Search tips
Search criteria 


Logo of hsresearchLink to Publisher's site
Health Serv Res. 2012 June; 47(3 Pt 1): 1137–1157.
Published online 2011 December 8. doi:  10.1111/j.1475-6773.2011.01355.x
PMCID: PMC3243108

Colonoscopist and Primary Care Physician Supply and Disparities in Colorectal Cancer Screening



To determine whether racial/ethnic disparities in colonoscopy use vary by physician availability.

Data Source

We used 100 percent Texas Medicare claims data for 2003–2007.

Study Design

We identified beneficiaries aged 66–79 in 2007, examined racial/ethnic differences in colonoscopy use from 2003 to 2007, and estimated the percentage of white, black, and Hispanic beneficiaries who underwent colonoscopy by level of physician availability and area income.

Principal Findings

For the 974,879 beneficiaries, colonoscopy use was higher in whites (40.7 percent) compared to blacks (35.0 percent) and Hispanics (28.7 percent, p < .001). For whites, increasing availability of colonoscopists and primary care physicians (PCPs) was associated with higher colonoscopy use. For blacks and Hispanics, colonoscopy use was unchanged or decreased with increases in colonoscopist and PCP availability. In multilevel models, the odds of colonoscopy were 20 percent lower for blacks (OR 0.80, 95 percent CI 0.79–0.82) and 32 percent lower for Hispanics (OR 0.68, 95 percent CI 0.66–0.69) compared to whites; adjusting for availability of colonoscopists or PCPs did not attenuate racial/ethnic disparities. We found greater racial/ethnic disparities in areas with greater colonoscopist and PCP availability.


Greater area availability of colonoscopists and PCPs is associated with increased use of colonoscopy in whites but decreased use in minorities, resulting in larger racial/ethnic disparities.

Keywords: Colorectal cancer, cancer screening tests, health care disparities, primary care physicians, colonoscopy

In 2010, there were an estimated 142,570 new cases and 51,370 deaths from colorectal cancer, making it the third leading cause of cancer-related deaths in men and women in the United States (Jemal et al. 2010). Blacks have higher incidence and death rates when compared to whites and other racial/ethnic minorities (National Cancer Institute 2006; Jemal et al. 2010).

Routine screening can provide early detection and improved survival for colorectal cancers (Selby et al. 2001; Hardcastle et al. 2010; Kronborg et al. 2010; Mandel et al. 2008, 2006). In March 2008, the American Cancer Society, the U.S. Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (Levin et al. 2010) jointly recommended screening for colorectal cancer beginning at 50 years of age by (1) high-sensitivity fecal occult blood test (FOBT) or fecal immunochemical testing annually, (2) flexible sigmoidoscopy every 5 years, (3) double-contrast barium enema every 5 years, (4) CT colonography (virtual colonoscopy) every 5 years, (5) colonoscopy every 10 years, or (6) fecal DNA at an unspecified interval. The report stated that approaches offering visualization of the colon were preferred to indirect methods.

According to the National Cancer Institute, approximately 1.6 million screening colonoscopies are performed annually (Brown, Klabunde, and Mysliwiec 2001). The increase in colorectal screening from <25 percent in the 1980s to 50–60 percent in 2006 can be attributed almost entirely to an increase in the use of screening colonoscopy (Holden et al. 1996). It is estimated that between 6 and 8 million colonoscopies would need to be performed every year to meet the steady-rate requirements for screening-eligible patients (Vijan et al. 2007).

There are racial disparities in colorectal cancer screening, with non-Hispanic whites being screened more frequently than blacks or Hispanics (Fenton et al. 2010; Maxwell and Crespi 1999; Doubeni et al. 2010). Lower screening rates among non-whites may be attributable to access to or characteristics of primary care physicians (PCPs; such as specialization, background, communication skills, or years into practice), geographic location, limited access to health care resources, reimbursement structures, population knowledge, cultural attitudes, health literacy, and socioeconomic disadvantage (O'Malley et al. 2005; Meissner et al. 2000; Levin et al. 2010).

Physician supply has been associated with increased use of screening tests in various cancers (Roetzheim et al. 2006; Campbell et al. 2008; Coughlin et al. 2003; Gorey et al. 2008). Although it has been shown that colorectal cancer screening is improved in patients who have a PCP (Brouse, Wolf, and Basch 2007), the effect of colonoscopist supply on overall colorectal cancer screening rates and screening with colonoscopy specifically is not fully understood. Several authors have suggested that colonoscopist supply might have an impact on the overall use of colonoscopy (Vijan et al. 2007; Meissner et al. 2000; Haas et al. 2005) and may increase (Meissner et al. 2000) or decrease (Haas et al. 2005) observed racial disparities in colonoscopy rates. Due to uniform coverage of colonoscopies in Medicare beneficiaries, and literature linking differences in access to health resources as a mechanism for racial/ethnic disparities (O'Malley et al. 2005; Meissner et al. 2000; Levin et al. 2010), we hypothesized that in this population an increase in colonoscopist supply would be associated with increased access to colonoscopies for whites, blacks, and Hispanics, which may be associated with changes in racial disparities.

