Comparison of barriers and facilitators of MIND diet uptake among adults from Northern Ireland and Italy

The theoretical framework underpinning this research is the COM-B model [37]. Changing behaviour involves changing one or more of the components of the COM-B model, which stand for, capability, opportunity, motivation, and behaviour (see Fig. 1). Capability can be either psychological (knowledge, psychological skills, or stamina) to perform the behaviour, or “physical” (having the physical skills, strength or stamina) to perform the behaviour. Opportunity can be divided into “physical” (what the environment allows in terms of time, resources etc) or “social” (interpersonal influences, social cues, cultural norms). Motivation can be divided into “reflective” (self-conscious planning and evaluations, beliefs about what is good or bad) or “automated” (wants and needs, desires, impulse and reflex responses) [37]. The Theoretical Domains Framework (TDF) facilitates understanding of health behaviours around evidence-based guidelines and provides a method to categorise behaviour [38]. Use of the TDF to identify factors influencing MIND diet behaviour can then be mapped onto the COM-B model for designing interventions. The TDF has 14 domains that may influence behaviour change [38] (see Fig. 1).

The COM-B model is at the core of an overarching framework called the Behaviour Change Wheel [37] which is a three-stage approach to designing a behaviour change intervention. This framework includes 9 intervention functions (education, persuasion, incentivisation, coercion, training, restriction, environmental restructuring, modelling, and enablement linked to the COM-B model. These are how an intervention might change behaviour and are linked to behaviour change techniques [37]. BCTs are considered the active component of the intervention designed to change behaviour, such as self-monitoring of behaviour and goal setting. The COM-B model and TDF have been used by several studies to explore barriers and facilitators to behaviour change in sexual health [39], physical activity in obese pregnant women [40] and reducing sugar [41].

Previous research found differences in dietary patterns of people who live in Rome and NI, with NI consuming more ready-made meals [42], snacked between meals more often than Italians [43] and consumed more takeaway food, sugary drinks and less fruits and vegetables that those living in other Mediterranean countries [44]. Although the MIND diet is a hybrid of the Mediterranean and DASH diets, it is a new diet that specifies foods such as berries, leafy greens, and wholegrains, which are not part of a traditional Mediterranean diet. Furthermore, the Mediterranean and DASH diets are not specific to the literature on nutrition and the brain. Therefore, the MIND diet was designed that incorporated many of the basic components of the Mediterranean and DASH diet, but with modifications that reflect the best scientific evidence on nutrition and prevention of dementia [17]. Comparing factors from the COM-B model (capability, opportunity and motivation) that may influence MIND diet behaviour across a Mediterranean and non-Mediterranean country, can reveal valuable insights that highlight diverse habits and beliefs across culture, which may be particularly informative in the development of behaviour change interventions.

The aim of the study was to establish and compare components of the COM-B model that influence the uptake of the MIND diet in a 40–55-year old Italian and Northern Ireland (NI) sample, that will inform a dietary behaviour change intervention.

Specific objectives were:

Part of the methods in this manuscript can also be seen in Timlin et al. [45]. Twenty-five participants from NI (Belfast, Armagh city, County Tyrone) (female 60%, male 40%) [45] and twenty-five participants from Rome, Italy (female 64%, male 36%) aged 40–55 years were recruited onto the study, to take part in either a focus group or an interview. Interview/focus groups took place in person (NI: 15 interviews, 2 focus group n = 6, n = 4; Italy 13 interviews, 2 focus group n = 6 n = 6). Ethical approval was obtained from the School of Psychology Staff and Postgraduate Filter Committee at Ulster University, which is in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). Participants were approached by email, Facebook and advertisement booklet, which included some brief information about the study. Interested participants were asked to contact the researcher by email and sent a participant information sheet (PIS), consent form and information booklet on the MIND diet. Questions asked to participants were the same for both NI and Italian populations. Before the Italian interviews began, questions were translated from English to Italian by a fluent Italian speaker (BG). Questions were then back translated to English to ensure the interpretation of questions [46]. Most of the interviews were spoken in English (18) and those that were not were translated during the interviews by one of the Italian researchers (BG), to allow the English-speaking researcher (DT) to transcribe and analyse data from all of the transcripts. All interviews/focus groups were recorded and transcribed verbatim. The oral recordings and transcripts were sent to the Italian speaking researcher to check for missing data. Interviews were conducted in a private room at either a research centre or a community facility such as a library, convenient to the participant. In accordance with the COM-B framework, both focus groups and interviews were conducted [37] using semi-structured questions and lasting between 30 and 60 min each (see Table 1). The interview and focus group questions were based on guidance using the COM-B [37] model and TDF [38] (Table 1). The interview schedule was developed using the COM-B model, and informed the content analyses, as seen in previous research conducted with the NI population of this research, as seen in previous research conducted with the NI population of this research [45]. All participants were asked to complete a personal information form and consent form before the interview/focus group began. The information form contained questions on participants diet at baseline (see Table 2) and showed that those living in NI consumed more red meat, fried food, butter, and sugary foods than those living in Italy. Participants were informed that the study was voluntary and that they could withdraw at any time. They were assured of confidentiality regarding any personal information they supplied to the researcher. It has been suggested by similar theoretical models, that 25 participants is the ideal sample size for qualitative research [47]. Also, similar to other qualitative studies using the COM-B and TDF [39, 40], twenty-five NI and 25 Italian participants were recruited onto the study, to take part in either a focus group or an interview.

