A qualitative analysis exploring preferred methods of peer support to encourage adherence to a Mediterranean diet in a Northern European population at high risk of cardiovascular disease

Participants were recruited from general practices via posters, through contact with community groups and networks, and identified from hospital outpatient clinic lists, which were scanned for suitable participants, who were contacted by letter to ask them whether they would like to opt in to study participation. Former participants in dietary intervention studies at the Centre for Public Health, Queen’s University Belfast, who had consented to future contact, were also invited to participate via telephone. Participants were eligible if they were aged over 50 and had two or more specified risk factors for CVD (overweight or obese, smoking, hypertension and hypercholesterolemia). Participants were excluded if they had existing CVD or type 2 diabetes.

Twelve focus groups were held between December 2012 and July 2013 with between 2 and 11 participants in each. This number of groups facilitated the recruitment of a diverse study population allowing us to access a diverse range of perspectives and achieve data saturation. Each focus group was similar in socioeconomic status and homogenous in gender, this was considered important to facilitate uninhibited discussion.

Focus group sampling was purposive, to ensure that people of different ages, genders, and from different socio-economic, domestic and geographical settings took part.

Prior to each focus group, demographic and dietary information (8-item MD food frequency questionnaire) was collected from all participants in a brief questionnaire. The MD questionnaire was developed based on a validated 14-item MDS questionnaire [25] and current guidance for MD consumption [26]. Points were allocated for high consumption of olive oil, fruit, vegetables, oily fish, wholegrains or nuts, moderate consumption of wine or low intake of red meat. Two forms were provided at the conclusion of each focus group, one for participants to indicate their personal preferred method of peer support and the other to indicate their personal preferences regarding peer supporter characteristics.

Information from the questionnaire was entered into IBM SPSS Statistics (v 21) and was analysed using descriptive statistics (Table 1). Preference for peer support approach was examined for all participants, and by gender (Table 2). Preference for peer supporter characteristics was collated (Table 3).

Vignettes were used to stimulate conversation and improve understanding of the different peer support methods for participants. Vignettes are concise stories explaining similar characters and backgrounds with only one differing feature under investigation, which in this case was the method of peer support provided. They were developed by the research team based on examples used in previous studies [27‐29]. Four different vignettes were utilised, each describing one of the four peer support methods: group, one-to-one, telephone and online. A unisex name was used in the vignettes therefore identifying with both genders during discussions (see Appendix).

Focus group schedules included discussion of dietary change, MD and heart health prior to the introduction of peer support as a method of encouraging dietary change towards a MD. Facilitators introduced each method in turn using the appropriate vignette followed by a discussion stimulated by questions such as: ‘do you think such support as I have just described might be useful in encouraging you to eat a Mediterranean diet?’ and ‘what do you like and dislike about it, and why?’. After all of the methods of peer support had been introduced, the facilitator asked what participants’ preferred method would be or whether a combination of methods could be helpful. Following this there was a brief discussion about the characteristics and skills required of a successful peer supporter. All discussions were tape-recorded and transcribed verbatim with all names changed to pseudonyms to ensure confidentiality and anonymity.

Qualitative data were analysed thematically using an inductive approach. Transcripts were read multiple times by CE to facilitate the development of codes indicating deducted themes and sub-themes with a clear focus on what participants preferred and why. The original coding framework was discussed with CMcE and further development of themes was discussed with MMcK. Consideration was given to contradictory or opposing views, and differences in acceptability of peer support methods for different groups. Transcripts were then coded using NVivo 10, a qualitative indexing software package, to manage the data and were further analysed to identify subthemes and illustrative quotations.

A total of 67 participants took part in focus groups, 27 males and 40 females. Three quarters of the focus groups took place in an urban setting (n = 9) and one third (n = 3) in rural settings.

