestaskforce.org/uspstf11/cervcancer/cervcancerrs.htm Accessed January 1, 2016 .

Saslow

, Solomon

, Lawson

, et al. 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. Am J Clin Pathol, 2012; 137:516–542.

Roland

, Soman

, Benard

, Saraiya

. Human papillomavirus and Papanicolaou tests screening interval recommendations in the United States. Am J Obstet Gynecol, 2011; 205:447 e441–e448.

Verrilli

, Winer

, Mao

. Adherence to cervical cancer screening guidelines by gynecologists in the Pacific Northwest. J Low Genit Tract Dis, 2014; 18:228–234.

Perkins

, Anderson

, Gorin

, Schulkin

. Challenges in cervical cancer prevention: A survey of U.S. obstetrician-gynecologists. Am J Prev Med, 2013; 45:175–181.

Roland

, Benard

, Greek

, Hawkins

, Manninen

, Saraiya

. Primary care provider practices and beliefs related to cervical cancer screening with the HPV test in Federally Qualified Health Centers. Prev Med, 2013; 57:419–425.

Teoh

, Marriott

, Isaksson Vogel

, et al. Adherence to the 2012 national cervical cancer screening guidelines: A pilot study. Am J Obstet Gynecol, 2015; 212:62.e61–e69.

Hawkins

, Benard

, Greek

, Roland

, Manninen

, Saraiya

. Patient knowledge and beliefs as barriers to extending cervical cancer screening intervals in Federally Qualified Health Centers. Prev Med, 2013; 57:641–645.

10.

Silver

, Rositch

, Burke

, Chang

, Viscidi

, Gravitt

. Patient concerns about human papillomavirus testing and 5-year intervals in routine cervical cancer screening. Obstet Gynecol, 2015; 125:317–329.

11.

Bureau of Primary Healthcare. Health Center Fact Sheet. Available at: www.bphc.hrsa.gov/about Accessed January 1, 2016 .

12.

Benard

, Saraiya

, Greek

, et al. Overview of the CDC cervical cancer (Cx3) study: An educational intervention of HPV testing for cervical cancer screening. J Womens Health, 2014; 23:197–203.

13.

CDC. National breast and cervical cancer early detection program. Available at: www.cdc.gov/cancer/nbccedp/about.htm Accessed January 1, 2016 .

14.

Smith

, Cokkinides

, Brooks

, Saslow

, Brawley

. Cancer screening in the United States, 2010: A review of current American Cancer Society guidelines and issues in cancer screening. CA Cancer J Clin, 2010; 60:99–119.

15.

Benard

, Saraiya

, Soman

, Roland

, Yabroff

, Miller

. Cancer screening practices among physicians in the national breast and cervical cancer early detection program. J Womens Health (Larchmt), 2011; 20:1479–1484.

16.

Berkowitz

, Saraiya

, Sawaya

. Cervical cancer screening intervals, 2006 to 2009: Moving beyond annual testing. JAMA Intern Med, 2013; 173:922–924.

17.

Habbema

, De Kok

, Brown

. Cervical cancer screening in the United States and the Netherlands: A tale of two countries. Milbank Q, 2012; 90:5–37.

18.

Dinkelspiel

, Fetterman

, Poitras

, et al. Cervical cancer rates after the transition from annual Pap to 3-year HPV and Pap. J Low Genit Tract Dis, 2014; 18:57–60.

19.

Huh

, Ault

, Chelmow

, et al. Use of primary high-risk human papillomavirus testing for cervical cancer screening: Interim clinical guidance. Gynecol Oncol, 2015; 136:178–182.

20.

Massad

, Einstein

, Huh

, et al. 2012 updated consensus guidelines for the management of abnormal cervical cancer screening tests and cancer precursors. Obstet Gynecol, 2013; 121:829–846.

21.

Kuppermann

, Sawaya

. Shared decision-making: Easy to evoke, challenging to implement. JAMA Inter Med, 2015; 175:167–168.