One of the local hospitals is undergoing a Medicare audit of Implantable Cardioverter Defibrillator (ICD) implants, and, in order to avoid fines, is preemptively paying back the government for “questionable” implants. ICD implant guidelines are quite strict and are based on the results of randomized controlled multi-center studies. Each hospital has a checklist that the implanting physician must fill out and sign, ensuring that every patient who gets an ICD that will be paid for by Medicare meets strict criteria. Unfortunately for both physician and patient there are gray areas — corner cases that don’t quite fit in to the guidelines. Unlike some insurance companies, you cannot call up Medicare and get pre-approval for a procedure. You can go ahead and do the procedure. Medicare reserves the right to review the case later on and make you pay back their money and possibly fine you or even prosecute you for Medicare fraud. On second thought, maybe you shouldn’t go ahead and do the procedure.
The problem is, sometimes the guidelines don’t make sense. One number that appears throughout the guidelines is an ejection fraction (EF) less than or equal to 35%. EF measures the percentage of blood that is pumped out of the heart with each beat. Most studies of ICDs enrolled patients with EF <= 35%, not because these are the only patients that might benefit from an ICD, but because these patients are at higher risk for sudden death, which the ICD prevents. The 35% cutoff was chosen for a very selfish reason from the point of view of the study designers. This cutoff decreases the number of patients that need to be enrolled in the study to find a difference between ICD therapy and no ICD therapy, which decreases the cost of the study. But because these are the only studies we have of ICDs the guidelines and checklists use this number 35% as an upper limit to qualify for an ICD for primary prevention of sudden death (some of the criteria use a lower number, 30%. As I said, the guidelines are complex). Yet there is no sharp cutoff in risk of sudden death at an EF of 35%. All studies show that there is a continuum of increasing risk as the EF decreases. But a line has to be drawn somewhere, especially when someone else (the taxpayer) is being asked to foot the bill for these (overpriced?) devices. So what does the physician do with patients who don't quite fit the guidelines? What to do with a patient who has long-standing ischemic cardiomyopathy with EF of 20%, who was scheduled for an ICD (and this does meet the criteria) but then has a myocardial infarction (MI), with drop in EF to 15% (and higher risk of sudden death)? By the Medicare criteria he must wait 40 days before he can get his ICD (yes it's true, look it up). Worse, if he was unfortunate enough have gotten a stent for his acute MI, he must wait 90 days before getting the ICD! Another example, a patient with nonischemic cardiomyopathy for years with EF of 25% on repeated measurements, is hospitalized at an outside hospital for heart failure 2 weeks before his scheduled ICD implant. An echo is done and is read (rather perversely, given the ICD guidelines) as showing an EF of 36%. What now? No ICD until another 3 months of drug therapy? Is an echo really accurate to one percentage point anyway? Or a patient who presents with low EF, nonischemic cardiomyopathy, heart failure and heart block. The guidelines would suggest putting in a pacemaker (too early for an ICD per the guidelines, but the patient needs something for the slow heart rate from heart block), waiting 3 months, and then upgrading to an ICD if the EF is still <= 35%. This seems insane. If the EF does not improve, you subject the patient to 2 procedures, with consequent increased cost and risk of infection and other complications. Yet the physician who implants an ICD in any of these cases is at risk for all the penalties mentioned above. It's that pesky "Art of Medicine" that the bean counters don't have the training or imagination to understand that is the problem here. It's true that physicians in the currrent American fee for service system of medicine have an inherent conflict of interest. Physicians have an obligation to their patients to do the right thing medically. At the same time they are paid for doing procedures. I know there are some unscrupulous MDs who do not follow guidelines in order to maximize their procedure volume and profits. This is to be condemned. But what about the physician who implants a device because he or she feels it is the right thing to do for the patient, and has a sound rationale for the decision, even if the indication cannot be quite shoehorned into the rigid Medicare guidelines? Unlike physicians, the administrators of the Medicare program have no potential conflict of interest. They don't care about the patient. They don't care about the physician. They just care about saving money. It's easy to go through a checklist and see whether or not it is filled out properly. It takes no medical training. And if the physician follows all the rules and the patient ends up dying while waiting 3 months for an ICD -- well, that saves money too.