Abstract
Purpose:
The Medicare Access and CHIP Reauthorization Act (MACRA) incentivized primary care practices to improve colorectal cancer screening rates. This study examined if colorectal screening rates improved among rural and urban primary care practices amid implementation of MACRA.
Methods:
Colorectal cancer screening data are from a national registry of 139 primary care practices. Repeated measures regression tested for rural/urban differences and changes in screening rates between 2016 and 2020, adjusting for county demographic factors and social deprivation.
Results:
Screening rates were 64% in both rural and urban practices in the first quarter of 2016 and increased to 80% and 83% in rural and urban practices, respectively, in the last quarter of 2020. In adjusted analyses, screening rates increased by 4% per year and there were no rural/urban differences. Lower screening rates were associated with higher county proportions of persons who were 45 to 74 years of age and Hispanic. Higher screening rates were associated with higher county proportions of persons who were White, Black, and Asian and higher social deprivation.
Conclusions:
Colorectal screening rates improved among rural and urban primary care practices during implementation of MACRA, but disparities persist among practices serving county populations that are relatively older, more Hispanic, and have higher social deprivation.
Introduction
Colorectal cancer accounts for an estimated 8% of new cancer cases and 9% of cancer deaths. 1 A study based on 2009-2013 data from the North American Association of Central Cancer Registries (NAACCR) found that colorectal cancer incidence rates were significantly higher in non-metropolitan (rural) than metropolitan (urban) counties (43.9 vs 40.1 per 100 000). 2 Screenings are highly effective for detecting colorectal cancer,3,4 but few contemporary studies have investigated potential rural versus urban inequities in colorectal cancer screening. Findings from the 2005 Behavioral Risk Factor Surveillance System (BRFSS) showed that residents of rural areas had lower adjusted rates of colorectal screening compared to residents of urban areas (48%vs 54%). 5 However, that study is based on data collected prior to Medicaid and private insurance expansions occurring after the Affordable Care Act (ACA). More recent estimates from the 2018 BRFSS found that colorectal screening rates were lower among non-metropolitan than metropolitan residents across age categories of 50-75, 50-64, and 65-75. 6 In addition, analyses of the National Health Interview Survey (NHIS) indicated that colorectal cancer screening rates rose at a faster pace between 2008 and 2018 among metropolitan (58.1%-66%) than non-metropolitan residents (48.5%-60.6%). 7
The Medicare Access and CHIP Reauthorization Act (MACRA) of 2015 may have contributed to increases in colorectal cancer screening rates, and attenuated rural/urban disparities, as MACRA established financial incentives for health care professionals to provide high-quality health services.8,9 The MACRA stipulates that health care professionals who receive outpatient Medicare reimbursement must participate in a Quality Payment Program.8-10 Health care professionals satisfy the Quality Payment Program requirement by participating in the Merit-Based Incentive Payment System (MIPS), unless they are eligible and decide to enroll in the Advanced Alternative Payment Model (APM) or are not required to participate because they have a low volume of patients with Medicare.11,12 Under MIPS, health care professionals have to choose to report data for a minimum of 6 clinical quality performance measures, which Medicare then uses to partially modify reimbursement. 12 One clinical quality indicator that is particularly relevant to primary care is the colorectal cancer screening rate, as primary care professionals (PCPs) routinely manage their patients’ colorectal cancer prevention screenings either through direct testing (eg, fecal occult blood testing, or FOBT) or, more frequently, referring patients for a colonoscopy/sigmoidoscopy. 13
In addition to rural/urban disparities, prior research points to racial and ethnic inequities in colorectal cancer screening rates.14-17 Two studies based on analyses of the 2008 BRFSS showed that persons who were Black/African American were more likely to receive colorectal cancer screening than person who were White,14,15 but one of these studies showed that the association between Black/African American race and screening was confined to those residing in an urban area. 14 Other research has found that persons who are of Hispanic ethnicity are less likely to receive colorectal cancer screenings than persons who are White or Black/African American.6,16 Whether racial or ethnic variations exist within primary care practices located across rural and urban counties has not been investigated.
