Wednesday, April 16, 2008

More Wooden-Headed Reimbursements and Perverse Incentives: Medicare Proposes Penalty for Supposed "Never Events"

The Associated Press just reported that the US Center for Medicare and Medicaid Services (CMS), which administers the Medicare single-payer insurance system for elderly and disabled Americans, just proposed a new policy for paying hospitals:


Federal health officials on Monday proposed adding dangerous blood clots in the leg and eight other conditions to the list of complications that Medicare won't pay to treat if they were acquired at the hospital.

Medicare set a new precedent last year by saying it would no longer pay hospitals for treating certain "never events" — conditions that occur as a result of hospital error
. For example, if a patient were given the wrong blood type, Medicare would not pay the hospital more for the subsequent care a patient required. Originally, eight conditions were covered under the new rules, which take effect Oct. 1.

The rules proposed Monday add nine conditions....


The full list of conditions was not in the article, but can be found in a CMS press release:


* Surgical site infections following certain elective procedures
* Legionnaires’ disease (a type of pneumonia caused by a specific bacterium)
* Extreme blood sugar derangement
* Iatrogenic pneumothorax (collapse of the lung)
* Delirium
* Ventilator-associated pneumonia
* Deep vein thrombosis/Pulmonary Embolism (formation/movement of a blood clot)
* Staphylococcus aureus septicemia (bloodstream infection)
* Clostridium difficile associated disease (a bacterium that causes severe diarrhea and more serious intestinal conditions such as colitis)


That press release further asserted,


The Centers for Medicare & Medicaid Services (CMS) today proposed additional steps to strengthen the tie between the quality of care provided to Medicare beneficiaries and payment for the services provided when they are in the hospital.

CMS is proposing to expand the list of conditions which are reasonably preventable through proper care and for which Medicare will no longer pay at a higher rate if the patient acquires them during a hospital stay.

Thus, this proposal would be a form of a pay-for-performance (P4P) system, and one based on outcomes. That is, if the patient develops one of the conditions above while in the hospital (the outcomes), the hospital's performance is deemed inadequate, and the hospital is paid less.

This, in my humble opinion, seems a remarkably wooden-headed way to implement P4P for hospitals. The problems here are:

  • The outcomes are at best only partially preventable. There are no clear way to completely, or even nearly completely prevent most of these outcomes
  • The outcomes more commonly afflict sicker and more complex patients.

Let us examine, for example, one outcome about which I know something more than the others, delirium occurring in the hospital.

A Cochrane review of "interventions for preventing delirium in hospitalised patients," published in 2007, included only six studies, all in surgical settings [Siddiqi N, Stockdale R, Holmes J. Internventions for preventing delirium in hospitalised patients. Cochrane Database of Systematic Reviews 2008. Link here. ]. The review found only one study that appeared to be adequately powered. The review's abstract described that study thus,


Only one study of 126 hip fracture patients comparing proactive geriatric consultation with usual care was sufficiently powered to detect a difference in the primary outcome, incident delirium. Total cumulative delirium incidence during admission was reduced in the intervention group (OR 0.48 [95% CI 0.23, 0.98]; RR 0.64 [95% CI 0.37, 0.98]), suggesting a 'number needed to treat' of 5.6 patients to prevent one case.

It is obvious that although this intervention reduced the incidence of delirium, it hardly totally prevented it.

The review concluded,


Research evidence on effectiveness of interventions to prevent delirium is sparse. Based on a single study, a programme of proactive geriatric consultation may reduce delirium incidence and severity in patients undergoing surgery for hip fracture. Prophylactic low dose haloperidol may reduce severity and duration of delirium episodes and shorten length of hospital admission in hip surgery. Further studies of delirium prevention are needed.


Furthermore, a well-known predictive model for in-hospital delirium [Inouye SK, Viscoli CM, Horwitz RI et al. A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Ann Intern Med 1993; 119: 474-481. Link here. ] suggested that sicker patients (that is, those with "severe illness," pre-existing cognitive impairment, and a high BUN/Creatinine ratio) are more likely to get delirium.

Is delirium a "never event?" Well, hardly ever...

Thus, it appears that the surest way to avoid incurring CMS' proposed financial penalty for delirium occurring in the hospital would be to avoid admitting sicker patients who are most likely to become delirious. This, of course, is a perverse incentive that could make care less accessible for those who need it the most, and would violate hospitals' fundamental mission to care for the sick.

Similarly, I would challenge the brainiacs who came up with this proposed rule to show how any of the supposed "never events" could be reliably prevented, short of turning away the sicker patients who are likely to suffer these events.

This is an example of wooden-headed and perverse incentives at their worst, perpetrated by government bureaucrats who apparently have no understanding of the practice of medicine in the hospital.

This comes from the same folks who rely on the secretive RBRVS Update Committee (RUC) to come up with a reimbursement scheme that does not allow physicians enough time to interact with, get information and actually think about their patients. (See post here.)

Paying physicians for the time it takes to gather information, think about it, and thoughtfully come up with the best possible plan for each individual patient would do a whole lot more to improve quality and patients' outcomes than penalizing hospitals (and physicians indirectly) for events that they could not have prevented.

Woe unto us for turning control of health care's largest and most powerful organizations over to wooden-headed bureaucrats and ill-informed, self-interested, and sometimes corrupt executives!

ADDENDUM (18 April, 2008) - See comments by DrRich on the Covert Rationing blog.

ADDENDUM (22 April, 2008) - See also comments on Retired Doc's Thoughts blog.

1 comment:

Anonymous said...

As a geriatric hospitalist who is part of a geriatric fracture program I could not agree more with Roy on delirium as a "never event" (I would say it is a never never event). Delirium and other items, including cdiff colitis, are not completely preventable problems. They are not cut-and-dry and are sometimes unavoidable or may stem from otherwise necessary care (e.g. treating a patient with UTI and sepsis with an antibiotic who then unfortunately goes on to suffer from cdiff colitis).

It makes me wonder if these "never events" are sloppily placed hood ornaments on the ugly vehicle of cost-reduction. Just as pior-auths and other tactics used by HMOs broadly function as disincentives for specific treatments and/or tests these "never events" may be designed similarly. CMS may be looking for a way to ding broadly while looking "good" (reducing errors) and knowingly lumping valid exceptions in with the targets. Further, these lists can appear to be "legitimate" to an uniformed press/populace. Our government has proven to be spin-masters in foreign policy matters; it would not surprise me in the least that similar tactics are now operating in the CMS.

The cost issue is obviously no small matter. Let's face it, CMS will be unable to pay for care of the pig-in-the-python boomers if we physicians are employing current care practices. We will simply be unable to sustain a 20+ GDP% healthcare system for a generation (as this demographic bulge moves through the snake). Cuts must happen. The question is not why, it who/how/where? I am afraid we haven't seen nothin' yet; humans and their governments react more effectively than pro-act. When it comes to money, action of any type will be bluntly and forcefully applied.