Welcome to part three of the Water Birth Series. If you find this information interesting, please be sure to check out the full meta-analysis. You can download it free from Midwifery until March 18, 2018: Neonatal outcomes with water birth: A systematic review and meta-analysis
Answering the question with a meta-analysis on water birth
In science, we figure out what is most likely true by figuring out what is not true. This is an important difference between scientific thought and the normal quick thinking we use to navigate our daily lives because it helps to prevent confirmation bias. Confirmation bias happens when we ask if something is true, and then only think about things that support what we already believe.
If you are inclined to think water birth is a great tool for laboring women, confirmation bias will lead you to think of “all the women and babies who use water birth without a problem.”
If you are inclined to think water birth is dangerous, confirmation bias will lead you to think of one or two times when someone gave birth in water and it didn’t work out well.
Confirmation bias can be an issue even when reading a single study. It may make us less skeptical of research that shows what we already believe, and it may make us overly dismissive of research that disagrees with our pre-existing thoughts.
A meta-analysis helps overcome confirmation bias because it includes data from several studies (in our meta-analysis we included 39) and usually synthesizes the findings from several outcomes (in our meta-analysis we included 10 outcomes). In this way, a meta-analysis helps us to see overall patterns in the research findings so we can understand the larger issue.
What we learned about water birth
We learned that there is no evidence in the existing scientific literature that there are an increased odds of poor neonatal outcomes with water birth. This is an important finding because it helps us interpret the case reports of poor outcomes with water birth — poor outcomes do occur with water birth, but they are no more likely to occur with water birth than with conventional hospital delivery.
But, we need to interpret these findings as water birth practiced according to a prescribed protocol. Obviously, this is not evidence that any woman can give birth underwater without increased risk. Every study indicated that women were required to meet certain criteria to demonstrate that they were at low risk of a problem. In addition, women were removed from the water if circumstances changed in their labor. This is no different than any other intervention used during labor. Women who desire to use an epidural for comfort must meet certain safety criteria before receiving an epidural, and the nurses, midwives, and physicians must continually assess the response of the woman and the fetus to the epidural so they can intervene if necessary. The same is true for induction agents, nitrous oxide, and even life-saving medications to prevent devastating effects of severe preeclampsia or post-partum hemorrhage.
The following are highlights of the specific outcomes reported in the meta-analysis.
1 Minute APGAR – no difference between water birth and conventional delivery.
5 Minute APGAR was reported by nearly every study. There was no difference between water birth and conventional delivery. More importantly, this finding has been stable since 1995 and additional research is unlikely to change the result.
Need for Resuscitation – Six studies, no difference between water birth and conventional delivery (stable since 1999)
Umbilical pH – Nine studies, no difference between water birth and conventional delivery (stable since first reported in 1997)
Respiratory Distress – Five studies, no difference between water birth and conventional delivery
Neonatal hypothermia – Four studies, no difference between water birth and conventional delivery (this finding did have a statistically significant difference, but this was due to the size of the sample. The difference found was so small you could never detect it by taking newborn temperatures.
Pneumonia – four studies, no difference between water birth and conventional delivery
Non-pneumonia infections – thirteen studies, a potential for lower odds of infection with water birth but no difference in sensitivity analysis.
NICU admission – Seventeen studies, a potential for lower odds of NICU admission with water birth but no difference in sensitivity analysis.
Neonatal death – four studies, no difference between water birth and conventional delivery
Shoulder dystocia – four studies, no difference between water birth and conventional delivery
Cord Avulsion – one study, no difference. This is an under-reported outcome overall. Anecdotal evidence suggests the risk increased in water birth due to the buoyancy of the water masking cord tension, but without descriptions of the frequency in conventional delivery, there is no way to compare the frequency.
What about bias?
If you read last week’s post about bias in water birth research, you may be wondering how that affects these outcomes. We used a few techniques to help understand how bias may alter the findings, these techniques are known as sensitivity analyses. Even correcting for bias, no outcomes were more likely to occur in water birth. This leads us to believe that bias may cause some studies to show better outcomes for a water birth, but the bias is not masking worse outcomes with water birth.
We had one more thing to test, and that is if the control group using analgesics made a difference. The studies were conducted throughout time and in many countries which created wide variation? So we thought we would use that to see if the differences in control groups were biasing the outcomes.
Eleven studies told us how many people in each group used an analgesic. Some allowed use in the in the water birth group. That is for three reasons 1) one study specifically allowed women in the water if it had been 4 hours since the medication had been used; 2) some prospective studies kept women in the water birth group for analysis even when they opted to leave the water for pain medication; and 3) many of the studies were done in countries where women can use nitrous oxide in labor – either in or out of the tub.
The use of analgesics for pain management occurs at a higher rate in the conventional delivery groups (which is often considered a benefit for water birth). We wondered if this biased our results – did it push our results to more favor hydrotherapy than it should?
The answer was no. Whether studies had similar or unequal use of mediations, the results were still no difference in neonatal safety with water birth.
We used the same analysis to examine bias from uneven use of uterotonics and found that it did not bias the results.
So, is water birth safe for the baby?
At this point, research demonstrates that there is no reason to suspect poor outcomes occur more frequently with water birth than with a conventional delivery. This does not mean poor outcomes do not happen, only that they occur as often as with women who are not delivering under water.