The goal of our study was twofold. First, we examined the association between the availability of colonoscopists and PCPs and receipt of colonoscopy, done for any reason, in Medicare beneficiaries. Second, we described racial/ethnic differences in colonoscopy use and examined the interaction between availability of colonoscopists/PCPs and race/ethnicity in colonoscopy use. We hypothesized that (1) colonoscopy use overall and within racial/ethnic groups would be higher in areas with increasing supply of colonoscopists and PCPs, and (2) that black–white and Hispanic–white differences in colonoscopy use might diminish as the supply of colonoscopists and PCPs increases.


Data Source

We used the claims and enrollment data for 100 percent of Medicare beneficiaries in Texas for the period 2002–2007. Patients’ demographic and enrollment characteristics were obtained from the Denominator File. This file is used to determine beneficiary demographic characteristics, entitlement, and beneficiary participation in Medicare Managed Care Organizations, including beneficiary unique identifier, state and county codes, zip code, date of birth, date of death, sex, race, age, monthly entitlement indicators (Part A/B/Both), and reasons for entitlement. Age was the original reason for entitlement in 92.44 percent of beneficiaries while the remaining 7.56 percent were originally entitled due to disability or end-stage renal disease. In addition, we used the outpatient Standard Analytic File (SAF), which contains claims submitted by institutional outpatient providers, and the Carrier SAF, which contains claims submitted by noninstitutional providers (physicians), to identify outpatient facility services and physician services. In addition, we obtained inpatient hospital admissions and claims data from the Medicare Provider Analysis and Review files. U.S. Census data for the year 2000 provided zip code–level income and population estimates.

Study Cohort

We selected Medicare beneficiaries aged 66–79 in the year 2007 (n = 1,504,783). From this cohort, we excluded beneficiaries who had an interruption in the continuous enrollment in parts A and B Medicare or who enrolled in a health maintenance organization (HMO) during the period 2002–2007 (n = 509,378). Finally, we excluded beneficiaries whose race was not white, black, or Hispanic (n = 20,526), yielding a final cohort of 974,879 beneficiaries.

Study Outcome

The study outcome was receipt of colonoscopy (for any reason) in the previous 5 years, including the year of cohort selection (2003–2007). Colonoscopies were identified by the following Current Procedural Terminology (CPT), Health care Common Procedure Coding System (HCPCS), and International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes (CPT: 44388, 44389, 44392, 44393, 44394, 45378, 45380, 45382, 45383, 45384, 45385; HCPCS: G0105, G0121; ICD-9-CM: 45.23, 45.25, 45.27, 45.41, 45.42, 45.43, 48.36). Beneficiaries with a Medicare claim for these codes during the study period were considered to have had a colonoscopy. Our study does not attempt to evaluate up-to-date screening colonoscopy rates. We evaluated overall colonoscopy use and included all procedures, whether their indication was for screening or for evaluation of any specific problem.

Physician Availability

Physician availability was defined as the number of physicians per 10,000 elders (65 years and older) in the Hospital Service Area (HSA). HSAs are local health care markets for hospital care. The Dartmouth Atlas research group defines HSAs by aggregating Medicare claims records at the zip code level to determine service areas where a hospital or group of hospitals provides a preponderance of care (Wennberg et al. 2004). Physicians who provided services in more than one HSA were considered to be available in each. For HSAs that straddled the border between Texas and an adjoining state, we only counted patients and physicians who were in Texas zip codes.

We created two measures of physician availability: colonoscopist availability and PCP availability. Individual providers were differentiated according to their Unique Provider Identification Number or National Provider Identifier. A colonoscopist was defined as a physician who performed more than five colonoscopies per year, based on the 100 percent Texas Medicare claims data. We chose this cut-point in an attempt to exclude providers who appear in Medicare colonoscopy claims simply due to coding error or who only rarely performed the procedure. Colonoscopists were identified as gastroenterologists (GI colonoscopists) and non-GI colonoscopists, including surgeons, family practitioners and other specialists who performed 5 or more per year. Results were obtained using the overall colonoscopist availability as well as stratified by GI colonoscopists and non-GI colonoscopists, with similar results for both analyses. Since the results did not differ, we report the results for overall colonoscopist supply, not stratified by colonoscopist specialty. A PCP was defined as a general practitioner, family physician, general internist, or geriatrician based on Medicare specialty codes on Part B claims in the Medicare Carrier files. There were a total of 1,264 colonoscopists in Texas in 2003, and this number increased to 1,399 colonoscopists in 2007. For PCPs, the number increased from a total of 8,787 PCPs in Texas in 2003 to 9,306 PCPs in 2007.

