Meet a doctor who wants us to say no more often when doctors want to do something to us.

Also, no to ourselves when we get the idea that we "ought to have" a prescription or test or treatment or even a physical exam, if it's too often.

He does not, this cautious doctor, want us to say no all the time, just some of it, based on more thought and information about the many things that can go wrong when doctors try to do right.

This therapeutic conservative is Dr. Eugene D. Robin, a Stanford University medical professor and practicing physician who has written one of the most provocative books of recent years about medicine's pluses and minuses: "Matters of Life and Death: Risks Versus Benefits of Medical Care" (Stanford Alumni Association/W.H. Freeman, $21.95; paperback, $11.95).

"No doctor knows everything," he said in a recent interview. "All doctors make many errors. You have a right to ask questions and come to an informed choice, with information from your doctor and any other advice or information you can get."

Robin tells a story. A man of 40 was hospitalized because of fever, stupor and neurologic problems. It was believed he might have herpes encephalitis, a brain infection caused by one of the ubiquitous herpes viruses.

His doctors were not sure of the diagnosis, or whether to give him the best drug available, one sometimes effective yet one with possible dire effects of its own. To try to pin down the diagnosis, they ordered a brain biopsy, opening the skull and removing some tissue.

They found no virus. On the reasonable supposition that herpes infection was still the cause of his illness, even though the biopsy had produced no solid evidence, they then gave him the drug anyway.

The drug had no effect. By then, however, the man had sunk into a deep and irreversible coma. The biopsy, not the illness, had apparently caused damaging bleeding into the vulnerable brain, turning the patient into a "vegetable."

"A patient suffered a horrible fate almost certainly as a result of a test," Robin writes. "Doctors then ignored the outcome of the test and treated the patient as if the test had never been done . . . Why was the test done if the results were ultimately ignored?"

The lessons, he says: "Many diagnostic tests are done despite an overwhelming probability" that the results will not help the patient. Never take a test that can do any conceivable harm, never even let your skin be punctured unless the result can give you a better chance of a treatment that can help you.

This should be true, he says, of both invasive tests, like skin punctures or mechanical probes or biopsies, and noninvasive, supposedly harmless ones, like urinalyses or ultrasound exams and other ways to visualize the interior of the body without injecting or inserting anything.

Most medical tests, he explains, produce many false positives, indicating trouble when there is none, as well as false negatives, indicating everything is okay when it is not. A false positive sometimes leads to a treatment with potential harm of its own.

Many doctors are either not aware of the potential risk of medical interventions "or may underestimate their magnitude" in their understandable zeal to help their patients, Robin believes. In some cases they have less attractive motives: Tests and interventions produce fees. In some doctors' minds, the faith in "doing something," ingrained in their training, may become inextricably entwined with the effect on their wallets.

Robin is not a doctor-baiter or medical heretic, however, or an opponent of medical science. He argues:

* That the "medical system," the way doctors are trained and the way "medical knowledge is introduced and applied" produce "a preoccupation with the process of diagnosis" and over-treatment. Thus, he says, far too many of the "disasters" and "horrors" like the case of the brain-biopsied "vegetable."

* That such horrors are not typical, but neither are they uncommon. "Each of the 50-odd lay people and doctors who have seen parts of this book have described comparable disasters attributable to medical care," and "when the doctor becomes a patient, his anxieties may be even more intense than those of a medically untrained person."

* That some of medical care's flaws and risks are unavoidable, but many can be corrected by more awareness of the problems by doctors, and more rigid research into the effectiveness of both old and new methods.

* That patients -- his main thesis -- "can reduce the risks and increase the benefits" of their treatment.

How?

By "restricting your medical encounters to those that are absolutely necessary."

By realizing that no doctor is all-knowing, none has ever practiced error-free medicine and some are not competent.

By understanding that most care involves both risks and benefits, and doctors cannot guarantee results.

And by asking questions like:

* Does this test or drug or treatment have any possible risks?

* Has its worth been established by good research? Has there been an adequate clinical trial in a large number of patients? What were the results?

* If the test turns out positive, will you do anything for me that you would not do anyway?

* What are my alternatives? What can happen to me if we do nothing?

One should also ask: Is doing nothing also a risk? One hears story after story of persons with unusual symptoms or pains who are told again and again they are normal, only to learn later they have a cancer. Every unusual symptom should be tracked down, if possible.

There are good doctors, therefore, who disagree strenuously with Robin's thinking. They believe warnings like his will dissuade patients from needed care.

There is indeed that risk, Robin concedes. But the risks of blind faith and overtreatment, he believes, are even greater.

Much of medicine does great, even extraordinary, good, he emphasizes. When sick, we must put our trust, finally, in the best doctor or doctors we can find.

But, remember, he says, ask them too some questions.

Next Week: When is it time to consult a second doctor?