Objective
Guidelines recommend when the human papillomavirus (HPV) and Papanicolaou tests are used together (HPV co-test) for routine cervical cancer screening, screening intervals can be extended to 3 years. We assessed HPV test practices and Papanicolaou test screening interval recommendations of US providers.
Study Design
Using a multistage probability design, we analyzed nationally representative data that were collected in 2006 through the Centers for Disease Control and Prevention’s National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey.
Results
Approximately 51% of providers ordered the HPV co-test; however, clinical vignettes found that <15% of providers who ordered the HPV test recommend the next Papanicolaou test in 3 years for women with concurrent normal HPV co-test results and a documented normal screening history.
Conclusion
Overall, annual cervical cancer screening continues to be a common recommendation, regardless of whether a screening history has been established or an HPV test has been ordered.
The Papanicolaou test is recognized as an effective and successful cervical cancer screening test. It has become an essential component of annual primary care for women with access to regular medical care. In addition, infection with high-risk human papillomavirus (HPV) has been identified as a principal cause of most cervical abnormalities and cervical cancers. HPV tests to identify high-risk types of HPV are available for both cervical cancer screening and management. Administering the HPV test at the same time as the Papanicolaou test, (ie, HPV co-testing) is an efficient approach with practical advantages. The co-test screening approach is recommended only for use among women ≥30 years old by many US health and professional medical organizations. Because of the negligible risk of invasive cervical cancer among women who have concurrent normal results on the Papanicolaou test and negative findings on the HPV test within 3-5 years, the traditional annual screening interval for cervical cancer can safely be extended to every 3 years.
Even before the approval of the HPV co-test strategy by the US Food and Drug Administration in 2003, the American Cancer Society and the American College of Obstetricians and Gynecologists recommended extending the screening interval for cervical cancer to 2-3 years for women ≥30 years old who had a history of normal Papanicolaou test results. Studies of primary care provider practices, which have been conducted from 2003 to the present, clearly show that providers are not incorporating these recommended screening intervals with the Papanicolaou test alone in clinical practice. To date, there are limited data on how the HPV co-test strategy is being adopted, whether providers are using it according to guidelines and extending screening intervals, and how it has affected patient care and resource use. With this study, we hoped to assess the state of HPV testing and co-testing practices and to document the recommendations for screening intervals as given by private office-based providers and hospital outpatient departments (OPDs).
Materials and Methods
Data sources
The Centers for Disease Control and Prevention (CDC), Division of Cancer Prevention and Control, commissioned collection of the Cervical Cancer Screening Supplement (CCSS) beginning in 2006. The objective of the CCSS is to obtain national data on providers’ self-reported cervical cancer screening practices. The CCSS is administered as a supplement to CDC’s National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS).
NAMCS collects information on visits made to private office-based physicians from a representative, randomly selected sample of physician offices in a 3-stage probability design that involves probability samples of primary sampling units, physician practices within primary sampling units, and patient visits within practices. Traditionally, only visits to the offices of non–federally employed physicians who were classified by the American Medical Association or the American Osteopathic Association as “office-based, patient care” had been included in the NAMCS. In 2006, however, NAMCS began collecting data from physicians and midlevel providers (physician assistants, nurse practitioners, and certified nurse midwives) who were employed in community health centers (CHCs), thus increasing the sample.
The NHAMCS, which uses a 4-stage probability design with samples of primary sampling units, hospitals within primary sampling units, clinics/emergency service areas within outpatient/emergency departments, and patient visits within clinics/emergency service areas. Information is collected on visits made to emergency and OPDs of noninstitutional general and short-stay hospitals, excluding federal, military, and Veterans Administration hospitals, that are located in the 50 states and the District of Columbia. The Census Bureau is the data collection agent for both the NAMCS and NHAMCS.
With the first administration in 2006, private office-based physicians, CHC providers, and OPD physicians from the NAMCS and NHAMCS sample were eligible to complete the CCSS if they provided Papanicolaou tests and specialized in general/family practice, internal medicine, obstetrics/gynecology, or general medicine. The CCSS was not sent to hospital emergency departments. The CCSS was shared with providers (either in the offices, CHCs, or hospitals) at the end of the established annual NAMCS/NHAMCS reporting period, when the interviewer was visiting. The CCSS was accompanied by an introductory letter that explained the importance of the information being collected and how responses would be protected. Completed surveys were mailed back or picked-up by the interviewer. Providers also had an option to complete the CCSS on a secure web portal that they could access with a supplied user name and password.
