Changes in Receipt of Cancer Screening in Medicare Beneficiaries Following the Affordable Care Act

Abstract

Background:

The Affordable Care Act (ACA) removed copayments for screening mammography and colonoscopy in Medicare beneficiaries, but its clinical impact is unknown.

Methods:

Using a 5% random sample of Medicare claims from 2009 through 2012 in individuals age 70 years or older who were due for screening, we examined claims for screening mammography and screening or surveillance colonoscopy for two years prior to ACA (2009–2010) and two years post-ACA (2011–2012). Receipt of the procedures at the patient level was compared across years using generalized estimating equations. Statistical tests were two-sided.

Results:

Compared with 2009, we found an increase in mammography uptake during the ACA coverage period, with multivariable odds ratios (MOR) of 1.22 (95% confidence interval [CI] = 1.20 to 1.25, P < .001) for 2011 and 1.17 (95% CI = 1.15 to 1.20, P < .001) for 2012 and less change in 2010 (OR = 1.03, 95% CI = 1.01 to 1.05, P = .01). In contrast to mammography, uptake of screening or surveillance colonoscopy decreased in 2012 (MOR = 0.95, 95% CI = 0.92 to 0.98, P = .002) compared with 2009, with no change in 2010 (MOR = 1.01, 95% CI = 0.99 to 1.04, P = .47) or 2011 (MOR = 1.01, 95% CI = 0.99 to 1.04, P = .34). Other factors associated with procedure receipt included younger age and prior preventive health visits. In an analysis restricted to patients age 70 to 74 years, colonoscopy use increased slightly in 2011 but was unchanged in 2012, and the findings by year for mammography were consistent with the main analysis.

Conclusions:

Following ACA implementation with concomitant reduction in out-of-pocket expenditures, there was a statistically significant increment in mammography uptake but not colonoscopy. This suggests that affordability is a necessary but not sufficient facilitator of preventive services.

The Affordable Care Act (ACA) has provided phased-in changes in health care delivery and reimbursement. As of January 1, 2011, the ACA eliminated copayments and co-insurance for recommended preventive services in Medicare beneficiaries including breast and colon cancer screening. The impact of this legislation is important because out-of-pocket expenditures are an important barrier to receipt of screening ( 1 ) and prior to the ACA almost two-thirds of beneficiaries routinely faced cost-sharing for preventive care ( 2 , 3 ). However, previous research on mammography ( 4 , 5 ) and colorectal cancer screening ( 6–8 ) has yielded variable results on actual receipt following changes in Medicare coverage.

The goal of the present study was to use a national sample of Medicare claims data to examine patterns of cancer screening, including mammography and colonoscopy after the ACA. We hypothesized that with elimination of copayments and coinsurance, rates of mammography and colonoscopy would increase following ACA implementation.

Methods

Medicare Data

The study was approved by the University Hospitals Institutional Review Board. We used a 5% random sample of Medicare beneficiaries from 2004 through 2012, within which the same beneficiaries are contained from year to year. The 2004 to 2008 data were used to exclude previous cancer diagnoses, to calculate the comorbidity score, and, in the case of colonoscopy, to determine receipt for the five-year period prior to study entry. The 2009 to 2012 files were used to determine receipt of screening procedures. Because of the 90-day extension that was used to satisfy criteria for receipt of testing (see below), we only included procedures through September 30, 2012.

Data included the carrier files, which include physician/supplier claims, the outpatient files, and the Medicare Provider Analysis and Review (MEDPAR) files. All files included diagnosis codes according to ICD-9-CM and procedure codes according to CPT-4 or ICD-9-CM. We also used the Medicare Beneficiary Summary Files, which contain demographic and enrollment information.

Study Sample

The study consisted of two separate cohorts, one for mammography ( Figure 1 ) and one for colonoscopy ( Figure 2 ). Because of incomplete claims, we excluded beneficiaries who were not enrolled in both Medicare Parts A and B at any time during the observation period. Beneficiaries enrolled because of end-stage renal disease or disability were also excluded. For mammography, patients with a history of breast cancer or carcinoma in situ and, for colonoscopy, those with a history of colorectal cancer in the five-year period prior to the index year were also excluded, as they would be subject to different follow-up guidelines. Finally, because of the need to obtain claims data from the preceding five-year period, the study was limited to Medicare beneficiaries age 70 years and older, and eligible patients entered the cohort at age 70 years.

