Medicare Policy and Colorectal Cancer Screening: Title and subTitle BreakWill Changing Access Change Outcomes?

Starting in July 2001, the Medicare Benefits and Improvement Act expanded Medicare coverage to include colonoscopy for colorectal cancer screening among average-risk beneficiaries older than 50 years.¹ This legislation has several important implications. From a clinical standpoint, colonoscopy is the most sensitive test available for diagnosis of early stage, curable colorectal neoplasms. Colonoscopy is also the only cancer screening test that is potentially preventive because it permits removal of premalignant lesions. Some have related the decrease in colorectal cancer incidence and mortality to screening of the distal colon (such as barium enema and sigmoidoscopy)² ; colonoscopic screening of the entire colon is intended to improve this further by addressing right-sided lesions.

From a policy standpoint, the expansion in Medicare coverage involved only fee-for-service (FFS) beneficiaries and not those enrolled in Medicare health maintenance organization (HMO) plans. Traditionally disadvantaged subgroups are less likely to be enrolled in Medicare HMOs, which tend to screen systematically. Among these subgroups, colonoscopic screening may be particularly worthwhile among women, the elderly, and African Americans, who are more likely to have right-sided-only neoplasms^{3 - 5} (beyond the reach of sigmoidoscopy), and among African Americans and poor individuals, whose higher colorectal cancer incidence and mortality rates⁶ are largely attributed to decreased access to care and more advanced cancer at the time of diagnosis.

In this issue of JAMA, Gross and colleagues⁷ examine whether liberalizing Medicare coverage of colorectal cancer screening has lived up to its promise. The authors approached the question in 3 parts using linked Surveillance, Epidemiology, and End Results (SEER)–Medicare data from 3 eras: 1992 to 1997 (before coverage of any colorectal cancer screening), January 1998 to June 2001 (during coverage of all other recommended screening regimens and colonoscopy for high-risk individuals), and July 2001 to December 2002 (universal coverage of all screening modalities). First, to assess utilization changes, Gross et al examined the frequency of colonoscopy during each time period. Second, to determine the relative effectiveness of screening, they identified the proportion of patients with stage I disease among all those diagnosed with colorectal cancer in each time period, and the likelihood of early stage cancer in the proximal colon. And third, to assess the impact of the policy change and to differentiate this from general changes over time in the way clinicians evaluate patients for cancer, they compared the relative proportions of patients with early stage colorectal cancer in Medicare FFS vs HMOs. Not surprisingly, colonoscopy utilization rates increased nearly 7-fold over the course of the study. Although there is no direct evidence that colonoscopy use increased due to the policy change, it is intuitive that improving reimbursement of screening results in more screening. The key questions are whether Medicare can afford such a screening policy and whether the policy change delivers more effective care.

In a previous study, Frazier et al⁸ determined that screening colonoscopy once every decade in average-risk individuals older than 50 years is cost-effective according to the usual definition of the term. What about colonoscopic screening among the 36 million elderly Medicare beneficiaries¹ in the United States? Assuming 50% compliance with the recommendation to undergo screening colonoscopy once every 10 years and an $800 charge per colonoscopy, up to 5% or 1.8 million beneficiaries could potentially undergo colonoscopy in any given year, costing $1.44 billion or about 0.4% of the entire annual Medicare operating cost.¹ If utilization continues to increase at a similar rate as the eligible population continues to increase, screening colonoscopy could soon cost Medicare several billion dollars per year.

Given these potential costs, is Medicare getting what it pays for? That depends partly on whether the goal is prevention of cancer or identification of disease at a presymptomatic and more curable stage. Based on the data set used in the current study, Gross et al⁷ were unable to assess changes in the frequency or incidence of colonic neoplasms. Others⁹ have reported that diagnosis of advanced neoplasms decreased by more than 20% when colonoscopic screening increased 3-fold in the general population. While this finding almost certainly resulted from screening a higher proportion of average-risk individuals, a similar decrease in incidence of colorectal cancer can be expected with colonoscopic polypectomies over time.¹⁰

Gross et al observed a small increase in the proportion of patients with stage I lesions after noncolonoscopic screening became available in 1998, but no further increase in early cancers overall with the advent of screening colonoscopy in 2001. The authors further assessed the impact of the Medicare policy change by comparing proportion of stage I lesions among Medicare FFS beneficiaries, to whom it was targeted, and HMO beneficiaries. Although there were no differences between groups in the first time period, before screening was available in the FFS plans, colorectal tumors in FFS patients were more likely to be identified at an early stage than tumors in HMO patients during each subsequent time interval. This temporal association implies that earlier tumor detection among FFS patients is due to expanded coverage.

The findings of Gross et al demonstrate that the change in Medicare policy was effective: a target population received screening at a higher rate and this resulted in an increase in the detection of early stage and right-sided cancers. It remains to be seen if future screening will continue to increase the rates of early identification of colorectal cancer. Given the costs of universal screening, if rates of colonoscopy continue to increase without additional benefit in overall diagnosis of early stage disease, policy makers, health care organizations, and physicians may have to devise a feasible rationing plan for broader colon screening. While increasing access to care and improving compliance with recommended care is an undeniable good, providing screening colonoscopy to all is not realistic. The onus is on physicians and other primary care clinicians to capitalize on the momentum associated with screening colonoscopy to encourage other effective but less expensive forms of screening, such as fecal occult blood tests with sigmoidoscopy.⁸ Screening colonoscopy could be made more effective by targeting subgroups with higher incidence and mortality from colorectal cancer, and those at greatest risk of right-sided or proximal neoplasms that would not be identified by sigmoidoscopy. Such a strategy, while difficult, would provide the best opportunity to judiciously maximize access to care without sacrificing outcomes.