Philadelphia Reflections

Computers and the Regulation of Medicine.

Computers and the Regulation of Medicine.

George Ross Fisher, M.D.

I am going to take chance in this essay that I can hold the attention of the reader through a preamble of theory, before addressing the consequences for the practice of medicine. That seems necessary because I believe that the consequences are different from what most readers would intuitively expect and persuasion lies in first convincing the reader of the theory.

CLOSED AND OPEN SYSTEMS

There is a growing body of endeavor known as the Theory of System, which acknowledge that all events are consequences of pre-existing conditions (like the consequences of adding acid to bicarbonate in a beaker), and are thus “closed” systems. However, most events in biology and sociology are so complex that it is only possible to deal with them as “open” system, for which we substitute wisdom for scientific certainty. “Wisdom” is a set of traditions, maxims, opinions, and strategies which allow you to make predictions about the inevitable outcome of events within an open system. The teleological nature of human events was once referred to as Manifest Destiny, and realists like Talleyrand spoke of diplomacy as the art of manipulating the inevitable.

Example:

Wisdom has it that in your choice of a practice location, you should remember that “you can’t make money where it ain’t.”

And now a conclusion about the computer revolution: Since computers increase the capacity to store and manipulate detail, the computer revolution increases the number of closed systems, and shifts the scope of wisdom in decision-making from traditional areas to new subject which was formerly incomprehensible.

HIERARCHIES

In dealing with open systems, managers and executives have evolved a basic strategy; they organize manageable subunits into hierarchies. Units are organized within departments, then organized within divisions, reporting to a policy-making body. Further, because the purpose of the organizational structure is to simplify management, each level of the hierarchy is oblivious to the techniques of the level below, and is only interested in the output of the level below.

Example:

The patient paying his bill is interested in the total amount that he has to write on the “bottom line,” which in his case is the dollar amount of the check he must write

The director of the x-ray department is concerned with a subtotal related to the x-ray department. The chief technician is concerned with individual studies. The dark-room attendant is only interested in pieces of film.

COMPUTER CONCLUSION: Primitive Computer Systems merely duplicate the pre-existing manual system. Their real power lies in the next step, which is to reorganize the reporting system. That is, they cause a reorganization of the hierarchy.

TRANSMISSION

The prediction is made that management system will be forced in the direction of a hierarchy of three: Those who are able to make decisions, those who cannot make decisions but are necessary for some task, and the computer. One function often seen in non-computer systems is simply to pass the information unchanged up to the line. There is little doubt this activity will vanish.

Example:

Robert McNamara (from Princeton via Ford Motors) was a computer expert who became Secretary of Defense under John Kennedy. By installing computers, McNamara was able to jump the Army reporting system (Sergeant to Captain to Secretary of Defense) and confront the generals with discoveries before the generals had received the news. We are told that the generals didn’t like it a bit. But can anyone doubt that Robert McNamara carefully filtered the data before h presented it to President Kennedy? The system of hierarchical reporting condenses the data to the next step up.

COMPUTER CONCLUSION: Many systems of management by delegation will soon be swept away by the computer revolution, and middle management will be the most threatened region. It will resist but it will lose.

MODIFICATION

Another function of delegated systems is to take raw information and reduce it to condensed form for the benefit of the next level upward. They do so by a process which is often a mystery to the next higher level, and hence a certain power is conferred on the lower subunit to modify the conclusion by modifying the system of manipulation. The method for controlling such activity is to produce a procedure manual which the next higher level must approve, but the inherent complexity which forced a delegated process to be created also obscures the power of the delegated subunit to modify the system.

COMPUTER CONCLUSION: Computer technology strictly defines and inflexibly follows defined procedures for steps in hierarchy. It thereby confers much stricter control power to the higher levels of hierarchy.

THE NEED TO KNOW

If for no better reason than to reduce programming costs, the computer process confers a new power to the lower levels of hierarchy. The higher level must now strictly define its reasons for asking for certain information. If it cannot demonstrate a need to know, it cannot justify the cost of knowing.

Example:

The PSRO, acting on behalf of the physician community, violently resists the inclusion of data elements in the reported tape sent to the Bureau of Quality Assurance. At the same time, it is anxious to acquire as much data as possible from units lower in the hierarchy, who in turn resist the process. It can be expected that this process will eventually settle out at roughly the best equilibrium for the community at large, although differing aggressiveness among the participants may cause temporary inequities. The weapons in the battle, which are at the disposal of the physicians are:

(1) Superior Claims on the decision-making process.

(2) The faith of the public in physicians as the most trustworthy custodians of their health privacy.

(3) A superior pool of talent, determination, and independent means committed to a vital issue.

COMPUTER CONCLUSION: If you have a good chance of being the winner in the reorganization of a hierarchy, it is better to participate to your utmost rather than hold back out of fear that someone else will be the winner, because only participants are winner.