A topic guide was developed using the TDF. An example of a question related to TDF knowledge was, “what is your understanding of the MIND diet”. A further question exploring participants skills was, “to what extent are you confident in cooking MIND diet foods”. The TDF represents an elaboration of the COM-B’s six components into 14 domains, covering a wide spectrum of behavioural determinants (see Table 1). A booklet containing information on the elements of the MIND diet, and the origins of the diet were given to participants. An in-depth discussion on the MIND diet components was discussed prior to interview and focus groups. All interviews and focus groups were audio recorded.

The data analyses has been described in full in Timlin et al. [45]. Two researchers (one English speaking and one Italian/English speaking) (DT&BG) independently read through the entire dataset and coded the data from each transcript and assigned initial “code names”. There was a 95% agreement on codes between the two main researcher, which demonstrates an acceptable level of agreement [48]. However, any differences in coding were resolved with discussion between the researchers. Summative content analysis [49] was applied as an additional step in the analysis following agreement of codes. This involved two researchers searching the text for occurrences of codes, that were counted to identify the frequency of each code. Using a common approach [50, 51], TDF domains were judged based on the frequency count of coding for each TDF domain, which had been aggregated from all the factors, beliefs or phrases mentioned that fell within that domain. For example, some participants reported that they believed the MIND diet would make them feel better generally. This belief statement is coded under the TDF domain “belief about consequences.” The frequency coding identified which TDF domains were most commonly reported, establishing the main barriers and facilitators to the uptake of the MIND diet.

Psychological capability was a COM-B component identified as a barrier to adherence to the MIND diet. Twenty percent of the barriers in the Italian sample fell into the psychological component of the COM-B model compared to 29% in the NI sample. These barriers fell into 2 of the TDF domains, behaviour regulation and knowledge. None of the Italian barriers fell into attention and decision process domain, unlike the NI sample, where 10% of barriers fell into this domain.

According to the COM-B model, for behaviour to occur, there must be a physical and social opportunity in the environment. Barriers relating to physical opportunity were the most commonly reported barriers in both the NI and Italian populations, with 29% of all utterances falling into this component in the NI sample and 33% in the Italian sample. The TDF domain related to this component is; environmental context and resources. Social opportunity was reported as being a key barrier and facilitator in both the NI and Italian samples, with 13% of all facilitators and 5% of barriers falling into this component from the NI sample, 15% of all barriers and 12% of facilitators from the Italian sample. The TDF domain related to this component is social influence.

Motivation is a component of the COM-B model and there must be strong motivation for the behaviour to occur [33]. Participants reported reflective motivation to be a barrier to the uptake of the MIND diet and 18% of barriers fell into this component of the COM-B model, compared to 15% in the NI study. More facilitators were reported under this domain with 33% from the NI sample and 37% from the Rome sample.

To our knowledge, this is the first study to develop a “behavioural diagnosis” of factors influencing the uptake of the MIND diet in a Mediterranean and non-Mediterranean country. This was the first study to apply the TDF to explore people’s attitudes towards a whole dietary pattern and compare these attitudes between a Mediterranean and non-Mediterranean country. The COM-B model provides a more comprehensive explanation of adherence than existing models [37], making it easier to identify appropriate interventions. The COM-B model was used as an additional step in the data analysis, increasing the efficiency of the study and showing the framework to be adequate for its purpose.

This study was undertaken in a small sample of Italian and Northern Irish men and women. Our findings in terms of barriers and facilitators reported are “perceived” and context based. Therefore, not only may the findings have limited value in predicting MIND diet behaviour, but also not be generalisable to the whole populations. However, generalisability was not the main aim of our study, rather to explore people’s attitudes and perceptions towards the uptake and adherence to the MIND diet, with the aim to inform an intervention. Another limitation of the study may be researcher subjectivity; however, two researchers identified the codes from the data, suggesting that the themes drawn have credence beyond the lead researcher’s interpretation. Focus groups run the risk of introducing bias [89], resulting from an individual’s desire to conform to social acceptability [90]. However, the focus group participants in this study were acquaintances, reducing the risk of social desirability. A limitation of this study is that the two samples differ in terms of socio-economic status, with all the participants from the Italian sample being of high socio-economic status and approximately one-third of the NI participants of low socio-economic status, which may make comparisons more difficult. Further research should include participants across different socioeconomic backgrounds. Furthermore, half of the Italian participants spoke in Italian and some of the richness of the data may have been lost in translation. However, the second researcher (Italian) translated, transcribed, and analysed the data to maximise interpretation and understanding of the data.