IT IS NOT what our patients tell us, but what their laboratory tests tell us that now seems to determine how we take care of them.

The frightening consequences of this development did not become clear to me until I recently returned to mainstream medical practice after 15 years as a teaching ophthalmologist. A few years ago, as a neurology resident in a large metropolitan hospital, I discovered that the "technification" of medicine has distanced us from our patients, unfairly raised their expectations, led us frequently to mis-diagnose their illnesses and cruelly prolong their lives when they are insensate or terminally ill. In providing spectacular options for the few who need them, we have lavished extraordinary resources on the many who don't.

I saw it first in the case of a 55-year-old Cuban woman who complained of a lingering headache and weakness in her legs. Her admitting doctor had found no signs of a physical ailment, and a CT scan of the brain was normal. But when her symptoms did not go away, the doctor ordered a magnetic resonance (MR) scan of the brain, a highly specialized and expensive ($1,000) study that can detect lesions not found on the CT scan. The MR scan returned a finding of "unidentified bright objects" (UBOs) interpreted as "compatible with a diagnosis of multiple sclerosis."

This created a dilemma for her doctor. These UBOs had been showing up on MR scans of many older people who seemed to have no symptoms or signs of MS. Were they simply a sign of normal aging? Trapped by the scan data, the doctor decided to admit the woman to the hospital for further tests. A standard inpatient battery of lab work turned up a borderline low thyroid hormone level. That triggered a consultation with Endocrinology, which sent in a team of doctors to undertake a three-day program of hormonal stimulation tests, including hourly blood samples. Meanwhile, a gastroenterologist came in to check out the patient's poor appetite. He had difficulty understanding her English, so to play it safe he ordered a "diagnostic four-pack": CT scan of the abdomen, barium enema, upper GI series and gastric endoscopy. The tests took a week to complete, and all were normal.

At this point, a psychiatrist concluded that her symptoms were caused by depression. The patient, however, was not about to be convinced. What about the abnormal MR scan and thyroid hormone level, she asked? From her relatives, I learned that she has been in and out of hospitals in several cities in the past few years with similar symptoms. It was becoming clear that we were dealing with a "somatizer," someone who habitually converts anxieties into physical symptoms. Such patients are not new on the medical scene, but in the past we had neither the resources nor the inclination to pursue the diagnosis so intensively.

High Tech, High Hopes

An experience with another patient convinced me that obsession with high tech can distract doctors from the real issues and can cause serious and irreversible harm. A 55-year-old man experienced a 30-second episode of blindness in his right eye. After examining him, I realized that he had probably had a "transient ocular ischemic attack," a brief interruption of blood flow to the eye. One possiblility was the carotid artery in the neck, which supplies blood to the eye. It may become clogged by atherosclerosis or spin off a traveling clot that plugs a key nutrient vessel in the eye.

Had I been evaluating him 20 years ago, the next step would have been to inject dye through a needle placed in the artery -- a relatively dangerous procedure. My account of its possible risks might well have scared him away. But today, one can obtain an exquisite anatomic picture of the carotid merely by bouncing sound waves off the vessel and by listening to the noises generated by blood flowing through it. As I watched this "carotid scan," I was captivated by the elegant pictures of the artery popping in and out of the oscilloscope screen against a background of blips and bleeps and computer graphics. The next day, his complete studies were delivered to me, showing 15 views of a partially occluded carotid artery. Should I recommend a surgical "roto-rooter" operation on the blocked vessel? I was aware that there is no rigorous scientific study that shows that the clean-out operation ("endarterectomy") works to prevent stroke. In fact, there is plenty of recent evidence that it is risky. Yet, as I reviewed the patient's elegant studies with him, I felt my objectivity wilting. One look at those pictures of the clogged artery, and he certainly wanted it cleaned out. I agreed to refer him for surgery.

Several hours after the operation, I went to see him. To my horror, he had developed a massive stroke that left him without language function or the ability to move his right arm and leg. Could it have happened 20 years ago? Certainly, perhaps even more often. But I wonder how much the advanced diagnostic techniques distracted me from anticipating that possibility today. I am not alone. Over 100,000 carotid endarterectomies are performed yearly in this country with a serious complication rate of 10 to 15 percent.

The high-tech atmosphere of modern hospital practice, with monitors beeping and computers generating up-to-the-minute data, can convey a false sense of security. At 61, a man developed severe Parkinson's disease and became so stiff he couldn't exchange enough air to avoid pneumonia. Had he been under my care 20 years ago, I could have chosen among two, perhaps three, antibiotics. Now there were five or six at my disposal, and I could accurately measure their blood levels. Where formerly it was a chore to get a bedside chest X-ray, now I could obtain high-quality film within minutes.

The man lay restlessly in an intensive care bed, a tube inserted in a neck vein to monitor the workload on his heart, a catheter inserted in a wrist artery to monitor his blood pressure, and EKG chest leads to monitor his heart. An ICU nurse who attended only two patients was measuring his vital signs every 30 minutes and drawing blood samples every three hours. On the daily log posted at his bedside, I could find not only temperature, blood pressure, pulse and respiratory rate, but pupillary reactions, fluid input and outgo, electrolytes, blood gases, clotting levels and measures of liver and kidney function. In fact, there was so much information that I usually spent more time studying the log than the patient himself.

