This is a single-blind, two-arm randomized controlled trial with randomization at the participant level. Participants will be randomly allocated to receive a HMI (control) or a SISMT (intervention). Primary and secondary study outcomes will be collected for both groups at baseline (T0) and 12 weeks(T1), 24 weeks post-intervention(T2). As the study involves the older adults with MCIs, a few specific ethical procedures should be followed and ensured. Only those who signed the consent form can be included in this study, and a guardian as the legal representative should be asked to co-sign the consent form. The trial has been registered at ChiCTR.org.cn (ChiCTR2200061991, date:16/07/2022).
Study aim, setting and ethics
To evaluate the effectiveness of a SISMT for DM-MCI will be conducted, with allocation concealment and outcome assessor blinding (Fig. 2).
Recruitment and screening
By convenient sampling, patients with DM will be screened for MCI and included in the study. This study is approved by relevant hospital departments and most patients were recruited from the endocrine clinic at a provincial Third-Grade Class-A General Hospital. The study researchers will met with DM patients who were visiting the clinic and asked them to complete the Montreal Cognitive Assessment (MoCA) , the Mini-Mental State Examination (MMSE) , the diabetes self-management scale (DSCS) , and the Activity of Daily Living (ADL) survey , along with a general information questionnaire. The questionnaires will be completed in separate rooms of the hospital. So far, 240 DM-MCI patients have been screened.
MCI will be diagnosed using Petersen diagnostic criteria as follows : 1) Chief complaint: memory decline; 2) MoCA score of 13–14 for illiteracy, 19–20 for patients with primary school education(1–6 years education), 24–25 for those with junior high school education(≥ 7 years education) ; 3) An MMSE score of 24–30 indicates the absence of dementia . (4) ADL will be used to measure the intact activities of daily living (Lawton–Brody ADL score < 18) .
1) Meet the diagnostic criteria for diabetes (The diagnosis of diabetes is based on the doctor’s diagnosis results in the hospital diagnosis and treatment system). 2) Meet the diagnostic criteria for mild cognitive impairment. 3) ≥ 60 years of age. 4) No obvious visual or hearing impairments. 5) Participants have good communication skills and can cooperate with researchers to complete the survey questionnaire. 6) Able to use Internet device dependently.
1) Serious physical diseases that hinder completion of the cognitive function screening. Such as severe Parkinson, paralysis and so on. 2) Drug or alcohol dependence. 3) Other nervous system diseases and/or serious medical conditions that can alter brain function. 4) Inability to use smartphones independently.
Sample size calculation
The required sample size is estimated using PASS v11.0 (NCSS, Kaysville, UT, USA) based on a completely random design for comparing the means of two independent samples. To our knowledge, there is no previous study on a stratified support pattern-based internet-assisted self-management therapy for DM-MCI in mainland China, so we estimate the effect size of one of the outcomes (HbA1c) based on previous research on patients with DM-MCI . This study find that HbA1c scores in intervention and control group are 6.49 ± 1.59 and 6.97 ± 0.73, respectively. A sample size of 34 participants per group was determined to be sufficient to detect an effect with a type 1 error rate of 5% (α = 0.05) and 90% power (β = 0.1). A total of 68 participants will be needed, with 34 participants per group. Considering a 20% loss of follow-up rate, the final sample size is 86, with 43 participants per group.
Randomization, blinding, and concealed allocation
Ensuring allocation concealment, participants will be randomized (after obtaining written informed consent, eligibility screening, and baseline assessment) to the intervention and control groups at a 1:1 ratio by a study staff member (who will not be involved in participant recruitment or outcome assessment) using Research Randomizer software (http://www.randomizer.org/). Participants in the intervention group will be assigned to different categories of the intervention group based on their cognitive function and self-management characteristics and will receive Intervention measures with different frequencies. Participants in the control group will be assigned to different categories of the control group based on their cognitive function and self-management characteristics, too, but there is no difference in intervention frequency. The participants will then be told their group assignment by the intervention staff. Due to the nature of non-pharmacological interventions, only the outcome assessors and data analysts (not the participants or intervention staff) will be blinded to group allocation.
In this study, the intervention group will adopt a SISMT. After the completion of the baseline survey, participants will formulate personalized intervention measures and optimize and expand their knowledge through hierarchical strategies to improve their self-management behavior. Based on our previous research in this study, the patients will be divided into four categories according to cognitive function and self-management ability: (1) high cognitive high management, (2) high cognitive low management, (3) low cognitive high management, and (4) low cognitive low management. Four WeChat groups will be formed to intervene separately to prevent communications between each category of participants in the intervention group. Self-management interventions targeted to patients with DM-MCI will be implemented, including diet, exercise, drug intake, psychology, and cognition. To achieve stratified intervention with different intensity, the content will remain the same for each group, but the frequency will differ. The intervention plan is shown in Table 1. All intervention content will be implemented through home self-monitoring and WeChat feedback. When patients experience suboptimal blood sugar monitoring, researchers will guide them in regulating self-management behavior. If behavioral regulation is ineffective, contact the attending physician for medication guidance. The outcome indicators will be collected at three time points: baseline, 12 and 24 weeks, respectively.
