Outcomes measures v. process measures –be careful what you wish for
Posted Jul 08 2010 8:29pm
Probably once a week or more I hear someone lament that most medical quality measures are process measures (e.g., whether someone was vaccinated) rather than outcomes measures (e.g., whether they got better). The assumption is that outcomes measures are much superior and that once the measurement field moves to such measures we’ll be in better shape. After all, we care about how things turn out rather than what processes were in place, right?
But as far as I’m concerned, properly designed process measures are actually better than outcomes measures. If we’re measuring the quality of a physician, medical office or hospital we should care about what they are able to achieve with the patients they have. Too often outcomes measures reflect other characteristics of patients such as socioeconomic status. This goes beyond the typical adjustments for severity of illness, which are already hard to do properly.
Think about an analogy with higher education. Harvard College has great outcomes measures. Its graduates earn high starting salaries, get into great graduate schools, etc. But Harvard starts with well-prepared, intelligent students, many of whom would do well even if Harvard expended little effort on their improvement. What we really want to know is what value Harvard adds to its inputs (aka incoming students); in other words what processes it undertakes and how well. Harvard has better outcomes than Podunk U, but how would Harvard’s outcomes look if it had Podunk’s students? And how would Podunk’s look if it had Harvard’s students.
Rather than looking at process measures as an interim step on the way to outcomes measures, I’d like to see more effort put into perfecting process measures. Get the processes right and the outcomes will follow as a matter of course.