The American College of Physicians Performance Measurement Committee announced their support of just one of five colorectal cancer screening and surveillance performance measures currently used in pay-for-performance, public reporting, and/or accountability programs, finding the others methodologically unsound or misaligned with clinical evidence.

Senior author of the position paper Rebecca Andrews, MD, Chair of the American College of Physicians (ACP) Performance Measurement Committee (PMC), explained the rigorous evaluation behind this decision in an exclusive interview with Conexiant. "Organizations are trying to do the right thing. They are trying to utilize data to see where there are gaps in care," she said. But good intentions aren't always enough. "Has it been tested? Is it valid? Is it reliable? And then the big one for me as a practicing clinician: Is it feasible?"

The committee's evaluation revealed fundamental issues with four of the five measures. "A study that came out in JAMA maybe 5 years ago looked at them [and] found some had no data to support that there was a gap in this area of care," Dr. Andrews noted.

The ACP believes performance measures should only be considered for inclusion in reporting, accountability, or payment programs if they are methodologically sound and have undergone appropriate statistical analyses at the level of attribution (ie, physician, group, or health plan) for which they are used. These principles are critical given the potential impact on physician administrative work and reputation,  reimbursement, and preventing unintended consequences of patient care.

The Measure That Made the Cut

In the position paper, published in Annals of Internal Medicine, the committee endorsed "Facility 7-Day Risk-Standardized Hospital Visit Rate After Outpatient Colonoscopy" (Centers for Medicare & Medicaid Services' [CMS] Measures Inventory Tool ID #253) at the facility level only. This facility-level measure doesn't have a Merit-Based Incentive Payment Program (MIPS) Quality ID, which is why the ACP doesn't rate it for use at the individual-physician, group-practice, or health-plan level.

The single measure that passed scrutiny—the facility 7-day postcolonoscopy hospital visit rate—succeeded for specific reasons. "It looks at something that really does need to be looked at, which is: What are the downsides to doing colonoscopy? What are the harms?" Dr. Andrews explained. "In one of those measures, you end up in an emergency room or hospitalized within a 7-day period after the colonoscopy," she added.

Dr. Andrews acknowledged the measure may not be perfect. "It's not 100% attributable to the colonoscopy, because there are confounders—you'd have to look back to determine that—but it does help capture complications." Critically, "it's not burdensome to the clinician: It can be pulled from Medicare claims," she indicated.

The emphasis on reducing clinician burden reflected a broader philosophy. "We want them in the room talking to the patient, seeing the patient, counseling the patient, treating the patient," Dr. Andrews underscored. "[This] addressed an overuse type of measurement—which, in my mind, is really where a lot of effort needs to go. Are we spending money on the right things that make a difference, and not just increasing the cost of health care with more, more, more—especially when there are downsides?"

The measure was designed to assess the rate of risk-standardized, all-cause, unplanned hospital visits within 7 days of an outpatient colonoscopy among Medicare fee-for-service patients aged 65 years or older. It supports patient safety by identifying unplanned hospital visits for outpatient colonoscopy and provides comparative results that can help hospitals and ambulatory surgery centers evaluate unplanned hospital visits following colonoscopy procedures.

There is sufficient evidence showing the importance of monitoring this outcome and the improvement efforts that facilities can implement to lower their unplanned hospital visit rates, for instance, by implementing quality-improvement efforts focused on patient, physician, and facility factors. The performance measure specifications are sensitive to the effect of patient-level variables on performance with a robust risk-adjustment approach to address these variables as well as health care disparities. There is a low reporting burden because the performance measure is based on CMS claims data.

However, unplanned hospital visits could be attributable to reasons other than colonoscopy complications. As a result, the interpretation of gaps identified using this measure requires thoughtfulness. The measure isn't designed or meant for use at the individual-physician, group-practice, or health-plan levels.

The Four Rejected Measures

Colorectal Cancer Screening (MIPS Quality ID #113)

The committee rejected this measure at the individual-physician, group-practice, and health-plan levels, citing its age range of 45 to 75 years as inconsistent with current evidence and ACP guidance to begin screening at age 50 years for asymptomatic average-risk individuals. Although some guidelines recommend screening at earlier ages, there is no evidence to support the practice. The incidence of colorectal cancer is lower in the average-risk population, and screening test performance data is available for adenoma rather than colorectal cancer detection.

The committee also noted that the measure may need testing at the individual-physician and group-practice levels and validation at the health-plan level. The incidence of colorectal cancer is 33.4 per 100,000 individuals in patients aged 45 to 49 years and ranges from 60.6 per 100,000 in those aged 50 to 54 years to 234.7 per 100,000 in those aged 85 years or older.

In addition, screening overuse is associated with preventable harms such as a perforated colon and bleeding. Expanding screening below the evidence-supported age increases demand, can worsen access for patient groups with proven benefits, and risks widening disparities—particularly among patients aged 45 to 49 for whom socioeconomic status and locality are known to drive uptake differences. A recent study found that socioeconomic status—incorporating insurance status and occupation—and residence in a metropolitan vs nonmetropolitan locality affected screening uptake and widened disparities in patients aged 45 to 49 years. Aligning the age range in screening guidance and performance measures in asymptomatic average-risk patients could be crucial to ensure that gaps in access to screening services aren't intensified.

