Comparison of an online versus conventional multidisciplinary collaborative weight loss programme in type 2 diabetes mellitus: A randomized controlled trial

Abstract Aim The aim of this study was to examine the effect of an online multidisciplinary weight loss management programme. Methods Between July 2016 and July 2017 this randomized controlled trial recruited patients in Nanjing, China who were living with type 2 diabetes mellitus and who were obese or overweight and randomized them to online versus conventional groups. All participants were managed by a multidisciplinary team. The experimental group was managed using the Why Wait WeChat Platform for Weight Reduction Management. Results There were 55 and 52 participants in the online and conventional groups, respectively. The decreases in fasting blood glucose (−4.26 vs. −2.99 mmol/L), 2‐h postprandial blood glucose (−4.48 vs. −2.68 mmol/L) and glycated haemoglobin (−22.11 vs. −6.21 mmol/mol) were more pronounced in the online compared to conventional group (all P < 0.05). After the intervention, self‐management ability parameters, including diet control, foot care and total score, were improved in the online group compared with the conventional group, as well as all indexes of quality of life (all P < 0.05). Conclusion The online multidisciplinary weight loss management programme improved blood glucose in obese or overweight patients living with type 2 diabetes mellitus. Self‐management ability parameters (including diet control, foot care and total score) and quality of life were improved in the online group compared with the conventional group.


Summary statement
What is already known about this topic?
• Multidisciplinary approaches have been introduced into weight loss management models for patients with type 2 diabetes mellitus and obese or overweight in outpatient clinics of third-level hospitals, but their effects have not been clearly described.
What this paper adds?
• The online multidisciplinary weight loss management programme improved blood glucose in obese or overweight patients living with type 2 diabetes mellitus.
• Self-management ability parameters (including diet control, foot care and total score) and quality of life were improved in the online group compared with the conventional group.
The implications of this paper: • This is the first trial of an online programme in the management of type 2 diabetes mellitus in China.
• Trial findings might have international applicability in the management of type 2 diabetes mellitus and could lead to improvement of prognosis.

| INTRODUCTION
With the development of the world economy and dietary changes, diabetes and obesity have become public health issues (Anders & Schroeter, 2015;Sami et al., 2017). Type 2 diabetes mellitus (T2DM) is related to overweight and obesity and results from insulin resistance (Al-Goblan et al., 2014;Wu et al., 2014). In China, 30-52% of patients with T2DM are overweight (Lu et al., 2021;Wang et al., 2015). Overweight/obesity has multiple risks for patients living with T2DM (Leitner et al., 2017;Piche et al., 2020).
The DiRECT trial showed that the study programme could sustain T2DM remission in about one-third of the patients, but sustained remission was associated with sustained weight loss (Lean et al., 2019).
Overweight or obese T2DM patients often fail to receive reasonable and standardized treatment and may even have therapeutic weight gain. Therefore, current guidelines emphasize that attention should be paid to the weight of patients living with T2DM (Belalcazar et al., 2010;Look et al., 2013;Look & Wing, 2010). How to appropriately reduce weight is the focus of current weight-loss management programmes. Multidisciplinary approaches (Zolotarjova et al., 2018) have been introduced into weight loss management models for patients with T2DM and obese or overweight in outpatient clinics of third-level hospitals. However, its effects have not been clearly described.
We hypothesized that an online multidisciplinary weight loss management programme could improve patients' blood glucose control, maintenance after weight loss, and improve patients' selfmanagement and quality of life (QOL). Therefore, the current study aimed to examine the effect of an online multidisciplinary weight loss management programme based on the Why Wait? programme (Grant et al., 2006) in patients living with T2DM and obesity or overweight.
The primary aim was to examine the impact of the intervention on the HbA 1c levels. Then, the study also aimed to examine the effect of the intervention on the other parameters of glucose metabolism, anthropometry and blood lipids. Our findings would help design optimal algorithms for weight loss management in the future. This is the first trial of an online programme in the management of T2DM in China.
Such a trial might have international applicability because the intervention described here would be easily exportable.

