Trends in Colorectal Cancer Test Use among Vulnerable Populations in the United States

Since the mid-1990s, multiple expert groups in the United States have recommended screening asymptomatic average-risk adults for colorectal cancer (CRC; refs. 1–3). In its most recent guidance published in 2008, the U.S. Preventive Services Task Force (USPSTF) strongly recommends screening average-risk adults aged 50 to 75 for CRC with annual high-sensitivity fecal occult blood test (FOBT), sigmoidoscopy every 5 years, or colonoscopy every 10 years (4). Although CRC screening rates in the United States have increased from 20% to 30% in 1997 to nearly 55% in 2008, millions of eligible adults are not screened (5). Moreover, certain population subgroups—including Hispanics, people with low income or low educational attainment, recent immigrants, and those without health insurance, a usual source of care, or recent physician contact—have shown especially low CRC screening rates (6–10). For purposes of this report, we have defined these 7 subgroups as vulnerable populations.

Evaluating trends in the use of CRC screening is important for understanding how screening is being implemented in practice in the United States; to assess whether the population groups targeted for screening are receiving it, consistent with guidelines; and to identify potential problems with underuse, overuse, and misuse of screening (11). Although some prior studies have examined trends in use of CRC screening, their focus has been on the Medicare population, or they have not included more contemporaneous data or comprehensively assessed trends among vulnerable populations (12–18). Following implementation by the Medicare program in 2001 of coverage for CRC screening with colonoscopy for average-risk enrollees, colonoscopy use has increased rapidly (5); of particular interest but relatively unexplored is whether vulnerable populations use colonoscopy at rates comparable with nonvulnerable groups. In the present study, we aim to (i) describe recent national trends in CRC test use overall and among 7 population subgroups that historically have shown lower rates of cancer screening; (ii) assess whether vulnerable subgroups are less likely to receive colonoscopy than their less vulnerable counterparts; and (iii) provide baseline data for assessing trends in CRC test use following implementation of health care reform measures.

We used data from the 2000, 2003, 2005, and 2008 National Health Interview Survey (NHIS) to examine national trends in the overall use of CRC screening and in use of FOBT, sigmoidoscopy, and colonoscopy. The NHIS is a multipurpose health survey sponsored by the National Center for Health Statistics, Centers for Disease Control and Prevention. Conducted annually since 1957, the NHIS is the principal source of information on the health of the civilian, noninstitutionalized household population of the United States. The majority of NHIS interviews are conducted in person by trained personnel from the U.S. Census Bureau. Response rates for the NHIS survey years included in this analysis were 72% (2000), 74% (2003), 69% (2005), and 63% (2008); more information about response rates is available in the appendices for each year's survey description (19–22). Survey responses were weighted to reflect the probability of selection into the sample and survey nonresponse.

Men and women aged 40 or older were asked whether they ever had a sigmoidoscopy, colonoscopy, or proctoscopy, and when they had their most recent examination. They also were asked whether they ever had a blood stool test using a home test kit, and when they had their most recent test. Brief descriptions of each test were provided. Respondents were also asked to report their age, race/ethnicity, educational attainment, household income, time in the United States, type of health care coverage, whether there was a place that they usually went when they were sick or needed advice about their health, and the number of times they saw a physician in the past year. Complete survey instrumentation for the NHIS is available from the National Center for Health Statistics (23).

Consistent with current USPSTF recommendations (4), we defined CRC test use as having had a home FOBT in the past year, sigmoidoscopy in the past 5 years, colonoscopy in the past 10 years, or any of these (up-to-date with CRC screening). We included in this definition tests done for any reason, not just as part of a routine exam, because the reported reason for having the test may not be accurate, and having the test within the recommended time interval, regardless of the reason, essentially means that the individual has been screened.

We used the racial/ethnic designations approved by the federal Office of Management and Budget (24) to categorize respondents as non-Hispanic white, non-Hispanic black, Hispanic, non-Hispanic Asian, and other race/ethnicity.

