Wrong Prescription

The Obama administration thinks it has discovered a magic bullet in the drive to lower health-care costs: electronic medical records (EMR). Getting the medical profession to switch from manual record-keeping to a national computerized system, boosters argue, will save money, reduce errors, improve quality and transform health care as we know it. President Barack Obama has proposed investing $50 billion over the next five years to help make it happen.

If only the cure were that easy. Any doctor will tell you the advantages of having lots of patient data on computers: it helps us avoid redundant tests, gather huge amounts of information for research, screen automatically for drug interactions — and spare others from having to decipher our illegible handwriting. I would be happy if every patient could give me a digital file of everything about him; it could really save time on first visits. But we must keep in mind that there will be a cost for computerizing patient records that could prove greater than the billions that would be spent setting it up. (See the top 10 medical breakthroughs of 2008.)

EMR is being pushed hardest by those who would profit financially from it — not just technology companies but also large hospitals and medical practices hoping to improve billing and control internal costs. With a digital chart, every test, diagnosis and treatment a doctor orders is instantly passed along to the billing side: Why give away that Ace bandage for free? This could make the billing bureaucracy more efficient. But communication the other way, from billing to medical, would take place too. And this is more insidious. In a digital system, doctors can’t simply write whatever they want: they generally must select from predetermined choices. That runs the risk of nudging them toward diagnostic decisions based on the computer’s choices. The structure of an EMR, in other words, can easily offer an open invitation to create hyped-up diagnoses and inflated bills.

When, for instance, does a urinary-tract infection become a pyelonephritis (an infection involving the kidneys and ureters)? There’s no clear-cut answer. A computer might remind the doctor that the hospital stands to make many thousands more if he simply clicks on pyelonephritis, the more serious condition. Or consider that nearly every patient who has a big hip or knee operation will run a fever for a while afterward. No one really knows why. But if a computer picks up the temperature elevation, it could prompt doctors to record a “fever of unknown origin” — a diagnosis that often triggers a bigger payment. EMR can inject more higher-paying codes into our patient contact, squeezing that much more money out of it, and quite innocently too. It is, after all, a computer forcing these choices. (See the most common hospital mishaps.)

Why, then, does everyone from the President on down seem to believe computerized medicine will help contain costs rather than inflate them? One argument is that having all that information available should make for better medicine and better medicine will be cheaper in the long run. But more information can also lead to less medicine. EMR can greatly increase insurance-company denials of the treatments doctors want. Might this eliminate unnecessary testing? Sure. But who determines what is necessary? When a white-blood-cell count isn’t high enough to “justify” hospitalization for IV antibiotics, the physician whose judgment says “this patient is sick and belongs in the hospital” is told his services, as well as the hospitalization, will not be paid for. So he has two choices: wait for the patient to get sick enough to justify his treatment plan or join the game — and lie about how sick the patient is. It’s just a matter of clicking a different item on a pull-down menu.

Using information technology to figure out which treatments are most effective seems eminently sensible. Certain heart patients, for example, do just as well with clot-busting drugs as they would with angioplasty procedures, which typically cost thousands more. Crunching huge amounts of data from a wide cross section of patients could help us do better research than we are doing now. But what will happen when the new computerized research turns up a treatment that works a little better but costs a lot more? Will the government-sponsored researchers tell us? What happens to the patient whose particular circumstances argue for a different treatment from what the computers and bureaucrats recommend?

Doctors and patients live in a world of painful, pressing questions. The great physicians I’ve known seek answers through personal commitment to each patient and judgment born of practical experience — neither of which I have found in a machine.

Haig is an assistant clinical professor of orthopedic surgery at the Columbia University College of Physicians and Surgeons

See pictures from an X-Ray studio.

Read “The Year in Medicine 2008: From A to Z.”