The effect of the smartphone app DiaCert on health related quality of life in patients with type 2 diabetes: results from a randomized controlled trial

The study design of the DiaCert-study has been described in detail previously [21]. Briefly, the DiaCert-study is a randomized controlled trial of patients with type 2 diabetes. A total of 181 participants were recruited between February 2017 and June 2019 from five primary care centers and one specialized medical center around Stockholm, Sweden. We ended the data collection after almost 2.5 years, since the technical infrastructure of the app could no longer be maintained. Participants were recruited by oral invitation from their treating physician or diabetes nurse during a healthcare visit at one of the participating health care centers. Thereafter, study personnel gave them more information about the study and scheduled them for a baseline meeting. The participants were randomized 1:1 to the use of the smartphone app DiaCert for a 3 month physical activity intervention in addition to routine care, or to a control group with routine care only, i.e., all participants continued their usual care as prescribed by their regular primary care physician and diabetes nurse in accordance with national guidelines. We randomized in blocks of ten by gender and within each health care center, using a random allocation list generated in STATA 14.0. Due to the nature of the intervention, participants were not blinded to their allocation. Inclusion criteria were: a diagnosis of type 2 diabetes, age above 18 years, ability to read and understand Swedish, being able to walk, and having access to and being able to use a smartphone. All participants provided written consent prior to participating in the study. The study was approved by the Regional Ethical Review Board, Stockholm, Sweden.

During the baseline meeting, participant responded to a questionnaire assessing background variables including educational level (≤ 12 years/ > 12 years) and smoking status (never/former/current). At baseline and follow-up after 3 months and 6 months assessment of the primary outcome physical activity (moderate to vigorous physical activity) and secondary outcomes including for example glycated hemoglobin (HbA1c), serum lipids, body weight, blood pressure, and HRQoL were performed. Participants’ weight, height, and waist circumference were measured by study personnel, physical activity was measured using accelerometers and blood samples were collected for measurement of HbA1c and blood lipids. All participants, both the intervention group and the control group, were offered to use the smartphone app DiaCert at the 6 month follow-up.

The smartphone app “DiaCert” promoted daily steps by registering and displaying the number of steps taken per day, measured by the smartphone. An individual step-goal between 1000 and 10,000 steps per day, based on the participants activity level prior to the study, was set together with the study personnel at baseline. The user could see if the goal was reached each day with the number of steps displayed, as well as a visually displayed circle going from empty (white) to filled (blue). Further, a bar chart with one bar for each day that the app had been used was displayed, where the bar became green if the goal was reached, and red if not. The user also received a positive feedback message in the app when the daily goal was achieved. Every second week, study personnel contacted the participant who was given the opportunity to revise the step goal by an even 500 steps.

HbA1c was measured at baseline, and after the 3 month long intervention and again at the 6 months follow up, i.e., it was not measured during the intervention. All participants, both in the control group and intervention group, received information about their HbA1c levels at baseline and follow-ups from study personal. Participants in the intervention group could also see their baseline HbA1c in the app. The app was compatible with both Android (version 4.1 or higher) and iOS (version 9.2 and higher).

The RAND-36-Item Health survey version 1.0 (distributed by RAND Corporation), referred to as the RAND-36 throughout the rest of the text, is a widely used, self-reporting questionnaire that measures HRQoL [22, 23]. The questionnaire was first published as SF-36 in 1992 [24], and as RAND-36 in 1993 [22]. It is a generic measurement, i.e., it is not population-specific or disease-specific, and it has previously been used in studies including patients with type 2 diabetes [8, 25‐27]. The RAND-36 and the SF-36 includes the same items. However, the scoring differs between them regarding the health concepts bodily pain and general health. The difference is often considered negligible as Hays et al. showed that the correlations between the scales were 0.99 in the MOS-study [22].

RAND-36 includes 36 questions (items) about physical and mental health, whereof 35 cover eight health concepts: 1. physical functioning (10 items), 2. role limitations due to physical health problems (4 items), 3. role limitations due to emotional problems (3 items), 4. social functioning (2 items), 5. emotional well-being (5 items), 6. energy/fatigue (4 items), 7. bodily pain (2 items), and 8. general health perceptions (5 items). One additional item measure change in perceived health status today compared to one year ago.

Each item in RAND-36 is scored on a range from 0 to 100 where the score represents the percentage of a total possible score of 100. For example, item 1 has five possible answers, where an answer gives a score of 0, 25, 50, 75, or 100, while item 13 has two possible answers, giving a score of 0 or 100, respectively. Items in the same health concept are averaged together to create the different health concepts scores. Since not all of the health concepts contain the same amount of items, the scores in each health concept are transformed to a 100-point scale ranging from 0 to 100, where higher scores indicate better HRQoL [23, 28].

Baseline characteristics are presented as mean (SD) and n (%) for continuous and categorical variables, respectively. To assess whether there were any statistically significant differences between the intervention group and the control group at baseline, the Mann–Whitney test were performed for continuous and Chi-square test was performed for categorical variables. Nonparametric tests were used, since they do not assume normally distributed data.

Only participants with complete baseline data on HRQoL were included in the analysis, fifteen participants were excluded due to missing RAND-36 data at baseline. Within group analysis comparing the HRQoL scores of RAND-36 within the intervention and control groups, respectively, at follow up after 3 months and 6 months were performed using the Wilcoxon signed-rank test. In these analyses, only participants with complete RAND-36 data at baseline and at each follow-up were included. The percentage of missing RAND-36 data at 3 months and 6 months were 11% (18/166) and 22% (37/166), respectively. A flow chart of participants with complete RAND-36 data is presented in Fig. 1.

For in between group analyses, to account for within subject correlations, statistical models for longitudinal data based on generalized estimation equations (GEE) were used to assess both the temporal trend and the intervention effect, and their eventual interaction. The difference in mean between the groups at follow-ups, as well as the intervention effect was analysed. The estimation of intervention effect was expressed in terms of change in HRQoL from baseline between the intervention and control group. We further conducted a sensitivity analysis by only including participants with complete RAND-36 data at baseline and at follow-up at 3 months and 6 months. Adjustments for multiple comparisons were not made [29]. Analyses were performed using Stata 14.2 (Stata Corporation, College Station, TX, USA). The significance level was set to α = 0.05.