Regulatory Approval
This study obtained ethical approval from the Office for Research Ethics Committees Northern Ireland (ORECNI) (16/NI/0224) and was registered with the (International Standard Randomised Controlled Trials Number) ISRCTN registry (ISRCTN66118345).
Design and development of the intervention (Phase 1)
A systematic approach was followed consisting of the analysis of dietary intake to design and develop intervention materials followed by qualitative research with the target population.
Analysis of the target population’s dietary intake
An analysis of the dietary intake of older adults (65+ years) was conducted using the National Data and Nutrition Survey (NDNS) years 1-to 6 combined, a publically available cross-sectional survey undertaken of a representative sample of people living in the UK (23,24). Survey results indicated that the majority of older adults were not meeting the 5-a-day recommendation for fruit and vegetables, the 18g/d of dietary fibre (non-starch polysaccharides) and the one portion (140g) of oily fish per week recommendation.
Intervention content
The analysis of older adult’s dietary behaviours identified three healthy eating domains for intervention content: fruit and vegetables, wholegrains and healthy proteins (e.g switching red or processed meat to leaner protein sources such as fish, chicken, beans or lentils etc.).This guided the design of the habits-based intervention materials which was based on a similar structure to a previous randomised controlled trial on healthy feeding habits for parents (known as the Healthy Feeding Habits study) (25). The intervention was delivered using an intervention booklet, where a researcher would talk through the content with the participant to establish three self-chosen habit-formation goals (based on the three targeted healthy eating domains). It was designed that each study visit focused upon one of these three domains separately where each time participants would identify a desired goal around the chosen domain to target (e.g. `to eat a portion of fruit with breakfast everyday’). In order to facilitate this, the researcher and participant would record potential barriers to accomplishing the habit goal, identify any preparations required and assign behaviour change start dates. Information on recommended daily servings, health benefits, food sources and portion guides around each of the three healthy eating domains were also included along with tips and strategies to achieving habit goals and self-monitoring sheets for goal tracking purposes (26–28). At subsequent visits, it was structured that healthy habit goals were reviewed as older adults continue in previous goals and set a new healthy eating goal in another domain.
Qualitative research with the target population
Following the design of the intervention, qualitative research with the target population using focus groups was used to investigate; the barriers and facilitators to healthy eating; the impact of factors that influence food choice; and the strategies to support older adults eating a healthy diet. Focus group discussions were also conducted for feedback on the proposed dietary intervention.
Phase 1 Focus group recruitment
Community dwelling, independent adults over the age of 65 years who were able to give informed consent were invited to participate in the focus group discussions. Participants were recruited by poster advertisement and also through snowball sampling from community, retirement and church groups across Northern Ireland. When required a member of the research team conducting the focus group interviews also attended groups and provided a short presentation about the qualitative study (aim of the study, what it involved and what it hoped to achieve). Participants were asked to take part in a focus group discussion lasting approximately 1 - 1.5 hours at the Centre for Public Health, Queen’s University Belfast or the researcher (LAM or SW) visited them if they were an existing community group.
Phase 1 data collection
Participants were required to complete a general demographics questionnaire before the focus group session. The focus groups were conducted in accordance with a standardised protocol consisting of semi-structured open-ended questions in order to ensure a consistent approach between groups. Examples of questions in the topic guide included: What are the main factors that influence your food choice?; What do you think is a healthy diet?; and What do you think prevents you eating a healthy diet? Participants were also given copies of the dietary intervention booklet as researchers explained the overall delivery process, followed by designated time administered to read through the booklet themselves for feedback. Groups were conducted until saturation of ideas and opinions were reached. Focus group discussions were tape recorded and transcribed verbatim, anonymising participants. Audio recordings of focus groups were destroyed as soon as verbatim transcripts were prepared.
Feasibility of the intervention (Phase 2)
Phase 2 consisted of a non-randomised (single-arm) feasibility study to test the dietary intervention developed during Phase 1. This allowed for appropriate adjustments and refinement of the dietary intervention to be carried out before a definitive RCT.
Phase 2 participants and recruitment
A non-randomised study was conducted at the Centre for Dentistry at Queens University Belfast between July 2017 to September 2017 on patients who has recently received dental treatment for their partial tooth loss. Patients had either been provided with removable partial dentures or restored to a functional dentition according to the principles of the shortened dental arch. All patients were now part of a 6-monthly recall programme to review their natural tooth replacement treatment and manage chronic dental disease. Dental notes of patients were screened using a screening questionnaire for eligibility. To be eligible patients had to be free-living, older adults (65+ years) who had a minimum of six natural teeth in at least one jaw with missing teeth replaced with removable partial dentures or restored to a functional dentition using fixed prosthodontics within the last 6 months at the Centre for Dentistry. Patients were excluded based on the following criteria that might have impacted their ability to fully engage or participate in the dietary intervention; clinically diagnosed dementia; Diabetes mellitus; history of alcoholism; an active treatment for psychiatric disorders; medical complication which contraindicate routine dental treatment; or were following a strict diet regime recommended by a physician in the prevention or treatment of disease.