That morning I was doing just that. All of the data were perfect -- what are sometimes referred to as "Harvard numbers." But for some reason, I decided to check him quickly before moving to the next bed. When I shook him, there was no response. I turned to the ICU nurse: "How long has he been this way?" As surprised as I was, she replied that he was alert only 15 minutes before when she took his "vitals." A CT scan later that morning disclosed the reason for his unconsciousness -- he had suffered a large brain-stem stroke. Late that night I was called to pronounce him dead. I glanced at his log: Half an hour earlier it had still reflected "Harvard numbers."

Dying by the Numbers

Technical refinements have given us the power to sustain life almost indefinitely by artificial means. Quite apart from the ethical issues, this new power has led to a major drain on resources, as another patient proved to me. A woman came under my care at 62 when she suffered a catastrophic stroke that left her comatose. A CT scan of her brain on the night of admission gave a dismal prognosis. This had all happened so quickly and she was in such good shape before the stroke that her family was unprepared for the news that she might not wake up. But I knew that even if she did, she would be forever bed-bound, unable to read or communicate. Even as I explained this to her family, they pleaded with me to "do everything -- only God can decide her fate."

Doctors have, of course, confronted this predicament many times. But 20 years ago we could not have played out the agonizing scenario that followed. The woman spent the first week in the intensive care unit, where she required artificial ventilation. By the seventh day, she had stabilized so she was moved to a "stepdown" unit where her respiratory devices could be monitored. There she developed pneumonia, eventually requiring two "high option" antibiotics. On the tenth day, her blood count began to drop, setting in motion a special panel of coagulation tests, and a gastric endoscopy which disclosed bleeding from a "stress ulcer." On the fifteenth day, she began to have protracted convulsions which were finally quieted with an antiepileptic medication. The next day she had a cardiac arrest, but a prompt resuscitation restored her heartbeat and blood pressure. Throughout all of these events, she never showed any signs of waking up. Finally, on the thirty-third day, she expired quietly. The bill for her hospitalization: $29,000.

The aura of endless possibilities that surrounds modern medicine obscures a sobering reality -- namely, that we are powerless to alter the ultimate course of most disease. Often we do not share that reality with our patients. That became clear to me in the case of a 71-year-old man admitted to the hospital because he was having hallucinations. The admitting doctor had told his wife that "some tests" would be done to "get to the bottom of the hallucinations." But it must have been obvious to him, as it was to me, that they were part of Alzheimer's disease, a dementing illness for which there is no treatment. The few tests pertinent to this diagnosis do not require hospitalization. But admission is a gesture that reflects added concern on the doctor's part. Unfortunately, it also engenders false hope. After three days of testing, the foregone conclusion was apparent -- he had Alzheimer's disease. When I confronted his wife with the news, she was crestfallen. Why wasn't I warned, she asked? Though I didn't tell her, I suppose it is because we sense that patients expect miracles, and nowadays we can postpone their disappointment almost endlessly with the array of laboratory tests available to us.

The family of an elderly woman had such expectations. She suddenly lost half of her field of vision at the age of 83. Her doctor had no doubt of the diagnosis -- a stroke -- and knew that nothing could reverse the damage. But her children beseiged him with telephone calls, so he brought her into the hospital where insurance would pay the costs of the workup. The diagnosis was confirmed by a CT scan within 20 minutes of her entry. But the doctor considered it awkward to discharge her immediately, so she underwent ophthalmologic examination, 24-hour electronic monitoring of her heart, brain-wave analysis and a variety of blood tests. Three days later she was still an old woman with a nondisabling stroke. Her bill was over $2,000. Her children were satisfied that "no stone had been left unturned."

I see no end to this wasteful and harmful spiral unless both physicians and patients make major concessions. Doctors must submit to greater accountability and, ultimately, limits in our decisions. The only surveillance system that will work is one that gives its guardians the power to withold reimbursement for medical services. This principle is already in practice at many HMOs, which invest the family practitioner, pediatrician, obstetrician or internist with the role of "gate-keeper," or arbiter of recommendations made by medical sub-specialists. Many insurers now require prior approval of hospitalization, a step which makes doctors think twice before admitting patients. Finally, reimbursement for medical procedures, including surgery, must be drastically reduced.

Sacrifice and Success

Our patients will have to give up some things, too -- including the privilege of having their physicians order whatever tests they choose. They will have to accept limits in malpractice litigation. Plaintiffs should be held to a tougher legal standard in proving negligence. They may have to forfeit a jury trial in favor of arbitration, and relinquish rewards for noneconomic losses. Moreover, we all must come to grips with the notion of rationing medical services, inevitably including the withholding of artificial life-support systems in irrevocably brain-damaged or terminally ill patients.

Two years ago, I had the occasion to examine Sandra M, a 35-year-old woman who had been unable to get pregnant. A recent blood test had shown a highly elevated prolactin level -- a hint that she might have a small pituitary tumor. Sure enough, high resolution CT scan located a tiny mass in the gland. With the combined expertise of an ENT surgeon and neurosurgeon, an incision was made in the gum above her upper front teeth, and under microscopic guidance bayonet-like instruments were used to carve a path up through the sinuses and base of the cranium. After gentle manipulation, the pea-shaped tumor came into view. It was carefully delivered through the surgical opening. Two hours after the operation, Sandra was alert and sitting up in bed. She had a small Band Aid over her nose. Shortly before writing this, I learned that she had had a baby girl. Since none of this would have been possible 15 years ago, I was reminded how far technified medicine can take us -- if it is used properly.