2.1 Study design
This was a randomized, single-blind (only the researcher knows the grouping of patients), parallel-group, controlled trial. The study included two aspects, the first was the SMS design, involving the SMS quality evaluation, conducted from October 2016 to April 2017. The second was the SMS intervention, which lasted for 12 months and conducted from May 2017 to April 2018.
2.2 Patient population
Patients with T2DM were all admitted to the Department of Endocrinology at the Affiliated Hospital of Inner Mongolia Medical University from October 2016 to March 2017 were selected for participation in the study.
Based on the results of a previous study [14], we expected an effect size of 0.26 in the plasma glucose level between baseline and at the end of intervention. The sample size was estimated with 80% power and a two-sided significance level of α = 0.05. We estimated a 10% dropout rate in the follow-up because of inconvenience or the time-consuming nature of the trial. We calculated a minimum sample size of 73 participants in each group.
The inclusion criteria for participants in the SMS quality evaluation were willingness and ability to participate in the SMS quality evaluation. Exclusion criteria were severe illness or inability to complete the SMS assessment.
For SMS intervention research participants, the inclusion criteria were as follows:
(1) Diagnosed with T2DM, and aged ≥ 18 years;
(2) Participants or a family member who lived together were literate and had a mobile phone;
(3) Clear consciousness and normal thinking; and
(4) Provided informed consent for the study.
The exclusion criteria for participants in the SMS intervention were as follows:
(1) Inability to complete the questionnaire because of serious illness; or
(2) Participants and their relatives had a lower level of education and could not read text messages.
The study protocol was approved by the Biomedical Research Ethics Committee of Inner Mongolia Medical University, and all selected participants signed informed consent forms. Over the course of the study, staff members protected the personal information of participants. Subjects were free to withdraw, and the researchers ensured that the participants’ rights were not affected.
2.3 Data Collection
Data were collected at baseline using the Diabetes Questionnaire, which was designed based on the Chinese Adult Chronic Diseases and Nutrition Monitoring Personal Questionnaire (2015). The content mainly included socio-demographic data, health-related behaviors, disease status, and dietary conditions, etc. And biochemical indicators, including fasting plasma glucose (FPG), postprandial glucose (PPG), glycosylated hemoglobin (HbA1c), Total cholesterol, triglyceride, high density lipoprotein-c (HDL-C), and low density lipoprotein-c (LDL-C), were assessed using discharge data.
2.4 Design and Methodology of SMS
2.4.1 SMS design
The messages were designed by a panel of experts, including endocrinology, chronic disease management, health education, disease prevention, etc. A total of 42 messages covering five main domains: health awareness, diet control, physical activities, living habits and weight control were designed, respectively. The content of messages was based on the report of the American Diabetes Association on Taking Care of Diabetes [15]. In terms of health awareness, we considered that participants need to have a full understanding of the etiology, process, and complications of diabetes, which is beneficial to the plasma glucose control. For diet control, this was considered to be the most direct factor affecting glycemic control [16,17]. The diet in this study included intake of vegetables, fruits, salt, protein, fat and food, and a low glycemic index, especially for cooking and eating habits. In terms of physical activity and weight control, as we all known that patients with T2DM can effectively reduce plasma glucose and lose weight through exercise. Therefore, we helped patients choose the best mode of exercise, and ensure the best time and frequency of exercise through SMS intervention. Smoking and drinking are almost all risk factors for chronic diseases including diabetes. Studies have shown that plasma glucose can be controlled by quitting smoking and limiting alcohol consumption [18]. Therefore, we intervened on unhealthy living habits. Each text message covered only one topic, and contained 70 words or fewer. The language expression was simple, direct, and easy to understand. Final list of SMS were shown in Appendix 1.
