In our recent paper criticizing how industrial Quality Improvement has been misapplied to primary care, we didn’t just complain, we made suggestions for a better way forward. This was under the assumption that regulators and payers will continue to insist on some kind of numeric reporting of outcomes by physicians or practices whether physicians like it or not, or whether it’s really useful and fair or not.
In classic industrial QI, single disease targets are essentially always 0 or 100. Zero inpatient infections represents high quality, though sometimes even the hospital overlords accept a small non-zero outcome, such as a 2% post surgical infection rate. They will look at other metrics such as a process measure that 100% of patients who received elective knee replacements got antibiotics prior to the first incision. The non-physician administrators kind of get it that infections are an unavoidable reality.
We found evidence that some organizations have already adopted this approach. Intermountain Health will often accept a 5% to 15% deviation in primary care, because it at least has some inkling of the complexity of primary care. However, this is for an employed insured patient population in one of the most health-conscious states in the U.S. (Utah). This target does no good for a physician caring for a population of vulnerable low-educational Medicaid patients in Mississippi.
Extremely crude risk adjustments have been used to try to give some credit to the doctor who cares for these high-risk patients, but the adjustment factors are too simplistic to be useful. For now, for the doctor who cares for a bunch of patients who are diabetic, schizophrenic, and homeless, the numeric targets should be thrown into the rubbish bin, and that doctor should receive a massive thank you for for caring for the patients no one else will.
Well said