Figure 1.

Flow chart for construction of mammography cohort. ESRD = end-stage renal disease; HMO = health maintenance organization.

Figure 2.

Flow chart for construction of colonoscopy cohort. ESRD = end-stage renal disease; HMO = health maintenance organization.

Mammography Sample

As screening mammography is typically performed on an annual or biennial basis ( 9 ), an individual could have multiple “index” procedures, with the next procedure date reset based on the most recent mammogram date. We considered someone due for mammography if there were no mammography claims during the preceding two-year period. In order to account for delays in obtaining screening, a 90-day extension from the beneficiary’s due date was used to satisfy the criterion for screening ( Figure 1 ). For example, if the index mammogram was performed in April 2009, a repeat procedure performed from January 2010 through July 2010 would be considered up to date.

Based on a previously validated algorithm ( 10 ), the mammography analysis was limited to procedures that were more likely to be performed for screening as opposed to diagnostic indications. Procedure codes were limited to CPT-4 76092 or G0202.

Colonoscopy Sample

For colorectal cancer screening and surveillance, the sample was limited to individuals age 70 years or older who were considered to be “high risk” for colorectal cancer based on diagnoses and procedures in the previous five-year period and were due for repeat procedures because of lack of colonoscopy in the previous five years ( Figure 2 ). High-risk diagnoses included colon polyps or polypectomy, Crohn’s Disease or ulcerative colitis, family history of gastrointestinal neoplasm or previous colonoscopy designated as “high risk.” The rationale for this limitation was two-fold. First, guidelines recommend colonoscopy as the preferred screening procedure in this group ( 11 , 12 ), and second, a five-year look-back period would capture patients who were due for colonoscopy as a five-year screening interval is generally recommended ( 12 , 13 ). In order to account for delays in obtaining screening, a 90-day extension from the beneficiary’s due date was used to satisfy the criterion for screening.

Analogous to mammography, we also limited the sample to colonoscopies performed for screening or surveillance as opposed to diagnostic indications. As previously described ( 14 ), the associated diagnosis codes were identified and if a code indicating gastrointestinal signs or symptoms was present the procedure was coded as diagnostic. If none of the codes were present or the procedure was coded as screening (G0105, G0121), it was categorized as screening/surveillance.

Measures

Demographic characteristics included age, sex, and race. A previously validated, weighted comorbidity index was included for the 12 months to one month period prior to the first date that the patient met study eligibility criteria ( 15 ). Geographic region of residence was divided into Northeast, Midwest, South, and West.

The mammography analysis also included breast cancer risk status. All claims for the previous five years were searched for the following high-risk criteria: benign breast neoplasm, abnormal findings on radiological and other breast examinations, complications because of breast prostheses, Cowden syndrome, personal history of chest irradiation, and family history of breast or ovarian cancer. In addition, women without one of these diagnosis codes but who received one or more breast MRI examinations during the previous five years were also included in this group.

Because socioeconomic status was not available at the patient level, using the 2010 US Census data, we included county-level median household income and proportion of high school graduates among adults age 25 years and older. We also used the 2010 AMA Masterfile to calculate physician supply at the county level, including the number of primary care providers and for colonoscopy, gastroenterologists per 100 000 population.

Another change instituted by the ACA was the annual wellness visit. Since 2005, beneficiaries within the first year of enrollment have been entitled to an initial preventive visit known as the Welcome to Medicare visit and under the ACA; this was supplemented by a yearly wellness visit. Among other features, this visit required providers to establish a written screening schedule for the next five to 10 years. As with other recommended preventive services, it was not subject to coinsurance or deductible. Receipt of the “Welcome to Medicare Benefit” (G0402) and the yearly wellness visit (G0438, G0439) during the calendar year of interest were measured and summarized as preventive visits.

