Musings on Computers in Medicine

By Philip R. Alper, M.D.

The place of computers in my practice and, indeed, in my life just came into sharper focus. In September I attended Medinfo 2004, the annual meeting of the American Medical Informatics Association (AMIA). This year it was in San Francisco in conjunction with the meeting of the International Medical Informatics Association.

Like most physicians, I had my first contact with computers in practice when I adopted computer billing many years ago. It allowed the office to manage with one less billing clerk and also eliminated the racket caused by the old-fashioned billing and accounts machine. I liked the clear activity summaries and aged accounts receivable. I also liked the automatic collection of data for accounting and tax purposes.

There was no doubt that this was the wave of the future. Still, I embraced computer billing only half-heartedly. The business part of medicine was a necessary evil, as I saw it. It wasn’t medicine itself and, as such, was best done outside the office. The staff gathered demographic information, and the doctors completed super-bills, one copy of which went to the outside billing service. In a similar fashion, our transcription was also done outside the office, with regular pickup and delivery.

Those were more genteel times and whatever was lost in efficiency by outsourcing certain office functions was more than made up in fewer distractions from attending to patients. From the present perspective, this sounds quaint indeed. Yet I persisted with it well into the time when most practices adopted in-house computer systems for billing, scheduling, and other functions. And because I believe that patients prefer to pay a doctor rather than a P.O. Box, checks continued to come to the office where my secretary could see them and send necessary reminders.

In retrospect, the old way had certain advantages. Main-taining duplicate records in two sites protected against loss of critical financial information by fire or vandalism. There was no need to cart around daily or weekly copies myself. More important, I became less vulnerable to staff disruptions, whether from illness, pregnancy, or simple turnover. The bills went out no matter what. And the hiring of staff, by narrowing the needs of the office, was simplified. Not to mention that greater attention to payments than to billing and tracking helped maintain a 95 percent collections ratio in those pre-managed care days.

As computers and billing software became cheaper, however, the disadvantages of my approach began to outweigh the advantages. For example, I could have asked the billing service to provide limited clinical information on all my diabetic or thyroid patients, but I rarely did, even though there was no extra charge for “reports.” I had many impressions about my practice, but little actual data. And the duplication of effort–entry of the same source information on encounter slips, registration cards, and bills by both the office staff and the billing personnel–began to look silly.

Jump to the present and the power of computers in medicine makes such a discussion sound even sillier. Obtaining drug information before ePocrates (and others) was such a chore that it often went undone, leaving prescribing highly error-prone. Who yearns for the library card catalogue and journal rack in place of Medline, specialized Web sites, and even Google for medical articles and texts? Medical informatics itself has become an enormous field that may soon be splitting into “bioinformatics” and “clinical informatics.”

At the same time, practicing physicians have long been criticized for sluggishness in adopting beneficial new tech-nologies, with failure to adopt the electronic medical record (EMR)–alternately called the Electronic Health Record (EHR)–in substantial numbers as the key example. Estimates of actual EMR use by U.S. physicians are usually 12 percent or less, though one Harris poll in 2001 found (exceptionally) that 17 percent of U.S. physicians “sometimes” use EMRs. Curiously, even this relatively high adoption rate is quite low when compared with the 52 percent and 59 percent rates the same study found in New Zealand and Great Britain. Whatever the numbers, it is generally agreed that primary care is the place where the EMR has the greatest potential to improve medical practice.

It took time, but it is now apparent to academia, government, and the computer software industry that doctors who hesitate to make the considerable time and dollar investment in introducing or upgrading computerization of their practices beyond billing and scheduling are not acting irrationally. There is no standardized language to use for the reliable exchange of information. There are no widely accepted electronic standards to allow seamless contact with other practices and facilities. The Veterans Administration, currently the largest medical information technology effort in the country, is on version 26 of its physician interface, which admittedly would be impossible with private vendors. Privacy issues are unresolved. Computer scientists report that even with de-identification of individual patients, skillful data mining may still reveal their identity.

Cost is another huge barrier. Kaiser-Permanente reputedly has spent $1 billion working with IBM in a failed attempt to create a fully electronic patient information system. True, there are 11,000 Permanente physicians, but that still works out to nearly $91,000 per physician for starters. (Of course an EMR for use throughout the Kaiser system involves more than its physicians, but I have no easy way to allocate the expense. In fact, the total price of the system now being rolled out has risen to $3 billion, which would work out to $275,000 per physician or however that sum can most appropriately be allocated.)

The American College of Physicians estimates the cost of setting up an EMR in community practice to be about $30,000. Other estimates of up to $70,000, plus ongoing expenses, are reported by the AMA in amednews. com (7/5/04).

The question of who benefits from the investment is just as daunting as the cost. Though a modest amount of time and money may eventually be saved by using an EMR, it is likely to occur only after a major commitment of human and financial resources. The average doctor who is not an informatics crusader is likely to become the investor; and the government, insurer, or employer will be the beneficiary. Of course, this is not a sound business basis for moving forward.

At the AMIA meeting, I learned that the Institute of Medicine, upset by the fragmentation of existing patient information systems and blaming it for many of the errors made in practice, hired a public relations firm to spread the word that 98,000 patients die every year in U.S. hospitals because of medical error. That made certain that the 1999 report, “To Err Is Human,” was featured on all the major TV networks and  printed in nearly every newspaper. No matter that the numbers were found to be egregiously exaggerated in subsequent analyses, the point was made: from an information and communications standpoint, medicine is in the dark ages in comparison with nearly every other segment of the economy.

This message still reverberates and draws near-universal approbation. Senators Bill Frist and Hillary Clinton, in a bipartisan consensus, jointly stated in a Washington Post editorial (8/25/04) that U.S. health care needs a “major transformative change” in which information technology is used to “improve care, lower costs, improve quality, and empower consumers.” Employers and health insurers, not surprisingly, echo these sentiments, as does the AMA, though in less rousing language. The editorial cited above states in its sub-title, “With EMRs becoming an inevitability, it’s essential for the physician’s voice to be heard in how the technology gets implemented.”

The Department of Health and Human Services, buoyed by 2003 Medicare reform legislation, has funded multiple demonstration projects promoting EMRs as part of a $50,000,000 National Health Information Initiative. (The President has asked for a similar sum for this year.) A scattering of Pay for Performance projects funded by insurers also reward office computerization.

But the money allocated to create truly interoperable–meaning effective across different computer platforms and software systems–office computerization in the United States pales into insignificance in comparison with the sums in-vested by Great Britain, Australia, and Canada,  where the investment is in the billions of dollars.

Thus, the American EMR, is supported more by talk than by substance. In response, primary physicians might do well to copy the inimitable 1980s advertising mascot, Clara Peller’s words and ask, “Where’s the beef?”

Until it’s produced, the American version of the universal EMR is certain to limp along as an attractive but imperfectly realized concept. 

Dr. Alper is an endocrinologist in Burlingame.