The age of focus group participants ranged between 44 and 86 years (mean 64.0, SD = 10.03). The majority of participants were overweight or obese: BMI values ranged from 20 kg/m² to 41 kg/m² (mean 28.8 kg/m², SD = 4.53). MDS ranged across groups from 0.6–3.5 out of a possible total of 8 (mean MD score 2.3, SD = 1.57). Previous awareness of MD ranged from 10 to 100% across groups; 40.9% (SD = 0.50) of the total study population were aware of the MD. Overall, 98.4% (SD = 0.13) of participants considered themselves open to changing their diet.

As indicated in Fig. 1, based on the individual preference scoring sheet, the group peer support was the most preferred approach, followed by a combination of methods, a peer mentor approach, and the least preferred options were, equally, telephone and internet based support.

It was apparent from the focus group discussions that a group model of peer support was already familiar to many participants, as two commercial weight loss schemes which involve a substantial group peer support component were mentioned in every focus group. One participant mentioned previously taking part in a study involving mentor peer support and another discussed taking part in internet forums for weight loss.

Analysis of the transcripts elucidated some of the reasons why group peer support was the most preferred method. The reasons were encompassed in three themes: components of an effective peer support group, catalysing motivation and stepping stones of change and are described in more detail below. The last theme also included discussion of the merits of having individual choice or using a combination of approaches to deliver peer support.

This theme emerged from identification of the predominantly positive discussions surrounding the group peer support approach. Participants identified what they considered to be the useful components that would facilitate an effective peer support group. These are depicted in Fig. 2, each layer of the circle illustrating a necessary component of an effective peer support group, with each layer building upon the one before.

Figure 2 demonstrates that the core elements for an effective peer support group are sharing knowledge with a group face to face. Group peer support was considered preferable to telephone and mentor peer support as it was seen to enable participants to learn more from interacting with a number of people. It was suggested that this would be particularly beneficial in developing problem solving techniques.

Give people ideas and you learn from other people’s ideas. FG8 (female, less affluent).

I think with this sort of approach when you’re getting together with other people you’re able to sort of experience what other people are eating and see then, even swapping dietary advice, cooking tips and things like that there, ways to prepare food, even on a reasonable sort of budget, if you were trying to cook for a family but also for yourself. It would be a lot of knowledge from other people getting together. I always think you learn something which you maybe don’t know. FG5 (male, less affluent).

Face to face contact was one of the key perceived benefits of group peer support over internet and telephone peer support. It was suggested that this would increase honesty and accountability between peers and increase the perceived level of support.

I think it’s always better to see a person and have that sort of human contact to know that there is someone on your side. An anonymous voice at the end of a telephone is not as reassuring, and if you have a face to that voice behind the phone, that’s fine as an extra support. But when you think you’ve got somebody in your corner and you’re looking at them and you’re talking to them and that’s helpful; an anonymous person at the end of the phone, no. FG8 (female, less affluent).

People that need to make dietary changes and don’t see anyone face to face probably lie [laughter], because I would. “Yes, I’ve been really good I’ve been brilliant!” FG5 (male, less affluent).

Participants discussed that having face to face contact with a group, and sharing knowledge with each other would help to build trust, honesty and accountability. These elements were discussed as being essential for an effective peer support group.

It seems to me the effectiveness of the group would depend on two things. One of them would be the extent to which people are able to trust each other and be open and honest about what their habits really are, and secondly, if you’re going back and you know that somebody is going to say to you, “well, what changes have you made?” and you are going to be held to account. FG7 (male).

In the quotes below, phrases such as ‘collective message’ and ‘fellow strugglers’ indicate formation of a group identity through the group support model. During some focus groups (both those newly formed for the study and those formed from previously existing community groups) it was suggested that the peer support may be particularly effective within existing groups who already have a close bond.

I think groups are good because not only do they convey a message, a collective message, you get a sense of bringing something together. FG10 (male, more affluent).

I’m just trying to imagine it, and I think it’s why I would be more comfortable in a group environment than with this (peer mentor approach), I think in a group environment you’re with a bunch of, if you like, fellow strugglers. FG7 (male).