Educational attainment was based on the highest level of education achieved by the respondent and categorized as more than high school (HS), HS graduate, and not attaining an HS diploma.

Annual income was recorded for each family and reported as a percentage of the federal poverty level (FPL). Families with a computed FPL at or below 100% are considered impoverished. We categorized respondents according to the FPL of their interview year, as follows: <200% FPL, 200% to <300% FPL, 300% to <400% FPL, 400% to <500% FPL, and ≥500% FPL and higher. Multiple imputation was used to impute missing data for NHIS income questions (25).

Time in the United States was measured as born in the United States, immigrated to the United States 10 or more years ago, and immigrated to the United States less than 10 years ago.

We developed separate categories for type of health insurance for respondents aged 50 to 64 and 65 to 75. For the 50 to 64 subgroup, respondents were categorized as having private Health Maintenance Organization (HMO), private non-HMO, public (i.e., Medicaid, Medicare, and other government-sponsored insurance), or no insurance. For the 65 to 75 subgroup, respondents were categorized as having Medicare HMO; Medicare plus private supplemental insurance; Medicare with no supplemental insurance; Medicaid, military, or other government-sponsored insurance; or no insurance or Medicare Part A only.

Respondents who indicated that they had a place where they usually went when they were sick or needed advice about their health were defined as having a usual source of care. Those who either reported that they did not have a place or the place they went to most often was an emergency room (ER) were defined as having no usual source of care.

We used 3 categories to measure the respondent's number of physician visits in the past year: none, 1, and 2 or more.

We used descriptive statistics to characterize overall use of CRC tests and use by specific test modality. Consistent with current USPSTF recommendations (4), we restricted the analysis to adults who were 50 to 75 years of age at the time of the interview. To approximate an average-risk screening population, we excluded individuals who reported a personal history of CRC (n = 320). Respondents who reported being up-to-date for multiple tests were counted as receiving the test when calculating percentages for each test option reported. All percentages reported in the tables and figure were standardized to the 2000 U.S. population by 5-year age groups. The statistical significance of overall trends and each subgroup category trend was tested using logistic regression modeling; separate models were estimated for each test modality and subgroup category. The models included respondent age in 5-year categories to control for changes in the age structure of the U.S. population over time, and survey year for assessment of the statistical significance of the trend. Wald chi-square tests for association of independent variables with each dependent variable (i.e., had home FOBT in the past year versus did not; had sigmoidoscopy in the past 5 years versus did not; had colonoscopy in the past 10 years versus did not; and up-to-date with CRC screening versus not) were computed. Statistical tests were 2-sided, and a P value of 0.05 was considered to be statistically significant.

We conducted a subsidiary analysis among respondents who were up-to-date with CRC screening to assess whether there were subgroup differences in the type of test by which they were up-to-date (i.e., FOBT only, sigmoidoscopy only, colonoscopy only, or up-to-date by multiple-test modalities). We used the 2000 and 2008 NHIS data to calculate percentages and 95% CIs for each type of test and compared them across the 2 time periods to assess trends in test use patterns.

Survey weights were applied in the analyses to reflect the probability of selection into the sample and adjust for nonresponse and poststratification; the weighted data yield national estimates of the prevalence of CRC test use overall, by test modality, and by population subgroup. SUDAAN version 9.0 was used to analyze the data; this statistical package accounts for the complex design of the NHIS and allows for calculation of accurate standard errors, from which 95% CIs around point estimates were derived.

Sample size and U.S. population estimates for each year and by population subgroup are shown in Table 1. The U.S. civilian and noninstitutionalized population aged 50 to 75 grew from an estimated 60 million people in 2000 to nearly 77 million people in 2008. The population subgroups shown in the table also increased in size over this time period, except for people with less than an HS education, with private HMO or Medicare HMO health insurance, or who were aged 65 to 75 and uninsured or had only Medicare Part A coverage. The number of people aged 65 to 75 with Medicare plus private supplemental coverage remained essentially the same over the period 2000 to 2008.

