Another issue confounding measure developers is that today, unlike in years past, patients are much more likely to have multiple comorbidities.
Quality measurement science, Cassel said, "has understandably focused initially on high prevalence, high yield conditions like diabetes, hypertension and heart disease," looking at one disease across time, "and have not put as much investment into composite measures" that aggregate a patient's outcomes overall.
As a result, an individual quality measure may "kind of backfire because what you might want for someone with diabetes, who doesn't have any other problems, could be very different for a patient with Alzheimer's disease, and is suffering from two or three malignancies and other kinds of issues, perhaps in a nursing home."
Cassel appeared to make the panel uncomfortable when she pointed out the danger of "overpromising" of what a good measure set can actually do because misdiagnosis "is probably 15% to 20%" of what is considered a medical error.
"A big part is making sure patients get the right diagnosis," she said. "We have no measures that tell you. All the measures we have assume that the patient comes in the door with the diagnosis on their forehead."
That prompted Sen. Patrick Toomey, (R-PA), to remark, "It strikes me that we could have a real problem measuring the final outcome of a patient's care if we don't know how well we got the diagnosis straight in the first place."