evidence-informed midwifery 2:
using research in midwifery practice
sara wickham

This article was originally published in
Midwifery Today. Republished with permission.

In the first article of this series I discussed how research, although the dominant form of evidence considered by the medical model, is only one example of the myriad evidence available to women and midwives. Although we might like to see a framework where this is given less credence by all birth professionals, I suspect the process of scientific enquiry will always remain with us, although more people will recognise its limitations. However, it remains important for midwives to understand the research process and be able to evaluate research relating to birth.

One of the main criticisms of research is that it is never value free, even though this is one of the 'goals' of the scientific process. Because the research we consider in practice almost invariably involves women and babies, and because it is carried out by humans, it is not possible to hold all variables constant, replicate experiments and remain completely objective about the results. All research is biased - it is just a question of which direction this bias comes from, and to what degree it could have affected the results. Again, if we are to use evidence from research effectively, we need to be able to evaluate this.

One of the ways I have found to date of keeping any research in perspective is the following 'three-point plan', which can be used when considering any research paper. This paper then becomes the starting point for consideration of the evidence in general. Although this works best when applied to research considering an intervention or intercession, it can be used generally to evaluate any 'new' evidence.


The Three Point Plan

1. Evaluate the paper you have. Find its strengths and weaknesses. Consider whether you think it is 'good' research.

2. Evaluate the body of research on the subject. Find other papers that address the same topic and evaluate them. If this is too time-consuming, find reviews of the evidence and evaluate those instead.

3. Evaluate the other evidence on the subject. What do you know from your experience? What do other midwives think? What does your intuition tell you? How does this relate to practice?

This might be best illustrated with an example. One of the areas which is currently under debate in the UK is the way the third stage should be conducted. Most areas have a policy that the third stage should be 'managed' with an oxytocic drug and controlled cord traction to deliver the placenta. If women prefer to have a physiological third stage, they can 'opt out' of this management, but medical management is the norm.

Where units cite research to 'back-up' this policy, they invariably mention the 'Bristol Third Stage Trial' and, more recently, the 'Hinchingbrooke Trial'. Both of these trials conclude that giving an oxytocic (by intramuscular injection) following birth reduces bleeding and reduces the need for blood transfusion. So is this really the case?


Step 1 - Evaluating the Paper

The problems with both of these papers are numerous and the Bristol Trial is generally recognised as methodologically flawed, which was one of the reasons that the later Hinchingbrooke trial was carried out. However, even this trial did not address all of the issues; there are questions as to whether women were randomly allocated to groups effectively, and there is an issue with confounding variables, which may have biased the results. Women who were given opiates in labour , or who received an episiotomy, remained in the study. As both of these interventions may have an impact on the amount of bleeding, we may need to be more sceptical about the results.

Michel Odent (1998) highlights one of the biggest issues with these studies; that they focus on the issue of 'management' and define physiological management not in terms of the positive things we do to help women through the third stage (such as heating the room, or breastfeeding), but instead on the absence of intervention. This in itself may cause us to suggest that the results of the studies were not as valid as they might be
.

Step 2 - Evaluating the Body of Research

With the area of the third stage, this is a difficult step. There are few papers which consider this area, although smaller studies (some of which were carried out by independent midwives) show results which are diametrically opposed to the larger medical trials; that women have better outcomes with physiological third stage and that they do not generally suffer compromising blood loss. However, the numbers of women in these studies are low, which some would argue may also make them biased. Although many people have written about the third stage, there are few studies which have approached this topic from the physiological perspective which Odent, and many midwives, would argue is so important.


Step 3 - Evaluating the Body of Evidence

The experience of independent, holistic midwives is vital to this area. These midwives are focused on the physiology of birth and the third stage, and report that enabling the third stage to occur naturally is not a dangerous scenario, but safe and preferable to many mothers, with the proviso that oxyticic herbs and drugs (as well as a number of 'tricks of the trade') are at the midwife's disposal in case they are needed.

Knowledge from women also adds to this debate; many women find the administration of an oxytocic to be painful and unpleasant, and would prefer not to have this. There is also a theory on some British labour wards that some of the 'batches' of oxytocic drugs may be flawed. The reason for this is that some midwives have noted that retained placentas rarely happen singly, but often in moderate numbers. While this may be coincidence, some midwives believe that the occasional batch of imperfect oxytocic may be the cause of this.

My own experience of working nights on a postnatal ward led me to question one part of the received view; that oxytocic drugs cause less bleeding. Women would come to the postnatal ward an hour or two after birth, having had a managed third stage and a recorded blood loss of 100-200mls. When they first went to the bathroom, they often passed large quantities of clotted blood into the toilet. I wondered if the blood loss was really less with oxytocics, or whether they caused the uterus to contract so hard that some of the blood was retained until the uterus relaxed again and expelled it.

We also know from physiological evidence that exogenous oxytocin inhibits endogenous oxytocin, which may also mean the woman's body struggles to keep the uterus contracted in the first few hours after birth (and the oxytocic), which may also lead to increased blood loss in the initial postpartum period. The fact that we generally do not record the amount of postpartum blood loss may have led us to overlook this possibility.

So it can be seen that not only is there a great deal of evidence which is outside the capacity and remit of the research evidence, but that this evidence (as well as the initial evaluation of the studies) throws doubt on the conclusions offered. However, in some areas, other forms of evidence concur with the results of research, or add important knowledge to that which it offers. This is why I would suggest it is vital that we look at all of the evidence available, in order to give us the 'bigger picture'.