| Participants
This parallel randomized controlled trial was conducted from July 2016 to July 2017 at the Affiliated Changzhou No. 2 People's Hospital of Nanjing Medical University. The convenience sampling method was used to recruit overweight and obese T2DM patients. The patients were recruited mainly via the internet, radio and television stations and outpatient visits. The patients contacted the project leader, and the investigators evaluated their condition of the patients to determine whether they could be included in this study based on the inclusion and exclusion criteria.

| Ethical considerations
This study was registered at http://www.chictr.org.cn/showproj.aspx? proj=44368 (No. ChiCTR2000029321) and was approved by the Ethics Committee (Approval No. 2016-C-16-01). All the patients enrolled signed informed consent forms. There were no important changes in the methods after trial commencement.

| Sample size calculation
Sample size was calculated according to the formula for sample size calculation (Li, 2013), comparing the means of two samples according

| Randomization
A third-party statistician used SPSS 18.0 (SPSS Inc., USA) to generate a random number table, which was used by the specialist nurses to assign the participants to the online (multidisciplinary collaborative management) and conventional (conventional management) groups.
Each group was further divided into subgroups of 10 participants, according to the order of enrolment.
Although the participants were told they were divided into two groups when signing the informed consent form, the investigators did not tell the grouping of the participants, and the participants were not informed of the exact content of the management plan. They were simply told what they needed to know. This blinding method was approved by the ethics committee. All investigators and personnel involved in data collection and evaluation were blind to grouping. All personnel involved in data entry, monitoring and analysis were blind to grouping until data lock.  included an overview of diabetes and obesity, diet therapy, exercise therapy, drug therapy, psychological adjustment and so on and occurred every 2 months for 30 min (six times in total). Each class was followed by a group discussion for 30 min.

| Follow-up
Diabetes specialist nurses entered health records, and telephone calls were made at the first, second, fifth, seventh, eighth, ninth, tenth, and eleventh months after admission. Outpatient follow-up was conducted in the outpatient clinic by the weight-loss management team during the third, sixth and twelfth months after admission.

| Endpoints and data collection
The primary endpoint was glycosylated haemoglobin (HbA 1c ) levels 1 year after intervention initiation.
The basic data of the two groups were collected before the intervention and at 3, 6 and 12 months after the intervention, including general demographic data, daily lifestyle, FPG, 2PPG, HbA 1c , BMI, waist circumference, hip circumference, SBP, DBP, total cholesterol (TC), triglycerides (TGs), high-density lipoprotein (HDL) and lowdensity lipoprotein (LDL). The body mass of the participants was measured using a smart Bluetooth body fat scale, and BMI was calculated.
FBG and 2HBG were detected during an oral glucose tolerance test (OGTT). Venous blood was taken to measure the biochemical indexes.
A mercury sphygmomanometer was used to measure blood pressure.
The waist-hip ratio, body fat percentage, target weight and health were assessed by the Inbody 720 human body composition analyzer before the intervention and at 6 months and 1 year after intervention initiation. There were no changes to the endpoints after trial commencement.
Two questionnaires were completed at baseline, 3, 6 and 12 months. One was the Diabetes Self-Management Behaviour Scale, and the other was the Diabetes Quality of Life Scale. Two investigators (specialized diabetes nurses and deputy chief nurses) used unified instructions to introduce the purpose and methods to fill out the survey to participants. After obtaining informed consent, the participants filled in the information by themselves. If the participants could not fill in the information by themselves, the investigators filled in the information on behalf of the participants according to their answers.
The Diabetes Self-Management Behaviour Scale revised by Toobert et al. (2000) and translated into Chinese by Wan et al. (2008), with good homogeneity, was used in this study. In terms of scoring, the Likert-like scoring method was adopted (Li et al., 2011;Zhu, 2014). The first two items (exercise and foot care) were assigned average numbers of days (scores); the following two items, including diet (measuring special dietary conditions, e.g., vegetables and highfat foods) and smoking status, were scored separately. Each subscale had a maximum of 7 points, for a total score of 28 points. A total score >23 points (single item >5.6 points) was considered to be 'good'; 17-23 (single item, 4.2-5.6) and <17 (single item < 4.2) points were rated as 'general' and 'poor', respectively (Wan et al., 2008).
The higher the score, the stronger the self-management ability. The Cronbach's α of self-behaviour management scale was 0.84, including 11 items of the four subscales of dietary control (four items), exercise compliance (two items), monitoring compliance (two items) and foot care (two items); the Cronbach's α values range from 0.71 to 0.93, all greater than 0.70, with good homogeneity.
The Diabetes Quality of Life Scale was developed by Liao et al. (2000) and included physiological, psychological (spiritual), social relations and treatment dimensions, with a total of 27 items. The Likert scoring system was adopted, with scores of 1 to 5 per item. The higher the score, the worse the QOL (Brown et al., 2004). The Cronbach's α of the Diabetes QOL Scale is 0.76, and the split half reliability coefficient is 0.87.