Age-Appropriate Screening Colonoscopy (MIPS Quality ID #439)

The committee also rejected this measure at individual-physician and group-practice levels. Although the measure evaluates the percentage of screening colonoscopies performed in patients aged 86 years or older, there is no documented performance gap for screening patients over the age of 86 years.

The committee suggested reducing the numerator age to 75 years or older to be more impactful and align with the ACP's guidance to cease screening at age 75 years among asymptomatic average-risk patients. The peak incidence of colorectal cancer occurs in the 65 to 69 age group. There is limited data on the appropriate age to stop offering colorectal cancer screening, indicating that more research is needed per age group. Other organizations also recommend selectively screening adults aged 76 to 85 years on the basis of the patient's overall health and screening history. The measure lacks testing data at individual-physician and group-practice levels and isn't designed for use at the health plan level.

Appropriate Follow-Up Interval for Normal Colonoscopy in Average-Risk Patients (MIPS Quality ID #320)

For this measure, the committee identified uncertain validity at the individual-physician level. It evaluated the percentage of patients aged 45 to 75 years receiving a screening colonoscopy without biopsy or polypectomy who had a recommended follow-up interval of at least 10 years for repeated colonoscopy documented in their colonoscopy report.

The measure specifications are unclear and could be interpreted to suggest that an interval longer than 10 years is appropriate, which may result in unintended consequences if the follow-up occurs significantly later than the recommended 10-year interval. The follow-up interval should be documented transparently for primary-care physicians.

The PMC recommended revising the measure so that it requires the follow-up to occur (not merely be recommended) and clarifying documentation expectations for primary care. As currently specified, the measure may not directly impact or account for the time needed to obtain the appropriate follow-up care. The measure is not tested at the group-practice level and isn't designed for use at the health-plan level.

Colonoscopy Interval for Patients With a History of Adenomatous Polyps—Avoidance of Inappropriate Use (MIPS Quality ID #185)

The committee criticized this measure for uncertain validity at the individual-physician level. This performance measure evaluates the percentage of patients receiving a surveillance colonoscopy who have a history of prior adenomatous polyps in previous colonoscopy findings, with an interval of 3 or more years since their last colonoscopy.

Although this measure directly addresses when the colonoscopy was completed (rather than focusing on the recommended follow-up interval), it doesn't specify the outer range of the acceptable time interval. In addition, the measure intends to address colonoscopy overuse, but current guidelines recommend longer surveillance intervals on the basis of polyp characteristics, such as number and size, which would further encourage reduction of colonoscopy overuse.

Variability in documentation for adenomatous polyps complicates the discrete data elements needed to transition the performance measure to electronic specifications and, as a result, the performance measure would be burdensome to report. It isn't tested at the group-practice level and isn't designed for health-plan use.

What Happens Next?

When asked about immediate actions for CMS and measure developers, Dr. Andrews was clear: "It would be great for the companies and the organizations that create or maintain measures to stop and say, 'Let's talk to some clinician groups and find out what is burdensome about this. Let's look at the data that is out there. Has this actually been tested?'"

The issue of testing at the appropriate level is critical. "If we're holding individual physicians accountable for those outcomes, has it been tested at a single-physician level to say it is an accurate, reliable measure of the quality of your care? If the answer to that is 'no,' then you can't use it at that level—or you have to go back to the drawing board," she stressed.

Dr. Andrews offered a pointed example from another area: "Hypertension: you're measuring high blood pressure control, and the only measure is the last blood pressure of the year, while the calendar year coincides with many weird factors [like] vacations, college breaks, school breaks, holidays, increased indulgences, salty foods and sweets, a lot of stress." This illustrated how poorly designed measures fail to capture clinical reality.

"We don't need 300 measures," Dr. Andrews suggested. "What we do need is a few that really address gaps in health care."

The Review Process

The 12-member committee used a modified RAND/UCLA Appropriateness Method to evaluate measures across five domains (importance, appropriate care, clinical evidence base, measure specifications, and feasibility/applicability). Members represented various US regions and practice settings, from solo to large specialty practices.

The Stakes

The stakes extend beyond administrative efficiency. "When measurement is done poorly, it increases costs, drives burnout, and pushes physicians to chase unattainable goals that make no sense," Dr. Andrews warned. "Then there's the burden of trying to report this—checking boxes or creating weird workarounds—which impacts patients, because they want to know I'm seeing them, not trying to figure out how to report out the quality of their care," she concluded.

About 70% of adults are up-to-date with colorectal cancer screening schedules in the United States. Colorectal cancer is the fourth highest in incidence and the second leading cause of cancer-related deaths in males and female patients. The PMC emphasized that aligning performance measures with strong evidence is essential.

Although there are five performance measures for colorectal cancer screening and surveillance, the ACP PMC found only one performance measure valid for use. The other four measures didn't meet the ACP's standards for appropriate use, high-quality evidence, and scientific acceptability. Some measures aren't aligned with current evidence. In addition, improvements are needed to the performance measure specifications. The ACP encouraged measure developers to refine and test these performance measures that aim to improve health outcomes for patients.

Source: Annals of Internal Medicine