Theme 1 brings together the explanations offered by participants for their preference for group support to identify the necessary components of a successful peer support group.

Discussion regarding peer support groups was predominantly positive, although there was some mention of the potential that it may be inconvenient due to its time, date, location or commutability. There was also some discussion that group support may not be suitable for everybody as people may not feel comfortable in group situations. This issue is discussed further in theme 3.

Table 2 provides exemplar quotes for this theme. Participants discussed how motivation was an important factor in helping people to change to a MD and that different models of peer support might provide different types and levels of motivation. Participants felt that the group model of peer support would use interpersonal factors to catalyse motivation through support and encouragement between members, peer pressure, competitiveness and feeling accountable. Measurements of weight or blood pressure at group meetings were also seen to provide evidence of success, which participants also considered to be motivating.

Participants discussed how individuals would need to have an initial motivation to attend the peer support group, such as being overweight, having high blood pressure or cholesterol or having a friend or family member who had suffered heart disease. Likewise, they acknowledged the challenges of sustaining motivation such as having a busy lifestyle, being distracted by television, not wishing to travel in poor weather or having difficulty forming positive relationships within the group.

Theme 3 (stepping stones of change) comprised discussions about how a combination of methods may be the most effective way of encouraging behavioural change and maintenance. It was evident in discussions that, whilst group peer support was overall the preferred approach, participants also felt that other methods of peer support might be useful at different points in time or for different individuals. Participants conceptualised change as a process, and considered that, at the beginning of the process, more intense support might be required such as would be provided by the peer mentor approach. They suggested that this could be followed by face to face group support and that internet or telephone approach could be utilised to assist in the final stages or maintenance of change depending on personal preference.

Mentor peer support was seen as the most intensive intervention and it was suggested that this may be most useful when a person was starting the process of changing as well as for those who dislike attending groups.

The one to one is nice but it could be a bit intense after a while. FG8 (female, less affluent).

But I think, getting back to this whole situation, it does boil down to group support, and for some for a period of time maybe one to one just to keep you going. When you’re starting something new you need encouragement and that is when you might need it and then when you get going, if that person who you had as a one to one would always be at the end of the phone line if you needed to contact them to say help. FG8 (female, less affluent).

I think it very much depends on the individual. If you get a real shrinking violet they’ll not want a group session, whereas us four here, we can all talk the best and amongst each other but it’s what an individual’s character, what their makeup is too. FG10 (male, more affluent).

Participants suggested that a group-based approach was the most preferable method to follow mentor peer support. To sustain change, it was considered useful to have a website to refer back to or a peer supporter to call when experiencing difficulties.

It’s like a personal trainer, I suppose, that type of idea. But I think the group idea is better than the one to one, I would prefer that myself, because if it’s a one to one it’s complete focus on each other and I don’t know, it just wouldn’t … I would prefer the group, I wouldn’t like the one to one basis, I wouldn’t really. FG4 (female).

I think the one to one to start with and then the group and maybe then that site that I can dip in and out of, because if shops ask me for my email and things like that I just couldn’t be bothered receiving all those. But if I knew there was a site there that I could go into and get some bits and pieces then, yes, I would do that. FG6 (female, more affluent).

Yeah, I like being in a group session and looking up a website and seeing different menus, easy steps to follow through. FG12 (male).

The group’s the best. Through the group you can expand all sorts of options, you could have ones meeting up for coffee to see how they’re getting on. If you had to phone Joe Bloggs, you know who Joe Bloggs is and what they look like and you know what they’re talking about, but not a complete anonymous voice. FG8 (female, less affluent).

Participants also suggested that different people may require or prefer different models, so having the option of when to move to each model or which to use might be very useful.

Surely the lesson of life is that it’s neither one nor the other that might work but it’s usually a combination of things. FG10 (male, more affluent).

