In a 2004 Pediatrics article, Dr. Ken Schroeter of the Stony Brook University School of Medicine summarizes reports of near drownings, infant pneumonia, water intoxication, and infections resulting from water births. He reviews the evidence on them, finds that claims of safety and efficacy are unsupported, and that the practice is unable to pass any sensible risk-benefit test.

In the world of evidence-based medicine, his paper doesn’t count.

The British National Institute for Health and Excellence (NICE), the government bureaucracy devoted to developing clinical guidelines for the National Health Service, released new clinical guidelines on water births in September 2007. CG55, Intrapartum Care, studies the studies of water births. It concludes that “[t]he opportunity to labor in water is recommended for pain relief” and that “[t]here is evidence of no significant differences regarding adverse outcomes when comparing labours with and without the use of water.”

Since Dr. Schroeter’s paper is not a statistical trial, it isn’t even referenced.

Statements of “evidence of no significant differences” do nothing to address avoidable risk. They do show that common sense is often the first casualty of too heavy a reliance on statistics. One child drowning in a birthing tank is a small sample, too small a sample to be statistically significant in a study of risks. But any risk of drowning at birth is easily avoidable. So is the risk of water intoxication. In fact, as soon as one agrees that water births could increase infant risk, it becomes difficult so see how a randomized trial of water births could be ethically conducted.

All of this would be of minimal interest for Americans except for the fact that evidence-based medicine requirements almost always appear in state health care reform proposals. They generally attract little notice and virtually no discussion. Disguised as a quality initiative, evidence-based medicine sounds good and the difficulty of discussing statistical results helps insulate it from public debate. People often do not realize that evidence-based medicine can be used to govern medical practice in ways that may please the politically powerful but harm patients.

In Massachusetts, Section 16K of the law that imposed individual mandates and created the Connector Authority also created an independent “health care quality and cost council.” It had an initial budget of $10,000,000 for 2007. Under Section 13B, Massachusetts hospital rate increases are to “be made contingent upon hospital adherence to quality standards and achievement of performance benchmarks” designed in consultation with the health care quality and cost council.

Water births may please those who prefer nontraditional birth techniques. In the absence of evidence, lobbying groups may succeed in getting them approved. The infants affected by them have little political power. They must rely on their parents and their doctors for protection. If Massachusetts makes NICE, that protection may be compromised by standards of evidence-based care.