On the Importance of Information in Health Care
The most significant social fact about the US is that the number of elderly people is going to increase dramatically. The effects won't be seen in the next 10 years, because not many babies were born in the late '30s and early '40s. In 25 years, however, the peak of the baby boomers will reach retirement age, and society will become predominantly age 65 or older. Obviously, a much bigger fraction of the gross domestic product will have to be spent on health care then. This might not be altogether a bad thing because the percentage of elderly people who are not infirm will probably also increase, and their contributions to society may be increasingly significant.
Unfortunately, the business model for US medicine is broken, and it may not be able to meet the challenges presented by an aging population. Physicians are lamenting because they have to deal with each patient's insurance coverage in planning a treatment, and there are about 1600 different insurance carriers. Patients are also generally unhappy because their doctors are too harried to spend enough time explaining the rationale for the treatments that they recommend. Finally, the people who pay for health care, in their roles as stockholders and tax payers, are alarmed because there is no apparent limit to the demand for health care which, however, produces relatively little increase in productivity and happiness at the margin.
The growing significance of information in health care, which currently accounts for about 30% of every patient dollar, or about $300B each year, is hard to overstate. Nevertheless, the most obvious benefit of modern information technology - the ability to electronically capture information as it is created and subsequently reuse that information at negligible incremental cost - has hardly begun to be realized in health care. It is remarkable that, although people own their own medical records in principle, only half of all states even guarantee the right of citizens to receive their own records from health-care providers.
Electronic information technology is especially good at sorting and compiling voluminous data and determining which hypotheses about the data are statistically valid. The bigger the sample, the more certain the conclusions can be. For the really hard and important questions, it may be necessary to use all the known facts. Current Internet usage and practice shows that it would not be difficult to allocate server space for everyone's medical records. The number of US adults using the Internet in the first quarter of 1997 was estimated to be 40 to 45 million, which is already a substantial fraction of the population. Morevoer, many Internet service providers throw in 5 or 10 megabytes of server space for free so that their customers can experiment with personal web pages if they want to.
Ten megabytes is the equivalent of about 2 million words, which could fill a small reference library. If that much data per person were online for the entire population, it might be possible to answer some difficult but very important questions. For example, randomized controlled trials have been conducted to assess the effectiveness of breast-cancer screening. An NIH panel reviewed the available data and concluded that there is still no proof that screening reduces mortality for women in their 40's, even though breast cancer is the leading cause of death for them. The implication is that the treatment that usually follows a positive screen is ineffective in this group, which is doubly regrettable because the treatments compound the misery of the disease. Clearly more experimental treatments and better statistics are urgently needed until better results are achieved. There are also problems with the statistics associated with the detection and treatment of prostrate cancer in men.
Medicine needs a repository segment that acts as an agent of the people for whom it keeps records of medical services. The records should be available anytime and anywhere there is a legitimate need for them. The Internet can provide this availability at reasonable cost. Of course, people will be sensitive about their privacy at first. However, networking medical records is merely an extension of computerizing them. Privacy can be preserved with cryptographic techniques. The real issue is who's interest does the holder of the records ultimately represent?
The best way to make sure that complete records are kept may be to require health-care providers to file them as a condition of being reimbursed by insurance carriers. This might require legislation, but many benefits would flow from the completeness of the records that would result. One of the considerable direct benefits would be that people could be reminded to make appointments for routine or follow-up care or to have their prescriptions refilled. The repository segment would also be an ideal source of data for compiling statistics without compromising individual privacy.
(last updated January 18, 1998)
home, biography, home-theater
large-area display, quantum realism,
a potter's book, against Peter Singer
Copyright 2003, Terence J. Nelson