Analysis

For each procedure, among individuals who were due for testing, we summarized the proportion of eligible beneficiaries in a given calendar month who received that service. Univariate analysis was then used to determine the association of calendar year with the use of mammography and colonoscopy. Because individual patients were eligible for screening in more than one year, patient-level generalized estimating equation (GEE) logistic regression was used to account for within-patient correlation. We then used multivariable GEE models to determine the independent association of demographic, socioeconomic, and clinical measures with receipt of testing. Because there was seasonal variation in procedure use, calendar year quarter was also included in the models. Also, because US Preventative Services Task Force (USPSTF) guidelines do not endorse universal mammography or colonoscopy in individuals older than age 75 years ( 9 , 11 ), in a secondary analysis we limited the sample to those age 70 to 74 years.

Beneficiaries were censored at the month of death or disrenrollment from fee-for-service Medicare plans based on the Beneficiary Summary File, and their eligible months were included prior to censoring. Because the subgroup of patients who disenrolled may be systematically different from the others, we also performed a sensitivity analysis after excluding these individuals. For all models, individuals were censored at the time of cancer diagnosis. Statistical significance was measured using two-sided testing, with P values of less than .05 considered to be significant.

Results

Using the 5% random sample of Medicare beneficiaries from 2009 through 2012, we identified 2 196 160 eligible women for the mammography cohort. From this cohort, we excluded 1 332 893 for the following non-mutually exclusive conditions: younger than age 70 years (n = 716 554), enrollment in Medicare managed care plans (n = 658 196), lack of enrollment in Medicare Part B (n = 531 961), and enrollment because of end-stage renal disease or disability (n = 19 417). The final sample consisted of 862 267 women. For colonoscopy, we identified 3 811 023 eligible beneficiaries and excluded 2 431 401 for the following non-mutually exclusive conditions: younger than age 70 years (n = 1 413 529), enrollment in Medicare-managed care plans (n = 1 100 811), lack of enrollment in Medicare Part B (n = 967 546), and enrollment because of end-stage renal disease or disability (n = 42 605). Of the remaining 1 108 228 individuals, 326 503 had one or more increased risk factors for colorectal cancer and were included in the study.

Patient characteristics at cohort entry for the mammography and colonoscopy groups are shown ( Table 1 ). In the mammography sample, the mean age was 79.7±6.9 years and most patients were white and had a comorbidity score of 0. More than 16% had evidence of increased breast cancer risk and at study entry, and more than half underwent one or more mammograms within the previous five years. In the colonoscopy sample, the mean age was 77.1±6.1 years, 59% were female, and most were white. At study entry, there was evidence of receipt of colonoscopy in the past five years in 68%.

Table 1.

Demographic characteristics of the mammography and colonoscopy samples

Characteristic	Mammography No. (%)	Colonoscopy No. (%)
Age group, y
70–74	242 155 (28.1)	101 151 (31.0)
75–80	215 952 (25.0)	97 135 (29.8)
˃80	404 160 (46.9)	128 217 (39.3)
Sex
Female	862 267 (100)	193 420 (59.2)
Male	----------------------	133 083 (40.8)
Race
White	745 775 (86.5)	293 429 (89.9)
Black	70 706 (8.2)	20 406 (6.3)
Other/unknown	45 786 (5.3)	12 668 (3.9)
Comorbidity score
0	364 481 (42.3)	118 616 (36.3)
1	225 569 (26.2)	86 232 (26.4)
2	112 353 (13.0)	46 233 (14.2)
≥3	159 864 (18.5)	75 422 (23.1)
Procedure in past 5 y
Yes	493 130 (57.2)	222 740 (68.2)
No	369 137 (42.8)	103 763 (31.8)
Geographic region
Northeast	99 161 (11.5)	36 043 (11.0)
Midwest	235 658 (27.3)	86 461 (26.5)
South	370 860 (43.0)	145 065 (44.4)
West	156 588 (18.2)	58 934 (18.1)
Preventive service visit*
Yes	38 884 (4.5)	33 467 (10.3)
No	823 383 (95.5)	293 036 (89.7)

* All characteristics were measured at time of first procedure except for preventive service visits, which were included if received any time during the study period. For patients without procedures, characteristics were measured at the end of the first 12 months of continuous enrollment, beginning in 2009.

