Over the last 15 years, the Food and Drug Administration has approved numerous high-risk women's health devices based on weak clinical studies, including some that showed no effectiveness, according to an analysis by Northwestern University researchers.

Using the FDA's database, the researchers looked at all 18 high-risk devices approved between 2000 and 2015, including products for fetal monitoring, contraception, and reducing abnormal uterine bleeding. Of the 18, four devices did not demonstrate effectiveness, six were not required to conduct postmarketing safety studies, and three have been withdrawn from the market by makers because of safety or other problems.

Another device - Essure sterilization coils - has been hit by lawsuits and a petition to remove it from the market. Earlier this year, the FDA responded by proposing label warnings that the coils can cause persistent pain, organ perforation, and allergic reactions.

The FDA has long been faulted for relatively lax approval standards for medical devices in general. This review focused on high-risk obstetrical and gynecological devices because, researchers wrote, recent controversies have pointed to regulatory shortcomings. In addition to reconsidering Essure, the FDA has issued warnings that transvaginal mesh can cause pelvic organ injury and that electric surgical tools called morcellators can spread undetected uterine cancer.

Under FDA rules, devices posing the highest potential risk to patients, such as pacemakers, must submit enough clinical data to show safety and effectiveness. (Transvaginal mesh and morcellators were approved under less rigorous pathways.)

Even with this most stringent approval pathway, the FDA does not require that clinical data come from randomized trials, which are the gold standard in clinical research. Nor do the data have to be evaluated by the FDA's expert advisory committee or bolstered by postmarketing safety studies.

"Device regulation is just so much weaker than drug regulation, and it doesn't make any sense," said Steve Xu, a Northwestern physician and health-policy researcher, who coauthored the analysis in the June issue of Obstetrics and Gynecology. "The argument is that more rigorous regulation would decrease innovation. But I would say that has not been demonstrated."

In response, an FDA representative said: "As a general policy, we do not comment on individual studies."

Among the findings in the analysis:

Intergel, a goo that aimed to prevent postoperative internal scarring called adhesions, failed to do that in clinical testing. Although the FDA's advisory committee recommended against approval, the agency gave its OK in 2001 and did not require any postmarketing studies. In 2003, the maker yanked Intergel from the worldwide market amid reports of adhesions, pain, and foreign body reactions.

OxiFirst, a fetal oxygen saturation monitor, failed in its stated aim of reducing cesarean deliveries. Not only was the device approved, but a similar system by another firm, Corometrics 120F Maternal/Fetal Monitor, was not required to show any clinical benefit - just that it was equivalent to OxiFirst. Yet another system, the Stan S31 Fetal Heart Monitor, was approved even though the expert advisory committee gave it thumbs down.

The LUMA Cervical Imaging device was intended to enhance existing technology for biopsying cervical tissue containing precancerous or cancerous cells. While the device improved detection of abnormalities, it also increased "false positives." The FDA approved it after an advisory committee rejected it.

Xu and his four coauthors are concerned that the 21st Century Cures Act - legislation that has passed the House and is being considered in the Senate - would make device regulation even weaker by broadening the definition of "valid scientific evidence."

"We believe both government and organizations representing obstetrician-gynecologists should require more rigorous regulatory standards," they wrote.