DiAlert: a prevention program for overweight first degree relatives of type 2 diabetes patients: results of a pilot study to test feasibility and acceptability

We aimed to recruit sufficient participants (approximately 20) for two groups to be able to test the intervention twice. Approximately 250 flyers and information leaflets were posted in the hospital building and outpatient clinic of the VU University Medical Center in Amsterdam plus an advertisement run in a local newspaper in October 2010. In addition, announcements were posted on the project website (http://​www.​dialert.​nl) and the website of VU university medical center (http://​www.​vumc.​nl). To trigger attention, all recruitment materials included the sentence ‘Does diabetes run in your family?’ along with information about the pilot study and inclusion criteria. Participants were allowed to bring relatives who met inclusion criteria to participate in the pilot study. Participants received a small incentive and reimbursement of travel expenses.

DiAlert is offered in two sessions of 150 minutes, and delivered over a period of two weeks by a trained health educator (WH), henceforth referred to as the trainer. The trainer was instructed according to a standardized training program for the PRISMA (Dutch DESMOND) program[9].

Participants are encouraged to explore possibilities and resources for prevention in a positive atmosphere and using a constructive didactic approach. In view of the fact that participants are simply overweight and symptom free, with a family history of T2DM, we assume an interest in the program but perhaps not a strong readiness for change, as one might expect in those who are medically ill. Therefore, DiAlert puts emphasis on promoting risk awareness and intrinsic motivation for changing lifestyle, while avoiding inducing psychological distress. The empowerment philosophy supports the educational process to develop awareness and autonomy to effectively assume responsibility for their decisions in relation to lifestyle behavior. In line with this philosophy, the trainer is guiding rather than teaching. Following the format of PRISMA, a group size of 8 to 10 participants was considered to be ideal with ample opportunity for participants to interact with the trainer and the other group members.

The development of DiAlert was guided by the HAPA framework (Figure1). HAPA builds on social cognitive theory, distinguishing two stages of behavior change: motivation, and action and maintenance. The basic assumption underlying HAPA is that motivation is a necessary condition for behavior change, but that goal-setting and action planning are required for the change to actually occur. Throughout the process of behavior change, feelings of self-efficacy play a key role.

In the first session, adequate risk perception is promoted by providing information and discussion of modifiable and non-modifiable risk factors for T2DM. To increase positive outcome expectancies for weight loss and physical activity, the trainer introduces the topic of combined insulin resistance and loss of beta cell function as the underlying pathophysiology of diabetes, using simple analogies and schemes. The benefits of weight loss and physical activity are discussed in that context. In relation to weight loss, energy balance (intake and expenditure) is discussed. At the end of the first session, participants are asked to record nutrition and physical activity behavior in a diary on two separate days during the week in between the two group sessions. Self-monitoring of their health behaviors will help participants set personalized and realistic goals.

In the second session, self-efficacy beliefs, strategies to lose weight, coping strategies, goal setting and action planning are addressed.

Patients are provided with a written curriculum highlighting the key points and including the diary and action planning sheets. During group sessions, worries, discussions and clarifications put forward by the participants and trainer are written down on large flip-over sheets, visible for all participants, that stay there for the whole program. During the modules about healthy food choices, wrappings and containers of commonly used products are displayed and discussed. A more detailed description of the intervention development and outline is given elsewhere[7].

Participants were asked to fill out a questionnaire twice, one week before the first group session and four weeks later (that is, one week after the last group session). See Tables 1 and2 for an overview of the measures and times of data collection. In addition, the flip-over sheets produced during the sessions gave information about main interests, questions and worries of the participants.

All determinants were assessed with questionnaires at baseline and 4-weeks follow-up. Performed analyses were Wilcoxon signed rank tests, and t-test (^a) in case of normal distributions. ^bCronbach’s alpha of sum scores ≥0.8, ^cCronbach’s alpha of sum scores of 3 items 0.67; ^dCronbach’s alpha of sum scores >0.9, ^e P <0.05. HAPA: Health Action Process Approach.

Characteristics of the study population were assessed by self-report, including sociodemographics, family history of diabetes in first and second degree relatives, body weight and lifestyle behavior.

To assess whether the intervention was delivered consistently with the underlying theory and philosophy and to what extent the intervention was delivered as planned, the sessions were observed and findings recorded on a checklist. The checklist was created based on the objectives of the program and trainers instructions. Two independent observers attended the group sessions and were instructed to check whether all modules were delivered and all objectives for participants were covered (see Additional file1); to report on the engagement of participants by looking at interactions between trainer and participants and among participants; and to observe whether the sessions were delivered in a constructive, empowering atmosphere (that is, the trainer listens and is respectful and empathetic to all participants).