Monthly and yearly trends for mammography ( Figure 3 ) and colonoscopy ( Figure 4 ) are shown graphically. For mammography, the frequency of eligible women who were tested in a given month increased from 2009 to 2010 and increased further to 2011. The proportion remained relatively stable between 2011 and 2012. For colonoscopy, the proportion remained relatively stable over the first three years and decreased somewhat in 2012. Of note, for both procedures, the monthly frequencies were higher in the initial part of a calendar year and declined toward the end of the year.

Figure 3.

Trends by calendar month and year in the use of mammography among eligible women. The graphs show an increment for mammography use following Affordable Care Act (ACA) implementation (2011–12) compared with the pre-ACA period (2009–10).

Figure 4.

Trends by calendar month and year in the use of colonoscopy among eligible men and women. The graphs show no appreciable change following Affordable Care Act (ACA) implementation (2011) and a decrement in 2012, compared with the pre-ACA period (2009–10).

The factors associated with receipt of screening mammography are shown ( Table 2 ). In both univariate and multivariable analyses, increased use of mammography was associated with younger age, lower comorbidity, and especially receipt of previous mammography and preventive health visits. We found an increase in mammography during the ACA coverage period, with multivariable odds ratios (MORs) of 1.22 (95% CI = 1.20 to 1.25, P < .001) for 2011 and 1.17 (95% CI = 1.15 to 1.20, P < .001) for 2012 compared to only a minimal increase in 2010 (MOR = 1.03, 95% CI = 1.01 to 1.05, P = .01).

Table 2.

Receipt of screening mammography among women without mammography in the past two years*

Parameter	Univariate OR (95% CI)	Multivariable model-adjusted OR (95% CI)
Age, y	0.90 (0.90 to 0.91)	0.91 (0.91 to 0.91)
Race
Caucasian	Referent	Referent
African American	1.00 (0.97 to 1.03)	1.13 (1.10 to 1.17)
Other	0.89 (0.86 to 0.92)	1.04 (1.00 to 1.08)
Comorbidity score
0	Referent	Referent
1	1.01 (0.98 to 1.04)	1.00 (0.98 to 1.02)
2	0.92 (0.89 to 0.95)	0.92 (0.90 to 0.94)
3+	0.84 (0.83 to 0.86)	0.87 (0.85 to 0.99)
Increased risk
No	Referent	Referent
Yes	2.40 (2.35 to 2.45)	1.16 (1.13 to 1.19)
Prior mammography
No	Referent	Referent
Yes	6.02 (5.92 to 6.13)	5.27 (5.17 to 5.37)
Prior preventive visit
No	Referent	Referent
Yes	2.21 (2.03 to 2.42)	1.72 (1.57 to 1.88)
Geographic region
Northeast	Referent	Referent
Midwest	1.14 (1.11 to 1.18)	1.01 (0.98 to 1.05)
South	1.36 (1.32 to 1.40)	1.17 (1.13 to 1.20)
West	1.34 (1.30 to 1.38)	1.12 (1.09 to 1.16)
Income quartile
1 (lowest)	0.79 (0.78 to 0.81)	0.92 (0.89 to 0.95)
2	0.91 (0.89 to 0.93)	1.01 (0.98 to 1.04)
3	0.95 (0.93 to 0.98)	1.01 (0.99 to 1.04)
4 (highest)	Referent	Referent
Education quartile
1 (lowest)	0.78 (0.76 to 0.80)	0.78 (0.75 to 0.80)
2	0.85 (0.83 to 0.87)	0.84 (0.81 to 0.86)
3	0.91 (0.89 to 0.93)	0.89 (0.87 to 0.91)
4 (highest)	Referent	Referent
Primary care physician density
1 (lowest)	1.01 (0.99 to 1.04)	1.00 (0.98 to 1.02)
2	1.00 (0.98 to 1.02)	0.99 (0.96 to 1.01)
3	0.98 (0.96 to 1.01)	0.94 (0.91 to 0.96)
4 (highest)	Referent	Referent
Calendar quarter
1	Referent	Referent
2	0.98 (0.97 to 1.00)	0.98 (0.97 to 1.00)
3	1.01 (0.99 to 1.02)	1.01 (0.99 to 1.03)
4	0.90 (0.88 to 0.91)	0.91 (0.89 to 0.93)
Calendar year
2009	Referent	Referent
2010	1.01(0.99 to 1.03)	1.03 (1.01 to 1.05)
2011	1.21 (1.18 to 1.23)	1.22 (1.20 to 1.25)
2012	1.18 (1.15 to 1.20)	1.17 (1.15 to 1.20)

* CI = confidence interval; OR = odds ratio.