Population size and physician counts in an HSA change from year to year. Therefore, physician availability for the study period was estimated using population estimates and physician counts for the years 2003 and 2007. Population estimates of the number of elders in the HSA in 2003 were obtained from the 2000 U.S. Census, and estimates of the number of elders in 2007 were obtained from 2006 population estimates from the Health Resources and Services Administration ( We calculated physician availability in 2003 and in 2007 using the respective physician counts and population estimates for the HSA, weighted these availability estimates by the relative population size in the HSA for that year, and summed the weighted availability measures. Thus, physician availability in the HSA during the study period was defined by the following formula:

equation image


equation image
equation image

Measures of colonoscopist availability and PCP availability were stratified into quartiles for analysis.

Study Variables

We categorized beneficiaries by age, gender, and race/ethnicity using data from Medicare enrollment files. Age was categorized as 66–69, 70–73, 74–76, and 77–79 years. Medicaid eligibility in 2007 (based on the state buy-in variable) was used as a proxy of low socioeconomic status. Median income of the zip code was obtained from 2000 U.S. Census data based on zip code of residence in 2007 and stratified into quartiles. Approximately 5 percent of beneficiaries lived in zip codes missing income information.

Risk of colon cancer was estimated using ICD-9-CM codes obtained from claims in the year before monitoring. Beneficiaries with a family history of gastrointestinal cancer (V16.0), personal history of adenomatous polyps (V12.72), colorectal cancer (V10.05), rectal cancer (V10.06), Crohn's disease (555.0-9), or ulcerative colitis (556.0-3, 556.8-9), noninfectious gastroenteritis and colitis (558.2, 558.9) were classified as high risk. For each patient the presence or absence of each comorbid condition included in the Elixhauser comorbidity index (Elixhauser et al. 2008) was determined using inpatient and outpatient claims. The Elixhauser index includes the following conditions: congestive heart failure, cardiac arrhythmias, valvular disease, pulmonary circulation disorders, peripheral vascular disorders, hypertension, paralysis, other neurologic disorders, chronic pulmonary disease, diabetes uncomplicated, diabetes complicated, hypothyroidism, renal failure, liver disease, peptic ulcer disease excluding bleeding, AIDS/HIV disease, lymphoma, metastatic cancer, solid tumor without metastasis, rheumatoid arthritis/collagen vascular diseases, coagulopathy, obesity, weight loss, fluid and electrolyte disorders, blood loss anemia, deficiency anemias, alcohol abuse, drug abuse, psychoses, and depression.

Statistical Analysis

The percentage of beneficiaries who had a colonoscopy in 2003–2007 was estimated for the total sample and separately for whites, blacks, and Hispanics by physician availability and area income. Chi-square tests were used to determine the statistical significance of differences in colonoscopy use between groups. Cochran-Armitage trend test was used to examine trends in colonoscopy use with increasing physician availability or income.

The effects of race/ethnicity and physician availability on colonoscopy use were estimated by hierarchical generalized linear mixed models with adjustment for patient characteristics and clustering of patients within HSAs. The odds ratios for black race and Hispanic ethnicity indicated the presence of disparities in receipt of colonoscopy. These odds ratios were compared across models before and after adjustment for colonoscopist and PCP availability to determine whether racial/ethnic disparities were attenuated after accounting for physician availability. We were also interested in whether the association between race/ethnicity and colonoscopy use varied by physician availability. Interactions between race/ethnicity and both measures of physician availability were tested and found to be statistically significant. Therefore, we also presented the odds ratios for racial/ethnic disparities separately for each quartile of colonoscopist availability and for each quartile of PCP availability to show how racial/ethnic disparities in colonoscopy use vary by level of physician availability in the HSA.

The follow-up period available for each patient varied by age. For example, beneficiaries aged 70 or older in 2007 had 5 years of claims data available for analysis, while those aged 66 in 2007 had only 1 year of claims data. As a result, differences in the age distributions of blacks, whites, and Hispanics could potentially introduce bias in colonoscopy estimates. Therefore, a sensitivity analysis restricted to patients aged 70–79 years with five continuous years of claims data was performed. The multivariate analyses described in the previous paragraph were repeated using this restricted cohort. Results were substantively equivalent to those for the full cohort in regard to racial differences in receipt of colonoscopy, the effect of physician availability on the odds of receiving a colonoscopy, and the effect of physician availability on racial/ethnic disparities. Thus, we report only the results for the full cohort. A detailed version of the sensitivity analysis is available online (Appendix SA2).

All analyses were performed with SAS version 9.2 (SAS Inc., Cary, NC, USA).


Cohort characteristics are summarized in Table 1. The majority of patients were white (88.9 percent), between 66 and 73 years old (63.4 percent), female (55.6 percent), without high-risk indicators for colorectal cancer (93.4 percent), and with at least one comorbidity (53.6 percent). Table 2 shows the racial/ethnic composition of the HSAs by quartile of provider (colonoscopist and PCP) availability and income. HSAs in the lowest quartile of colonoscopist availability were 88.8 percent white, 5.9 percent black, and 5.3 percent Hispanic, while HSAs in the highest quartile were 91.7 percent white, 5.5 percent black, and 2.8 percent Hispanic. Blacks were somewhat more likely to live in an area with better physician availability while Hispanics were less likely to live in areas with better availability. Lower income zip codes had much higher percentages of blacks (12.0 percent) and Hispanics (8.7 percent) compared to higher income zip codes (3.6 and 1.3 percent, respectively).