The worries, questions and reasons for participation discussed in the first session provided insight into the extent to which the goals of the program matched those of the participants. In addition, we measured worry about diabetes risk, feelings of personal control and perceived consequences by means of questionnaires at baseline and follow-up[13]. These outcomes could also show whether DiAlert had any adverse effects on these perceptions.

Information on feasibility is essential before embarking on a RCT in a larger sample. For the aim of this pilot, we explored recruitment strategies, by asking participants how they knew about the intervention and why they were participating in the intervention. Length of the modules was timed with a stopwatch by the observers and recorded on the checklist and we examined whether all information could be delivered in two sessions of 150 minutes. Group size was informed by PRISMA and evaluated to confirm acceptability and feasibility (evaluation form: too small, just right, too large), observations (were all participants involved in the intervention?) and experiences of the trainer (was it feasible to deliver the intervention as intended with this number of participants?). Length of the sessions was assessed in the follow-up questionnaire.

Participants’ views and experiences with the DiAlert program were assessed using a short evaluation form at the end of the second group session and with the questionnaire at follow-up.

The evaluation form asked participants to give an overall grade between 1 and 10 for the whole intervention and to rate usefulness of each module of the intervention and the homework assignment on a five-point Likert scale (see Table 1).

At follow-up, we assessed whether participants would recommend the DiAlert program to others. Participants were asked to respond on two statements about the manual: ‘In my opinion the information in the manual of DiAlert is… clear/instructive’ and to evaluate the action plan ‘I managed to create personal goals’ and ‘I think it is useful to create a personal action plan’.

In line with the HAPA framework (Figure1) determinants of behavioral change were made operational by questions at baseline and follow-up, derived from existing measures. Risk perception for diabetes was assessed on three different domains: causal beliefs, using a validated questionnaire[14]; comparative risk; and risk estimation using questionnaires adopted from former studies in the field of family risk information[13]. To assess the other determinants we adopted questionnaires from Schwarzer et al.: perceived self-efficacy for healthy eating and physical activity was assessed by 10 questions[16]. Outcome expectancies for a healthy diet and increasing physical activity were measured with eight questions adopted from Schwarzer et al.[15]. Intentions and action planning to change health behaviors were examined, asking participants whether they plan to consciously eat healthier, exercise more or lose weight and if they have formulated a detailed action plan (what, when, how) for changing diet and physical activity[15]. See Table 2 for exact wording of the questions.

In total, 22 people signed up for the DiAlert pilot study and participated in two different groups, 10 and 12 participants respectively. Twenty participants had a first degree relative with T2DM, one participant had no first degree relatives but did have a number of second degree relatives with diabetes and was allowed to participate. Of the participants with a first degree relative with T2DM, two had a sibling with diabetes; all others reported parental family history of T2DM. One participant appeared not to have any relative with T2DM, and was therefore removed from the analyses, leaving 21 participants for baseline analyses. Characteristics of the participants are described in Table 3. The majority of participants was female (86%) with a mean age of 47.9 ±9.7 years. Most participants were from Dutch descent (N = 15), others reported Surinamese, Moroccan, Hindustani, Indian, Polish and Chinese ethnicities. All participants spoke Dutch fluently. Mean self-reported body mass index was 29 ±6.3 kg/m². More than half of the participants (N = 13) had attempted to lose weight in the past five years with a mean number of 5.9 ± 4.5 attempts.

One participant dropped out after the first session, due to family circumstances.

All topics of the intervention were covered in the two sessions as planned, in both groups, and all materials developed were used. All participants received the participant manual and took it home. In general, engagement of the participants was high in both groups - all participants were actively involved in both sessions. Participants were particularly engaged in the module discussing calories of displayed food products. Observers confirmed that the intervention was delivered in an empowering atmosphere.

As shown in Table 1, examination of the flip-over sheets showed that family risk information was an important topic of discussion and participants expressed concerns about their own risk of developing diabetes. Also, concerns were expressed about relatives, in most cases the parents who were having problems controlling their diabetes. In addition, the risk of diabetes in own offspring was raised in one group. Two main themes emerged from the listed ‘burning questions’ at the beginning of the sessions. First, questions about causes of diabetes and its relation to lifestyle, and secondly, both groups raised questions about T2DM treatment and complications, see Table 1 for examples of quotes from participants.

No significant changes were found for the items on personal control and worry about personal risk, indicating that the intervention had no effects on these determinants. Perceived consequences of T2DM slightly increased at follow-up and participants more often disagreed with the statement: ‘I think I have little influence on getting type 2 diabetes’.