Factors associated with screening or surveillance colonoscopy are shown ( Table 3 ). In the univariate and multivariable analyses, colonoscopy was more common in younger individuals and in men, as well as in individuals with lower comorbidity and prior preventive visits. In contrast to mammography, uptake of screening or surveillance colonoscopy was unchanged in 2011, with an MOR of 1.01 (95% CI = 0.99 to 1.04, P = .34) in 2011, and decreased in 2012 (MOR = 0.95, 95% CI = 0.92 to 0.98, P = .002). There was also no change in 2010 (MOR = 1.01, 95% CI = 0.99 to 1.04, P = .47).

Table 3.

Receipt of screening colonoscopy among patients at increased cancer risk without colonoscopy in the past five years*

Parameter	Univariate OR (95% CI)	Multivariable model-adjusted OR (95% CI)
Age, y	0.95 (0.95 to 0.95)	0.92 (0.92 to 0.93)
Sex
Male	Referent	Referent
Female	0.91 (0.90 to 0.92)	0.92 (0.90 to 0.94)
Race
Caucasian	Referent	Referent
African American	1.00 (0.97 to 1.02)	1.08 (1.03 to 1.13)
Other	0.95 (0.92 to 0.99)	1.02 (0.96 to 1.09)
Comorbidity score
0	Referent	Referent
1	0.98 (0.96 to 1.00)	0.98 (0.96 to 1.01)
2	0.99 (0.96 to 1.01)	0.92 (0.89 to 0.96)
3+	0.89 (0.88 to 0.90)	0.82 (0.79 to 0.84)
Prior preventive visit
No	Referent	Referent
Yes	1.24 (1.20 to 1.27)	1.37 (1.31 to 1.44)
Geographic region
Northeast	Referent	Referent
Midwest	0.84 (0.82 to 0.86)	0.98 (0.94 to 1.02)
South	0.95 (0.94 to 0.97)	1.03 (1.00 to 1.07)
West	0.85 (0.83 to 0.87)	0.93 (0.89 to 0.97)
Income quartile
1 (lowest)	0.90 (0.89 to 0.92)	0.92 (0.88 to 0.96)
2	0.94 (0.93 to 0.96)	1.01 (0.97 to 1.05)
3	0.97 (0.95 to 0.98)	1.03 (0.99 to 1.06)
4 (highest)	Referent	Referent
Education quartile
1 (lowest)	0.93 (0.91 to 0.94)	0.86 (0.83 to 0.90)
2	0.93 (0.92 to 0.95)	0.89 (0.86 to 0.92)
3	0.98 (0.96 to 0.99)	0.96 (0.93 to 0.99)
4 (highest)	Referent	Referent
Primary care physician density
1 (lowest)	0.96 (0.94 to 0.97)	1.05 (0.97 to 1.13)
2	0.94 (0.92 to 0.95)	1.04 (0.98 to 1.11)
3	0.94 (0.92 to 0.95)	1.01 (0.96 to 1.06)
4 (highest)	Referent	Referent
Gastroenterologist density
1 (lowest)	0.93 (0.91 to 0.95)	0.99 (0.92 to 1.07)
2	0.93 (0.92 to 0.95)	0.99 (0.93 to 1.05)
3	0.92 (0.90 to 0.94)	1.02 (0.97 to 1.07)
4 (highest)	Referent	Referent
Calendar quarter
1	Referent	Referent
2	1.02 (1.00 to 1.05)	1.02 (1.00 to 1.04)
3	0.99 (0.96 to 1.01)	0.98 (0.95 to 1.00)
4	0.83 (0.81 to 0.85)	0.82 (0.79 to 0.84)
Calendar year
2009	Referent	Referent
2010	0.99 (0.98 to 1.01)	1.01 (0.98 to 1.04)
2011	0.98 (0.96 to 0.99)	1.01 (0.99 to 1.04)
2012	0.94 (0.92 to 0.95)	0.95 (0.92 to 0.98)

* CI = confidence interval; OR = odds ratio.