Use of home FOBT among U.S. adults aged 50 to 75 declined during 2000 to 2008 by 6.5 percentage points; in 2008, 11% of adults reported that they had completed a home FOBT in the past year (Table 2). Statistically significant declines ranging from 3.6 to 13.6 percentage points occurred for non-Hispanic whites and non-Hispanic blacks; all education and income groups; people born in the United States; people with private health insurance or Medicare coverage; those reporting a usual source of health care; and those who had seen a physician at least once in the past year.

Similar trends were observed for sigmoidoscopy (Table 3). The proportion of U.S. adults aged 50 to 75 who reported having a sigmoidoscopy in the past 5 years declined by 7 percentage points during 2000 to 2008, to 2.4%. Statistically significant declines ranging from 1.9 to 10.9 percentage points occurred for all population subgroups, with the exception of non-Hispanic Asians, immigrants who had been in the United States less than 10 years, the uninsured, and people without a usual source of health care.

In contrast to the trends for home FOBT and sigmoidoscopy, use of colonoscopy increased markedly among U.S. adults aged 50 to 75: in 2000, 19.0% reported having a colonoscopy in the past 10 years, whereas in 2008, 47.5% indicated that they had undergone this procedure (Table 4). Statistically significant increases in colonoscopy use over the period 2000 to 2008 were seen for all population subgroups, with 1 exception: people aged 65 to 75 who were uninsured or reported having only Medicare Part A coverage. The sample size for this group, however, was small (n < 90).

Although colonoscopy use increased significantly among nearly all population subgroups, there were differences in use rates (Table 4). The percentages of Hispanics reporting colonoscopy use were consistently lower than those of non-Hispanic whites, and the percentage point increase in use among Hispanics was less than that of non-Hispanic whites. Similar patterns were observed for people with less than an HS education (versus those with more than an HS education); recent immigrants (versus individuals born in the United States); people with family income less than 200% of FPL (versus those with family income ≥500% of FPL); the uninsured (versus people with health insurance coverage); those with no usual source of care (versus individuals who reported having a usual source of care); and people who had no physician visits in the past year (versus those with 2 or more visits).

The proportion of the U.S. population aged 50 to 75 that is up-to-date with CRC screening increased by nearly 16 percentage points over the period 2000 to 2008 (Table 5). In 2000, 38.6% were up-to-date; by 2008, nearly 55% reported being up-to-date. Statistically significant increases in the proportion up-to-date ranging from 1.8 to 28.3 percentage points occurred for all population subgroups, with the exception of people aged 65 to 75 who were uninsured or reported having only Medicare Part A coverage.

Although the percentages of people up-to-date with CRC screening increased significantly among nearly all population subgroups, there were subgroup differences (Table 5). The percentages of up-to-date Hispanics were consistently lower than those of non-Hispanic whites, and the percentage point increase in being up-to-date was lower for Hispanics than non-Hispanic whites. Similar patterns were evident for people with less than an HS education (versus those with more than an HS education); people with family income less than 200% of FPL (versus those with family income ≥500% of FPL); individuals aged 50 to 64 who are uninsured (versus people with health insurance coverage); those with no usual source of care (versus individuals who reported having a usual source of care); and people who had no physician visits in the past year (versus those with 1 or more visits). The percentages of recent immigrants who were up-to-date were consistently lower than those of individuals born in the United States, as were the percentages of individuals aged 65 to 75 who have Medicare but no supplemental coverage compared with those covered by Medicare plus private supplemental insurance.

Among those up-to-date with CRC screening, there were few subgroup differences in the type of test they reported using (Fig. 1). In 2000 and in 2008, no significant differences were evident in the percentages of non-Hispanic whites, non-Hispanic blacks, Hispanics, and Asians who were up-to-date by FOBT, sigmoidoscopy, or colonoscopy (Fig. 1A). By education subgroup, there were no significant differences in the percentages up-to-date by FOBT, sigmoidoscopy, or colonoscopy in 2000, although in 2008, people with less than an HS education were more likely than those having more education to be up-to-date by FOBT only (Fig. 1B). For those aged 50 to 64, there were no significant differences in either year in the percentages up-to-date by FOBT, sigmoidoscopy, or colonoscopy by type of health care coverage (Fig. 1C). Similarly, in 2000 and in 2008, there were no significant differences in the percentages of people with or without a usual source of health care who were up-to-date by FOBT, sigmoidoscopy, or colonoscopy (Fig. 1D). Among those up-to-date, for all population subgroups, the most notable trend was the substantial increase in colonoscopy use.

