Last week, Medicare proposed nine additional “do not pay” conditions, several months before implementing the first eight. I like the concept of not paying for preventable adverse events, but this new list is a case of too far, too fast.

In my previous review of the new policy (here and in this article), I described four conditions that should be met for an adverse event (AE) to be included on the “do not pay” list:

• Evidence demonstrates that the AE in question can largely be prevented by widespread adoption of achievable practices.
• The AE can be measured accurately, in a way that is auditable.
• The AE resulted in clinically significant patient harm.
• It is possible, through chart review, to differentiate an AE that began in the hospital from one that was “present on admission” (POA).

Take, for example, catheter-related bloodstream infections (CR-BSI), one of the eight original AEs slated for “do not pay” implementation by the Centers for Medicare & Medicaid Services (CMS) later this year. CR-BSI has a generally accepted definition and causes significant mortality and morbidity. POA is only an issue for patients transferred from other healthcare facilities; most will have fever or leukocytosis as clues to an infection. Most importantly, a group of relatively straightforward interventions has been demonstrated to nearly eliminate these infections, both in small studies and in Michigan’s Keystone project involving over 100 ICUs. Given this, there is strong justification to withhold additional payments when a patient suffers a CR-BSI.

But check out the new list of nine proposed “no pay” diagnoses:

• Surgical site infections following certain elective procedures
• Legionnaires’ disease
• Extreme blood sugar derangement (including hypoglycemia and diabetic ketoacidosis)
• Iatrogenic pneumothorax
• Delirium
• Ventilator-associated pneumonia (VAP)
• Deep vein thrombosis/Pulmonary Embolism
• Staphylococcus aureus septicemia
• Clostridium difficile associated disease

Before you get indignant about this list, remember that the idea is that the hospital will not be able to claim the AE as a “complicating condition,” which would increase its Medicare DRG payment. The hospital will still get paid for the hospitalization – it’ll just be as if the AE never happened for the purpose of DRG calculations.

Of course, that’s small consolation when the treatment of the AE costs the hospital tens of thousands of dollars, as it does in cases of pulmonary embolism and staph septicemia. And it is a near certainty that any AEs that Medicare puts on “no pay” lists will be the subject of public reporting, making the hospital with high rates of delirium and DVT look unsafe (even if it's not).

With that in mind, as I look over the list of nine, I can’t find a single entity that meets my four conditions. Yes, many surgical site infections are preventable with perfect technique and antibiotic prophylaxis, but they suffer from non-standard, highly variable definitions. Ditto VAP. I can’t say I’ve seen a lot of hospital-acquired diabetic ketoacidosis, but I can live with the hospital not being paid extra if I ever do. Hypoglycemia – this measure will discourage hospitals from trying to achieve tight glucose control. As I noted previously, a more holistic quality measure might assess the amount of time that patients are kept in normoglycemic range, with points deducted for hypoglycemic episodes.

That’s it for the reasonable ones. What’s up with Legionnaire’s disease, which is usually community acquired? If it ends up on the list, you can bet that every pneumonia patient will have a Legionella antigen checked on admission to catch POA, a real waste. Delirium -- expect it to vie with “early decubitus ulcer” for the title of America’s most common admission diagnosis. And does anybody believe that all cases of DVT or C diff colitis are preventable? I did a quick PubMed search on "Clostridium difficile" and "prevention" and could not find a single intervention trial demonstrating that the rate of this infection could be lowered in hospitalized patients.

I can’t argue with the premise – many of the AEs on this list are no doubt partly preventable with more religious implementation of certain safety practices (for example, for C diff, avoiding unnecessary antibiotics and adhering to strict infection control practices with suspected cases). But they are nowhere near ready for prime time. Adoption of this new list will lead to all kinds of gaming, POA shenanigans, wasted effort on preventive strategies with no supportive evidence, and nasty unintended consequences.

Most importantly, while the initial list of eight appeared to represent a good faith effort by CMS to move the safety ball down the field, the new list looks like a Medicare cost-cutting effort clothed in the garb of patient safety. It is a major league overreach, and CMS should hit the pause button before it goes too far.