In a secondary analysis, we limited the sample to individuals age 70 to 74 years. The results of procedure use by calendar year were similar as in the entire cohort, though the use of colonoscopy did not decline. For mammography, the MOR for 2010 was 1.07 (95% CI = 1.04 to 1.11, P < .001), for 2011, 1.30 (95% CI = 1.26 to 1.35, P < .001), and for 2012, 1.30 (95% CI = 1.25 to 1.35, P < .001). For colonoscopy, the MORs for 2010, 2011, and 2012 were 1.03 (95% CI = 0.98 to 1.08, P = .21), 1.07 (95% CI = 1.02 to 1.12, P = .005), and 1.00 (95% CI = 0.95 to 1.05, P = .85), respectively.

We also performed a sensitivity analysis excluding those who disenrolled from fee-for-service Medicare plans at any time. In the GEE models, we found similar results as in the primary analysis for mammography (2010 MOR = 1.07, 95% CI = 1.04 to 1.11, P = .01; 2011 MOR = 1.30, 95% CI = 1.26 to 1.35, P < .001; 2012 MOR = 1.30, 95% CI = 1.25 to 1.35, P < .001), and colonoscopy (2010 MOR = 1.03, 95% CI = 0.98 to 1.08, P = .21; 2011 MOR = 1.07, 95% CI = 1.02 to 1.12, P = .005; 2012 MOR = 1.00, 95% CI = 0.95 to 1.05, P = .85).

Discussion

The use of mammography and colonoscopy for the prevention and early detection of breast and colorectal cancers, respectively, is recommended by all professional societies ( 1 , 9 , 11–13 ), and out-of-pocket expenditures, which were removed by the ACA, were thought to be a barrier to screening. In an analysis of 2009 pre-ACA Medicare data, we estimated that only one-third of beneficiaries had dual Medicare-Medicaid eligibility or Medigap plans, both of which typically include co-insurance. Thus, the majority was subject to out-of-pocket expenditures for mammography and colonoscopy ( 2 , 3 ). Using a population-based sample of Medicare enrollees for two years prior to and two years after ACA implementation, we found that there was a statistically significant increase in mammography uptake, but colonoscopy use remained stagnant or decreased, even in high-risk individuals. However, for both mammography and colonoscopy, we found a statistically significant association with wellness visits, which may facilitate referral for screening procedures. In addition, we found that, in contrast to the entire sample, among individuals age 70 to 74 years colonoscopy use remained stable over time, suggesting that practitioners may have targeted the most age-appropriate patients for screening.

Two other recently published cross-sectional studies found divergent results of post-ACA changes in screening in Medicare beneficiaries ( 16 , 17 ). One study observed a modest increase in colonoscopy but not mammography ( 16 ), and the other found no change in the use of any cancer screening procedure ( 17 ). However, our study design provided an opportunity to measure screening in the same individuals longitudinally pre- and post-ACA.

Although screening colonoscopy has been reimbursed under Medicare since 2001, and every two years in high-risk beneficiaries since 1998, prior to the ACA the beneficiary was responsible for 25% of the Medicare-approved amount, which was eliminated following ACA implementation. An important exception that remains is that if biopsy or polypectomy is performed, the procedure is considered diagnostic regardless of the original indication and the Medicare Part B deductible applies. If no biopsy or polypectomy is performed, the deductible is waived. It is unknown if this loophole serves as a deterrent to colonoscopy. Previous studies have documented increases in colonoscopy use following initiation of Medicare reimbursement ( 6 , 18 ), but it is unknown if these were a direct result of the policy. In addition, two studies have examined the impact of financial barriers on receipt of colonoscopy in commercially insured individuals and have found only a modest association ( 19 , 20 ), suggesting that other barriers to receipt of colonoscopy, including lost wages, bowel preparation, and fear of complications, likely remain as deterrents.