I sort of see a combination developing for me, just going through my head listening to the scenarios and different things. Initially you start off in a group. Some people are that wee bit more sort of confident and then can go off and do their own thing, some people aren’t just as confident maybe, they’re new to it and they need to learn skills, give them that extra help. And then the third step being … you were talking about links with computer, email and different things and then sort of stem off for that. So people get what they need out of it; if you need a lot you can get a lot. FG1 (male, less affluent).

There was some discussion during focus groups about how the face-to-face peer support groups could look from a more practical perspective. There was preference for an informal environment, such as a community centre. Weekly or monthly meetings were suggested as the ideal frequency, with some participants suggesting a tapering of the frequency may be helpful to achieve dietary change.

I would say you probably need more meetings at the start … to change and then less after that. But initially you would need more. FG6 (female, more affluent).

The sharing of personal experiences was considered useful for making dietary changes, this has been discussed in theme 1. Practical sessions, such as cookery demonstrations, were also considered useful.

Measurements at group sessions were discussed as being helpful in that they could provide an indication of success and also provide a justification for changing diet.

If someone’s very much overweight then there’s a good chance that there’s a greater risk of heart disease. I think that’s well accepted in society. If you’re overweight you’re putting more pressure on your heart, and so I think they already have an indicator. But someone like myself, I’m not skinny and I’m not overweight but I’m somewhere in between, people would be saying to me ‘what do you want to do that for? You’re all right.’ But if I had some kind of indicator that said ‘oh...’ then I might say ‘yeah, definitely, this is a good idea. Go for it.’ FG10 (male, more affluent).

Male focus group participants discussed that it could be beneficial to bring a partner to the group meetings, particularly if their partner was the main cook in the household.

If there was somebody living long term in the house you would want them involved for moral support as much as anything else. FG10 (male, more affluent).

Focus group participants were asked whether the background, age and gender of peer supporters would be an important factor in a questionnaire (Table 3). The majority did not think it was important that the peer supporter would be a similar age (75%), gender (75%) or from the same area (77%). Rated most important were that the peer supporter would have successfully made changes towards a MD (83%), they would have expert dietary knowledge (72%) and the group members would feel the peer supporter is like them and wanted to make similar dietary changes to them (71%).

Focus group discussions revealed that a good peer supporter should be empathetic, encouraging and have a good sense of humour, plus they should have personal experience of eating a MD, good knowledge of health and budgeting and strong communication and listening skills. See Table 4 for exemplar quotes regarding preferred qualities, attributes and skills of peer supporters.

Focus group discussions in a Northern European population at high risk of CVD found that there was a strong preference for group peer support to encourage adoption and maintenance of a MD.

This study is one of few in the literature to employ a thorough approach to peer support intervention design, allowing the target group to shape its format, content and delivery and contributing to behaviour change literature. The ‘Football Fans in Training’ Intervention [30] developed an expert multidisciplinary team to decide on their preferred method of peer support delivery. Although it focused on changing physical activity patterns rather than diet, there are some key similarities to our findings; aiming to focus on practical and experiential learning which authors believe strongly contributed to its success.

Theme 1 identified the predominantly positive viewpoint taken in discussing the group peer support model and how participants considered some of the key ingredients for an effective peer support group. One of the key ingredients for an effective peer support group was to the opportunity to form a group identity, this has been discussed as important in literature [31‐33]. Interestingly, group identity is often discussed in the literature in accordance with groups developing a ‘sense of ownership’. Such an issue was not discussed by participants during focus groups in this study, where the focus was more on the support, commonality with other participants and motivational aspects gained by giving and receiving support in a group setting.

The benefits of group peer support in catalysing motivation have also been found in a study of mental wellbeing in which group members reported feeling ‘uplifted by exchanging emotional and practical support; they gained self-esteem, knowledge and confidence, thereby increasing their control over their situation’ [34]. Similar suggestions of challenges to motivation, including inconvenience, lack of interest and poor relationships, were found in the literature [35].

