Medical care providers in the past have faced the challenge of trying to provide adequate care based on inadequate data. Now that they have billions of bytes of data, the challenge is more complex and not necessarily any easier.
Each year hundreds of thousands of new scientific articles are added to the Library of Congress; the medical world is “awash in new information,” according to a New York Times essay. While each study has unique perspectives and strengths and weaknesses, in theory the data can be studied in aggregate. Systematic reviews such as those produced by the Cochrane Collaboration, the author argues, provide “a verdict on unsettled medical debates based on a painstaking reassessment of all the relevant research.”
It is tempting to believe that information, if only there is enough of it and it is analyzed properly, can provide a “final word” on our most important decisions. Another article argues that data could cause our lives to become overly prescribed, constraining us to always choose the correct course of action – the one most likely to produce a “good” outcome. Freedom could be eroded by data’s ability to always predict the best course of action for a given person in a given situation.
But this argument is embedded in two big assumptions. One is the assumption that evidence of outcomes is the primary guiding motivation for human actions, and the other is that evidence is always potentially available, even if we haven’t discovered it yet. Those promoting evidence-based medicine often fall into this trap: if only the evidence were known, they say, then doctors would follow it. But the data is often missing, or impossible to acquire, and optimal decisions are not clear.
Take four obstetric practices that are constant for nearly all women giving birth in the US. There should be tons of data on these practices, and care providers should practice according to evidence of best outcomes – right? Here is what Cochrane has to say about these interventions, practiced nearly universally here:
Vaginal exams to assess dilation and labor progress. Cochrane says: “It is surprising that there is such a widespread use of this intervention without good evidence of effectiveness, particularly considering the sensitivity of the procedure for the women receiving it, and the potential for adverse consequences in some settings.”
Intravenous fluids. Cochrane says: “The evidence from this review does not provide robust evidence to recommend routine administration of intravenous fluids.”
Giving birth in the hospital. Cochrane says: “There is no strong evidence from randomised trials to favour either planned hospital birth or planned home birth for low-risk pregnant women.”
Continuous fetal heart monitoring. Cochrane says: “Continuous cardiotocography during labour is associated with a reduction in neonatal seizures, but no significant differences in cerebral palsy, infant mortality or other standard measures of neonatal well-being. However, continuous cardiotocography was associated with an increase in caesarean sections and instrumental vaginal births. The challenge is how best to convey these results to women to enable them to make an informed choice without compromising the normality of labour.“
The experts at Cochrane have done a great deal of work finding the studies, evaluating their quality, and analyzing those that qualified. In three out of these four cases, however, the only conclusion they can draw is “the data are not clear” or “there is not enough high quality data,” while in the fourth case the conclusion is “the data indicate something, but it’s complicated.”
If the overwhelming proliferation of data provided clear answers, and if those clear answers drew a straight line to physician practice, we might expect to see something different than we see in American obstetrics. Perhaps physicians would not give IV fluids, or only give them sometimes; perhaps they would use noninvasive methods to assess labor progress; or perhaps they would discuss the complexities of evidence for and against these interventions with patients. The widespread use of these interventions without discussion indicates that doctors’ decisions may be driven by something other than evidence that they will improve the health of the mother or baby.
The other factors driving these decisions include tradition, inertia, malpractice fears, hospital policy, culture, insurance reimbursement, and a host of other reasons that are not easily washed away with data. I believe that evidence, especially when it is equivocal or complex, has less power than we think, and it is not necessarily making our lives any easier or our health any better.
Even this argument is complex, though: Cochrane recently published an analysis stating that delayed cord cutting can benefit the infant, and in my experience I have seen doctors becoming more open to a discussion about it. So – data can sometimes drive change.