Message design was based on the Trans-Theoretical Model (TTM) [19], which is a widely used theory to promote health behavior. For diabetic patients, the pre-contemplation stage is almost non-existent, and combined with the SMS quality evaluation, we directly intervened from the contemplation stage. Therefore, in our study, the intervention was divided into four stages: the contemplation stage, the preparation stage, the action stage, and the reinforcement stage. The contemplation stage focused on education about basic knowledge about diabetes. This stage aimed to make participants aware of the importance of integrated diabetes management and to develop awareness of health management. The preparation and action stages focused on providing tips and suggestions for changing behaviors and on guiding participants to gradually develop the attitudes and actions necessary to change their behaviors. The reinforcement stage aimed to consolidate the changes in behaviors and transform healthy behaviors into habits. Examples of message in different stages were shown in Appendix 2.
2.4.2 SMS Quality Evaluation
The overall quality of short message was assessed using the SMS Quality Assessment Questionnaire, which included three aspects: the understanding of the text messages, the willingness to act on the text messages, and the current status of the response to the text messages. Specifically, five questions were asked for regarding each short message: Questions 1 and 2 reflected the participant’s understanding of the text message (yes=1, no=0), questions 3 and 4 reflected the participant’s willingness to act on the text message (yes=2, no=0), and question 5 reflected the status of the participant’s response to the text message (known and have done=1, unknown but have done=2, known but not have done=3, unknown and not have done=4). The setting of the score was based on the principle of scoring the scale and the patients’ demand for message, the higher the demand, the higher the score setting [20]. The total score of each message was calculated as the sum of the scores of the five questions, reflecting the overall quality of the message, with a maximum total of 10 points. We classified >7 as high quality, 4-7 as medium quality and <4 as low quality. The SMS Quality Assessment Questionnaire was shown in Appendix 3.
2.5 SMS Intervention
The intervention period was from May 2016 to May 2017. A total of 171 discharged participants with diabetes were randomly divided into two groups using random number. Participants in the intervention group (IG, 85) received evaluated messages, which were sent twice a week, only one message at a time. The content of the text messages during the same stage included five different domains. Combined with the SMS evaluation results, for some text message, more than 60% of participants reported that they did not take action, we added the transmission frequency and repeated every four weeks. Simultaneously, in order to reduce the rate of loss to follow-up, participants in the control group (CG, 86) were sent regular messages, mostly limited to general theoretical knowledge. Examples of regular message that CG received are shown in Appendix 4.
Telephone follow-up after each stage of the intervention was conducted. The purpose of the telephone follow-up was to ensure participants compliance and thus to ensure the quality of the study. The follow-up questionnaire was shown in Appendix 5.
2.6 Outcomes and Measurement
The main outcomes were the changes of plasma glucose and the control rate of each index. The Guidelines for the Prevention and Treatment of Type 2 Diabetes in China (2017 edition) [2], defines the FPG control rate as the proportion of all managed participants with FPG from 4.4 to 7.0 mmol/L and PPG control rate as participants with PPG < 10.0 mmol/L as a percentage of all managed participants. The secondary outcomes were changes in diet control (vegetables and fruits consumption), physical activities (more than 30 minutes per exercise, three to five times per week), living habits (smoking and drinking) and weight control. Each outcome measure was obtained by telephone follow-up.
2.7 Statistical Analysis
Data were recorded using EpiData 3.1 software. Data analysis was performed using IBM SPSS Statistics, Version 19.0 (IBM Corp, Armonk, NY, USA) software. Continuous variables were expressed as means (standard deviation, SD), and t-test was used for comparisons between the two groups. Categorical variables were expressed as percentages, and c2 test was used for comparisons between the two groups. P ≤ 0.05 was considered statistically significant.
To compare the health behavior of the two groups over time, we converted the behavior change including vegetable and fruit intake, weight control, and physical activities into score at baseline and 12 months. Each instance of behavior change from healthy to unhealthy status was assigned a score of -1, change from unhealthy to healthy status was assigned a score of 1, and no change was assigned a score of 0. The scores for each of the four behaviors were then summed, generating a composite change score that ranged from -4 (when all four behaviors changed from healthy to unhealthy) to 4 (when all four behaviors changed from unhealthy to healthy after the intervention). We then compared the difference in total scores between the two groups.