| Statistical analysis
All data were double entered by two specialist nurses. A third-party statistician conducted all analyses. Per-protocol (PP) analyses were performed by excluding the participants who dropped out of the study or failed to comply with the study protocol. The ITT analyses were performed using all participants according to their original grouping, regardless of whether they complied with the intervention protocol or not (e.g., missing at follow-up assessments and dropouts).
For the participants missing follow-up assessments or dropped out, the Department of Cardiology. Therefore, 120 participants were enrolled in this study, including 60 each in the online and conventional groups. The participant flowchart is shown in Figure 1. The demographic and other baseline data such as sex, age, disease course and education level were comparable between the two groups, as shown in Table S1.

| Body composition indexes after intervention
The waist-hip ratio, body fat percentage, target body weight and health assessment in both groups met the normality and spherical symmetry (P > 0.05) criteria. The results indicated significant differences in target body weight and health assessment between the two groups at various times (P < 0.05) ( Table 3).

| Self-management ability after intervention
All dimensions of the Diabetes Self-Management Behaviour Scale (self-management diet control, exercise adherence, foot care and monitoring adherence) as well as total scores in both groups met the normality and spherical symmetry (spherical symmetry test P > 0.05) criteria. Both groups were tested by repeated-measures analysis of variance at each time. The results showed that diet control, foot care and total scores were significantly higher in the online group compared with controls (P < 0.05) ( Table 4). Before the intervention, all dimensions and total scores of Diabetes QOL Scale in both groups met the normality and spherical symmetry criteria (spherical symmetry test P > 0.05). Repeated measures ANOVA showed that all four QOL dimensions and total scores significantly decreased over the 12 months (all P ≤ 0.001).
The intervention group had lower scores for all four dimensions and total scores than the conventional group at 3, 6 and 12 months (all P < 0.01). The time Â grouping interaction was significant for the physiological dimension (P < 0.001), psychological dimension (P < 0.001), therapeutic dimension (P = 0.009) and total score (P < 0.001) but not for the social relation dimension (P = 0.051) (Table 5).