Research is being increasingly carried out by midwives and generally seems to be proving the things that midwives have known in their hearts all alone; that birth is better left alone, that intervention should be occasional, individualised and appropriate and that most routine interventions cause more harm than they do good. This may be one of the reasons that it is important for us not to 'throw out the baby with the bathwater' when it comes to research; I would suggest that it is a very effective tool for birth change.


Evaluating Research

There are two main types of research; the main aspects of which are outlined in table 1. My focus here will be on the evaluation of quantitative research, which is the type most commonly used to explore and investigate aspects of birth. I should add that this is only a broad overview to help the reader who may be new to the subject - there are several good books which explore this process further for midwives who want more detail or are thinking of carrying out research themselves.

Table 2 offers an outline of the process of research, or the stages which researchers should go though when carrying out a study. The reason for including this here is that it is important to understand what researchers should be doing in order to evaluate how well they have done it! Although there are deviations from this 'norm', the stages which have been carried out during the research should be explained for the reader, and any deviations from this should be explained and justified in the paper.


Table 1: Types of Research

Quantitative Qualitative
Deductive ~
Testing theory 		Inductive ~
Generating theory
Experimental Observational
'One-shot' ~
measure as much as possible 		Process-orientated ~
aim for richness
Large samples ~
Data in numbers 		Small samples ~
Data in words
Medical model ? Midwifery model



Table 2: The Research Process (Quantitative Research)

1. Formulate / define the research problem
2. Review the literature
3. Develop a theoretical framework
4. Formulate a hypothesis
5. Select a research design
6. Specify the population
7. Select methods to measure research variables
8. Conduct a pilot study
9. Select the sample
10. Collect the data
11. Prepare the data for analysis
12. Analyse the data
13. Interpret the results
14. Communicate the findings


The first stage of evaluating any paper is to read the abstract of the paper (if one is included). What are your first impressions? How do the conclusions relate to what you already know or believe? It is at this point that we need to be honest about our preconceptions and consider the need for objectivity! However, we also need to assess the objectivity of the researchers; who are they? Might they have a personal bias? What is their experience, qualifications, role? Was the research funded, and could the sponsors have an interest in the outcome? Finally, where is the research published, and could this tell us anything about the study?

The other important aspect to consider are the ethical issues; was the research fair to participants? All research should follow the guidelines set out in the Nuremburg Code, which underpins the Declaration of Helsinki. However, we know this is violated by some researchers, particularly where the testing of pharmaceutical drugs is concerned. Was the research taken to an ethical committee for approval, and was this approval given?

A number of midwives have asked me what they should do when evaluating research. Although there is no right answer to this question, my personal approach is as follows. As I read through the study and look at various issues, I make pencil notes in the margins, often phrased as questions. This allows me to record all of my thoughts, which I can then come back to later and consider more fully. Often the questions I asked at the beginning of the paper are answered by the time I reach the end; in other cases they remain unanswered and may highlight gaps in the research. Making notes from the first time I read the research enables me to remember my first impressions, even when I have read the study a number of times. If I am writing an evaluation or precis of the research, I then have these notes to use as the framework for my response.

Having read through the paper and jotted down first impressions and questions, the easiest way to begin to evaluate the research is to use the framework of the research process and consider each aspect of the process as it relates to the paper. Table 3 shows some of the questions which might be asked at each stage. Clearly, it is important to identify the positive as well as the negative aspects of the research in order to make a fair evaluation. This evaluation can then become the starting point for discussing the other evidence and putting the research in the context of experience, philosophy and all of the other forms of evidence upon which we rely.


Table 3: Questions to ask when Evaluating Quantitative Research


1. Formulate / define the research problem

Does the question relate well to practice? Is the question relevant? Is it asked in an appropriate manner? Is other (previous) research relating to the area discussed?


2. Review the literature

Was an appropriate literature search carried out? Is the literature cited up-to-date? Does the author give a fair evaluation of other people's findings? Are both sides of any argument presented?


3. Develop a theoretical framework

Is the framework used appropriate to the study? Is it suggestive of the conclusions? Could it suggest bias?


4. Formulate a hypothesis

Is the wording of this appropriate? Is it suggestive of bias?


5. Select a research design

Does the design fit the question and aims of the research? Is it the design most likely to give the 'correct answer'? Are the pros and cons of this discussed? Is justification given for the use of the chosen design?


6. Specify the population

Are the participants suitable? Who was not included in the research, and why?


7. Select methods to measure the research variables

Are these methods appropriate? Would other measures be better?


8. Conduct a pilot study

If a pilot study was used, were the results used to improve the design?


9. Select the sample

How did the researchers decide who would be eligible for entry? Was a control group used and was randomisation used appropriately? Were the groups studied large enough to make a fair assessment of the effectiveness / side effects of whatever is being studied? Was double blinding used? (where neither the researcher or the participant knows which group a person is in) Did all participants give informed consent?


10. Collect the data

Does the paper tell you how, when and where this was done? Do you feel this was carried out appropriately? Who collected data? (Some kinds of data may be more subjective and therefore more open to bias - eg results of vaginal examination.)


11. Prepare the data for analysis

Does the paper explain how this was done?


12. Analyse the data

Is the method of analysis explained and justified? Are reasons for using particular statistics or other tests given? Is the probability of getting the results by chance included?


13. Interpret the results

Is this process intelligible and relevant? Are raw figures given as well as percentages? Are data given in graphs and tables explained? Could pictoral representations of data be misleading? Is the discussion of results appropriate and relevant? Are conclusions appropriate and does the research 'fit together' well?


14. Communicate the findings

Is the research 'reader-friendly' and well-presented? Is there a logical progression to the paper? Are there any gaps? Is anything 'glossed over'? Are the limitations and areas of potential bias discussed honestly?