• E-mail
• Print
• RSS

Measuring process vs. outcome: Is the pendulum swinging back?

Patient Safety Quality Monthly, November 15, 2007

Want to receive articles like this one in your inbox? Subscribe to Patient Safety Quality Monthly!

At our recent seminar about public accountability for quality in Palm Beach, I was teaching the attendees about the Donebedian quality measurement triangle of structure, process, and outcome. Many of you are familiar with this. But for those who are not, the concept is that although we would like to measure quality by measuring the outcomes of care, this is often difficult to do. In the 1960s, Dr. Donebedian proposed that instead of only measuring outcomes, one can measure either the structures or the processes that have been proven to strongly influence those outcomes to evaluate quality.

This idea has provided the basis for many of the publicly available quality measures today. The outcomes data movement in the late 1990s ran into difficulty with the acceptance of risk-adjusted outcomes data from claims databases and with actually getting the real outcomes of care that occurred outside the hospital, such as inpatient wound infection rates. As a result, much of our healthcare data collection turned to measuring processes based on evidence-based medicine, such as core measures, to change medical practice. The advantage of process measures is that they are easier to collect than outcomes and do not need to be adjusted for severity of illness because all patients need the process performed reliably. These publicly available process measures are evaluated based on how you perform compared to the rest of group (i.e., percentile rank). 

During our discussion, an astute medical director of quality of a multihospital system raised the issue that perhaps the pendulum had swung too far and may need to start swinging back to a greater emphasis on outcomes data. He pointed out that because core measure compliance for some measures are at such high levels, there is little real differentiation between the organizations when it comes to the process measure, so only outcomes data will show if the process really makes a difference. As a result, we will need to refocus our investment of obtaining real outcomes data beyond just the easily available claims data.

I agree with his analysis, but with one caveat. We are beginning to see more specialty-specific databases beyond the cardiac surgery and cardiology areas, such as the NSQIP database developed through the American College of Surgeons that is designed to provide normative and severity-adjusted data. There is also a movement among some states to collect actual postoperative infection data. Finally, as more hospitals increase their electronic medical records, it will be easier to obtain outcomes data beyond just that found in claims data.

However, the caveat is that process measures that have achieved their goal of high compliance will probably be dropped and replaced by new measures of evidence-based practices measures that need our attention. And, unfortunately, there is no shortage of those.

So the pendulum may be starting to swing back, but I don't believe it will return to where it started.

Want to receive articles like this one in your inbox? Subscribe to Patient Safety Quality Monthly!

• E-mail to a Colleague
• Print
• RSS
• Archive