Colorectal cancer screening quality measures – beyond colonoscopy
Yet, outside of selected clinical settings (e.g., the Veterans Affairs Healthcare System), no existing quality measure focuses on colonoscopy completion after positive FOBT. Other examples of important performance gaps beyond colonoscopy technical performance include appropriate use of monitored anesthesia care (data suggest overuse) and appropriate pathology interpretation (data suggest variation in the interpretation of serrated lesions, some of which are associated with increased risk of CRC).
Second, existing measures fail to consider individual patient factors that are known to be predictive of screening benefit, such as life expectancy and personal or familial cancer risk.8
Returning to the Healthcare Effectiveness Data and Information Set measure as an example, this measure relies on age alone to identify patients who are candidates for screening. Yet, it is widely recognized that life expectancy is a critically important factor when considering screening, particularly in older patients. However, existing measures fail to incorporate estimates of life expectancy. As a result, clinicians are encouraged to screen based on age alone. Data suggest that this leads to overuse in older, sicker patients younger than age 75 (the age cutoff suggested by guidelines and incorporated into the Healthcare Effectiveness Data and Information Set measure) and underuse in healthy and unscreened patients older than age 75.9
Furthermore, existing measures fail to systematically consider patient preferences and undervalue shared decision making. Data suggest, for instance, that patients who are offered a “menu” of options for screening rather than a single screening test are more likely to complete a screening test.10
But existing measures do not consider whether a physician assessed a patient’s preferences, or offered more than one screening option. Of course, measuring the quality of decision making is an evolving science that presents challenges and requires creative solutions.11
Finally, existing measures do not capture the collaborative nature of modern health care. For example, most existing measures focus on the performance of gastroenterologists and GI practices. Although such a focus is appropriate for specialty-specific quality improvement efforts, it neglects to capture the important interactions and hand-offs between providers needed for successful completion of screening (and downstream reduction in cancer mortality). As a result, this specialty-focused approach fails to meet the needs of patients, who are increasingly interested in whether the care they receive is improving their health regardless of who provides it. It also fails to meet the needs of payers, who are interested in whether a network of providers is delivering high-value care for its members. Developing measures that capture the collaborative nature of modern health care becomes increasingly important as payers embrace alternative payment models and other value-based reimbursement systems.
Given these challenges with existing measures, we propose the following concepts for consideration as noncolonoscopy quality measures of CRC screening:
1. Appropriateness of referral for CRC screening (PCP): Measures in this domain should incorporate assessment of family history, prior screening history, and/or life expectancy, all of which are predictive of screening benefit and can be extracted from structured electronic health record data.
2. Shared decision making in choosing a CRC screening test (PCP): Currently, measures in this domain are (by necessity) relatively simple, such as assessment of documentation of whether stool testing was offered alongside colonoscopy during a clinic visit.
3. Percentage of individuals with a positive noncolonoscopy CRC screening test who complete a colonoscopy within 6 months (PCP and GI): A measure in this domain can be developed and implemented using existing infrastructure. The Veterans Affairs Healthcare System currently uses such a measure to monitor appropriate follow-up of positive fecal immunochemical tests.
4. Predicted reduction in CRC mortality: This measure, which provides an estimate of the effectiveness and value of a screening program, represents the “bottom line” that is ultimately of interest to both patients and payers. However, composite measures that use projected results require sophisticated statistical approaches and case-mix adjustment. They must also contend with “churn” in the covered population.
Although a more comprehensive view of screening is desirable, operationalizing measure concepts, such as those proposed previously, is challenging for several reasons. First, as we have outlined, CRC screening is truly a “team sport.” The stakeholders in CRC screening include public health, primary care, gastroenterology, colorectal surgery, oncology, anesthesiology, and pathology practitioners. As a result, screening has no natural and obvious steward. Rather, various components of care are adopted by different specialty groups, and this approach leads to unmonitored gaps in care. We believe that gastroenterologists can and should take the lead in developing comprehensive quality measures for CRC screening. Such comprehensive measures will be of value to patients and to payers. Second, assessing quality of care more comprehensively has traditionally been infeasible. For example, incorporating life expectancy into quality measures requires electronic assessment of health status.
