Benard, Vicki B. PhD; Lawson, Herschel W. MD; Eheman, Christie R. PhD; Anderson, Christa; Helsel, William MS
Widespread use of the Papanicolaou (Pap) test has helped to reduce incidence and mortality from cervical cancer since the test's introduction in 1943. Each year in the United States, approximately 50 million women undergo Pap testing.1 Of these, approximately 3.5 million (7%) are diagnosed with a cytological abnormality requiring additional follow-up or evaluation.2 Determining which women with cytological abnormalities are at risk for progressive cervical neoplasia, performing appropriate and timely diagnostic tests, and treating precursors to cancer together present a major clinical and public health challenge.
Providing appropriate follow-up for the 2 low-grade Pap test abnormalities, atypical squamous cells of undetermined significance (ASC-US, Bethesda 1991 classification) and low-grade squamous intraepithelial lesions (LSIL), is important because a woman with a ASC-US has a 5% to 17% chance of having cervical intraepithelial neoplasia (CIN) 2 3 confirmed by biopsy, and a woman with LSIL has approximately a 15% to 30% chance of having CIN 2 3 identified.3 In this report, we examine patterns of follow-up for women with 2 Pap test reports of ASC-US or LSIL within 2 years who were enrolled in the National Breast and Cervical Cancer Early Detection Program (NBCCEDP) during the years 1991–2000 to determine whether these women were followed up according to the clinical guidelines. The 1994 National Cancer Institute (NCI) interim guidelines4 specify that clinicians should follow up patients with ASC-US and LSIL with repeat Pap tests every 4–6 months for 2 years or until the patient has 3 consecutive negative Pap test results. If a second result of ASC-US or LSIL is reported within 2 years, the woman should have a colposcopic examination within 30 to 60 days of the second Pap test.
Having a low income or no insurance is associated with decreased use of screening;5 and the goal of the NBCCEDP is to increase access to cervical cancer screening, diagnostic, and follow-up services among low-income and uninsured women.6 Screening services are coordinated and provided free of charge within the NBCCEDP by a wide range of organizations, including local health departments, community health centers, the practices of private physicians, and hospitals. This analysis 1) provides insight into provider practices in following up women with abnormal Pap test results as measured by several factors, including age and race or ethnicity, 2) describes clinical outcomes among a unique population of medically underserved women, and 3) may indicate areas where additional education of providers is needed to improve the adoption of new guidelines.
SUBJECTS AND METHODS
Information was collected on the demographic characteristics and Pap test results of each woman screened for cervical cancer in the NBCCEDP, which is administered by the Centers for Disease Control and Prevention (CDC). For those with abnormal results, information on diagnostic procedures, final diagnosis, and an indication that treatment was started was also collected. Our study population was comprised of women with a first program Pap test result of ASC-US or LSIL followed by a second Pap test with an abnormality—either ASC-US or LSIL—within 2 years of the first test. The time frame examined, July 1991 to September 2000, ended well before the publication of updated recommendations for follow-up of abnormal Pap test results in April 2002.3 Institutional Review Board approval was obtained for this study from the CDC.
Each woman's age was calculated using the birth date she reported at enrollment, and this age was employed throughout the study. We grouped women into 5 age groups (in years): younger than 30, 30–39, 40–49, 50–59, and 60 or older. If a woman self-identified as Hispanic or Latina she was considered Hispanic or Latina regardless of any other designations. If she self-identified as white, black or African American, Asian or Pacific Islander, or American Indian or Alaska Native, she was classified as such. We classified women who did not self-identify in 1 of the 5 groups listed as “other.”
At the time of this study, the CDC required laboratories to use the 1991 Bethesda System categories7 to report Pap test results: normal, infection/reaction, ASC-US, LSIL, high-grade intraepithelial lesion (HSIL), squamous cell cancer, or other. The NBCCEDP recommends that all women with an HSIL or squamous cell cancer Pap result be followed up with colposcopy and a directed biopsy to detect any precancerous or cancerous conditions. Women in this study whose first Pap result of ASC-US or LSIL resulted in an immediate colposcopy (n = 3,101; 23.7%) were not included in the study population. Immediate colposcopy was an option after 1 ASC-US or LSIL result if a women was at high risk, which was defined by such factors as sexual history, smoking status, and infection with human papillomavirus; we were not able to determine risk status, however, because such information is not routinely collected by the programs.
Women who received follow-up with colposcopy after 2 ASC-US or LSIL diagnoses were compared by age and race or ethnicity with those women who received follow-up with a third Pap test or had no follow-up during the 2-year time period. We used the χ2 test to determine statistical differences across the type of follow-up. We also examined the percentage of women whose results on a third Pap test stayed the same (persisted), decreased in severity of dysplasia (regressed), or increased in severity of dysplasia (progressed), as well as the histologic results for colposcopy.
We identified 10,004 women with a low-grade Pap test result (ASC-US or LSIL) followed by a second low-grade result within 2 years during July 1991 to September 2000. Women aged younger than 30 years made up a higher percentage of those with any LSIL result than they did of the overall NBCCEDP population (Table 1). In addition, a higher percentage of women with at least 1 LSIL result were American Indians or Alaska Natives than was true in the NBCCEDP overall. The percentage of women who were followed up in 6 months or less ranged from 54% for women who had LSIL/ASC-US to 72% for LSIL/LSIL.
Of the 10,004 women with 2 abnormal Pap test results, 44% were followed up in accordance with the established guidelines: They underwent colposcopy (Table 2). Those women with an ASC-US result followed by a LSIL were more likely to receive the recommended colposcopy (68.9%) than were women in the other 3 groups, whereas women with an LSIL followed by ASC-US were the least likely to receive colposcopy (34.9%).
