Dr. Stanley Feld at Repairing the Healthcare System has been writing his blog for over a year and most would agree that he is a very thoughtful physician. Even without an MBA or an advanced management degree (that I can tell), he presents the kind of insights that non-MD executives and business people would do well to carefully heed.
His most recent series has been on physician ratings, especially the ratings produced by the insurance industry. He follows up on an article in the Washington Post that reported physicians’ reputations beings harmed by the inaccuracies of insurance company data.
Dr Feld writes,
“In our dysfunctional healthcare system tensions are building to great heights between the insurance industry, patients and physicians. The appearance of this article makes me think that the insurance industry subconsciously has a death wish to completely destroy the Healthcare system. The insurance industry is totally insensitive to patients and physicians. Its only concern is its bottom line and seducing employers to buy its product.”
Right on the money!
Of course, one could ask what a for-profit entity should try to accomplish if not increase profits? [Full disclosure: I currently own 100 shares of United Healthcare, a company that was raked over in the Post article.]
It is doubtful that United or the insurance industry is trying to destroy the healthcare system; it is much more likely that the corporations that run health plans are pandering to one or more of their stakeholders, especially Wall Street. Employer groups, who are much more vocal than individual patients, have been clamoring for more physician ratings and greater transparency in health care. They have nothing to lose publishing imperfect physician ratings:
- If the data are accepted they get kudos for being proactive and their efforts are appreciated by wall-streeters who judge it a positive development for the company. Investors be sold on the idea that it is a worthy attempt to reduce costs (meaning payments to providers). One of Dr. Feld’s point that deserves high-lighting is that the insurance company’s real customers are human resource directors and benefits managers. These people make their CEO’s happy when they take care of their employees at the lowest cost with the fewest complaints. After all most consumers don’t get to chose an insurance company unless it’s on their employer’s menu.
- If the data are rejected, the insurance companies can simply blame physicians. They can tell the world physicians are impairing transparency in healthcare. This effectively deflects that the lack of transparency is mostly a by-product of the insurance industry working with hospitals. Be clear about this: it is not in the insurance industry’s interest to promote transparency.
I happen to believe in data and performance improvement. Collecting data is never a bad thing, as long as you keep away individuals who do not understand the limitations of numbers! The only ‘bad’ data is the data are the data that policy-makers take inappropriate action with.
One of the major objections on the part of physicians is that they do not directly control patient behavior and they can’t be accountable for the health outcomes of their patients, most of whom do not follow physician direction, no matter how meticulously it is presented. These are valid arguments against pay-for-performance, which in turn, is the logical extension of physician ratings. Our understanding of what constitutes quality in healthcare is incomplete, definitions are not really operational except in defining negative incidents, the data is imperfect and the the measures would not meet the most superficial test for validity. Despite all this, there are forces afoot that would move the process forward because that is how systems learn. Systems make mistakes, adjust, refine and correct themselves.
Another counter-argument to physicians who feel they are being held accountable for outcomes that are beyond their ability to influence is the variation in outcomes. Some doctors simply do better at changing people’s behavior and achieving certain results.
For example, it is a reasonable argument that low-income patients with poorer health status and greater health literacy needs and enormous racial and ethnic disparities would have poor health outcomes. However, data from the Health Disparities Collaboratives would suggest that Federally Qualified Health Centers can blow away the competition in most measures of health outcomes according to the chronic conditions measures. [Full disclosure: the facility I help manage participates in the Collaboratives and I have only seen data on diabetes.]
Bottom-line: some physicians appear to get better outcomes from their patients, no matter the initial health and socio-economic status. I know one doc who has spent considerable time learning about motivational interviewing and even before this training had an uncanny connection with some patients. She was capable of getting patients to do things most physicians had enormous trouble with. She could get people to quit smoking and exercise and collect body fluids for 24-hour urine tests and stool collections where most patients produced contaminated or incomplete specimens. They would get their mammograms and even come back for repeat Paps when they were unspeakably scared of what they might find. They would take their pills and could be seen walking to the mall!
Health outcomes and health behaviors are multifactorial and -make no mistake – physicians can have a huge impact. This should be reflected in pay for performance, physician evaluations, ratings etc. However, it needs to be a relatively small component of the overall rating to reflect the fact that patient behavior is not entirely determined by the physician. The patient remains part of the picture. The bulk of the rating should reflect the process: i.e. on the referral for mammography, not how many patients actually followed through. (Process measures not outcome measures.)
There are quality measures that look at processes (did you do the test?), and quality measures that look at biological outcomes (did the patient’s test result improve?). These two categories are at opposite ends of what is directly in the physician’s control and what is not. Behavioral outcomes are probably somewhere in the middle. Each of the three should have different weight in a physician rating. For example, 75% on the referral and 20% on if the patient went and 5% on the actual outcomes, but that’s pretty arbitrary.
The conundrum is that all quality measures are really proxies for quality. Nobody really knows that the definitions of quality truly represent the notion of quality. Half a physician’s value, at least in non-procedural specialties, is due to intangibles. The practice of medicine is all about relationships: if a patient has a good relationship with their doc, they’ll do better. No performance measures can capture that. We also know that patient satisfaction measures are subject to the halo effect, affected as much by the age of waiting room magazines are and the attractiveness of the staff. These are not direct measures of how good the physician is! I suspect the main reasons physicians resist and are disengaged from the process of performance improvement is the intuitive understanding that both patient satisfaction and performance improvement activities have serious limitations. Data are good, but physicians question if the data is good enough to be actionable. Experience has taught them that such efforts are not often based on good science, but can be reflection of societal pressures and the political struggles behind them. It has progressively made their lives miserable. Data is intended to be used to make policy, but rarely does data drive the policy!
Unfortunately, physicians have not had much reason to embrace quality and performance improvement. The us wastes and incredible amount of resources on healthcare and gets very little outcome on a population basis. Physicians should engage; they are a better source of solutions than we have now. Physicians can point out the strengths and limitations of data on their own practices.
If physicians are disengaged because they are skeptical, outcomes are less likely to improve. My advice is get involved and back a single horse to smooth out the standards. (There are currently over 10 organizations and agencies in the US that encourage the use of one or another standard for measuring quality of care. The National Quality Forum is my fave. The Post article points out that there are subtle differences in the wording and cutoffs of the standards, which leads to confusion.)
Moreover, once the data are part of a performance improvement project, the are often used to reward high-performing physicians and punish another. One of the primary principles put forward by Donabedian in the 1960’s was not to use such data punitively or you’ll never be able to collect valid data again! The goal isn’t to make people look good or bad, it is to improve care overall. Everyone who is working to collect data knows the data isn’t perfect and that it doesn’t need to be perfect to be useful. But it better be accurate enough to be actionable.