The limits of using Medicare patient data for quality outcomes

Originally published in MedPage Today

by Todd Neale, MedPage Today Staff Writer

Most primary care physicians don’t see enough Medicare patients to determine statistically significant differences in cost and performance between practices, researchers found.

In fact, nearly two-thirds (65.7%) of primary care doctors worked in practices that saw too few Medicare patients to provide reliable comparisons with national benchmarks for cost and quality of care, according to David Nyweide, PhD, of the Centers for Medicare & Medicaid Services (CMS) in Baltimore, and colleagues.

“Our study suggests that rethinking the approach to performance measurement in ambulatory care may be necessary for the Medicare program,” the researchers wrote in the Dec. 9 issue of the Journal of the American Medical Association.

The limitations observed in the study could be overcome by increasing the size of the patient pool for statistical analysis, they said. This could be accomplished by combining information on patients from all payer sources, pooling patients across several measures of quality, instead of just one, and accumulating data over a two- or three-year period.

Sufficient statistical power might also be achieved by aggregating smaller practices into networks comprising 50 or more physicians, they said.

In an accompanying editorial, Donald Berwick, MD, MPP, of the Institute for Healthcare Improvement in Cambridge, Mass., said the study findings pose a dilemma.

“As illogical as it is to act as if all physicians were ‘above average,’ there is almost no choice but to do so if there is no way to discern differences among them,” he wrote.

“That is true no matter the intended reason for discerning these differences: to choose, to reward, to punish, or (my favorite option) to learn.”

In addition to the suggestions Nyweide’s group made for overcoming statistical limitations in comparing quality between practices, Berwick said directly asking patients about their care could be another approach to evaluating performance.

Nyweide and his colleagues said CMS has been overseeing a series of value-based purchasing initiatives to stimulate improved quality and lower costs in primary care.

One problem is that physicians care for a variety of patients, making it unlikely that their individual caseloads would be large enough to tease out statistically reliable performance data for specific measures.

Nyweide’s group examined the possibility that evaluating data on the practice level would overcome this obstacle.

They used claims data and a database containing physician information to determine whether the approach would allow for the detection of 10% relative differences in ambulatory costs, rates of mammography for women 66 to 69 years old, and hemoglobin A1c testing in 66- to 75-year-olds with diabetes.

The researchers also looked for between-practice differences in the preventable hospitalization rate and the 30-day readmission rate after discharge for congestive heart failure.

Out of the 71,980 physicians included in the analysis, 60.6% were solo practitioners and 93.8% worked in practices of fewer than six physicians.

Median caseloads ranged from a median of 170 Medicare patients for solo practitioners to 13,400 Medicare patients for practices containing more than 50 primary care physicians.

Looking at one-year caseloads only, virtually no practices with fewer than six physicians had a caseload sufficient to detect a 10% relative difference in costs or either of the quality measures.

The number of practices with sufficient caseloads grew as the number of practicing physicians increased.

For practices with 21 to 50 physicians, 94.4% had enough cases to determine differences in costs, 84.6% for mammography rates, and 64.3% for compliance with hemoglobin A1c screening.

All practices with more than 50 physicians had adequate caseloads to detect 10% relative differences in all three variables.

None of the practices in the analysis, regardless of size, had caseloads sufficient to detect meaningful differences in the two clinical outcomes, preventable hospitalization or 30-day readmission after discharge for congestive heart failure.

“In the absence of performance measurement approaches that amass larger numbers of eligible patients at the physician or practice level, the results from this study call into question the wisdom of pay-for-performance programs and quality reporting initiatives that focus on differentiating the value of care delivered to the Medicare population by primary care physicians,” Nyweide and his colleagues concluded.

Visit for more practice management news.

View 1 Comments >

Most Popular

✓ Join 150,000+ subscribers
✓ Get KevinMD's most popular stories