“The need for unbiased research struggles with the inherently biased sources of that information–human beings.”
—Kenneth Schulz (1998)
As the quotation above makes clear, all research is subject to bias, Childbirth U lecturers’ analyses of it not excluded. Objectivity isn’t possible, however much we may try, because we all come to the table with a point of view. The best we can do is to strive for transparency. By explicitly describing our philosophy and the methods by which we conduct our analyses, we can place all the cards face up on that table. Transparency will enable you to determine whether you share our point of view, find our reasoning compelling, and agree with the conclusions we reach. Here, then, are our beliefs about what best promotes safe, healthy birth and the basis on which we collect and evaluate the research evidence.
CBU’s guiding principle is the provision of optimal care. The concept derives from the American College of Nurse-Midwives’ Optimality Index-US. The Optimality Index-US measures positive maternal-child health outcomes within a system of care while factoring the care process into that evaluation by awarding a higher score to women of similar characteristics with similar outcomes who undergo fewer medical interventions.4 The index defines optimality as the “maximal perinatal outcome with minimal intervention placed against the dynamic context of the woman’s social, medical, and obstetric history” (p. 766). Adopting that definition and putting it more simply, optimal care becomes the least use of medical intervention that will produce the best outcomes given the individual woman’s case.
Optimal care, in turn, is rooted in physiologic care. Physiologic care is “the use of supportive care practices and low-technology techniques that facilitate the normal biological process of childbirth” (p. 3,).2 For healthy women experiencing uncomplicated labors, physiologic care is all that is necessary; however, even when problems arise, the escalation to high tech should not deprive women of high touch. Medical intervention should add to, not replace, providing emotionally and physically supportive care, having women participate to the extent safely possible in resolving the difficulties by their own actions, and respecting their right to informed consent and refusal.
CBU Review Methodology
The goal of our research strategy is to combine rigor with flexibility. CBU faculty support their positions with what amount to mini systematic reviews of the evidence. Systematic reviews are “studies of studies.” Unlike narrative reviews, which often provide no information on how sources were chosen, systematic reviews require attempting to locate all potentially relevant studies and determining the criteria on which they will be included or excluded ahead of time. (Systematic reviews may also, but don’t necessarily, include meta-analyses, statistical analyses that pool data among studies. We may cite data from meta-analyses, but performing them is beyond our scope.)
To ensure applicability and utility to CBU’s audience, we include only studies describing contemporary care that are conducted in countries with adequate medical resources and that report clinically significant outcomes, that is, outcomes of sufficient importance and impact that they would influence your decisions. In addition, studies must provide enough information to allow judging their quality. Accordingly, we exclude:
- studies published in languages other than English: the English-language abstracts that accompany some non-English-language studies do not provide sufficient information to judge study quality (exception: the lecturer can read the study in its original language);
- abstracts: studies published only as abstracts or brief reports also do not provide sufficient information to judge study quality;
- narrative reviews, commentaries, clinical guidelines: these are all opinion pieces and as such provide unacceptably weak evidence, although they may be cited or quoted to furnish additional support and credibility to our arguments and recommendations;
- studies reporting only surrogate outcomes, that is, laboratory values or imaging results believed to be in the pathway to outcomes of clinical importance but not in themselves of clinical importance or problems that may be part of normal recovery after birth: therefore we don’t, for example, report comparisons of Apgar scores, which evaluate the baby’s condition at birth, because low Apgar scores predict long-term adverse outcomes poorly or prevalence of urinary incontinence limited to the first few months after birth;
- systematic reviews that include other systematic reviews: we felt that this removes us too far from the original sources and therefore the ability to evaluate them;
- unpublished studies or those not published in peer-reviewed journals: publication in a peer-reviewed journal by no means guarantees a quality study, but the peer review and editing process at least sets a minimum standard.
We also have exclusion factors that don’t necessarily apply under all circumstances, specifically:
- studies published before 1990: we chose a cutoff date, albeit an arbitrary one, to ensure that care in the study was reflective of modern-day care (exception: lecturers may choose a different cut-off date if that makes sense for a particular topic, as, for example, when current treatment is based on a body of research conducted prior to the cut-off date);
- studies in low-resource maternity care systems, except in rare cases where we lacked data from high resource maternity care systems and the study’s methods section established that care in that setting provided access to appropriately trained and knowledgeable clinicians and modern technology, or, alternatively, lack of resources wouldn’t affect the outcome of interest;
- studies included in included systematic reviews, as doing so would duplicate results, thereby giving a misimpression of the strength of the data. Sometimes, though, a systematic review didn’t report on an outcome of interest. In those cases, we included the study for that outcome alone, but excluded it for remaining outcomes included in the review.
For some topics, we added exclusion factors unique to that issue. For example, we excluded studies comparing outcomes between cesarean surgery and vaginal or planned vaginal birth that failed to take into account factors that would increase both the likelihood of poor outcome and the likelihood of cesarean delivery.
Studies could be excluded from one mini-review but included in another, as it is possible for study weaknesses to affect reliability of some findings but not others. Conversely, some studies that met our general criteria were excluded for specific flaws or weaknesses. We sometimes deconstruct influential studies falling into that category.
We don’t adhere strictly to the study-design pyramid, which ranks study designs hierarchically according to the strength of evidence they are believed to produce, because, as has been pointed out,1 different study questions require different study designs. That being said, if systematic reviews or randomized controlled trials (RCTs) (studies in which investigators allocate participants by chance to one form of treatment or another or to treatment versus usual care) adequately address all outcomes of interest, we may exclude studies of other designs. When using systematic reviews, we add studies that met the review’s criteria but were published subsequent to the review in order to keep lectures up to date.
One final word
An inherent defect of research conducted under medical-model management is that we don’t know what outcomes would have been had women received optimal care. For example, among women planning vaginal birth after cesarean (VBAC), VBAC rates in hospitals average 74% whereas in a study of freestanding birth centers, where women were under the care of midwives, the rate was 87%, 13 more VBACs per 100 women.3, 5 Even among women with no prior vaginal births, which decreases likelihood of VBAC, the birth center VBAC rate was 81%, 7 more per 100 than the mean hospital rate. Keep the confounding effect of medical-model management in mind as you view CBU lectures.