In this month’s issue of the Joint Commission Journal of Quality and Patient Safety, I (with UCSF’s Adams Dudley and the American Hospital Association’s Nancy Foster) tackle this provocative question. The answer may surprise you: yes (probably).
The devil will be in the details.
I hope you’ll have a chance to read the full article (the Joint Commission Journal has made it freely available to readers of AHRQ Patient Safety Network, so you can get it here). But the bottom line is this: in our judgment, a policy of withholding payment for adverse events is reasonable when the following criteria are satisfied:
- Evidence demonstrates that the adverse events in question can largely be prevented by widespread adoption of achievable practices.
- The events can be measured accurately, in a way that is auditable.
- The events resulted in clinically significant patient harm.
- It is possible, through chart review, to differentiate adverse events that began in the hospital from those that were “present on admission” (POA).
Are the Adverse Events Preventable?
Well, how preventable is preventable? Complete preventability is too high a bar (virtually nothing is completely preventable); a reasonable threshold might be that the majority of the adverse events can be prevented with strict adherence to a set of safety practices. For some of the events on the Medicare list, particularly the infections (such as catheter-related bloodstream infections), there is good evidence that the vast majority of events can be prevented. For others, such as pressure ulcers and falls, although some commonsensical practices have been widely promoted (particularly through IHI’s 5 Million Lives campaign), the evidence linking adherence to “prevention practices” and reductions in adverse events is tenuous. These adverse events should stay off the list until the evidence is stronger.
The Events Can be Measured Accurately
This too is tricky stuff. For example, a recent article on ventilator-associated pneumonia concluded that the measurement challenges were so severe that VAP rates should not be used as quality or safety measures. Partly because of the absence of standard definitions, the opportunity for shenanigans is rising, and – once we link payments to error reports – it may soon be Halloween every day. As but one example, I received a call a few months ago from an Indianapolis newspaper reporter. Indiana has become one of 27 states that now requires hospitals to report any of the National Quality Forum “Never Events” to state authorities. What did I think, asked the reporter, of the fact that the two biggest hospital systems in the state had wildly disparate numbers of major errors: one had 15, the other none. “I wouldn’t be caught dead in the one with zero,” I said. “Either the hospital is lying or front line workers don’t report their errors to management.” Until standard definitions are coupled with the threat of audit, gaming is inevitable.
The Events Resulted in Significant Harm
One might argue that this is not a big deal when it comes to the new Medicare policy – since the “stick” is withholding the extra payment usually associated with a “Complicating Condition.” In other words, a fall that resulted in only a skinned knee shouldn’t have been coded as a complicating condition in the first place; a threat to withhold an extra payment should be moot. But this will be an important principle as the “No Pay Errors” list becomes the subject of public reporting, malpractice litigation, and more. All adverse events are not created equal.
Can We Accurately Determine Whether the Adverse Event Was Present on Admission (POA)
Just like the Hertz guy sniffs around your car before your drive off the lot, we’ll be doing a lot of “checking the fenders for dings” at the time of admission. This may be the most problematic aspect of the new policy – the one most subject to gaming and likely to yield inappropriate care. Take these two examples. First, is there any doubt that “early decubitus ulcer” will be come the nation’s fastest growing new diagnosis by 2009? And, how many nursing home patients will get a urinalysis on admission, be diagnosed with “UTI” based on some white cells, and receive unnecessary antibiotics (for their perennially colonized urine) in part to justify the diagnosis of UTI POA.
So, with all these caveats and cautionary notes, how did we come to the conclusion that the new Medicare policy might well be a good idea? Because – since the policy was announced last year – we’ve already seen hospitals putting far more resources into trying to prevent line infections, falls, and decubs than they were before. And remember that the dollars at stake are relatively small: the extra payments for “Complicating Conditions” (CC) are not enormous, and many patients who have one CC have more than one; in which case, the hospital will still receive the extra payment even if the adverse event-related payment is denied. So, in essence the policy is creating an unusual amount of patient safety momentum for a relatively small displacement of dollars – a pretty clever trick.
Moreover, there are certain aspects of the patient safety movement that just make sense to patients, politicians, and the press. Computers seem like an awfully good idea, particularly if you’ve seen doctors’ handwriting. Residents shouldn’t work 48 hours straight, to the point that they become human zombies. And paying hospitals more for their mistakes is just plain dumb. Unless fixing these “no brainers” is completely wrongheaded, we should save our bullets and accept the idea that the system will need to address these areas. That’s not to say that we should endorse silliness or substantial negative consequences – we do need to be sure that the details make sense and that the cure isn’t worse than the disease. But for this particular Medicare policy, I predict that the overall impact will be surprisingly positive.