Table 1
Characteristics of Medicare Beneficiaries in Texas, 2007
Table 2
Racial/Ethnic Composition of HSA by Quartile of Provider Availability and Income

Table 3 summarizes colonoscopy use based on colonoscopist availability, PCP availability in the HSA, and median income at the zip code level. Information is provided for the overall cohort and is also stratified by race and ethnicity. Overall, 39.8 percent of beneficiaries in the cohort had a colonoscopy between 2003 and 2007, inclusive. Also, 40.7 percent of whites had a colonoscopy during the study period, compared to 35.0 percent of blacks and 28.7 percent of Hispanics (p < .001).

Table 3
Colonoscopy Use in Texas Medicare Beneficiaries by Area-Level Characteristics, 2003–2007

Across racial/ethnic groups, colonoscopy use does not show a consistent pattern with increasing colonoscopist availability. Because of the size of cohort, most comparisons will be statistically significant. Therefore, it is important to focus more on the magnitude of any differences rather than their statistical significance. For whites there was a slight increase in colonoscopy use as colonoscopist availability increased, while for blacks and Hispanics colonoscopy use decreased. A similar pattern was seen between availability of PCPs and colonoscopy. For HSAs in the lowest quartile of PCP availability, colonoscopy use for the overall cohort was 37.3 percent compared with 41.6 percent in the upper quartile (p < .001). This overall trend was accounted for by increased colonoscopy use among white patients, while blacks and Hispanics had no clear pattern in colonoscopy use as PCP availability increased.

There was a strong association between colonoscopy use and zip code–level income for the overall cohort, with higher income quartiles associated with higher colonoscopy use. This pattern is maintained in the white and black subgroups, but there does not appear to be a clear association between income level and colonoscopy use in Hispanics. A positive association between socioeconomic status and colonoscopy use was also evident when receipt of colonoscopy for Medicaid eligible patients was compared with those who were not eligible (30.3 percent and 41.2 percent, respectively, p < .001).

Table 4 presents the results of a multilevel analysis of the association of race/ethnicity with receipt of colonoscopy, independent of individual and area factors. After adjusting for age, sex, comorbidities, presence of risk factors for colon cancer, and income, the odds of colonoscopy were 20 percent lower for blacks (OR = 0.80, 95 percent CI 0.79–0.82) and 32 percent lower for Hispanics (OR 0.68, 95 percent CI 0.66–0.69) compared to whites (model 1).

Table 4
Effect of Physician Availability on the Odds of Colonoscopy in Texas Medicare Beneficiaries by Multilevel Analysis

Model 2 adds colonoscopist availability. Patients living in HSAs within the highest quartile of colonoscopist availability had somewhat higher odds of colonoscopy (OR = 1.09, 95 percent CI = 1.02, 1.18), while availability of PCPs in the HSA had no effect on receipt of colonoscopy (model 3). Neither availability of colonoscopists nor availability of PCPs affected the racial ethnic disparities in receipt of colonoscopies. The odds for blacks versus whites or Hispanics versus whites in receipt of colonoscopy did not vary when availability of colonoscopists or PCPs was entered into the model, either individually or together (models 2, 3, 4 of Table 4 versus model 1).

In the multilevel analyses presented in Table 4, there were significant interactions between either colonoscopist or PCP availability and black versus white or Hispanic versus white odds of receipt of colonoscopy (all p < .001). Therefore, in Table 5 we present the odds of black-white and Hispanic-white differences separately for each quartile of colonoscopist or PCP availability. The black-white and Hispanic-white disparities actually increased as availability of colonoscopists in the HSA increased. For example, blacks living in HSAs in the lowest quartile of colonoscopist availability had 12 percent lower odds of colonoscopy, while blacks in the highest quartile of availability had 23 percent lower odds (Table 5).

Table 5
Effect of Colonoscopist and PCP Availability on the Racial/Ethnic Disparity in Colonoscopy Use by Multilevel Analysis

A similar pattern was seen for quartiles of PCP availability in Hispanics versus white comparisons. Hispanics living in HSA in the highest quartile of PCP availability had 24 percent lower odds of receiving colonoscopy, compared to whites, while Hispanics in the highest quartile of PCP availability had 49 percent lower odds. There were no clear differences in black versus white disparities by quartile of PCP availability (Table 5).


