A doctor described to a colleague a long-term patient who he had seen earlier in the week. She has diabetes that can be controlled but has failed to come into his office regularly even though he has contacted her many times. The doctor is highly ranked on quality measures that the local health insurer has laid out for evaluating and paying physicians to improve medical care and cut costs. Yet the doctor asked this colleague about his infrequently seen diabetic patient: “She just can’t afford to take that much time off from work. Does that make me a worse doctor?”
In pay-4-performance plans established by Medicare and private health insurers, the basic assumption is that, yes, what a doctor does has a direct effect on patients’ health. But what patients do or do not do is not part of that assumption. For example, many of the measures used to rank and reward doctors come from guidelines drawn from extensive research studies that are designated as “best” practices such as patients getting regular mammograms, Pap smears, screening for high cholesterol, diabetes, high blood pressure, and colon cancer. Health insurers want physicians to use these evidenced-based practices to standardize delivery of quality care, improve health of patients, and reduce medical costs. Health insurers, at the same time, also evaluate and pay doctors on the frequency in use of these evidenced-based practices. Billions of public and private dollars are now invested in evaluating doctors’ performance and paying them. Like so many businesses, the belief that were individual incentives distributed on the basis of performance employees and professionals will work harder and do the right things for their customers and clients.
For those familiar with the trajectory of well-intentioned policies aimed at changing individual and institutional behaviors, unintended consequences occur as predictably as windy days in Chicago. Policymakers are stuck, however. They just don’t know which unexpected consequences to anticipate. Although when it comes to using money as an incentive to change behavior, much literature exists on what happens when results for hospitals, surgeons, and medical procedures are reported publicly. Some astute decision-makers might have figured out ways that any such policy could be gamed by individuals and institutions (see WillP4PandQualityReportingAffectHealthDisparities).
No surprise, then, that unexpected consequences have stuck thumbs in the eyes of policymakers and insurers on pay-4-performance plans. For example, Medicare requires—as a quality measure—that doctors administer antibiotics to a pneumonia patient within six hours of arriving at the hospital. As one physician said: “The trouble is that doctors often cannot diagnose pneumonia that quickly. You have to talk to and examine the patient and wait for blood tests, chest X-rays and so on.” What’s worse, he continues, is that “more and more antibiotics are being used in emergency rooms today, despite all-too-evident dangers like antibiotic-resistant bacteria and antibiotic-associated infections.” He and other doctors know that surgeons have been known to cherry pick reasonably healthy patients for heart bypass operations and ignore elderly ones who have 3-5 chronic ailments to insure that results look good.
Also medical researchers know far more about the effects of low-income, under-insured, and non-English speaking patients when doctors are ranked on the quality of care they render especially if rewards or penalties follow such rankings. One study involving 125,000 patients revealed that who doctors cared for affected their rankings for pay-4-performance plans. Those doctors who cared for older or sicker patients were ranked higher (probably because of frequent follow-up, behaviors that received higher rankings) than those doctors who cared for minority and under-insured patients who saw doctors irregularly. Which patients are cared for, then, affects rankings.
In light of emerging evidence that untoward outcomes occur when cash incentives are put into place to evaluate, rank, and reward individual doctors, pay-4-performance schemes have raised disturbing questions not only about the basic assumption that what doctors do determine effects upon patients but also about the assumptions driving pay-4-performance plans for individual teachers based upon student test scores.
Although both doctors and teachers are in helping professions, many differences separate them from one another in their training, daily work, the scientific basis for what they do, societal respect, and accountability. Nonetheless, there are some striking similarities insofar as evaluation and pay-4-performance policies. The next post takes up these similarities.
Larry Cuban on School Reform and Classroom Practice 
I really enjoyed your article and agree that it’s nonsense to base pay on outcome when we don’t have control of that outcome for so many reasons. We have control of our own practice as teachers or as doctors, but not what our kids or patients do before and after they are with us. I was going to post this on my facebook, but you don’t have a facebook link. I’m sure there’s a way to do it without that, but I thought you might consider adding one! Thanks for your piece!