A sample size of 8-10 participants was considered a feasible recruitment target and was estimated to be sufficient to meet the aims of this single-arm feasibility study by the research team and dental clinicians. Patients were informed about the study and invited to participate either by letter or at their dental review appointment. If interest was expressed, further exclusion criteria was then assessed by one of two researchers (LAM or LM). Patients were excluded if they did not have a sufficient level of English literacy to read study materials and to keep a food diary; were not able to recite their understanding of the study back to the researcher; felt that they could not take any responsibility for diet changes discussed during the course of the study, for example, they should have been able to influence the type of foods bought and eaten in their household; did not feel that making changes to their diet was important to them; did not feel they were ready to make changes to their diet. Patients gave informed consent at their baseline study visit.
Phase 2 data collection
Eligible participants who received the tailored habit-based dietary intervention met with one of two researchers four times at fortnightly intervals (at the Centre for Dentistry at the Royal Victoria Hospital). Researchers (LAM and LM) were trained in habits and behaviour change methods prior to intervention delivery. The researcher delivered the dietary intervention using a standardised protocol (roughly 30 minute sessions) to discuss a new healthy habit to incorporate into the participant’s diet each time as identified from Phase 1 development (fruits and vegetables, wholegrains and healthy proteins). Participants were asked to attend assessments at baseline and at the end of the intervention (8 weeks later) as laid out in Figure 1. However, during the course of the intervention it became apparent that the 8 week visit was not necessary and could be easily combined with the 6 week visit. Therefore, to reduce participant burden, it was decided that length of follow-up would be changed from 8 weeks to 6 weeks, thus reducing the number of study visits by combining the last two.
Study visits lasted approximately 2 hours (baseline and 6 week visits) to allow sufficient time for the researcher to carry out the study visits in accordance with the intervention protocol. Although this paper does not report on these measures directly, the study visits consisted of obtaining informed consent (baseline only), habit automaticity questionnaire and the completion of various lifestyle and dietary behaviour change questionnaires, anthropometry measures (height, weight, BMI), discussion of food diaries (four-day), sufficient breaks if required and a qualitative interview (end of intervention only). In between fortnightly study visits lasted approximately 30 minutes as they were only to deliver the intervention.
Outcome measures
The first line feasibility outcomes were the usability and acceptability of the intervention to study participants with regards to intervention delivery, data collection procedures, study materials and executing the new healthy habit. This was assessed through feedback from the participants using a study-specific evaluation questionnaire after intervention delivery (6 week time point) to gather information on the participant’s opinions, thoughts and experiences. This was also assessed through one-to-one interviews using semi-structured open-ended questions from an interview topic guide (6 week time point). Interviews were conducted by the same researcher who delivered the habits-based intervention to the participant. The interview covered topics such as personalised dietary changes, intervention delivery, engagement of study materials, preferred intervention delivery settings and improvements for a definitive RCT. Interviews were tape-recorded and transcribed in order to inform on intervention refinement (29).
Supporting feasibility outcomes consisted of further dimensions of the study that would inform the structure of a RCT. This specifically included the feasibility of screening procedures; recruitment strategies; and retention/ attrition rates. Another supporting feasibility outcome was the participant’s compliance to the intervention. Daily compliance to each of their new healthy habits was assessed at each study visit when the researcher asked the participant how many times they completed their healthy habit out of the last 14 days. Compliance to the use of study materials was measured using tracking sheets which were sent home with participants and collected at subsequent study visits. Tracking sheets monitored new healthy habits by getting the participant to tick whether they did or didn’t do the new behaviour each day.
Phase 1 and 2 data analysis
Questionnaire data were summarised as appropriate using SPSS Statistics for Windows Version 22 (SPSS Inc, Chicago, IL). Descriptive analyses included means, standard deviation, median, frequencies and/or percentages where appropriate. After focus groups and interviews were tape-recorded and transcribed, a thematic analysis, as outlined by Braun & Clarke was conducted to code and identify key themes (30). Due to the exploratory nature of this study, a single suggestion analysis over a general theoretical framework was provided. This process involved the repeated reading of the transcripts followed by generating a list of key codes (by LAM) which led to the development of a coding scheme. Codes were then grouped into categories leading to key themes being constructed. In order to minimise the potential for bias of the coding process with respect to the themes identified, a second researcher (either SW or LM) from the team also performed the procedure outlined above independently. Anonymised transcripts were discussed in team meetings in a process of triangulation to decide on emerging key themes. Researcher notes were also taken throughout the course of the intervention in order to inform for an RCT.