Philadelphia Reflections

The musings of a physician who has served the community for over six decades

Related Topics

George (3)
It's often desirable to get live financial data and everyone knows. XML is the thing to use but actually writing programs that work takes a bit of trouble. Plus, once you've got the data you need to display it.

George Ross Fisher III M.D. : Memoirs
New topic 2018-08-23 16:15:31 description



Address to the Symposium on Computer Applications in Medical Care Washington DC, November 5, 1984

George Ross Fisher, M.D.

Almost everyone in this audience knows the PROs replaced the PSROs less than a month ago. If not, absolutely everyone does know that a prospective-pricing DRG system for paying hospitals replaced the old cost-plus system, up to twelve months ago. Prospective payments of hospitals, therefore, appear to be older than the physician review mechanism which provides oversight, but nothing could be more misleading than this time sequence. The payment system and its oversight system were created by the enactment of separate statutes. It is not possible to understand where the Medicare program stands without knowing that the PRO law was drawn up before the DRG (Diagnosis-related Group) law was passed, and reflects a context of uncertainty whether the DRG would be passed.

In a moment, I will return to the important consequence of this anomaly for the PRO data system. However, first, it seems useful to offer a brief historical review of hospital reimbursement by Medicare. In 1965 the driving force behind the creation of Medicare had been to improve access to care by removing the financial barriers to access. The Blue Cross system of cost reimbursement was ideally suited for that goal, since it allowed hospitals to spend whatever they thought they needed to spend, with a year-end retrospective audit and settle-up. The recklessness of this approach was quickly appreciated as hospital costs to the Medicare program rapidly climbed above expectations. By 1972 the matter was of such concern that Congress created the Professional Standards Review Organizations (PSRO) to identify and restrain what was felt to be the main cause of the cost escalation, namely unnecessary or excessive use of the hospital by doctors. From the doctors’ viewpoint, that did have an uncomfortably reasonable sound to it. We were reminded, and we agreed, that only doctors admit patients, only doctors write orders, and only doctors send patients home. After ten years more, however, it was evident that hospital costs had continued to soar in spite of the PSROs. Whether the PSROs had done a good job or not, was immaterial; it was clear that what they had done was not enough. PSROs protested that the blank-check reimbursement system made utilization review of the volume of hospital services quite futile. If doctors should cut the number of blood counts in half, all that was likely to result was that the price of blood counts would double.

By 1982, there were two main options for the government. Either make the PSROs program a whole lot tougher or else change the reimbursement incentives. As it turned out, the government adopted both options. The DRG prospective payment system was enacted; and a set of incredibly bloody-minded policies were prepared for the PRO, including quotas for overall cost savings, quotas for reduced admissions, and a mind-boggling intensification of the surveillance of hospital records. As one government official has written, “The medical record has now become the hospital bill.” It is not possible for an outsider to know how much, or by whom, these regulations and policies were changed during the long period they were held up by internal wranglers within the administration. What is clear is that when they did emerge, the first few months of the prospective payment system had already demonstrated that such an attitude was 180 degrees away from what was really needed. Under the old system of reimbursement, the PSROs urged the hospitals to reduce unnecessary services, whereupon the hospitals protested they meant to preserve the quality of care. Under prospective payments, however, it is the hospital which is cutting costs, and the PRO is the one which needs to protect the public from consequent deterioration in quality. This role-reversal has been very sudden since the impact of DRGs on length of stay has been far greater than generally anticipated. Many hospitals are already reducing staff in order to down-shift to an era of lower occupancy.

This Gilbert-and-Sullivan comedy came to a head and one hopes, came to an end on August 2, 1984 at a Senate Finance Committee hearing when Senator Durenberger (the chairman of the Health Subcommittee) confronted the top administration of HCFA and proclaimed for the public record in no uncertain words that the intent of Congress was that the purpose of the PRO was not to save money. Not to save money. The DRG prospective pricing system was what was to save the money; the PRO was to safeguard the quality of care from the inevitable pressures of reducing expenditures.

