Wide variations in Caesarean section rates among U.S. hospitals cannot be explained by the mothers' health and medical conditions, according to a new study that could redouble efforts to cut the nation's relatively high rate of the costly and risky procedures.

A woman should get the same treatment regardless of where she goes to deliver her baby, said Katy Kozhimannil, a University of Minnesota professor who spearheaded the study. But that's clearly not the case.

"Her likelihood of having a Caesarean varied between 11 percent and 36 percent across hospitals … regardless of her diagnosis," Kozhimannil said.

The study, published Tuesday in the influential journal PLOS Medicine, found wide variations even when mothers were at high or low risk for C-section deliveries.

"And so that means that we're not getting the right Caesareans to the right women," Kozhimannil said.

About 1.3 million C-sections are performed annually, making it the most common inpatient surgery in the United States.

A sharp increase in C-section rates, from 20.7 percent in 1996 to 32.9 percent in 2009, triggered a reassessment of the procedure's benefits and risks and prompted efforts by many hospitals and federal officials to reverse the trend.

Tuesday's study notes that the procedure can increase a woman's risk of infection, pain, re-hospitalization, breast-feeding challenges and future pregnancy complications, and that the infants have higher rates of hospitalization and breathing complications.

Cause is elusive

The study analyzed data from nearly 1.5 million live births at 1,373 U.S. hospitals between 2009 and 2010. The data is a representative sample of one-fifth of all U.S. hospitals, but shortcomings in the numbers meant that Kozhimannil and her co-authors at the Harvard School of Public Health were unable to pinpoint the exact cause of variations from one hospital to another.

Among the maternal factors they considered were diabetes, hypertension, hemorrhages, placental complications, fetal distress, fetal disproportion and obstructed labor. They also considered maternal age, race and ethnicity, insurance status, hospital size, location and teaching status.

But the data lacked key information about whether the women had given birth before and the gestational age at the time of delivery. As a result, it's possible that the apparent variability between hospitals could be exaggerated, the study notes. Kozhimannil said the database also lacks information about hospital policies and practices that might affect the type of care that women get.

"We can't answer the million-dollar question, which is: What is causing this variability?" she said.

Even so, Kozhimannil said she's nearly certain the answers won't be found in the medical conditions of the mother or the fetus.

"I think we need to do what's more uncomfortable, which is look at the broader health care system and see how it's serving women, families and clinicians," Kozhimannil said.

Noting that maternal mortality rates have doubled since the mid-1980s, she added, "We are going in the wrong direction, and this is why I think a broader systems-level look is so important."

Caesarean deliveries pay about 50 percent more than vaginal births for both hospitals and doctors. While that's not likely to play a role in the decisions of individual doctors, it might have an influence on policies at the broader systems level, Kozhimannil said.

"There needs to be clear, evidence-based protocols … so that providers and clinicians have good guidance about how to treat people with particular … conditions," Kozhimannil said. "Women who are clinically the same should have similar clinical care."

A subjective diagnosis

In an accompanying essay, Dr. Gordon C.S. Smith, professor and head of the Department of Obstetrics and Gynaecology at Cambridge University, lamented the fact that the analysis lacked certain basic information, such as whether it was a woman's first birth or gestational age.

"This limits our ability to understand from this analysis why the rates varied so much and how the variation might be addressed," Smith wrote.

Kozhimannil said prior studies show that the primary reason cited for C-sections is something called "non-reassuring fetal status," derived by doctors from fetal monitor readings during labor.

"It's subjective," Kozhimannil said. "That diagnosis has been used much more over time consistent with Caesarean delivery," she said. "But it's less likely that it's happening for reasons where it was absolutely medically necessary because we're not seeing the same kind of increase in child survival or decrease in infant mortality [rates]."

Dan Browning • 612-673-4493