| DISCUSSION
The results showed that FPG, 2-h PPG and HbA 1c were significantly improved in the online group compared with conventional group, although both groups showed some improvements over time; these results are supported by a study on the integration of hospital community management (Zhou & Lu, 2011). Waist circumference, hip circumference, blood pressure, blood lipid and other indicators were not improved in this year-long study. Although waist and hip circumferences provide important data on the shape of the patient, they might not reflect the changes in body weight completely. Although weight loss is known to lower blood pressure and blood lipids (Kawamoto et al., 2014;Winnicki et al., 2006), the impact of T2DM and inflammation on these parameters could be stronger than the impact of weight loss, and there might be some lag or delay in their improvement after weight loss. It will have to be examined in future studies with a longer follow-up.   et al., 2008), whereas the other study showed a smaller weight regain (2.5 kg) with face-to-face intervention than with online (4.7 kg) or control (4.9 kg) interventions (difference of 2.4 kg, 95% CI: 0.002-10.8) (Wing et al., 2006). A meta-analysis showed that there is no significant difference in the weight loss effect between the network and face-to-face intervention models, while the network intervention model is not as good as face-to-face or group communication for weight maintenance (Turk et al., 2009;Wing et al., 2006). Indeed, the network intervention models often play a less pronounced role in the initial maintenance after weight loss (Svetkey et al., 2008). This is not consistent with the current results. The BMI values in the online group at 3-12 months were significantly lower than the conventional group values (P < 0.05). In addition to network intervention, the MDT team played a very important role. The intervention team included individuals with the professional backgrounds necessary to deliver the required multidisciplinary collaborative model, and personalized weight loss programmes were designed for patients, increasing compliance. From the perspective of whole-person medicine, the MDT provides patients with perfect and full-course professional team care, achieving shared decision-making between doctors and nurses and improving their self-management ability, as shown above (Table 4).
This helps increase the weight loss effect and maintain weight after loss. However, as suggested by a recent meta-analysis, further evidence is needed to demonstrate the effectiveness of web-based interventions for weight loss management (Neve et al., 2010).
The research population was mainly young and middle-aged people, which could be explained by the online element, with which an older population might be less comfortable with. Still, it is also important as there is a shift in the age of diabetes onset towards a younger age and as a younger age at diagnosis will lead to a higher risk of poor outcomes due to the morbidity from the chronic complications of diabetes. Therefore, developing management strategies that specifically target younger patients is relevant.

| Study limitations
The limitations of this study should be mentioned. First, it was a single-centre study with a relatively short follow-up. Secondly, cardiovascular benefits during the follow-up period were not comprehensively assessed. Thirdly, the research population was mainly young and middle-aged patients living with T2DM, and elderly patients were not well-represented. Fourthly, information security should be enforced to maintain data integrity and promote the patient's use of the platform. The intervention should be tested in more than one site before it would be generalizable to the wider population. A multicentre study with a follow-up of 48 months is currently being planned.
Finally, there were some differences between the results of the PP and ITT analyses. The sample size of this study was not large enough, and seven patients were missing or dropouts before the first followup (3 months after intervention), including two in the online group (one moved and one dropped out due to disease) and five in the conventional group (three were lost to follow-up and two withdrew their consent). For these patients, no data could be collected besides those at baseline, so only the data of those 113 patients were enrolled in ITT analysis. Considering the small sample size, this difference in the number of patients may cause an important bias in the results, and the true effect of the online intervention cannot be reflected. Still, the two analyses reported similar outcomes for glucose control.

| CONCLUSION
The online multidisciplinary weight loss management programme improves blood glucose in obese or overweight patients living with T2DM. Self-management ability parameters (including diet control, foot care and total score) and QOL were improved in the online group compared with the conventional group. Given the substantial resource required by this programme, one of the implications of the findings is the importance of a future health economic analysis. Such a trial might have international applicability in the management of T2DM and improvement of prognosis. Zhu, W. (2014). Reliability and validity of Chinese version of diabetes selfmanagement behavior scale. Chinese Journal of Nursing, PLA, 16, 5-8. Zolotarjova, J., Ten Velde, G., & Vreugdenhil, A. C. E. (2018). Effects of multidisciplinary interventions on weight loss and health outcomes in children and adolescents with morbid obesity. Obesity Reviews: An Official Journal of the International Association for the Study of Obesity, 19 (7), 931-946. https://doi.org/10.1111/obr.12680

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.