Within the context of managing colorectal cancer screening among their patient populations, PCPs may also need to consider community-level social and economic factors that influence colorectal cancer screenings. One study found that a greater percentage of the population living below the federal poverty level was associated with greater odds of never receiving colorectal cancer screening. 18 In a study based on analyses of metropolitan respondents to the 2002 national BRFSS, area-level poverty rates were associated with higher odds of not receiving colorectal cancer screening. 18 Other analyses of the 2006 BRFSS in Missouri found that a higher area-level poverty rate was associated with lower odds of receiving a colorectal cancer screening. 19 A more recent study based on data from primary care practices in 3 midwestern states found that worse area deprivation (as based on the area deprivation index derived from the American Community Survey) was associated with lower odds of patients receiving recommended colorectal cancer screening. 20 Even when adjusting for area deprivation, patients residing in rural areas had lower odds of colorectal cancer screening than those residing in urban areas. 20
This paper provides new insight into the degree to which rural and urban primary care practices manage their patients’ colorectal cancer screenings amid implementation of MACRA within the context of area-level racial/ethnic, social, and economic factors. Our specific hypotheses are that colorectal cancer screening rates are lower among primary care practices that are (1) located in rural relative to urban counties, (2) in counties with higher percentages of the population that are racial or ethnic minorities, and (3) in counties with greater social deprivation.
Methods
Overview
We assembled and analyzed longitudinal, practice-level colorectal cancer screening data from the American Board of Family Medicine (ABFM) PRIME Registry coupled with county-level contextual indicators of social deprivation and demographic characteristics. To test for changes within practices across time, we restricted the sample to practices that had complete colorectal screening data for each quarter of the year from 2016, or the year prior to MACRA implementation, to the last quarter of 2020. A total of 139 practices had complete quarterly practice-level screening performance data for the study period. The project received IRB approval.
Data Sources
The PRIME Registry 21 is a Centers for Medicare and Medicaid Services (CMS)-certified Qualified Clinical Data Registry open to all primary care clinicians and captures electronic health record (EHR) data from more than 2500 clinicians in approximately 800 practices in 47 states caring for 5.4 million patients. PRIME practices are disproportionately rural (39%), small (average less than 3 clinicians), and independently owned, compared to all U.S. primary care practices. Less than half (46.2%) of the patients are White, 7.3% are Black, 20.5% are Hispanic, 0.4% are American Indian/Alaska Native, and 0.2% are Native Hawaiian or Pacific Islander (the remaining balance is a mixture of multiple, other, or not specified). Additionally, PRIME includes patients of all ages with at least 40% of patients meeting age requirements for colorectal cancer screening. We combined PRIME data with county-level demographic information from the 2019 to 2020 Health Resources and Services Administration (HRSA) Area Health Resources File (AHRF) and the social deprivation index from the Graham Center.
Dependent Variable
The dependent variable is the proportion of patients in a practice that received guideline recommended colorectal screenings (ie, the percentage of adults 50-75 years of age who had appropriate screenings for colorectal cancer, or CMS quality indicator #113).
This measure was developed by the National Committee for Quality Assurance (NCQA) and is used by CMS (see https://qpp.cms.gov/docs/QPP_quality_measure_specifications/Claims-Registry-Measures/2018_Measure_113_Registry.pdf for details). Briefly, the denominator is the number of patients who were 50 to 75 years of age on the date of a visit and during the period of assessment. Patients who had colorectal cancer or a total colectomy were excluded. The numerator is the number of patients with at least one colorectal screening over the period of assessment (the screening could be provided by the primary care physician or a specialist, such as a gastroenterologist). Colorectal screenings are defined as fecal occult blood test (FOBT) during the period of assessment, flexible sigmoidoscopy during the period of assessment or the prior 4 years, a colonoscopy during the period of assessment or prior 9 years, a computed tomography (CT) colonography during the period of assessment or 4 prior years, or a fecal immunochemical DNA test (FIT-DNA) during the period of assessment or prior 4 years.