I think a better analogy would be nursing. Comparing doctors, a highly esteemed and well-paid male-dominated field (might I add with the strongest professional lobby in the country) with teachers (a low-esteemed, female-dominated field with a union that is looked down upon everywhere and especially now and with loads of male administrators and policy makers and researchers overseeing her actions in the classroom) might be less strong than one with nurses.
Sondra,
Your comment on nurses got me thinking about female-dominated helping professions including social workers, librarians, and, of course, teachers. As you suggest, their status is lower than physicians, lawyers, and academics. Evaluating these professionals and pay-for-performance plans for their effects on patients, clients, and college students have really stirred up passions about the narrow measures used and the simplistic formulas to determine “success” in even these fields.
Comparing teachers to nurses is sensible insofar as the similarities you noted but hospital and group practice nurses cannot be singled out for their effects (except in instances of neglect) because they are carrying out doctors’ orders in caring
for patients.
Excuse me, but the strongest professional lobby in the country is the lawyers. Doctors have been regulated up the ying yang (I am not one, by the way), and lawyers still go on their merry way destroying businesses, suing doctors, and taking home millions while their clients get the scraps. Every new regulation the government creates – including in the healthcare legislation – is a new goldmine for the lawyers.
A similar analogy was made in EPI’s report on using student test scores to evaluate teachers last August: http://www.epi.org/publications/entry/bp278
Thanks for pointing this out to me.
On the whole, I agree that the doctor pay-for-performance discussion has a lot to learn from the similar debate that has been going on (for a substantially longer period of time) in the education world. However, a few differences worth noting:
1) Doctors are paid on a fee-for-service basis, unlike teachers who are salaried. Therefore, there is absolutely no benefit besides satisfaction of personal integrity, a nebulous professional ethic, and provider reputation (which is far from a transparent process) that would lead doctors to even take the time out of their busy schedules (when they already have many other “non-problem” patients that are coming in for their appointments — which is what doctors get paid for) to follow-up on the problem patient that is not coming in for their appointment. So, pay for performance is a way of incentivising that sort of follow-up and holistic care that would be better for patients and the type of care that many physicians would like to provide but can’t afford because of other market pressures demanding their time.
2) Much of the “gaming” of the system that you worry about already occurs, even without a pay for performance system. It’s called cherry-picking and everybody does it in order to keep their practice afloat. Because public insurance, for example, tends to reimburse so much less (and even below cost) of private insurers, most hospitals and clinics that are not federally-subsidized to care for low-income/problem patients will try to attract more of the higher-income/less-problematic patients. This already happens, regardless of pay for performance. Whether pay for performance exacerbates this or not is complicated, but it’s hard to say that it necessarily makes things worse when the secular trend is to already tend towards cherry-picking.
All of that makes me sound incredibly negative about doctors and the health care industry and I’m sorry about that. But the fact is that if you look at the incentive structure that is currently in place, it’s actually a miracle and a matter of great personal and professional integrity that health care is not more profit-driven than it already is.
Emily,
Thanks for comments on fee-for-service vs. salaried (although more and more physicians are becoming salaried but not yet the majority as in Canada and European health care systems). The issue of “cherry picking” and your point as to its existence prior to pay-for-performance is one I had not given much consideration to. Thanks. In the next post, I will compare doctors and teachers so your comments are most helpful.
Pingback: Cuban on pay for performance | Chris Osmond PhD
If it’s truly pay for outcome, and they’re punished financially if they treat a patient who dies, doctors will refuse to treat the sickest patients. This sets up a conflict of interest. Most doctors I know care about their patients, but this would be a huge disincentive to try to save someone who has a low chance of survival.
Pingback: Perfidy | Chris Osmond PhD