There is, therefore, the reason to hope we will soon see major shifts in emphasis in the PRO program, one hopes even including total renegotiation of all the contracts. However, the data system will not be so easy to change, and if that is not changed the people in OMB may yet win the battle. Any program which deals with the analysis of data becomes dominated by the data it receives far more than it is dominated by the intention of its leadership. Present policy mandates the PRO receive its data from the fiscal intermediary, who derives it from the hospital’s reimbursement request. While there may be minor cost savings in such a third-hand data system, it is totally inadequate for the purpose of assessing the quality of care. We have several intermediaries in Pennsylvania and one of them has yet to give us a scrap of data for months Pennsylvania has been on prospective payments. The other intermediaries have indeed given us data, but I would not always describe their performance as speedy. The reason for the delay is not always the fault of the intermediary. In reviewing cost outliers, we have encountered at least one 300-bed hospital which declares it is totally unable to produce an itemized bill. The other hospital seems to produce bills, but the inaccuracy of what we have checked is very ominous.

Now, assume that threats, penalties, and fulminations can coerce the data to be accurate and timely. I would not count on it, but it might happen after a year; the previous speaker, dr. Ertel can tell you better than I can. But the content of the data found on the reimbursement invoice is completely barren of information useful for assessing the quality of care, as indeed it should be if a parsimonious design has limited the data elements to those necessary for the single purpose of issuing and auditing reimbursement. As for the chances of enhancing the data set for additional purposes, the system of passing the data through the intermediary creates a permanent adversary system. Naturally, intermediaries can be expected to rebel at incurring key-entry and data processing costs for purposes which are irrelevant to the reimbursement function, for which they are paid, and on which they are judged. The intermediary can scarcely be expected to become an enthusiastic agent in imposing new data collection requirements on hospitals, while hospitals will predictably resist proposals that they supply data to prove they have been skimping on their services. If past experience is any guide, every proposed addition to the data set will be denounced as costly, unnecessary and unconstitutional, and these denunciations will emanate from the parties who are expected to report it accurately and quickly.

It seems clear to me that HCFA urgently needs to convene some planning task forces. It needs to create a permanent public negotiating arena for the various data combatants; for this, the Prospective Payment Commission would be a good model to copy. HCFA needs to commission some public-domain software. It also needs to stimulate a very large number of demonstrations projects, because the PROs are embarked on a totally new venture, with no existing models to copy.

If you please, I would like to make a proposal right here. We have heard a great deal about provoking or preventing unnecessary deaths, as an index of the quality of medical care. Everyone eventually dies, however, so it is more precise to say: medical intervention shortens or prolongs the interval between the intervention and the subsequent date of death. In a statistical sense, we improve on or fall short of, the individual’s adjusted life expectancy. The advent of disease adjusts the life expectancy, while the therapy or therapist modifies the adjustments. Over on Security Boulevard in Baltimore, a vast computer empire regularly keeps track of the date of death of every Medicare recipient. The data is supplied by undertakers who not file a death certificate without centrally reporting the social security number; the purpose is to stop sending out those green checks as soon as possible after death.

What that system means to me is that HCFA already has in existence two data systems which, if linked together, would permit calculation of actual subsequent life spans of every Medicare patient, or specified groups of Medicare patients, regardless who treated them, how, or for what. It boggles my mind to consider the potential for assessing the average impact on adjusted life expectancy by one hospital or one therapy. And in the aggregate, if the overall “effects on adjusted life expectancy” should begin to deviate from historical values, there would be cause to investigate whether changes in the financial environment of medical care had been the underlying cause.

Let me close by describing our situation as an exercise in macro-economics. Economics have wondered whether 11.5% of the Gross National Products is too much to spend on health. Others say we can and should afford even more than that. Let me tell you how every government always answers such questions of resource allocation. The method is simple. You cut 11.5% down to 10.5%, and watch to see if anything bad happens. If not, you cut it some more. Eventually, when something bad does happen, you restore a little money and declare the issue is closed. If you agree with me that the DRG is just such a methodology, then you do not say the role of the PRO is to preserve the quality of care. Rather, you say the role of the PRO is to detect the first signs of inevitable deterioration in quality at the very earliest possible moment. If our data system is poor, it will take longer to detect problems, and it will be harder to convince people you have detected them. With a poor data system, the budget cuts will go deeper than they ought to go, and care will get worse than it has to get. And that will be a shame.

Originally published: Wednesday, August 15, 2018; most-recently modified: Monday, June 03, 2019