Independent Variables
The main independent variable is rurality of the practice location, which is defined according to U.S. Office of Management and Budget (OMB) classifications of metropolitan (urban) and non-metropolitan (rural) counties. A metropolitan county has an urban core with more than 50 000 persons, and all other counties are considered non-metropolitan. 22 We also included the practice’s Census Region, as prior research has indicated colorectal cancer screening variations across states 6 and regions. 14 Because we did not have information on the demographic characteristics of patients in a primary care practice, we included from the AHRF 2010 Census estimates (2010 Census numbers are the most recent age distribution estimates in the AHRF) of the proportion of the county population that was 45 to 74 years of age as this reflects potential demand for colorectal cancer screening (we note that that the AHRF does not report the proportion with a cut off 50 years of age and thus we used 45 years of age as the cut off). We also obtained from the AHRF 2018 estimates (2018 represents the “center” of our sample) of the proportions of the county population that were White, Black, Asian, and Hispanic because prior research has indicated racial and ethnic variations in colorectal screening rates.14-17 Gender, race, and ethnicity were derived using 2018 Census estimates as that year is the “center” of our sample.
County-level social deprivation was assessed using the Social Deprivation Index (SDI). 23 The SDI is based on the 7 factors contained in the American Community Survey (ACS): the percentages of population living in poverty, with less than 12 years of education, residing in a single parent household, residing in rental housing, residing in overcrowded housing, and without an automobile and the percentage of the population under 65 years of age that is unemployed.
We created a discrete numeric variable reflecting years since 2016, thereby testing for changes in colorectal cancer screening rates in the year leading up to the implementation of MACRA and the years following its implementation on January 1, 2017. We also included dummy variables for each quarter within each year.
Analysis
We first descriptively compared colorectal screening rates across rural and urban practices and time (quarters within years). Next, we conducted repeated measures regression analyses using SAS Proc Mixed to account for correlations in the repeated quarterly colorectal cancer screening rates within practices. We tested for an interaction between rurality and year to determine if changes in colorectal cancer screening rates differed by rurality, but the interaction term was insignificant and therefore removed from the final model.
Results
Table 1 shows the means and standard deviations (SD) for colorectal screening rates among all practices and by rural/urban location for each quarter in 2016 through 2020. Figure 1 graphically illustrates changes in colorectal screening rates among rural and urban practices. Colorectal screening rates were approximately 64% in rural and urban practices in the first quarter of 2016. Colorectal screening rates increased to approximately 80% and 83% in rural and urban practices, respectively, in the last quarter of 2020.
Colorectal Screening Performance Rates by Rural/Urban Practice and Time.
Colorectal cancer screening rates by rural/urban practice and time.
Table 2 shows findings from the repeated measures regression analysis. When adjusting for other factors, a lack of a rural/urban difference in screening rates persisted. Colorectal screening rates increased by approximately 4% per year, as indicated by the estimate of 4.22 for the years since 2016. The quarter within year was not associated with screening rates. A higher SDI score, indicating worse social deprivation, was associated with a slightly higher screening rate. Higher percentages of the population aged 45 to 74 years of age and Hispanic were associated with lower screening rates, whereas greater percentages of White, Black, and Asian were associated with higher screening rates. Practices located in the Northeast U.S. had higher screening rates and those in the Midwest and West had lower screening rates relative to practices in the South.
Repeated Measures Regression of Colorectal Screening Rates.
Discussion
This study yielded new information about rural and urban primary care practices’ management of their patients’ colorectal cancer screenings during implementation of MACRA. Specifically, our findings suggest that primary care practices in rural and urban counties improved their colorectal cancer screening performance approximately 4% per year since 2016, or 1 year prior to the MACRA implementation. We did not detect any differences between practices located in rural and urban counties, suggesting that rural/urban disparities reported in prior literature may have attenuated in recent years. Our study’s findings regarding rural and urban location could also differ from prior research as we used practice-level data from a national primary care quality improvement registry, as opposed to individual-level data from population-based surveys.5-7
Our hypothesis that primary care practices in counties with greater (worse) social deprivation would have lower screening rates was also not confirmed. Rather, in contrast to prior research indicating that greater area deprivation and poverty are associated with lower odds of colorectal cancer screenings,18-20 we found that greater area-level social deprivation was associated with higher screening rates. One explanation for these divergent findings is that the current study relied on practice-level screening rates, whereas the aforementioned studies had information about individual persons’18,19 or patients’ 20 screening histories. Another explanation for the current study’s finding is that counties with worse social deprivation have better primary care resources, which then contribute to higher screening rates. Related to social and economic resources, some patients may fall into a health insurance gap or “doughnut hole” in which they do not qualify for public insurance coverage and are unable to purchase private health insurance. However, our study’s use of practice level data precluded our ability to examine how individual patients’ insurance coverage may have influenced their receipt of colorectal cancer screenings.