NETWORK

We have spoken thus far of hierarchy as the only manageable approach to the complexity of open systems. A more general description would be “modularity” since modules can interact in a lateral direction as well as vertically. When they do, the result is a network of modules in three dimensions. Since computers increase the ability to cope with complexity, they increase the ability to work in three dimensions. Hierarchy is the last resort of manual management; just as three-dimensional chess is beyond the ability of people who are not even very expert at two-dimensional chess. In this sense, the computer revolution provides some hope for the American System, which presents hierarchy and naturally prefers networks when feasible. This is to some extent a philosophical preference and does not seem to be true of the Japanese social system, or the German mentality, or the Communist method. The natural American instinct for lateral equality is thus an ally in Medicine’s conflict with Government, but a hindrance when it encourages Nurse independence or unrealistic consumerism.

Whether lateral or vertical, the interaction of modules in a complex open system is the same: delegation of a method, the output from one module as the sole input to other modules, and resistance to the need to know.

COMPUTER CONCLUSION: The organization of modules into vertical hierarchies or horizontal networks is largely a political process, with three-dimensional networks as the last resort of compromise, and with strict vertical hierarchies as the last resort of inadequacy.

CONFIDENTIALITY

Complexity is itself a major defense of confidentiality; since computers reduce complexity, they also destroy the smoke screen. Computer System which stumbles ahead or is manipulated into breaches of confidentiality is certain to raise a great uproar about the need to know and the right to conceal. In the PSRO system, the issues balance between the duty of accountability and the patient’s right to privacy. When reduced to these terms the physician community has a clear advantage in the mind of the public, if the advantage can be effectively exploited. The latter can, however, be overturned by speedy pre-emption of the turf. it can be predicted that special pleaders will insist on accountability when all they really want is power and satisfaction of envy; it can fairly be predicted that some will weaken their claim to privacy by overextending its bounds.

Example: The system of peer review on Medicaid prescriptions in Pennsylvania has turned up a number of instances of patients who obtained multiple prescriptions for “controlled” drugs from multiple doctors, filled at multiple drug stores, probably for resale on the streets. When the doctors and pharmacists were notified, they were universally grateful and took steps to curtail the problem. However, the computer vendor learned of the problem (regardless of the fact that all reports are shredded after review) and persuaded the state government to institute a system of restricting problem patients to a single physician. It may now be impossible to dislodge this meat-ax reaction, in spite of the fact that the computer peer-review system is probably able to cope with the problem without invoking hierarchical power.

A Second Example: The United States Navy recently developed a system of computer protection so elaborate that they boasted of it in the newspapers. Two computer scientists read of it, and in a month’s, the time had broken into the system via telephone. The Navy was then agitated to read of its disarmament in other newspaper articles.

COMPUTER CONCLUSION: There is no present foreseeable technical method of protecting the confidentiality of computerized data, except by physical ownership and physical protection of the machine itself and all of its activity.

CONCLUSIONS

The most significant event in the Twentieth Century is the Computer Revolution, just as the Industrial Revolution was the major event of the Nineteenth Century. By the greatest good luck for medicine, the computer revolution is capable of solving the four major problems which now threaten the American Medical System.

1. THE FAILURE OF THE PRE-PAYMENT INSURANCE MECHANISM. The removal of cost restraints on the patient (and thus the provider) has had a predictable upward effect on costs. The overwhelmed system has reacted in a typical hierarchical manner: try to convert insurance companies into regulatory bodies, and if that fails, into rationing systems. The computer revolution (if we are agile) has the potential of drastically reducing the information costs which are now 40% of hospital and insurance company costs. It also has the ability to control utilization abuses, and expose power abuse to public decision.

2. THE VAST INCREASE IN PARAMEDICAL PERSONNEL. Middle management is most vulnerable to computer replacement, and middle management now costs 20% of the hospital dollar. Physicians in complex medical centers are most alarmed about this problem, which they can easily identify by comparing the hospital parking problem with what it was, twenty years ago. Surgeons are typically least concerned since their role at the center of procedures is least threatened by aspirants. But surgeons are hearing of “unnecessary” surgery, and even the small-town solo practitioner has to hire girls to fill out forms. The complexity of our system must be reduced, and computers can do it. The best way to thwart the claims of aspirants to power is to eliminate jobs.

3. THE EXPLOSION OF SCIENTIFIC INFORMATION. No one would wish to reduce the output of the research community, but ways must be found to organize and transmit the information without resort to fragmentation by sub-sub specialists. The computer is ideally suited to the problem. THE MALPRACTICE CRISIS. Physicians are uncomfortable with the idea that peer review may soon become entangled in the malpractice system, as indeed it inevitably will. The matter comes down to biting the bullet, armed with a statistic. Surely consent for an arteriogram is more threatening if it is couched as “you might lose your leg” than if you are told, “you have one chance in five thousand” of such an occurrence. Realistic insurance premiums can be set when the risks are defined. Juries can be provided with realistic statistics on normal risks and normal expectation of benefits.

Through all of these four problems runs a common theme: The cost of medical care. The PSRO seems to be the last best hope of curtailing the cost threat to medicine, and so the PSRO can be expected to be the vehicle for the computer revolution’s resolution of the issue. Senator Bennett probably had no idea of what he was doing, but he did it, and the problem is now our problem.

Originally published: Wednesday, August 22, 2018; most-recently modified: Wednesday, May 15, 2019