The results of this study demonstrate clear preference for a group peer support approach. However, the findings reported under Theme 3 suggest that a combination of peer support approaches may be most likely to suit the widest range of people. Other research teams have faced the challenge of developing an intervention suitable for a number of people, one utilised telephone support for people with diabetes and found that some participants preferred to speak to a diabetes nurse than a peer [36]. Other studies have found that a structured group-based support does not appeal to all patients or populations, and that drop-in groups or telephone contact may work better for those participants [21, 32, 37]. In this study, participants also discussed the challenges of constructing an intervention that could serve everyone, and recommended a combination approach, involving a group support component. Due to the complexity of behaviour change, it is prudent to test one method thoroughly; the process evaluation of which will determine whether additional peer support methods are required. Each additional layer of approach is likely to incrementally increase complexity and cost, and these are important considerations for implementation of public health interventions.

The peer supporter role has been discussed in the literature with regards to the need for training, the personal benefits to peer supporters of participating and the benefits that a peer supporter with personal experience of relevant issues brings to the role [38‐42]. This study adds further information regarding the qualities, attributes and skills that a population at high risk of CVD consider to be beneficial in peer supporters. Key necessary characteristics identified by participants were empathetic, encouraging, having personal experience of MD, having knowledge of disease risk factors and the MD and having strong communication and listening skills. It was considered less important that the peer supporter would be a similar age or gender. This highlights the question of how best to recruit and also train peer supporters, to get a pool of people with these characteristics.

A particularly strong aspect of this study was the inclusion and exclusion criteria, as this helped to ensure that the population would possess similar characteristics to future intervention participants. Furthermore it is valuable that the focus group participants were of different genders, varying levels of affluence and from urban and rural areas as this ensured a range of opinions were considered. There were no clear differences between participants differing in the aforementioned characteristics.

Vignettes were used in this study to help participants ‘visualise’ the different peer support approaches and encourage discussion. This may be particularly important when participants have little prior experience of the concepts or scenarios being discussed. The use of vignettes in qualitative and quantitative research is reasonably common throughout a number of disciplines. The vignette scenarios must be considered authentic to be effective and aim to reduce limitations of participants personal experience or circumstances [43, 44]. It has been suggested that vignettes are limited in their effectiveness as the relationship between belief and action is not often straightforward. It was considered in this study that they would be beneficial as an ‘icebreaker’ to stimulate discussion, and their use in conjunction [45] with a structured interview would mediate any negative aspects.

The differing familiarity of different support methods could perhaps be seen as a weakness to the study. ‘Weightwatchers’ and ‘Slimming World’ were regularly discussed as reference points which may have introduced familiarity bias in discussions. It may be possible that a group method received the most favour as it was the most familiar and readily understood method of peer support, although these effects were potentially lessened by the use of vignettes.

It could be suggested that collecting data using focus groups, a face-to-face group setting, may have influenced the preference for face-to-face group peer support. Participants may have enjoyed taking part in the focus group discussion which could have influenced their decisions in the preference scoring sheet, administered at the end of the focus group. It could also be possible that those who accepted invitations to participate in a focus group would be more likely to prefer a group method of peer support or that recruitment of people from existing community groups inadvertently recruited those who particularly enjoy participating in face-to-face group formats. However the use of focus groups was key to access group beliefs about the different methods of peer support.

This research has been used to inform the development of TEAM-MED, a MD peer support intervention. It will also be valuable during the final evaluation of the intervention as it may be useful to explore whether the perceived benefits of group peer support were apparent or whether any particular types of motivation were identified as more or less useful by participants.

Further studies are recommended into effective peer supporters, the development and optimisation of an appropriate training programme for peer supporters and to explore the possibility and benefits of utilising existing groups to deliver a lifestyle behaviour change intervention, rather than ‘creating’ groups for this purpose. In line with the MRC framework for the development and evaluation of complex interventions [23], this work highlights the value of conducting exploratory research with the target group to inform intervention design, and provides important information on preferred peer support approaches in a Northern European population that will be useful to other researchers who plan to develop behaviour change interventions.