There has been growing interest in a significant change in payment methodology in health care, one called pay for performance. It’s a simple and alluring concept: physicians and other health care organizations will be paid better when their care meets or exceeds certain as-yet-unspecified quality indicators. The medical blogs, such as Medpundit, DB’s Medical Rants, and Galen’s Log, have been discussing this issue for a while. In general, there is some interest among physicians in the idea, although the wariness is high. Most concerns revolve around the variables in the physician-patient relationship (the non-compliant patient, the patient who cannot afford their medications), the specifics of the quality standards, and concerns about another bureaucratic boondoggle adding mountains of paperwork to physicians’ desks in return for meager increases in reimbursement–or no reimbursement, with cuts for those not meeting the QA standards. Access issues have also been discussed: will physicians start avoiding or dumping their higher-risk or less compliant patients in such a system? The law of unintended consequences (a corollary of the Law of Rules) seems certain to result in unexpected–perhaps disastrous–sequellae to such a large systemic change in payment for health care services.
A recent situation raises another aspect for concern under this payment system: the problem of systemic variables. Consider the following scenario which recently affected the care and outcome of one of my patients:
- Local hospital–a member of a small non-profit chain–has been contracting with a single service provider for mobile surgical laser services. High quality equipment, superb technicians, prompt service and excellent outcomes are the norm for this business. Business provides service at only two of the chain’s three hospitals, however, because of logistical and distance constraints.
- Hospital middle-managers decide, for reasons of contract pricing and simplicity of negotiations throughout the chain, to contract with a larger competitor for laser services. This process occurs at the administrative level only, with no input from the clinicians who utilize the services. It is presented as a fait accompli. Numerous physicians complain, but are assured that this is a high quality service and that their feedback is “valuable” and “appreciated”. But no trial period is written into the multi-year contract; it’s a done deal.
- Uncomplicated cases appear to go smoothly with the new service: so far, so good. Techs are spotty but generally competent.
- Mr. Jones (not his real name), an 81-year-old gentleman, presents to me with blood in the urine; evaluation reveals an enlarged prostate and a very large bladder stone–larger than a golf ball. Surgery is scheduled, requesting the laser service, which will be used to break up the stone, avoiding open surgery. He is a relatively high-risk surgical candidate, with a history of recent heart attack and angina.
- The OR scheduler notifies the laser service, but does not mention the stone is a large bladder stone, rather than a much smaller ureteral stone. The laser service also does not ask about the type or size of stone.
- Patient undergoes anesthesia, and laser service arrives–with a woefully underpowered laser. After a few minutes of use at full power, the laser overheats and shuts down, requiring 5-10 minutes to cool before restarting. This happens repeatedly. Not dangerous as such–but very time-consuming.
- The procedure–normally taking 30-45 minutes, takes over 2 hours to complete. Consideration for early termination of surgery is given, but this would require another surgery–with the risks of another general anesthesia in a high-risk patient–and the risk is judged to outweigh the benefit of stopping surgery.
- Because of the markedly prolonged surgery, the patient absorbs a large amount of hypotonic fluid–a problem which never occurs in shorter cases. This results in low serum sodium and other problems of blood electrolytes, hemolysis (breakdown) of red cells, acute renal failure, mental status changes, several consultants required to manage the resulting problems. The patient leaves the hospital 12 days later–rather than the next morning, as would be the norm. Fortunately, there were no permanent medical impairments.
- Cost of hospitalization is at least 10-12 times the norm. In any QA system, this qualifies as a very poor outcome. Had I terminated surgery early, he would have gone home the same day, but required another surgery for the same problem–another adverse outcome–and perhaps had cardiac complications under a second anesthesia. Checkmate.
Herein lies one of the many difficulties with the concept of pay for performance: not merely the variables of physician or hospital care, or patient compliance, or co-morbidities (other health problems affecting the management of the primary problem), but systemic problems which are entirely outside the control of the health care provider, and which are notoriously difficult to track down and eliminate.
