Universal health insurance is a kind of financial design. Naturally, the first step in properly managing its finances is to keep clear accounting records. Expenditures should be clearly delineated according to specific districts, hospitals, even medical departments and patients. The second step is to seek out unaccounted-for medical information in a timely manner. For instance, which institutions are being profligate, or are wasting medical information resources? And which people have not paid their insurance premiums? Waste should be eliminated, and debt should be collected; only then can financial holes be filled in.

It sounds simple, but if these goals are to be realized, what is called for is an exceedingly complex health care information network.

Computerized filing the first step

In the context of an overall health care network plan, which the ROC Department of Health (DOH) has continuously promoted, a "national medical information network" figures as a crucial element. But in previous years when public health insurance was principally administered under labor insurance, limitations in terms of equipment, manpower, and even information about the total number of monthly registered cases in all the clinics and hospitals, made it completely unattainable. Computerized statistical analysis and cross-referencing of the vast amounts of information was even farther outside the realm of possibility.

When universal health insurance was put into effect, the NHIB and the medical community came to a deadlock in discussing payment rates. Wu Kai-hsun, professor of public health at National Taiwan University, points out that after the rate was established, annual income from insurance payments is generally fixed. Therefore, the first demand of the NHIB was to "control the total of medical payments based on income." The number of individual cases in any category of treatment can easily be calculated with the formula: "Individual costs times total number equals total costs." The individual costs are readily known, but the problem is that the total number of treatments in the various categories has never been assessed. For instance, on the average how many appendectomies are performed in the nation every year? How many computerized tomographic scans have been taken? An overall accounting has never been made. When a concrete number is unobtainable, both sides of the dispute can only resort to haggling over the price.

In order to prevent repeating the mistake of the era of labor-insurance-subsidized health care, when "the hospitals went on brazenly abusing the system, and the labor insurance bureau went on silently paying," promoting the computerization of medical institutions is a major policy of the universal health insurance program. But in the hasty course of health insurance implementation, even the NHIB's own computer program was not fully prepared in time; to this day it remains in a disheveled mess and has not actually gone on-line. Of the more than 13,000 medical institutions that have already signed their insurance contracts, exactly how many have completed computerizing their operations is utterly unknown.

Though the actual state of affairs is somewhat behind schedule, a clear blueprint can still be seen within the NHIB's plan. Because the biggest bottleneck in the development of the health care information network is that the amount of data which must be keyed in is too vast, in the future, in addition to the conventional hardcopy declaration forms which hospitals and clinics submit, they will be encouraged to submit softcopy records, as either "stored media declarations," in such forms as diskette or tape, or "on-line declarations." A system of punishments and rewards will be used as an incentive.

"For medical facilities to file by computer will increase the overall cost, so the National Health Insurance Bureau has already factored in the expense," says Li Ling-ling, manager of NHIB's computer department. "In the future, for those who still make their declarations by hardcopy, there will be a regulation deducting NT$5, which counts as private institutions' data registration expense." In addition, those who use computerized declarations will be given a preferential rate in terms of "temporary remittances" (money advanced to hospitals and clinics before a case-by-case audit has been completed). Temporary remittance rates can be as high as 100% of the full post-audit remittance. Those who submit hardcopy declarations can receive at most a rate of 90%.

Information "census"

In the past an employee audit could not perform many management functions. But with a computerized information network, it can easily be carried out. Currently, methods are being devised to isolate "irregular indicators." For instance, after the beginning of April when large hospitals began to give extra fees to people who do not come by referral, rumor began to circulate that certain unethical clinics were issuing large numbers of medical referral slips, making a brisk trade as "referral stations." In the future, computer cross-referencing may well shed some light on this matter.

"For the most part, if the rate of medical referrals for any given clinic in a single month exceeds 5%, or if the rate of examination referrals exceeds 10%, we will consider their behavior suspicious," says NHIB director Yeh Chin-chuan. By the same token, which hospital has performed the most cesarean sections in a given month, which doctor sees more than 300 patients in a single day, etc., all becomes clearly visible from computerized accounts.

"In the future we will implement a Biggest Offenders List to show which hospitals are wasting the most medical resources," Yeh Chin-chuan says with extreme confidence. Punitive regulations will not be necessary; when we have numerical evidence, hospitals will naturally be more vigilant.

The NHIB will also be able to observe the average citizen's use of health insurance resources, with a census drawn from information on insurance registration cards.

"When any person who is insured goes to see the doctor, his or her ID number, the date, a description of the ailment, treatment history and medical costs will all be recorded on the declaration form. If we notice that a patient has suddenly gone to the doctor three times in one day, or gone to the same clinic 10 times within the same month, or if a person with a chronic illness goes to three separate hospitals and picks up 28 days' worth of medicine at each, these illegal actions can all be traced," points out Yeh Chin-chuan. Within the context of the insurance registration card "census," they can even ferret out clinics that collude with patients, so that the former problem of "living off insurance" can be curbed.

Computerized bloodhound

From the midst of the current dispute, in which citizens claim to be laden with extra burdens and the medical community decries a drop in the standards of payment, a new problem is beginning to emerge--how to ensure that insurance premiums are paid. In particular, the more than 980,000 unemployed and self-employed who have registered for insurance under the miscellaneous "regional population" category will be far less easily accounted for. When a patient enters a clinic with insurance card in hand, how can they judge if he is in arrears on his insurance payments? Should the NHIB cover his expenses? If hospitals and the insurance bureau are not effectively linked via computers, these kinds of loopholes will be very difficult to seal up.

Looking at matters along these lines, a health care information network certainly possesses amazing capabilities. Groping and out of control before completion, or orderly and neat afterwards, it is now viewed as a crucial element in the health insurance system. But when will the network be up and running? Yeh notes that 240 to 300 million insurance cases are transacted annually (more than 21 million insurance holders seeking medical help an average of 12 times per year), a much smaller figure than the number of transactions processed by the postal service. If the postal service can institute full on-line computerization in order to make its mail handling more accurate and error-free, then the universal health insurance program should "have no great problem accomplishing it."

Nevertheless, some management personnel in the medical community dare not be so optimistic; some even estimate, "If we can get a computer network more or less set up within a few years, that would be pretty good!"

Do we really have to wait that long? In the area of management, will health care resources be used up within the short term, drop by drop? Currently, everyone is hoping that a fresh review three to six months after implementation, with accurate and concrete statistics, can yield a decisive analysis.

It seems this is yet another race against time.

[Picture Caption]

After a health care information network is up and running, management an d investigative work will perhaps become more practical.