Our hypothesis that practices located in counties with higher percentages of the population who are racial or ethnic minorities would have lower screening rates was confirmed, at least in regard to Hispanic ethnicity. This finding is consistent with other studies that have found that persons who are Hispanic are less likely to receive colorectal cancer screenings than persons who are White or Black/African American.6,16 In contrast, higher percentages of the population that were white, Black, and Asian were positively associated with screening rates. One other area-level demographic characteristics, the proportion of the population 45 to 74 years of age and male, was negatively associated with screening rates. We posit that primary care practices located in counties with higher proportions of persons ages 45 to 74 years may not be able to keep up with demand for colorectal cancer screenings. Prior research also indicates that individuals who are male have a higher likelihood of colorectal cancer screening than those who are female. 24 This could similarly translate into greater unmet demand for colorectal cancer screenings in counties with higher proportions of males.
Lastly, like prior research making use of individual survey responses to the 2008 BRFSS, we found regional variations in colorectal screening rates. 14 Because colorectal cancer mortality rates are particularly high in the southern U.S., 25 one might conclude that colorectal cancer screening rates would be lower in the same region. However, our findings suggest that primary care practices located in the Midwest and West have lower screening rates than those in the South. One potential explanation for the latter finding is that public health attention has focused on increasing screening in the South. Other regions may now need to further promote or colorectal screenings.
The current study focused on primary care practices in rural and urban areas, but the findings may differ for other physician specialties. Family physicians may deliver endoscopy services, but prior research showed that fewer than 5% actually do so and that the percentages declined between 2014 and 2016. 26 In a study based on 2006 and 2007 Medicare Beneficiary Survey and Medicare claims, residents of metropolitan areas whose usual physician was a general internist had higher odds of colorectal screening than those whose usual physician was a family physician. 27 However, there was no association between physician specialty and colorectal screening among non-metropolitan residents. 27 Future research should further explore colorectal screening performance by usual source of care and primary care clinicians, including nurse practitioners and physician assistants.
The U.S. Preventive Services Task Force (USPSTF) recently revised its colorectal cancer screening recommendations in 2021 to include persons 45 to 49 years of age, but it graded the evidence supporting screening among this lower age group as a B compared to an A for 50 to 75 years of age. 4 The CMS has not commented about whether it will modify its colorectal screening measure in the MIPS to include this younger age group. Thus, further research should also address how any change in age-based screening recommendations are effectively adopted in primary care practices.
Limitations to our study include lack of detail on patient level differences and the screening method. For instance, differences in colorectal cancer screening rates for women in rural and urban areas have been found, 28 but our practice-based analyses could not address this. Past work has revealed differences in screening methods between rural and urban populations, with increasing use of FOBT in rural areas compared to urban areas. 5 Proximity of endoscopy services is also a barrier to rural populations utilizing this method. 29 Lastly, because all of the practices were “exposed” to MACRA, the study had no non-exposure comparison group. It is possible that the annual increases in colorectal cancer screening are attributable to factors beyond MACRA, such as public health messages encouraging colorectal cancer screening.
Conclusion
Among both rural and urban primary care practices participating in a national quality improvement registry, colorectal screening rates improved dramatically from 2016 to 2020. These increases may be partially attributed to the implementation of MACRA, as well as overall public health messaging aimed at colorectal cancer screening awareness. Although we found no rural versus urban differences, other demographic and social disparities were evident. To improve their colorectal cancer screening performance, and increase their Medicare reimbursement, practices should consider collaborating with community-based organizations to tailor and target colorectal cancer screening messages, modify clinic office hours, or develop other strategies to better accommodate populations who are relatively older, Hispanic, and have higher social deprivation.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by the Federal Office of Rural Health Policy (FORHP), Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services (HHS) under cooperative agreement # U1CRH30041. The information, conclusions and opinions expressed in this document are those of the authors and no endorsement by FORHP, HRSA, HHS, or the University of Kentucky is intended or should be inferred.
