Is There an Answer to the "Medical Review Policy" Nightmare?

by Steven J Baumrucker, MD
This article appeared in the American Journal of Hospice and Palliative Care in a similar form.

"Through its responsibilities in administering Medicare and Medicaid, HCFA pays for the care associated with approximately 65% of all deaths that occur each year in the United States." So begins the 1995 Health Care Financing Administration position paper on end-of-life care.* Financing the care that leads to 65% of all deaths leads to some staggering numbers; it is obvious why HCFA takes such an interest in Hospice care.
Over the last few years practitioners and bureaucrats alike have attempted to reconcile personal and fiscal realities; how does one decide when a patient is appropriate for hospice care? Previously, the game (and it was, sadly, best described as such) included finessing recertification periods; if someone stayed in the program into the last recert, he could never again receive hospice care if discharged for any reason from the program. This was probably designed to prevent long stays as hospice administrators were expected to rush to discharge patients before they hit the last period. What actually happened, however, was that patients sometimes ended up in the last recert period anyway, and languished there; hospice workers were reticent to discharge someone who would probably need the service later, but could never be readmitted due to the rules of the game.
Recently, common sense was amazingly added to the mix. Gaming theory is less useful now that patients can be discharged in any recertification period and readmitted at will if they meet the criteria. All we need to do now is agree on the criteria. Since the Supreme Court's common-sense "I know it when I see it" technique is not acceptable to HCFA, we are required to use criteria that are tangible, reproducible, and therefore essentially rigid. For cancer diagnoses, which currently have a well known pattern of progression, the job is difficult enough. For non-cancer diagnoses, the problems are nearly insurmountable.
In 1996, the NHO published the Medical Guidelines for Determining Prognosis in Selected Non-cancer Diagnoses, Second Edition. The guidelines were written in part to " helpful in determining patient eligibility under the Medicare/Medicaid Hospice Benefit by defining a population that may have a life expectancy of approximately six months."* Referencing the best of the literature on disease prognosis, the NHO criteria are far from perfect. Many of the studies referenced used pooled data, or were on non-generalizable populations. The authors also astutely point out that palliative care of non-cancer symptoms is often identical to life prolonging care, e.g., diuretics for symptoms of congestive heart failure. The NHO criteria are, however, arguably the best we have, and should improve over time with further research.
There was some initial excitement that Medicare would adopt the NHO criteria en masse. This has not occurred, apparently. New proposed criteria are felt by hospice practitioners and administrators as being far too rigid and restrictive. Before the hospice movement could argue the case effectively, however, an objective demonstration of the perceived problems was required. Schonwetter, et al, in their "Review of Medicare's Proposed Hospice Eligibility Criteria for Select Non-cancer Patients"* may have provided the hospice community with enough evidence to at least get the dialog started.
The authors of this article analyzed the charts of one-fourth of the patients admitted to LifePath Hospice in West Central Florida who met the criteria of a survival time of less than six months, and who died within the first 10 months of 1997. Doing a retrospective chart review, the proper Medicare criteria were applied to each patient and eligibility status under the guidelines determined. The results are equally interesting and frightening.
The Medicare criteria did reasonably well in the patients with diagnoses of stroke and coma, correctly identifying 94% of the sample group as eligible. However, the next best score was 44% of patients with dementia. None of the six renal, fifteen pulmonary, or ten HIV patients met criteria, according to the study. (What was truly depressing about the results was that the mean time in hospice care for all patients was 30 days, not nearly the 180 days that is allowed. Late referrals to hospice is an ubiquitous problem, however, and is grist for a different mill.)
Using the guidelines for Renal Disease as an example, one notes that the definitions of renal failure are the same for the NHO and Medicare guidelines (and indeed are culled from the same HCFA form #2728); the difference lies in the small words. Where the NHO says "may", HCFA generally says "must." The Medicare guidelines state that patients are hospice eligible if they are not seeking dialysis or transplant, AND have a creatinine clearance of <10cc/min (<15cc/min for diabetics) AND have a serum creatinine <8.0 mg/dl (>6.0 mg/dl for diabetics).* Obtundation due to uremia can only be used as supporting documentation of the above; if any one of the three is missing, the patient does not qualify for eligibility. It is obvious to most practitioners in the field that a patient referred to hospice with a creatinine of 7.1 who is obtunded, oliguric, and has a GI bleed will have a very short hospice stay indeed. However, this same patient would not meet the currently proposed Medicare criteria.
The criteria for determining prognosis in Liver Disease are another story. The NHO guidelines and the Medicare guidelines should in practice qualify the same patients, as they are identical. Anyone claiming that the NHO guideline is somehow "better" in this case shows inherent bias.
Medicare does include a way to circumvent these problems. Patients who do not automatically qualify under the criteria will be eligible for review, and "...such patients may be eligible for the Medicare hospice benefit if the documentation by the physician shows sufficient evidence for a prognosis of six months or less in the absence of meeting the criteria in these policies."* This sounds reasonable, but in an age where hospice workers are asked to admit patients to their service nearly DOA due to chronically late referral patterns by the medical community at large, taking time out for a Medicare review is an unwelcome thought.
Asking Medicare to just adopt the NHO guidelines is not the answer, either. One reason that the guidelines seem more friendly is that they are intrinsically fuzzy. For example, the guidelines for HIV Disease state, "It is important to discuss a patient's clinical course with a physician who is experienced in caring for persons with HIV disease...the course over the previous month may reflect the patient's prognosis" [emphasis added]. The rest of the guideline gives well referenced indicators of early mortality that "may be helpful when evaluating a patient for terminal care..."[emphasis added]. This approach assumes a caring practitioner is seeking guidance to whom to admit to hospice in an enlightened system that allows for the knowledge and personal experience of the practitioner, as well as that practitioner's expert opinion based on personal contact with the patient. This assumption is fine if we're using the guidelines just to confirm our opinion that a certain patient would benefit from end-of-life care. Medicare wants guidelines that will ensure that it only has to pay for patients that are appropriate for the Medicare Hospice Benefit. The difference in purpose is subtle, but significant in the end. Bringing reimbursement into the picture turns a point-of-view difference into a potential battleground.
If there is an answer to the problem, it lies in further research. Studies like Schonwetter et al illustrate dramatically that we are operating within a flawed system. It is doubtful that there will ever come a time when hospice professionals will be able to gauge the time-to-end-of-life with any precision, except at the very end, when it is usually obvious to everyone. More research into prognostic indicators may well prove that the goal of accurately predicting the six month pre-terminal mark will be forever out of reach. If this is so, then the current system of certification and recertification based on evidence-based criteria will have to be radically changed. Perhaps in the future Medicare will trust and accept the opinions of experienced hospice professionals that, when it comes to the appropriateness of certain patients to hospice care, "they know it when they see it." Nah.