The 2006 CCSS was a 9-item self-administered questionnaire that took approximately 15 minutes to complete. The NAMCS and NHAMCS protocols were approved by the Research Ethics Review Board of CDC’s National Center for Health Statistics. The Research Ethics Review Board waived requirements to obtain informed consent from patients and patient authorization for health care providers to release medical record data.
Response rates were calculated according to Office of Management and Budget Guidelines, which dictate that response rates for cross-sectional sample surveys are calculated as the product for 2 or more unit-level response rates. In 2006, of 495 eligible NAMCS and CHCs providers, 387 providers responded to the CCSS (61.1% unweighted response rate); however, we excluded 11 NAMCS records from our analysis that were identified as specializing in pediatrics or “other”, which made the total NAMCS records to be analyzed 376. A total of 255 NHAMCS providers were considered eligible to participate in the CCSS, of which 216 providers responded (84.7% unweighted response rate).
Data measures
For office-based and CHC providers, demographics were collected through the American Medical Association and American Osteopathic Association master files. Variables that were collected included specialty, age, sex, race, and type of medical school graduate (United States vs foreign). Practice characteristics for office-based and CHC providers were collected through the physician induction interview. Variables that were collected included the number of providers in the practice, geographic region of the practice, and county population of the practice. Data that were collected on hospital OPD physicians included clinic specialty (limited to general medicine and obstetrics/gynecology), whether the OPD was affiliated with a teaching hospital, geographic region of the practice, and county population of the practice.
For this study, we examined the type of Papanicolaou test that had been used to screen for cervical cancer (conventional Papanicolaou test, liquid-based cytologic test, or both), if the practice orders the HPV test (which could include the reflex/recall or co-test approach), orders the HPV co-test, and recommendation for time until next Papanicolaou test screening.
Our primary outcome measures were (1) HPV co-test use and (2) recommended screening intervals for next Papanicolaou test among providers who order the HPV test, as assessed by clinical vignettes. HPV co-test use and practices were examined by asking whether the provider or clinic routinely ordered the HPV test at the same time as the Papanicolaou test. Providers who answered “Yes” were then asked for the age group of the patient for which they ordered the co-test.
Recommended screening intervals for the next Papanicolaou test were examined by asking whether the provider or clinic routinely ordered or provided the HPV test. Providers who answered “Yes” to ordering the HPV test were then asked their recommendation for next Papanicolaou test according to 3 clinical vignettes. The vignettes describe a woman between the ages of 30 and 60 years with a current normal Papanicolaou test with (1) no current HPV test results and a history of 2 consecutive normal Papanicolaou test results, (2) a current negative HPV test result and a history of 2 consecutive normal Papanicolaou test results, and (3) a current negative HPV test result and no history of Papanicolaou test provided. Note guidelines at the time of survey administration (2006) supported extending the screening interval up to 3 years in each of the vignettes described.
Analysis
To provide a nationally representative sample of the providers and their practices, a sample weight that also accounted for nonresponse was assigned to each of the responders. We used SAS software (version 9. 2; SAS Institute Inc, Cary, NC) and SUDAAN software (release 10; Research Triangle Institute, Research Triangle Park, NC) to apply sampling weights and account for stratified survey design in our analysis. CCSS sample sizes are small and therefore results are limited to descriptive analysis only. We computed weighted percentages and 95% CIs to examine demographic variables that were related to providers and their screening and co-testing practices and stratified by setting (office-based vs hospital OPD). We further stratified our analysis by obstetrics/gynecology vs other primary care specialties within the setting because we are interested in cervical screening practices according to provider specialty. In NAMCS analysis, the non–obstetrics/gynecology category included general/family practice, internal medicine, and midlevel CHC providers. In NHAMCS analysis, non–obstetrics/gynecology specialty included general medicine clinics only. Relative standard error (calculated as the standard error percent divided by the estimated percent × 100) ≥30% is considered unstable and should be interpreted cautiously. Additionally, estimates based on <30 sample cases are also considered to be unreliable.
