Capturing Meaningful Measures in Cancer

Joseph Alvarnas, MD: Is there an easy fix for the dearth of robust quality measures? My bias is that—based upon how we’re willing to measure things and the amount of administrative burden that we’re willing to take on to do so—I don’t think we’re going to get to measures that matter. When you look at capturing patient-reported outcomes that matter to patients—asthenia, insomnia, pain, depression, hopelessness, functionality—you have to think differently. EHRs and billing data don’t capture any of those things in a meaningful way. Even if you look at survival data, there’s no way that tumor registry data get connected to billed and paid data. So, even if you’re asking something as fundamental as “Do more people survive when they’re treated with X?” we’re hamstrung from doing this. It either requires a radical re-envisioning or someone in the payer space to say, “This is what we need, and we are willing to pay more to get it.”

Bhuvana Sagar, MD: Yes, I think there are lots of important points of data that we’d like to get—date of diagnosis, date when the patient was staged, and date of death. There are a lot of complicated administrative details involved there, as you just mentioned. I think to start with, process measures are great. At least we’re trying to get the providers to focus on simple things first. There are some PQRS measures. There are some pain measures. There are some depression screening measures. Are they even doing that? Are they addressing the pain? Are they filling a prescription for pain? I think we should start there, and eventually we need to get to the measures that really matter, and we need to be able to get the data. We need to be able to measure them and go forward.

Robert Carlson, MD: I would look at this slightly differently. I think defining quality measures is crucial if you have a metric that you want to move. I think the real question is, not how you define quality, but what quality measures you choose to measure so that you can feed back to the system—to the physician, to the pharmaceutical industry, and to the payer. Then you can modify or change behavior to improve things. I don’t think any of us want to measure something just to measure something. We want to measure something because we want to be able to figure out where we need to put resources and where we need to modify things.

It’s a little bit of a modification of the Heisenberg uncertainty principle in physics, where the more precisely you measure something, the more you perturb the system. If patient-reported outcomes are what we want to perturb, then we should measure those very intensely. If survival is what we want to measure or be our quality metric, then we want to focus on therapies and what therapies are going to optimize or maximize survival. I think the decision is what metric we want to move, and whatever metric we want to move is where we should focus.

John Fox, MD: I agree with you. We have to be very thoughtful about what we measure. Does it make a difference to patients? Do patients care about that measure? I think one area that we as a delivery system—and, in fact, a region—have focused on is end-of-life care and ensuring that patients have a conversation. It’s not necessarily with their physician, who may not be good at having those conversations. But they have a quality discussion with somebody trained to have that discussion, along with their family members, and measure—for those patients who died—if the care that they received was consistent with their state of preferences. We all know the kid who flies in from California not having seen their mom for the last 6 months. Did we get care that was concordant with what the patient wanted? I think that would be another critical element: not only having assessed what patient preferences are but also, did we provide care that was concordant with those preferences?