A sore throat, that commonest of medical complaints, can pose a dilemma for both patient and doctor: To treat or not to treat.

Many physicians tend to diagnose ordinary sore throats as strep infections, a new study shows, and begin to treat them with antibiotics even though laboratory tests eventually show there is no strep.

The study points up the role of guesswork in many medical decisions, says Dr. Roy M. Poses, who headed a team of researchers at the University of Pennsylvania School of Medicine.

"A lot of times, in medicine, you just don't know," says Poses, now assistant professor of medicine at Rutgers Medical School in Camden, N.J. "These are judgment calls.

"Physicians, if you press them to make quantitative estimates, don't do a really good job in terms of figuring out which patients have a strep throat and which don't."

The year-long study, published last week in the Journal of the American Medical Association (JAMA), looked at how 10 physicians at a major university health service treated 308 patients with sore throats.

The doctors overestimated the probability of strep infection in 81 percent of the patients studied. They prescribed oral penicillin for about one third, or 104 patients, while waiting for throat culture results, but found later that only eight of those had a strep infection requiring treatment with penicillin.

The study, Poses emphasizes, was designed not merely to evaluate treatment of sore throats, but to find out how doctors make decisions when the information available is incomplete.

"Physicians, like experts in all fields," he says, "are often called upon to make judgments when they don't have adequate information. We have to deal with a lot of uncertainties in this business."

He says it's important for doctors to realize the uncertainty of much of what they do and to acknowledge that uncertainty to patients, who naturally would prefer a black-or-white answer about how to treat their illness.

"I'm most uncomfortable," says Poses, "when I know I'm not sure and yet I am pressed by a patient to say or do something definite."

That dilemma can occur with medical problems as serious as an unidentified form of cancer or as mundane as a sore throat.

Sore throats can be caused by viruses or bacteria. Viral sore throats, like other viral infections such as the common cold, cannot be treated effectively with antibiotics. They must be allowed to run their course.

Bacterial sore throats are caused mainly by the group A beta streptococcus bacteria, commonly known as "strep." A strep throat must be treated with an antibiotic such as penicillin, because in rare cases strep can lead to serious complications, including kidney infections and rheumatic fever.

Strep throat is much more common in children than in adults, but often is passed from one family member to another. An ordinary sore throat that occurs as part of a cold or flu is unlikely to be strep.

"It's a longstanding dilemma," says Dr. Stephen Cochi, an epidemiologist with the Centers for Disease Control's division of immunization in Atlanta. "There's no universal approach to deciding whether a throat is strep."

No matter how careful the diagnosis, Cochi says, "you're going to end up treating some people who don't have strep."

As a guide to whether immediate treatment with antibiotics is warranted, doctors look for classic strep symptoms -- fever of at least 101 degrees, tender and swollen lymph glands in the neck and irritation of the lining of the tonsils and throat. But the only way to make a specific diagnosis of strep is to take a culture by swabbing the throat for a sample of throat secretions for laboratory testing.

To treat every sore throat with antibiotics without waiting for a positive throat culture would expose patients needlessly to the risk of allergic reaction to the drugs. And antibiotics have been shown to reduce only the complications, not the discomfort of a strep throat.

Many doctors delay treatment with antibiotics until the throat culture results are available -- usually about 48 hours. The delay does not increase the patient's risk of getting rheumatic fever.

But even the throat culture, Cochi says, is a "far less than perfect" diagnostic tool.

"It tells you with about 90 percent likelihood whether or not the strep bacteria is in the back the throat of that patient, but not whether it is actually causing the sore throat," Cochi says.

Between 5 and 15 percent of people may carry the strep bacteria in their throat without showing any symptoms, Cochi says. When they get a sore throat, they test positive for strep even though it's really a cold or other virus causing the sore throat.

"No one really knows what to do with those people," Poses says. "My own judgment is that if the throat culture is positive and the patient has a sore throat, you ought to treat."

The risks and costs of needless treatment of sore throat are small, Poses says. For most, it means only the mild inconvenience of taking the relatively cheap, relatively safe drug penicillin needlessly for a couple of days before arrival of the negative throat culture.

Although there is growing concern that overuse of antibiotics will encourage development of drug-resistant bacteria, Poses says the practices of the doctors in his study pose little threat.

A national survey of doctors by the Centers for Disease Control in 1982 found a wide disparity of opinion about use of throat cultures for patients with sore throats.

About 25 percent of the doctors surveyed by the CDC always used a throat culture on patients with sore throats, and about 23 percent relied on physical diagnosis without a throat culture. The other 52 percent selectively used a throat culture on patients with symptoms of strep or those considered at above-average risk of strep.

But of those doctors who used a throat culture, 42 percent always started the patient on antibiotics before the results came back. And nearly half of those had the patient continue with the 10-day regimen of antibiotics, even after the test results came back negative.

In other words, Cochi says, for a "fairly sizable percentage of practicing physicians the throat culture results didn't affect either the decision to start antibiotics or to stop them after a negative test result came back."

There's no way of knowing how typical the physicians in the JAMA study are, says Poses, who is an internist with a special interest in clinical epidemiology and medical decisionmaking.

"But I do know this," he says. "They're darn good doctors. They're experienced, smart people."

The most important finding of the study, he says, is not that the doctors were often wrong, but that no one can tell right away whether a sore throat is a strep throat.

"It's impossible to be sure, given our current technology," Poses says. "You have to use informed guesswork."