Recruitment resulted in a sufficient number of participants for two groups within a relatively short period of three weeks. Twenty-five people showed interest in participating in the pilot study and contacted the research team by email or telephone. Three people decided not to participate after receiving more detailed information. Both younger and older individuals showed interest in the DiAlert intervention, therefore we decided to include participants from 25 to 65 years old. Although the DiAlert intervention was targeted at overweight relatives, four participants had a normal body mass index (<25 kg/m²), and were included because the main aim of this pilot was to evaluate the process and feasibility of the group sessions. Most participants signed up after reading about DiAlert in flyers and the newspaper advertisement (63.7%). Main reasons for participating in the DiAlert pilot study were: ‘prevention of T2DM’ (N = 9) and ‘learning about the personal risk of diabetes due to a family history’ (N = 5). In addition, the three motives for joining the program that were mentioned at the beginning of the first session were risk awareness and worry, information seeking and motivation to change behavior.

The feasibility of the group sessions was confirmed in terms of duration of the modules and group size. All modules were delivered within the scheduled 150 minutes. Some modules exceeded the planned time with a maximum of 10 minutes, while other modules took less time.

The group size was evaluated as ‘just right’ by participants in both groups. Most participants (90%) stated that the duration of the interventions was good.

At the end of the sessions, observers and trainer confirmed that delivery of DiAlert is feasible in line with the empowerment philosophy and theoretical framework. In both groups, all participants were able to formulate goals and create a personal action plan to lose weight. The homework assignment was completed by almost all participants (N = 20) in between the two sessions.

Following the underlining empowerment philosophy, participants seemed to be encouraged to play an active role in the intervention, and came up with examples and questions to be answered during the sessions. High engagement of participants was noticed in both groups, especially sections with activities, where all participants were involved and came to the table to discuss calories and food choices together. After one week of follow-up, participants gave an overall mark of 8.0 ±1.0 on a scale from 1 (lowest) to 10 (highest). All participants would recommend the DiAlert program to others. Overall, evaluation of the usefulness of the intervention modules showed a mean score of 1.5 ±0.4 (scale, 1 = very useful to 5 = totally not useful). The module ‘development of diabetes’ got the highest rating 1.3 ±0.6. In this module the development of diabetes is discussed with participants using a drawing to explain insulin resistance, loss of beta cell function and the positive effects of body weight loss and physical activity. Evaluation of the information in the manual and its clarity was good, 3.4 ±0.5 (scale, 1 = totally disagree to 4 = totally agree).

Because the main objective of DiAlert is weight loss, post intervention analyses of the determinants of behavior change were performed for the overweight participants only (N = 16). Analyses of baseline questionnaire scores showed that participants were aware of the main risks for developing T2DM at baseline, with mean scores of >4 (scale, 1 = definitely not to 5 = definitely) on the items nutrition, heredity, aging and lack of physical activity (Table 2). Not surprisingly, we found a relatively high baseline score for heredity (4.3 ±0.7). Participants perceived their risk (comparative risk and risk estimation) somewhat higher than average. Sum scores for outcome expectancies for a healthy diet and for physical activity were 16.0 ±2.5 (scale 1 to 20) and 12.2 ±1.4 (scale 1 to 15), respectively. This suggests that participants had quite positive outcome expectancies already for eating healthy foods and doing exercise.

With regard to action planning, significant positive changes were seen for both diet (P = 0.008) and physical activity (P <0.001). This means that the participants were more able to formulate concrete plans to change their dietary habits and physical activity pattern, including addressing coping plans to anticipate difficult situations in the future.

Some issues relating to the conduct and management of the future RCT have been raised by this pilot study. First, interest in the topic was confirmed based on the finding that recruitment efforts proved effective to reach a sufficient number of eligible participants. However, mainly women were reached; therefore, in the RCT, we should take into account possible strategies to include both men and women and from a broad range of socioeconomic classes. In the RCT, we will apply a mixed recruitment strategy, involving general practitioners and diabetes specialists together with use of media and brochures to recruit participants with a positive family history of T2DM directly. As result of a direct recruitment approach (through general practitioners) we may expect participants with lower perception of risk, less positive outcome expectancies and lower self-efficacy for lifestyle changes. Another issue in relation to recruitment was that, although the DiAlert study was initially aimed at relatives 29 to 55 years of age, younger and older people showed interest and were enrolled. Therefore, the inclusion criteria for the upcoming RCT will be changed to 25 to 65 years of age to certify validity of the intervention. We plan to deliver a culturally-sensitive version of the DiAlert intervention to relatives of Turkish origin living in the Netherlands. The intervention will be pretested in this target group before we conduct the RCT in this group.

Second, discussion of risk information did not increase worries about personal risk. However, we should keep in mind that participants did express concerns and worries about their own family members with T2DM. Some participants clearly were in need of information on diabetes and its management with regards to their relatives rather themselves. In the RCT, the focus of DiAlert should stay on prevention and risk of developing diabetes due to family history, rather than discussing problems that may occur in their relatives with T2DM in the future.

From this pilot, we learned that no changes to the intervention modules are anticipated before embedding DiAlert in a RCT.