Our study evaluates the influence of colonoscopist and PCP supply on racial disparities in receipt of colonoscopy. We hypothesized that overall colonoscopy use in Medicare beneficiaries would increase with improved colonoscopist and/or PCP capacity and that black-white and Hispanic-white disparities in receipt of colonoscopy would diminish as supply increased. The results of our analyses do not support either of those hypotheses. Consistent with previous studies, we identified significant racial disparities in receipt of colonoscopy (Meissner et al. 2000; Ananthakrishnan et al. 2007; Fenton et al. 2010; Doubeni et al. 2010). While increasing colonoscopist and PCP capacity was associated with increased colonoscopy use in whites, there was a decrease in blacks and Hispanics. As a result, the black-white and Hispanic-white disparities in receipt of colonoscopy actually increased as availability of colonoscopist and PCPs increased.

These data have important public health implications with regard to colorectal cancer screening. Other studies have demonstrated that capacity has little effect on health care access. Zuvekas and Taliaferro (2008), using data from the Medical Expenditure Panel Survey, demonstrated that physician capacity had little effect on access to health care. Furthermore, the increase in colorectal cancer screening rates over the last decade are largely driven by increased rates of colonoscopy (Meissner et al. 2000). Therefore, Meissner and colleagues (2000) suggested that promotion of colonoscopy as the preferred method of screening may actually serve to widen socioeconomic disparities, which is consistent with our findings.

Our data imply that increasing colonoscopist capacity alone may not improve colonoscopy use and may be associated with increased racial/ethnic disparities. This finding takes on more significance given that our study population was uniformly covered by Medicare Parts A and B. Previous studies have demonstrated disparities in the use of screening by educational attainment, income, health insurance coverage, receipt of routine medical care, characteristics of minority PCPs, geographic location, access to health care resources, population knowledge, cultural attitudes, health literacy, and socioeconomic disadvantage (Breen et al. 2007; Zuvekas and Taliaferro 2008; Seeff et al. 1999; Meissner et al. 2000). These barriers may be even more difficult to overcome if we continue to move toward colonoscopy as the preferred screening method (Meissner et al. 2000).

It is unclear why increased provider availability was associated with wider racial/ethnic disparities; however, it is likely that socioeconomic factors played a role. There is evidence that socioeconomically disadvantaged populations are slower to adopt complex medical innovations than more advantaged populations (Goldman and Lakdawalla 2009; Glied and Lleras-Muney 1998). Colonoscopy is invasive and time consuming, and involves a difficult preparation. In addition, Medicare coverage of screening colonoscopy for average-risk patients was only implemented in 2001, and even with coverage, colonoscopy is a costly procedure. Therefore, this phenomenon may be responsible for widening racial/ethnic disparities in the setting of increased provider availability, if more advantaged groups are more likely to follow through with this complex and costly screening procedure.

Another possibility is that the different associations between provider availability and colonoscopy use among whites, blacks, and Hispanics operated through individual-level income and income inequality within an HSA. Our analyses adjusted for a crude measure of contextual socioeconomic status, namely zip code–level income; however, residual confounding from individual-level differences in socioeconomic status is likely. Minority and white beneficiaries who reside in the same HSA may have very different socioeconomic status and live in neighborhoods and communities with very different income or education levels. Descriptive results from our study confirmed that lower income zip codes had much higher percentages of blacks and Hispanics compared to higher income zip codes. Minorities living in HSAs with higher provider availability may still lack access to the providers because they reside in poorer neighborhoods or communities or because of individual-level socioeconomic disadvantage. The increasing disparities noted with increasing colonoscopist availability may be driven by the relationship between income and colonoscopy use. We evaluated a three-way interaction between income, race, and colonoscopy availability. Within income quartiles, the pattern of increasing colonoscopist availability and increasing disparities is not as apparent while within provider availability quartiles there is a very apparent pattern of increasing disparities with increasing income. However, the interaction between race and availability remained significant after including the income/availability interaction. In addition, we also performed an HSA fixed effects model controlling for income, and the racial/ethnic differences persisted.

In contrast to our study, a recent study by Haas et al. (2005) reported a reduction in racial disparities for colorectal cancer screening with increasing colonoscopy capacity. Similar to our study they reported no association between PCP supply and colorectal cancer screening. There are methodological differences between the Haas study and our own that may explain the different results. Their study used data from the National Health Interview Survey, using a positive response to the question “Have you ever been screened for colorectal cancer, using FOBT, sigmoidoscopy, or colonoscopy?” as having been screened. In addition, they define capacity based on colonoscopies performed rather than providers. They use Medicare data based on the number of colonoscopies and sigmoidoscopies performed in the county in Medicare beneficiaries who are, by definition, 65 years or older. They then established a ratio per 100,000 residents aged 50 and older, including the subgroup of individuals between 50 and 64 years old, who were not counted in the total number of colonoscopies/sigmoidoscopies. Using their definition, the screening capacity for a county will vary based on the age (50–64 versus 65+) and racial distribution of the county. In counties with higher proportions of the older population who are aged 65+, there will be more colonoscopies recorded (because colonoscopies are only measured for adults older than 65). Counties with very different actual Medicare screening rates could appear to have similar rates depending on the age distribution of the counties.