This analysis of nationally representative data for adults aged 50 to 75, the age group targeted by the USPSTF as benefiting the most from CRC screening (4), showed that the proportion up-to-date with CRC screening recommendations increased from 39% in 2000 to 55% in 2008. Moreover, most CRC screening in the United States during 2000 to 2008 was undertaken with colonoscopy, whereas use of FOBT and sigmoidoscopy declined. Despite the 16 percentage point increase in the proportion of U.S. adults up-to-date with CRC screening, though, it is important to note that in 2008, 45% of the population aged 50 to 75—or nearly 35 million people—were not current with screening as recommended by the USPSTF.

A key aim of our study was to examine recent national trends in CRC test use among 7 population groups that historically have had lower rates of cancer screening: Hispanics; people with low income or low educational attainment; recent immigrants; and those without health insurance, a usual source of health care, or recent physician contact. Over the period 2000 to 2008, we observed statistically significant increases in the proportions of people in these vulnerable populations who were up-to-date with CRC screening. The 1 exception was individuals aged 65 to 75 who were uninsured or reported having only Medicare Part A coverage. However, because of small sample sizes for this group, we may have lacked sufficient statistical power to detect a significant increase.

Nevertheless, it is important to highlight that there are still disparities in CRC screening rates. Among the 7 vulnerable populations examined in our analysis, all showed lower rates of colonoscopy use and in being up-to-date with CRC screening, compared with more advantaged populations. For example, in 2008, 39% of Hispanics were up-to-date with CRC screening, compared with 57% of non-Hispanic whites; 32% of Hispanics reported having colonoscopy in the past 10 years, compared with 50% of non-Hispanic whites. Other groups with CRC screening rates lower than 40% in 2008 included people with less than an HS education (38% up-to-date), immigrants who had been in the United States less than 10 years (26% up-to-date), the uninsured (20% up-to-date), people with no physician visits in the past year (18% up-to-date), and those with no usual source of care (16% up-to-date).

Our analysis further revealed that, even among vulnerable subgroups, most CRC screening during 2000 to 2008 was undertaken with colonoscopy, whereas FOBT and sigmoidoscopy use declined. Although concerns have been raised that individuals from vulnerable populations might not have access to colonoscopy equivalent to that of more advantaged individuals (26), we found no significant subgroup differences in test use patterns among individuals who were up-to-date with CRC screening. In other words, in 2008, colonoscopy use predominated among all groups, including racial/ethnic minorities, people with low educational attainment, the uninsured, and those lacking a usual source of care.

Our study has limitations. Data are self-reported; validation studies have shown that adults may over-report screening behaviors in surveys (27). Institutionalized and noncivilian populations are not captured in the NHIS. American Indian, Alaska Native, and Pacific Islander samples were too small to analyze separately and therefore were excluded from the study.

Despite its limitations, our study provides important national benchmark data for monitoring progress toward improved delivery of CRC screening, particularly following the 2010 National Institutes of Health State of the Science Conference on Enhancing Use and Quality of CRC Screening (5, 28)—which outlined research and public health agendas for addressing underuse, overuse, and misuse of CRC screening—and implementation of national health care reform legislation, which includes provisions for reducing or removing barriers to preventive services, including CRC screening (29). Demonstration of improved rates of CRC screening among historically underserved groups will be a critical marker of success for public health efforts that target CRC screening. As our study documents, although CRC screening rates in the United States increased during 2000 to 2008, considerable improvement among all population groups is still needed.

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the National Cancer Institute or the Centers for Disease Control and Prevention.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.