Measuring and Paying for Performance and Outcomes

Measurement is the first step that leads to control and eventually to improvement.  If you can’t measure something, you can’t understand it.  If you can’t understand it, you can’t control it, And if you can’t control it, you can’t improve it.

Dr. H. James Harrington (born 1929),  Statistician, entrepreneur, and improvement guru

It all sounds so logical, compelling, and scientific.  Compile enough measurement data,  measure physician performance,  measure patient outcomes,  and measures of quality of care will improve. 

In the words of George Halvorson, CEO of Kaiser Permanente,  “"We are on the cusp of the golden age of healthcare delivery,” He went on to explain that the toolkit to improve patient care, "is getting better every day. We have better technology, better connectivity, better databases, and better science. We have better opportunities to interact with patients to help them improve their health.”

This is very persuasive rhetoric, at least until you get the details, namely,   what measurements, what performances, what outcomes, and what rationale? 

Let’s take hospitals first.    The Centers of Medicare and Medicaid (CMS) has seized upon the idea if we could only measure the rate of 30 day readmissions  among American hospitals,  we could improve care before and after discharge.  

How?  Well,  we could give patients more instructions on  what to do once out of the hospital,  we could coordinate the transition to home and rehab facilities.  We could have the hospital take charge of care once patients have left the hospital. And we could punish those hospitals with high readmission rates by lowering their Medicare payments.  

As I see it, there are a couple of problems with this approach, rational and admirable as it may be,  Most discharged patients who are readmitted have serious chronic diseases only temporarily altered by their hospitalization.   Most cash-squeezed hospitals do not have the resources to take care of patients outside the hospital environment.     Patients,  many of them elderly,  return to the environments and behaviors  that caused their illnesses in the first place.

You can use big data to measure these problems, which are well known, but measurement does not necessarily change the realities and dynamics on the ground.

Consider physician performance.  As   Regina Herzlinger, PhD, professor of business administration at Harvard Business School and the “godmother of consumer driven care,” in Who Killed Health Care (McGraw Hill, 2007),  comments, “ Congress is now practicing medicine. Its pay-for-performance (P4P) initiatives enable governments to tell health care providers how to practice medicine. The higher the performance, the higher we pay. The health care system lacks metrics of performance. Despite its name, P4P does not pay for performance – the attainment of improved care at a reasonable price. Instead, it pays for conformance – the adherence to a government-dictated recipe for the provision of health care. The government pays for adherence for its recipes for the process of delivering health care rather than for outcomes.”

And as of a January 26, 2001 report in Reuters, pointed out, “ Paying doctors financial rewards to meet targets for improving the care of patients made no discernible difference to the health or treatment of people with high blood pressure, a study has found.”

“The findings suggest governments and health insurers across the world may be wasting billions of dollars on doctor incentive schemes but getting no improvement in patient care, researchers who conducted the study said.”

“Researchers from Britain, the United States and Canada assessed the impact of incentivised targets on quality of care and health outcomes in around 470,000 British patients with hypertension and found that they had no impact on rates of heart attacks, kidney failure, stroke or death.”

As it turns out,  Medicare and Medicaid, and  medical societies,  can’t yet figure out what clinical outcomes to measure.  Two obvious measure targets are diabetes  and obesity.   You can measure hemoglobin A1C,  which declines  when  patients lose weight, exercise, take their medicine, and eat properly.  All of these activities depend on how patients behave as outpatients,   not necessarily on instructions they receive in doctors’ offices,  Obesity is similar.

Measurements of hemoglobin A1C and body mass indices could serve as signposts of improvement ,  and  teams of primary physicians,  dieticians, nurses, and others visiting homes or in telephone contact could improve outcomes.  But diabetes and obesity are only two clinical conditions and account for a  fraction of chronic diseases requiring improvement.   

Studies to date have shown only modest improvement when doctors are paid for  improvement.  Maybe with big databases , studies of population health,  closer interaction with patients based on information technologies,  these numbers will improve.

Tweet:  Measuring and paying for performance and outcomes promise   to improve healthcare but they have yet to significantly improve results.