The Quality of Quality Data

Outcomes data can be confounded by patient factors like burden of illness, and they are time consuming and expensive to collect.

That’s why many quality measurement programs focus on process quality and cross their fingers that the link between process and outcomes actually holds true.

Uh, guys…can we huddle up a minute?

Leslie Kernisan and colleagues at UCSF just finished a study to determine whether performance on a set of process measures known as Safe Practices for Better Healthcare could predict hospital mortality.

They couldn’t.

Hospitals with higher scores on the Safe Practices indices did not have lower inpatient mortality than those with lower scores. Progressing from the worst to the best quartile on Safe Practices, inpatient death rates bounced around from 1.97% to 2.04%, to 1.96% to 2.00%.

Data from more than 1,000 hospitals were analyzed. These data were adjusted for patient and hospital variables that could impact the results.

“It is possible that inviting hospitals to self-report on their patient safety practices and then assigning them to quartiles of score is not an effective way to assess hospital quality and safety,” stated the authors.

That may be, but it’s more likely that inpatient mortality rates are not sensitive enough as measures of quality. Before throwing in the towel on these process measures, scientists would be wise to see how well they predict other, more sensitive outcome measures like readmission rates, functional status at discharge and 6 months, and so forth.

Then again, they could be lousy process measures, or the data collection tools could have been flawed.

The Safe Practice Guidelines have been endorsed by the National Quality Forum. The write-up appears in JAMA.

 Print This Post

 Email This Post