Providing the results of a trial to its participants: qualitative study

How do the participants of studies react to receiving a summary of the results of the trial in which they took part? Christiane Rehwagen critically appraises a qualitative study

This month's paper is Dixon-Woods M, Jackson C, Windridge KC, Kenyon S. Receiving a summary of the results of a trial: qualitative study of participants' views. BMJ 2006;332:206-10. You can read the paper by going to studentbmj.com and clicking on the link.

Why did the authors do the study?

Since the 1990s, the idea that research participants should be given the results of studies in which they have participated has gained ground. In 1997, the UK National Childbirth Trust and the Association for Improvements in Maternity Services stated that participants in studies in pregnancy should have the right to see the results. A recent edition of the Department of Health's research governance framework (2005) recommends that findings from research work should be disseminated promptly and fed back as appropriate to participants. Despite these strong positions, little research has been done, and according to the authors there is no detailed understanding of the outcomes and implications of providing such feedback and how it should best be managed. This understanding is needed so that decision makers have evidence on which they can base their decisions. In view of this, the authors thought it timely to investigate how study participants respond to receiving results of the trial.

What did the authors do?

To answer their research question, the authors did a qualitative study. Qualitative research often analyses verbal data collected in interviews. It is good at asking why something is the way it is (compared with quantitative research—for example, case-control studies or randomised controlled trials, which set out to measure something and investigate to what extent a phenomenon exists rather than why). Here the authors wanted to find out what trial participants thought about receiving results of the trials in a leaflet at their homes. They investigated the views of a subset of women who had participated in the ORACLE trial, a large randomised controlled trial of antibiotics in pregnancy. It studied women less than 37 weeks gestation to see whether broad spectrum antibiotics prolonged labour and reduced neonatal morbidity in those who were in preterm labour or had prelabour rupture of the membranes.

Data collection: semi structured interviews The authors did face to face semi structured interviews with 20 trial participants. Questions in semi structured interviews are open ended but are based on a predetermined list of topics. Usually, as in this study, an interview prompt guide is used, which is developed after literature review and discussions within the project team to structure the interviews.

The interview is then used flexibly in response to the way in which participants want to direct the interviews. This is important because qualitative interview data are often used in explorative research, where little is known about the research problem. The interviewing author of this study also kept a diary to record contextual details of the research. All interviews were transcribed verbatim.

Analysis of the data: themes Qualitative data require careful and systematic analysis. In this study the authors analysed the interview responses using the “constant comparing method.” This means that they studied the ideas expressed by the participants to find themes (“codes”) while comparing these with the original interview data. As with statistical analyses, computer software can be used to help organise the data. However, all final decisions about what themes and patterns are present within the data are made by the researcher.

What did they find?

All 8941 women who were recruited into the ORACLE study in the United Kingdom were offered the opportunity to request the trial results, but only 1803 (20% of all participants) requested this information. This low percentage of trial participants wishing to receive a summary is already one part of the results.

The authors then looked at participants from the Trent region (qualitative research often studies a small group of people), where 193 women requested the results; all the women received a letter inviting them to participate in an interview about the results leaflet. Only 22 women agreed to be interviewed, and 20 were interviewed. They were between 22 and 33 weeks pregnant at the time they were recruited into the ORACLE trial. No participants had babies who had died after participating in the trial. The sample was socially (although not ethnically) diverse. The authors presented the patients' responses to the study in four themes: getting the results, positive reactions to the results leaflet, negative reactions to the leaflet, and knowledge of the results.

Getting the results Of the participants who spoke about the method of sharing the results with participants, most found receiving a results leaflet through the mail satisfactory or preferable to personal contact. This is how one patient expressed it: “I was fine with the letter. You can, because...you can go over and over and over it, like a phone call you can't always remember what was actually said and things like that, and, you know, and even if somebody comes to your door, if you've got a letter and you're on your own you can take it how you want to take it.” Positive reactions to the results leaflet For most women the leaflet had a positive impact or little impact. Half expressed feelings of pleasure on receiving the leaflet, particularly at what they saw as the success of the trial, or felt that taking part had been worthwhile. Negative reactions The most common reason for disappointment on receiving the results, experienced by most interviewees, was the lack of personalised information. Participants wanted to know the arm of the trial to which they had been allocated and the possible consequences of this. Several participants found the leaflet difficult to understand. One negative consequence of receiving the results was that for some women it revived memories of a difficult time. Knowledge of the results After reading the leaflet, most participants showed at least some knowledge of the study results, although there was also evidence of confusion. For example: “No, well ... it went into what the trial was about and things like that, but it didn't really give you any results or any concrete evidence, information—‘the trial worked,' ‘the trial didn't work,' ‘it was a success.'”

What are the implications of the findings?

After summarising and evaluating all the responses, the authors conclude that feedback to participants in a trial is not harmless and that further research is needed. The finding of this study that only a fifth of women who participated in the ORACLE trial wished to receive the trial results is different from previous small studies that have found much interest in receiving research findings among participants. Even though reactions to receiving the summary of study results were generally positive, there were signs of negative implications for women who had adverse outcomes. The bioethical debate about the need for respect for autonomy should take this aspect into account. In that way our new understanding of trial participants' views might be helpful in evaluating the recommendation that results should be routinely given to participants. It also shows that we need to know more about appropriate methods for disseminating trial results to participants.

PHOTOS.COM

One important finding of the study is that leaflets giving only a summary of results seem to disappoint participants. Many of the trial participants interviewed would have preferred to find out in which arm of the randomised controlled trial they were, and this was the main reason why they requested the results. This is understandable, but leads to difficulty for researchers in terms of how to provide feedback to participants, particularly if they were in the least effective arm of a trial or had adverse outcomes.

A number of possible conclusions can be drawn from this study. One is to abstain from feeding back until more data on how to feed back trial results to participants are available. Another conclusion might be to continue to give feedback and accept that there will be a downside for some trial participants. A third conclusion might be to provide individual feedback to all participants. The best solution may be to vary the level and format of feedback by type of study and study population.

Is the study valuable?

When reading a research paper it is a good idea to ask if the findings are interesting and important and whether they make a meaningful contribution to understanding the field or to patient care. A related idea is the generalisability of the results. In this case, can the findings be applied to other participants of this trial and to participants from other trials?

This study was of a subset of women who had participated in a randomised controlled trial when they were pregnant. This probably limits the applicability of the study results to other trial populations participating in different kinds of trials.

Of the 193 women who had requested the results of the trial, only 20 were eventually interviewed. Presumably those women requesting the results and agreeing to be interviewed were the ones who felt strongly about something and are not representative of all participants. Design and content of the leaflet may also have influenced the results.

Another problem with this paper is that it doesn't clearly distinguish between giving a general summary of the trial results to a study participant and providing them with their own result, which is an important difference. This study adds to our knowledge on the response of women receiving a summary of trial results in pregnancy, but this constitutes only a small aspect of a broad question. We are left with the message that further studies are needed. Finally, one characteristic of qualitative research is that it is good in exploring areas where there isn't much evidence available.

 

Christiane Rehwagen, Roger Robinson editorial registrarBMJ
Email: crehwagen@bmj.com


studentBMJ 2006;14:177 - 220 May ISSN 0966-6494