In contrast, we defined colonoscopist and PCP availability according to the number of physicians per 10,000 adults aged 65 and older in the HSA. These methodological differences may explain the difference in results. We included non-GI and GI colonoscopists in our final results, but the analysis was performed for GI colonoscopists only and the results were identical. Evaluating the two studies together, it may be that racial/ethnic disparities decrease with increasing numbers of colonoscopies (i.e., overall utilization rate) but not by the number of colonoscopists and PCPs (i.e., provider capacity).

There is a large body of evidence that suggests that patients with PCPs are more likely to receive colonoscopy and other important screening tests (Roetzheim et al. 2006; Selvin and Brett 2004; Liang et al. 1996; Meissner et al. 2000; Brouse, Wolf, and Basch 2007; Cardarelli, Kurian, and Pandya 2003). Our data support previous findings, demonstrating that increasing PCP supply does not increase colonoscopy use (Haas et al. 2005). Despite increasing PCP capacity, racial/ethnic minorities may not receive routine screening tests. In addition, when they do receive usual care, minority patients have been shown to be less likely to see PCPs who routinely perform screening (Bao, Fox, and Escarce 2007).

Our study has several limitations. We evaluated overall colonoscopy use without differentiating between screening and diagnostic colonoscopies. Medicare beneficiaries who had a colonoscopy before their enrollment in Medicare at 65 years old would be classified as not having received a colonoscopy in our analysis. Therefore, colonoscopy use for blacks might be under-reported as they have a higher proportion of beneficiaries in the younger age strata. However, sensitivity analyses conducted among patients aged 70–79 with 5 years of follow-up found similar results to those reported here for patients aged 66–79. Our estimates are not meant to represent overall “up-to-date” rates for colon cancer screening, which would require a longer time line for data acquisitions, as well as information on receipt of screening modalities such as sigmoidoscopy. Therefore, we focused on differences in colonoscopy use. In addition, we did not consider the number of colonoscopies performed by each colonoscopist, which may contribute to the understanding of the observed trends. Also, Hispanics prefer to receive colorectal screening tests other than colonoscopy when compared to other racial/ethnic groups, which might decrease the total number of colonoscopies performed in this group (Sheikh et al. 2003; Meissner et al. 2000; Shokar, Carlson, and Weller 2011, 2004). Our cohort is restricted to Medicare beneficiaries older than 65 years old in the state of Texas only, and results might not be applicable to a national population or areas with younger populations. Texas is a state with recognized geographic variation in both utilization of health resources and health care costs. Moreover, the racial/ethnic composition of Texas is different than the national population, with 37.6 percent of Texas residents identifying as Hispanic according to 2010 U.S. Census data. There is evidence that disparities in colorectal cancer screening between blacks and whites (but not Hispanics and whites) vary across U.S. regions (Semrad et al. 1992). Disparities were driven by variation in screening rates among blacks, whereas whites had similar screening rates across regions. These results suggest that the underlying determinants of disparities, such as colonoscopist capacity or screening barriers, may differ regionally (Semrad et al. 1992). We included Medicaid eligibility as a crude measure of low socioeconomic status; however, a recent study showed that the state buy-in variable in the Medicare data does not adequately capture Medicaid eligibility (Koroukian et al. 2010). Finally, study findings were based on the cross-sectional observation of correlations between provider availability and colonoscopy use; therefore, it is not possible to make causal inferences or establish the direction of the association.

Our study has implications for public policy. An increased supply of providers may have little beneficial effect on ethnic and racial disparities in receipt of colonoscopy, as shown in this population of Texas Medicare beneficiaries. Indeed, there is the potential for disparities to widen. As more colonoscopists become available to satisfy the predicted increase in demand for screening colonoscopies in the future, further studies need to be performed to establish better allocation of colonoscopists. Finally, further explanation of the etiology of the increase in racial disparities associated with increased colonoscopist supply is required to avoid further increases in disparities as capacity expands to satisfy current needs.


Joint Acknowledgment/Disclosure Statement: This study is supported by following grants: NIH R01CA134275; K05CA134923; K07CA130983; and Cancer Prevention Research Institute of Texas (CPRIT) RP101207. The study uses 100 percent Texas Medicare data for which we have obtained a data use agreement. We are not required to get approval from the Center for Medicare and Medicaid Services prior to publication, and the manuscript has not been reviewed by CMS.

Disclosures: None.

Disclimer: None.


Additional supporting information may be found in the online version of this article:

Appendix SA1: Author Matrix.

Appendix SA2: Sensitivity Analysis Comparing Results between the Full Cohort and Patients Aged 70–79 Years.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.


