Birth Planning, Research

The Water Birth Series Part One – What types of research evidence do we have?

The Water Birth Series Part One Audio File


You probably are familiar with the theory of “levels of evidence” where a study is considered more likely to be providing information closer to “truth” the higher it is on the evidence pyramid. In the evidence pyramid a randomized controlled trial is the highest form of primary research, and a synthesis of studies provides the best evidence. And this preference for randomized controlled trials makes sense for some parts of health care because randomized controlled trials can reduce some of the biases that interfere with our seeing the truth. A randomized controlled trial can eliminate some of the outside variables and let you see the effect of making just one change.

But this doesn’t make sense for everything we do in health care, and we know the information we get from randomized controlled trials doesn’t always translate well to real life because 1) the restrictions we put on eligibility for studies rarely allows the results to be generalized to the population as a whole and 2) Adherence to the intervention tends to be higher in a randomized controlled trial than in real life which makes the effect seem “better” than it is. Research is starting to address these issues by using different types of trials that do a better job simulating “real life”, and these pragmatic trials look a lot like prospective observational studies.

Hydrotherapy for labor in second stage is one of the areas where a randomized controlled trial doesn’t make much sense. First, women are not always willing or able to be randomized to water or not water. Medical ethics dictate other factors take priority to randomization. For example, you would never force a woman to remain in the water to deliver if she meets one of the criteria to remove her from the water. You should never force a woman to remain in the water if she wants stronger pain management. Even if you do randomize to water or not water, you have the problem of attrition.

Attrition is the name we give to drop out from a study, and attrition can be a problem when it occurs unevenly. This is because reasons a person leaves a study are often related to the intervention being studied. When you have unequal attrition you lose part of the information for one group of study participants and this means when you analyze the information for the people who are left in your study you get a skewed picture of what is happening.

With hydrotherapy we have built in attrition. We remove women from the water when certain things happen during labor – like thick meconium or category II fetal heart tracings. This is not a clinical problem, it is actually good practice. It is a research problem because we only remove people from the water group. We don’t remove people from the conventional delivery group for meeting these same criteria. With built in attrition the research may be biased by not comparing equivalent groups.

So we have the problem of attrition, lets say you try to correct for this by keeping the women in the study in the water birth group. Now you are reporting on the effects of “water delivery” using a group of women that maybe only 50% did actually deliver under water. Can you really say the results you found were due to delivery under water?

Another problem for randomized controlled trial with hydrotherapy is the inability to “blind” the study. Blinding means that we don’t know who gets what treatment so we cannot influence the outcome by treating one group differently. But with water birth there is never an ability to blind the study – everyone evaluating the woman and the baby – the people making clinical decisions about the other care that they will receive — knows if this labor and delivery occurred in water. So clinicians be more likely to say “oh, this baby is breathing fast and born in water, better send her to the NICU for observation” or “oh, this woman is delivering in water I don’t need to do an episiotomy”. In this way, even in a randomized controlled trial, it is difficult to determine what is the effect of water, and what is the effect of changing our practice because the woman was in water.

Because of these challenges, the bulk of evidence for hydrotherapy during second stage comes from non-randomized studies, we call this observational research. These are the studies that might look back at the last year at the birth center to see if there were differences between women who delivered in water and those who did not. But these are also the studies that say, for the next year we are going to pay attention to every woman who comes to our hospital and is eligible to use hydrotherapy whether she uses it or not.

Some of our evidence comes from what can be considered case studies. These are the papers that describe one or two unusual outcomes to alert other practitioners to a potential problem. Case studies of a sudden cluster of rare skin cancers were how the world suddenly became aware of HIV. They have an important place in medical literature, but they do not replace or over-ride comparative evidence.

Similar to case studies are papers that describe the outcomes with water birth at a facility without any comparison data. These types of papers were common when water birth practice was first developing and at the time they were appropriate methods to share practice details – kind of like reporting an implementation project we would do today.  In this type of research the author is more interested in talking about why and how they made the change with limited outcome reporting.

For water birth, we have evidence in all four tiers of the evidence pyramid.  We do have systematic reviews and meta-analyses and none of these syntheses have found a reason to suspect there is an increase in poor outcomes with water birth. We have randomized controlled trials and observational studies.  We also have case reports.


Jennifer (Author)