Mammography has had longstanding Medicare coverage with reimbursement for women age 40 years and older every 12 months, including digital technologies. Although prior to the ACA mammography was covered without a Part B deductible, a 20% co-insurance or copayment applied. Therefore, out-of-pocket expenditures may have been a limiting factor for mammography as was demonstrated in Medicare-managed care enrollees ( 21 ). In the current analysis, we found that mammography uptake did increase following ACA implementation. We also found that receipt of previous mammography was associated with subsequent screening, suggesting that women who are already engaged in screening practice are more likely to continue with such.

We also found that receipt of preventive visits was associated with increased use of both colonoscopy and mammography. However, a recent study has documented that as recently as 2013 only 32% of fee-for-service beneficiaries received this benefit ( 22 ). Another study found that prior to the ACA no more than 5% of women received a Welcome visit and it was not associated with subsequent uptake of breast or cervical cancer screening ( 23 ). Conceivably, in conjunction with reduced out-of-pocket expenditures, receipt of these visits may have a greater impact.

We recognize several limitations to the study design and data sources. First, although we used previously developed algorithms to identify screening mammography and colonoscopy ( 10 , 14 ), we could not definitively differentiate screening vs diagnostic indications. We were unable to measure patient and physician preferences regarding screening, both of which are associated with screening receipt. In addition, although claims data are collected for billing purposes and not research, their accuracy for measuring procedure use is thought to be high ( 24 ). The study did not measure other colon cancer screening tests such as fecal immunochemical testing (FIT). However, the change in out-of-pocket expenses for FIT under the ACA would be lower than for colonoscopy, and colonoscopy is currently the dominant screening strategy ( 1 ). The study also did not examine beneficiaries who were enrolled in Medicare advantage plans but focused on the impact of reform changes in fee-for-service enrollees. The study was also limited to an older patient population, and thus the impact of reform legislation in younger, privately insured individuals could not be measured. Although the USPSTF does not recommend routine screening colonoscopy in individuals older than age 75 years ( 11 ), other guidelines do not have an upper age cutoff ( 1 , 12 ), and Medicare reimbursement is not contingent on age. Moreover, the USPSTF recommendations do not necessarily apply to the high-risk individuals who were the subject of this analysis. For mammography, the USPSTF guidelines state that there is insufficient information as to whether screening should be performed in women older than age 75 years ( 9 ), and American Cancer Society guidelines do not specify an upper age limit ( 1 ). Of note is that similar results for mammography were observed when the sample was limited to individuals age 70 to 74 years. Although individuals age 65 to 69 years are contained in Medicare files, we did not include them because of the inability to measure previous colonoscopy use and cancer diagnoses. The colonoscopy cohort was limited to patients at increased risk for neoplasia, including polyp surveillance, and although the use of colonoscopy is likely higher in this group, it is unknown whether the findings would be observed in average-risk individuals. Another limitation was the lack of data on supplemental or Medigap insurance. Also, because patient-level socioeconomic status was not available in claims data, we used small area measures, an approach that is commonly used in studies of Medicare data. Finally, the observational design demonstrated associations with ACA coverage but could not ascertain cause and effect.

In summary, we found that following ACA implementation with concomitant reduction in out-of-pocket expenditures, there was a statistically significant increment in mammography uptake but not colonoscopy. This suggests that affordability is a necessary but not sufficient facilitator of preventive services and that other barriers to uptake should be considered for colonoscopy.

Funding

This work was supported by the American Cancer Society (RSGI-12-218-01-CPHPS), the Case Clinical and Translational Scientific Collaborative, National Center for Advancing Translational Sciences at the National Institutes of Health (UL1TR000439), the Case Comprehensive Cancer Center, the National Cancer Institute at the National Institutes of Health (P30 CA043703), and the Cleveland Digestive Disease Research Core Center (P30 DK097948).

The study sponsors had no role in the design of the study; the collection, analysis, or interpretation of data; the writing of the manuscript; or the decision to submit the manuscript for publication.