  • Ananthakrishnan AN, Schellhase KG, Sparapani RA, Laud PW, Neuner JM. Disparities in Colon Cancer Screening in the Medicare Population. Archives of Internal Medicine. 167(3):258–64. [PubMed]
  • Bao Y, Fox SA, Escarce JJ. Socioeconomic and Racial/Ethnic Differences in the Discussion of Cancer Screening: “Between-” versus “Within-” Physician Differences. Health Services Research. 42(3 Pt 1):950–70. [PMC free article] [PubMed]
  • Breen N, Wagener D, Brown M, Davis W, Ballard-Barbash R. Progress in Cancer Screening over a Decade: Results of Cancer Screening from the 1987, 1992 and 1998 National Health Interview Surveys. Journal of the National Cancer Institute. 2007;93(22):1704–13. [PubMed]
  • Brouse CH, Wolf RL, Basch CE. Facilitating Factors for Colorectal Cancer Screening. Journal of Cancer Education. 2007;23(1):26–31. [PubMed]
  • Brown ML, Klabunde CN, Mysliwiec P. Current Capacity for Endoscopic Colorectal Cancer Screening in the United States: Data from the National Cancer Institute Survey of Colorectal Cancer Screening Practices. American Journal of Medicine. 2001;115(2):129–33. [PubMed]
  • Campbell RJ, Ramirez AM, Perez K, Roetzheim RG. Cervical Cancer Rates and the Supply of Primary Care Physicians in Florida. Family Medicine. 2008;35(1):60–4. [PubMed]
  • Cardarelli R, Kurian AK, Pandya V. Having a Personal Healthcare Provider and Receipt of Adequate Cervical and Breast Cancer Screening. Journal of the American Board of Family Medicine. 2003;23(1):75–81. [PubMed]
  • Coughlin SS, Leadbetter S, Richards T, Sabatino SA. Contextual Analysis of Breast and Cervical Cancer Screening and Factors Associated with Health Care Access among United States Women, 2002. Social Science & Medicine. 2003;66(2):260–75. [PubMed]
  • Doubeni C, Laiyemo A, Klabunde C, Young A, Field T. Racial and Ethnic Trends of Colorectal Cancer Screening among Medicare Enrollees. American Journal of Preventive Medicine. 2010;38(2):184–91. [PMC free article] [PubMed]
  • Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity Measures for Use with Administrative Data. Medical Care. 2008;36(1):8–27. [PubMed]
  • Fenton J, Tancredi D, Green P, Franks P, Baldwin L. Persistent Racial and Ethnic Disparities in Up-To-Date Colorectal Cancer Testing in Medicare Enrollees. Journal of the American Geriatrics Society. 2010;57(3):412–8. [PubMed]
  • Glied S, Lleras-Muney A. Technological Innovation and Inequality in Health. Demography. 1998;45(3):741–61. [PMC free article] [PubMed]
  • Goldman DP, Lakdawalla DN. A Theory of Health Disparities and Medical Technology. Contributions to Health Economics and Policy. 2009;4(1):1–30.
  • Gorey KM, Luginaah IN, Fung KY, Bartfay E, Hamm C, Wright FC, Balagurusamy M, Holowaty EJ. Physician Supply and Breast Cancer Survival. Journal of the American Board of Family Medicine. 2008;23(1):104–8. [PMC free article] [PubMed]
  • Haas JS, Brawarsky P, Iyer A, Fitzmaurice GM, Neville BA, Earle C, Kaplan CP. Association of Local Capacity for Endoscopy with Individual Use of Colorectal Cancer Screening and Stage at Diagnosis. Cancer. 2005;116(12):2922–31. [PMC free article] [PubMed]
  • Hardcastle J, Chamberlain J, Robinson MH, Moss SM, Amar SS, Balfour TW, James PD, Mangham CM. Randomised Controlled Trial of Faecal-Occult-Blood Screening for Colorectal Cancer. Lancet. 2010;348(9040):1472–7. [PubMed]
  • Holden D, Harris R, Porterfield D, Jonas D, Morgan L. Enhancing the Use and Quality of Colorectal Cancer Screening. Rockville, MD: Agency for Healthcare Research and Quality; 1996. Report No.: 10-E002.
  • Jemal A, Siegel R, Xu J, Ward E. Cancer Statistics, 2010. CA: A Cancer Journal for Clinicians. 2010;60(5):277–300. [PubMed]
  • Koroukian SM, Dahman B, Copeland G, Bradley CJ. The Utility of the State buy-in Variable in the Medicare Denominator File to Identify Dually Eligible Medicare-Medicaid Beneficiaries: A Validation Study. Health Services Research. 2010;45(1):265–82. [PMC free article] [PubMed]
  • Kronborg O, Fenger C, Olsen J, Jorgensen O, Sondergaard O. Randomised Study of Screening for Colorectal Cancer with Faecal-Occult-Blood Test. Lancet. 2010;348(9048):1467–71. [PubMed]