Results
Provider and practice characteristics
The most common specialty for private office-based providers was general/family practice, followed by obstetrics/gynecology, and internal medicine ( Table 1 ). Most providers were 35-54 years old, male, white non-Hispanic, a US medical school graduate, and worked either in a small group (3-5 physicians) or in solo practice. The South and metropolitan areas represented the highest percentage of private office-based providers to be sampled.
| Characteristic | Office-based providers (n = 376) | Hospital outpatient departments (n = 216) | ||
|---|---|---|---|---|
| n | % (95% CI) | n | % (95% CI) | |
| Specialty | ||||
| Obstetrics/gynecology | 90 | 26.8 (22.7–31.2) | 108 | 24.0 (17.4–32.1) |
| General/family practice a | 170 | 46.0 (40.6–51.5) | — | — |
| General medicine | — | — | 108 | 76.0 (67.9–82.6) |
| Internal medicine | 57 | 23.9 (18.5–30.3) | — | — |
| Midlevel provider b | 59 | 3.3 (2.4–4.7) | — | — |
| Age, y c | ||||
| <35 | 48 | 8.0 (4.9–12.9) | — | — |
| 35 to <45 | 123 | 33.4 (26.7–40.9) | — | — |
| 45 to <55 | 106 | 31.0 (25.0–37.6) | — | — |
| 55 to <65 | 70 | 22.8 (17.6–29.0) | — | — |
| ≥65 | 11 | 3.1 (1.4–6.6) d | — | — |
| Sex c | ||||
| Male | 201 | 58.6 (51.7–65.2) | — | — |
| Female | 173 | 41.3 (34.7–48.3) | — | — |
| Race | ||||
| White non-Hispanic | 208 | 49.3 (42.1–56.5) | — | — |
| Black non-Hispanic | 27 | 4.0 (2.5–6.4) | — | — |
| Asian | 30 | 10.1 (6.2–16.1) | — | — |
| Hispanic | 5 | 2.1 (0.8–5.7) d | — | — |
| Native American/Alaskan | 4 | 0.01 (0.0–0.04) d | — | — |
| Other | 9 | 4.6 (2.0–10.4) d | — | — |
| Unknown/missing | 93 | 29.8 (23.9–36.5) | — | — |
| Medical school graduation e | ||||
| Foreign school | 59 | 23.6 (17.5–30.9) | — | — |
| US school | 219 | 65.4 (57.5–72.5) | — | — |
| Unknown/missing | 39 | 11.1 (7.7–15.6) | — | — |
| Practice size c , f | ||||
| Solo practice | 85 | 31.2 (24.1–39.2) | — | — |
| 2 physicians | 62 | 12.9 (9.2–18.0) | — | — |
| 3-5 physicians | 133 | 32.2 (25.6–39.7) | — | — |
| 6-10 physicians | 53 | 14.6 (10.1–20.7) | — | — |
| 10 physicians | 42 | 9.0 (5.2–15.2) | — | — |
| Affiliated with a teaching hospital c | ||||
| Yes | — | — | 145 | 46.9 (34.3–59.9) |
| No | — | — | 68 | 52.2 (39.2–65.0) |
| Geographic region of practice g | ||||
| Northeast | 72 | 16.9 (12.7–22.1) | 74 | 27.0 (17.0–40.1) |
| Midwest | 97 | 24.2 (18.8–30.5) | 61 | 28.6 (17.5–43.0) |
| South | 113 | 37.1 (31.2–43.3) | 47 | 33.0 (20.5–48.4) |
| West | 94 | 21.9 (16.4–28.6) | 34 | 11.5 (5.4–22.6) d |
| County population of practice area | ||||
| Metropolitan (>250,000) | 284 | 77.0 (68.9–83.5) | 177 | 59.1 (39.7–76.0) |
| Nonmetropolitan (<249,999) | 92 | 23.0 (16.5–31.2) | 39 | 40.9 (24.0–60.3) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