Let’s take the above case as an example. In any well-run business, the process would go something like this: You are evaluating a change in contracted service, required in the production of your primary product, to see if you can cut costs while improving–or at least not damaging–the quality of your core product. You take historical information about the costs of your existing service, both direct and indirect. In other words, you look not only at the direct costs of paying for the product or service, but at the costs incurred by using this service–how many defective items need to be returned? How often did the contracted business fail to provide timely service, or provided inferior quality or defective goods? What are the costs associated with such shortfalls, and do they fall within an acceptable range?
In health care, however, evaluation mechanisms such as these do not exist–or when they do, function imperfectly, examining at only a small portion of the larger process. The mechanisms are further constrained by the nature of health care. You may accept a 1% rate of defective widgets components, knowing that your internal QA will catch most of those and your core widget product will remain at acceptable quality. How many additional deaths or complications are you willing to tolerate when changing medical services or supplies, in order to save money which may be used to provide care to other patients? Who is even willing to do this kind of analysis, knowing that hordes of lawyers are hovering, panting at the opportunity to find evidence that the evil greed of your hospital contributed to Aunt Nellie’s early demise?
My patient’s case illustrates this problem in multiple ways. The decision to change surgical laser services was made at an administrative level–no physician who actually uses the service was consulted. The contract change was made on the basis of up-front costs (how much is paid for each instance of the service) and administrative efficiency (having a single contract covering multiple hospitals rather than multiple, separately-negotiated contracts). The contract will likely have some clauses paying homage to “quality of care”–a pure CYA maneuver, which does not specify what constitutes “quality” care. Good thing, too–since the laser service has little or no access to outcome data for the services it performs. In over 15 years of using contract laser services at multiple hospitals, I have never once been asked by these services–or by the hospital–to provide outcome data, such as complications, mechanical problems during surgery, re-operation rates for poor outcomes, etc. Even if they started today demanding such accountability from contract services, there would be no historical data with which to assess the quality of a new service provider in contrast to the old.
At the clinician and hospital level, quality feedback breaks down again. Several years ago, a local hospital decided to change brands of urinary catheters–a medical supply used in great quantity in hospitals, where the savings of even a dollar or two per unit can result in millions of dollars annually in cost reductions. On a day-to-day basis, the minor reduction in product quality was insignificant: they worked fine for Aunt Millie when she need a catheter for a few days to monitor her fluids accurately while managing her congestive heart failure. But when Uncle Ernie had prostate surgery, and developed a few small clots from bleeding, the inexpensive soft-walled catheters failed miserably, their sides collapsing like a shantytown in an earthquake when irrigated–greatly increasing the chances that old Ernie would have to be returned to surgery to be rid of his clots, rather than have a simple procedure by the nurse in his hospital room. It’s the outliers that kill you in these situations, and they’re devilishly hard to predict when a middle-manager sits down to save a few cents on a high-volume supply item. And even when they’re detected, they are easy to blame on other factors–the patient’s condition, nurse or physician error or neglect, etc. And they’re even harder to change once problems arise.
The urologists who need the higher-quality catheters on a regular basis screamed bloody murder from the outset: their patients’ care was hugely affected by this seemingly benign change–but no one was listening. Or perhaps more accurately: no one was responding. The hospital set up an 800 line for physician complaints–but years later, the lower-quality catheters remain a stock item. Like the button on a streetlight at the crosswalk, pressing it made you feel like you were doing something–but the light changed at the same time regardless of how often you pressed. The physicians who did complain were dismissed as troublemakers, or whiners. No one knows how many patients suffered needless discomfort, or unnecessary returns to surgery, because of this simple, seemingly cost-effective change–which has almost certainly cost the hospital far more than it saved in hidden costs, not to mention harming patient care quality. And that’s exactly the problem: no one knows.
So there you have it–pay for performance is a noble idea destined to fail in the health care system. The necessary means and mechanisms do not exist to monitor quality in a whole host of areas–and in areas where they do, conflicting factors such as patient privacy, medical liability risk, and the inertia and nearsightedness of large organizations will thwart any progress toward actual quality improvements using financial incentives. Slapping a “free market” reform as a band-aid on the huge, micromanaged, economically-handcuffed health care system dinosaur is an invitation to create a new Jurassic Park. But of course, that is unlikely to prevent it from taking place.
Coming soon, to a hospital near you…