  • Levin B, Lieberman D, McFarland B, Smith R, Brooks D, Andrews KS, Dash C, Giardiello FM, Glick S, Levin TR, Pickhardt P, Rex DK, Thorson A, Winawer SJ. the American Cancer Society Colorectal Cancer Advisory Group, the US Multi-Society Task Force, and the American College of Radiology Colon Cancer Committee. Screening and Surveillance for the Early Detection of Colorectal Cancer and Adenomatous Polyps, 2008: A Joint Guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA: A Cancer Journal for Clinicians. 2010;58(3):130–60. [PubMed]
  • Liang SY, Phillips KA, Nagamine M, Ladabaum U, Haas JS. Rates and Predictors of Colorectal Cancer Screening. Preventing Chronic Disease: Public Health Research, Practice, and Policy. 1996;3(4):A117. [PMC free article] [PubMed]
  • Mandel J, Church T, Ederer F, Bond J. Colorectal Cancer Mortality: Effectiveness of Biennial Screening for Fecal Occult Blood. Journal of the National Cancer Institute. 2008;91(5):434–7. [PubMed]
  • Mandel J, Church T, Bond JH, Ederer F, Geisser MS, Mongin SJ, Snover DC, Schuman LM. The Effect of Fecal Occult Blood Screening on the Incidence of Colorectal Cancer. New England Journal of Medicine. 2006;343(22):1603–7. [PubMed]
  • Maxwell AE, Crespi CM. Trends in Colorectal Cancer Screening Utilization among Ethnic Groups in California: Are We Closing the Gap? Cancer Epidemiology, Biomarkers & Prevention. 1999;18(3):752–9. [PMC free article] [PubMed]
  • Meissner H, Breen N, Klabunde C, Vernon S. Patterns of Colorectal Cancer Screening Uptake among Men and Women in the United States. Cancer Epidemiology, Biomarkers & Prevention. 2000;15(2):389–94. [PubMed]
  • National Cancer Institute. 2006. SEER Cancer Statistics Review 1975-2006 [accessed on August 20, 2010]. Available at:
  • O'Malley A, Forrest C, Feng S, Mandelblatt J. Disparities Despite Coverage: Gaps in Colorectal Cancer Screening among Medicare Beneficiaries. Archives of Internal Medicine. 2005;165(18):2129–35. [PubMed]
  • Roetzheim RG, Pal N, Gonzalez EC, Ferrante JM, Van Durme DJ, Ayanian JZ, Krischer JP. The Effects of Physician Supply on the Early Detection of Colorectal Cancer. Journal of Family Practice. 2006;48(11):850–8. [PubMed]
  • Roetzheim RG, Gonzalez EC, Ramirez A, Campbell R, van Durme DJ. Primary Care Physician Supply and Colorectal Cancer. Journal of Family Practice. 2006;50(12):1027–31. [PubMed]
  • Seeff LC, Nadel MR, Klabunde CN, Thompson T, Shapiro JA, Vernon SW, Coates RJ. Patterns and Predictors of Colorectal Cancer Test Use in the Adult U.S. Population. Cancer. 1999;100(10):2093–103. [PubMed]
  • Selby J, Friedman G, Quesenberry C, Weiss N. A Case-Control Study of Screening Sigmoidoscopy and Mortality from Colorectal Cancer. New England Journal of Medicine. 2001;326(10):653–7. [PubMed]
  • Selvin E, Brett KM. Breast and Cervical Cancer Screening: Sociodemographic Predictors among White, Black, and Hispanic Women. American Journal of Public Health. 2004;93(4):618–23. [PubMed]
  • Semrad TJ, Tancredi DJ, Baldwin LM, Green P, Fenton JJ. Geographic Variation of Racial/Ethnic Disparities in Colorectal Cancer Testing among Medicare Enrollees. Cancer. 1992;117(8):1755–63. [PubMed]
  • Sheikh RA, Kapre S, Calof OM, Ward C, Raina A. Screening Preferences for Colorectal Cancer: A Patient Demographic Study. Southern Medical Journal. 2003;97(3):224–30. [PubMed]
  • Shokar NK, Carlson CA, Weller SC. Prevalence of Colorectal Cancer Testing and Screening in a Multiethnic Primary Care Population. Journal of Community Health. 2011;32(5):311–23. [PubMed]
  • Shokar NK, Carlson CA, Weller SC. Factors Associated with Racial/Ethnic Differences in Colorectal Cancer Screening. Journal of the American Board of Family Medicine. 2004;21(5):414–26. [PubMed]
  • Vijan S, Inadomi J, Hayward RA, Hofer TP, Fendrick AM. Projections of Demand and Capacity for Colonoscopy Related to Increasing Rates of Colorectal Cancer Screening in the United States. Pharmacology and Therapeutics. 2007;20(5):507–15. [PubMed]
  • Wennberg J, Fisher E, Goodman D, Skinner J. Tracking the Care of Patients with Severe Chronic Illness: The Dartmouth Atlas of Health Care 2008. Lebanon, NH: Bonner; 2004.
  • Zuvekas SH, Taliaferro GS. Pathways to Access: Health Insurance, the Health Care Delivery System, and Racial/Ethnic Disparities, 1996-1999. Health Affairs (Millwood) 2003;22(2):139–53. [PubMed]

Articles from Health Services Research are provided here courtesy of Health Research & Educational Trust