he practice of delivering babies has long been influenced by traditions, based on the theories of popular obstetricians rather than on scientific proof.

Recently, though, some routine practices have been put through the rigors of scientific testing and have been found ineffective at best and harmful at worst.

One such procedure is the episiotomy, an incision that widens the vaginal opening to allow the baby's head to pass through.

For decades, a vast majority of women delivering babies were routinely given episiotomies, which involve painful recoveries and come with great risks of complications, including incontinence, later in life.

Now, at some leading teaching hospitals, doctors have virtually abandoned the practice. At other hospitals, episiotomy is still used, but the number of procedures has declined drastically over 20 years.

Among woman having vaginal deliveries in hospitals, 64 percent had episiotomies in 1980 compared with 39 percent in 1998, a study published in the September issue of the journal Birth found. The rates are higher among first-time mothers.

Some hospitals fall well below the average, with rates of less than 10 percent. At the University of California at San Francisco, for instance, the surgery is performed on 2 percent to 3 percent of women, said Dr. Bill Parer, director of perinatal medicine and genetics there.

"The science supporting routine use of episiotomy is marginal at best," said Dr. Andrew W. Helfgott, director of maternal fetal medicine at Sacred Heart Hospital in Pensacola, Fla. "It's a procedure that has been overutilized, whose benefits are not as pronounced as some would claim."

Recent studies have found that women who have episiotomies have greater risks of complications than those who do not.

An episiotomy is a one- to two-inch incision of the perineum, the area of skin and muscle behind the vaginal opening. First advocated in the United States in the 1920's, the episiotomy was believed to help prevent serious third- and fourth-degree lacerations to the perineum as the infant's head pushed through the birth canal.

Third- and fourth-degree lacerations tear into the anal sphincter and can lead to temporary or long term fecal incontinence.

Episiotomy was also thought to prevent relaxation of the pelvic floor muscles, a complication of childbirth that also contributes to urinary and fecal incontinence.

Last, because episiotomy speeds the pushing stage of labor, doctors believed that it would prevent injury to the baby.

The procedure became widespread in the 1940's, when childbirth shifted to hospitals from homes. Until the late 70's, nearly all women delivering in hospitals had episiotomies. But then, women, midwives and obstetricians started questioning its use.

In 1983, two authors did a systematic review of the literature and found no good evidence for its use and considerable risks associated with it. Since then, other studies have backed these findings.

"Most of the best data in the literature suggest that routine use of episiotomy only increases the incidence of serious lacerations involving the rectum and the anal sphincter," said Dr. Michael Greene, director of maternal-fetal medicine at Massachusetts General Hospital in Boston.

The birth study found that the incidence of serious tears was 3.6 percent in women who did not have episiotomies, but was 7.8 percent, more than double, in women who had episiotomies. Other studies have shown even higher rates of lacerations.

The problems association with the episiotomy call be illustrated by picturing what happens when a bed sheet is torn. The sheet will not rip easily until it is first snipped with scissors. Likewise, as the baby's head pushes against the surgical incision and stretches it open, the cut is likely to extend farther back, often as far back as the anal sphincter.

Nonetheless, when many obstetricians see a woman start to tear, their instinct is to control it with an episiotomy, said Dr. Keith Johansen, an obstetrician at the Mayo Clinic in Rochester, Minn.

"You're going to prevent a lot of smaller tears and have a clean surgical incision that's easier to put together," Dr. Johansen said.

But most researchers agree that the risk of severe lacerations outweighs the benefits.

When doctors do perform an episiotomy, they have a choice of two cuts. The first cut, called a midline episiotomy, goes straight back and is more likely to tear into the anal sphincter. The second, a mediolateral cut, angles to the side and helps prevent tears to the anal sphincter and rectum in addition to long-term complications.

One study found that 11.6 percent of patients who had midline episiotomies had lacerations to the anal sphincter, compared with 2 percent of those who had mediolateral incisions.

The trade-off is that the mediolateral is more painful for the patient, and she faces a longer recovery.

Over the decades the two types of incisions have reversed places, with mediolateral now ranking as the preferred procedure.

In addition, experts have found little to show that episiotomy prevents pelvic relaxation, a condition linked to incontinence.

"On the contrary, it would appear that by cutting those muscles, you're damaging the integrity of the muscles and you'll be doing more damage to the pelvic floor," Dr. Helfgott said. Nor does the procedure reduce the pain for a woman or protect the baby.

"This was a lot of old dogma that was taught. When I was a resident, virtually everyone had an episiotomy because that's what everyone thought was the right thing to do," Dr. Greene said.

There are times when cutting an episiotomy is warranted.

"We've gone from routine use to selective use," Dr. Johansen said. Some doctors recommend episiotomy when forceps or vacuum suction are needed, when the baby is very large, when the baby's shoulder gets stuck or if the baby is in distress.

Others will not automatically do an episiotomy until they have exhausted all other options.

"I use forceps and vacuum, and I can't remember the last time I cut an episiotomy," Dr. Helfgott said.

"If you really pushed hard, you could get the episiotomy rate down to 10 percent," Dr. Johansen said. At his hospital, the Mayo Clinic, the rate is about 60 percent for first babies, whereas it was almost 100 percent 20 years ago.