Among women with 2 ASC-US results, those aged younger than 30 years were more likely to receive colposcopy than other age groups, whereas women aged 60 years and older were more likely than other age groups to be followed up with a third Pap test (Table 3). In the 3 other Pap test result groups, no significant associations were seen between type of follow-up and age, but the smaller numbers in these other 3 groups reduced the statistical power of these comparisons. Still, some evidence was seen of age differences in these groups; for example, a pattern similar to the finding for ASC-US/ASC-US women was found in the LSIL/ASC-US and ASC-US/LSIL groups, in which women aged younger than 30 years were more likely to receive colposcopy than women aged 60 years or older, a group followed up more frequently with a repeat Pap test.
For each of the 4 result groups, American Indians or Alaska Natives had the highest percentages of a third Pap test, whereas black or African-American women had a higher percentage of no follow-up than other racial or ethnic groups. In all but the ASC-US/LSIL group, where Asians or Pacific Islanders ranked highest for the recommended follow-up by colposcopy, Hispanics or Latinas and white women had the highest rates of follow-up with this procedure.
Of the 3,437 women who received a repeat Pap test after 2 abnormal tests (Table 4), 70% had a normal result (range, 63–78%), indicating a high level of regression. Of the 4,397 women who underwent colposcopy after 2 abnormal tests, 35% had a normal result (range, 24–48%). Women with 2 LSIL results had the highest percentage of CIN 1 (60%). For women with LSIL as the second Pap test result, the percentage with CIN 2 or worse was 15% or 16%, versus 7% or 9% when the second result was ASC-US.
Only 44% of women in the NBCCEDP who had 2 low-grade abnormalities on the Pap test within 2 years were followed up with colposcopy, which the interim guidelines specified. The remaining 56% were followed up with a third Pap test or were not followed up. Women aged younger than 30 years and Hispanics or Latinas were among those most likely to be followed up by colposcopy, but differences were often not significant. Black or African-American women were more likely not to be followed up than other racial or ethnic groups. Women in this study whose second Pap result was LSIL had colposcopy more frequently (63.3%) than those whose second result was ASC-US (38.2%). This more aggressive follow-up of LSIL is encouraging, because 15–16% of women who had a colposcopy after a second LSIL Pap result were diagnosed with CIN 2 or worse, essentially twice the rate of 7–9% when the second Pap was ASC-US.
In 2002, new recommendations were published for managing women with abnormal cervical cytology.3 These new guidelines introduce screening for human papillomavirus and subdivide ASC-US into 2 categories, ASC-US (undetermined significance) and ASC-H (cannot exclude high-grade intraepithelial lesions). According to these new guidelines, women with a single ASC-H or LSIL should be referred for immediate colposcopic evaluation. The relatively low level of appropriate follow-up in our study raises serious questions about future adherence to these new, more demanding guidelines.
In the present study, for each woman receiving a screening examination for cervical cancer in the NBCCEDP, information was collected on demographic characteristics, but additional information on risk factors for cervical carcinoma was not obtained. Thus, one limitation of the present study is the lack of information about risk, which could influence the physician's chosen path of follow-up. In addition, many women apparently did not have any follow-up. Because of eligibility criteria for the program, some women whose income or uninsured status changed may have received subsequent screening services outside of the program, but we have no way of tracking these women to assess their follow-up status.
There is very limited information on patterns of follow-up as viewed against the interim guidelines for 2 low-grade abnormalities on the Pap test, and thus this study provides new information for the literature on cervical cancer from a large population of women. The women served in the NBCCEDP have low incomes and no health insurance, but are cared for in a variety of clinical settings throughout the nation. Although the data are not population based, this program provides important information about the detection of cervical cancer in a traditionally underserved population.
The NBCCEDP has evolved over the past decade, with many changes occurring since its beginning in 1991, including a rise in program grantees from 8 in that year to 70 in 2000. Early on, the primary aim was to screen women; as the program matured, the focus shifted to rescreening women. The NBCCEDP got its start 3 years before the interim guidelines in 1994, and thus adherence to these guidelines in 1991–1995 (the programs would have received the guidelines in 1995) might not be expected. Accordingly we plotted the data for receipt of colposcopy after 2 abnormal Pap tests by year (Figure 1) to see whether inclusion of the earlier years my have affected our results. An increase can be seen in the proportion of women receiving colposcopies until 1995, followed by a decrease and then another increase. This analysis suggests that inclusion of data from 1991 and 1992 may have had some effect on our analysis, but we should note that the number of women participating in those years was very small. In addition, the figure provides no evidence that the guidelines affected the clinical practice.
Our findings have 3 major clinical implications. First, many health care providers no doubt believe that follow-up colposcopy is unwarranted after low-grade findings on the Pap test, thinking that most such findings are transient and will either regress or go away fairly quickly. Our results might give them some support, because 70% of women who had a third Pap test rather than colposcopy regressed to a normal result. On the other hand, 23% had persistence and 7% progression, underscoring the value of proceeding to colposcopy. Second, our finding that CIN 2 or worse was diagnosed in 15–16% of women whose second abnormal Pap test was LSIL reinforces the need for ongoing surveillance of women with abnormal findings to ensure that they receive appropriate follow-up. Finally, because the rate of follow-up with colposcopy was well below 50%, we can speculate that many, perhaps more than one half, of health care providers today are not following the 2002 guidelines. To ensure that providers embrace and apply newly established guidelines for screening and follow-up published by credible professional sources, creative educational interventions may be required, with measurement of use before and after the interventions.