WeChat is the intervention medium for this study. The researchers will establish WeChat groups and created WeChat official account before the intervention. The researcher will send the completed content of the WeChat official account to the participants on the WeChat social platform. Participants can use the chat box function of the WeChat social platform to provide feedback and exchange on content of the WeChat official account, and can also interact with other participants in the WeChat group to share their self-management experience and personal insights on health knowledge. And they are required to use this application to inform researchers of their fasting blood glucose level, postprandial blood glucose level, hypoglycemia, diet and exercise status (Fig. 3).
The control group will adopt the HMI. After the outpatient service, health education will be given to patients. We will provide patients with disease-related health manuals to help with the self-management of patients. The content of the health manual revolves around the epidemiological status, diagnostic methods, clinical manifestations, influencing factors, and intervention methods of cognitive impairment. Distribute a health manual and explain it to each participant in the control group. Outcome indicators will be collected at three time points: baseline, 12 and 24 weeks, respectively.
Criteria for discontinuing or modifying
During the intervention period, if the participant is unwilling to continue due to personal reasons, the intervention will be terminated. Participants who are unable to continue but are willing to undergo follow-up will be modified for intervention measures and replaced with a control group.
Adherence of intervention
Firstly, WeChat is adopted as the intervention medium and a multifunctional social media application in our study. Researchers can request participants to provide feedback and clock in on the intervention content to test the adherence of the intervention implementation. Secondly, Wechat official account has background data (the number of times and number of readers read), which can monitor the learning of participants. When the data does not match the number of participants, it will remind them in the Wechat group. Thirdly, establishing good cooperative relationships with participants fundamentally ensures adherence.
During the intervention process, participants should continue to take medication related to the disease. And inform patients before the intervention begins that they cannot participate in other intervention activities during the participation process. However, if participants participate in other interventions targeting outcome indicators such as blood glucose control and cognitive function during the participation process, their test data will be excluded.
Outcome measures timeline
To compare the outcomes of the self-management interventions provided under the SISMT and HMI, a baseline (W0) survey will be administered to both the experimental and control groups before the start of each pattern. Outcome indicators will be collected 12 weeks (T1) and 24 weeks (T2) after the intervention to compare the effects of the two patterns (Table 2). The indicators will be assessed by experienced staff members who will be blinded to the group allocation. Outcome indicators and measurement tools are as follows (Domains /characteristics of tools see Table 3):
The secondary outcome measures will include several commonly used measures of specific domains of cognitive function, psychological indicators, blood glucose control (FBS, HbA1c, PBS) and other relevant indicators. The measures of specific domains of cognitive function (memory, language, attentions) will comprise the Auditory Verbal Learning Test (AVLT) , Symbol Digital Modalities Test (SDMT) . Psychological status will serve as the secondary index of this study. This will include anxiety, depression, well-being, quality of life, self-efficacy, and self-esteem. The Self-Rating Anxiety Scale (SAS) will be used to measure anxiety  and the Geriatric Depression Scale (GDS) will measure depression . Well-being will be assessed using the Memorial University of Newfoundland Scale of Happiness (MUNSH) . Quality of Life-Alzheimer’s Disease (QoL-AD) will be used to assess patient quality of life , the General Self-efficacy Scale (GSES) will be used to measure self-efficacy , and the Self-Esteem Scale (SES) will be used to measure self-esteem . Other relevant indicators include the healthy literacy and self-management behavior which can be used to investigate the patient’s acceptance of the intervention content. Health literacy and self-management behavior will be determined using the Health Literacy Scale for type 2 diabetes and the diabetes self-care scale (DSCS).
FBS, HbA1c, and PBS will be measured and collected from participants’ self-reported data. Other data that need to be measured using a scale will be invited participants to the hospital again for measurement and collection.
All data analysis will be performed using SPSS Statistics v21.0 (IBM). To ensure that they are blinded to group assignment, the outcome assessors and data analysts will not interact with the participants except during data collection (in the case of outcome assessors). Demographic and other baseline characteristics will be summarized using descriptive statistics. If necessary, the results will be adjusted for potential confounders, such as age, gender, and education. Quantitative data will be expressed as frequencies or rates and statistical analysis will be carried out using the χ2 test. If these data are normally distributed, they will be expressed as (x ± s). The effects of the interventions (between-group differences) will be calculated with linear mixed models using interaction terms (group allocation versus time), which are equivalent to the between-group differences. The within-group differences will be calculated using repeated measures ANOVA. P-values < 0.05 will be considered statistically significant. The data will be reviewed before analysis. The Statistical Package for the Social Sciences (SPSS) software will be used for all analyses. The clinical relevance of the results will be confirmed by calculating the effect size (Cohen d) of the significant differences found between the assessments. The following effects will be considered: small: 0.00–0.49; medium: 0.50–0.79; high ≥0.80 (Cohen, 1988). Comparison between the four subgroups within the intervention group using ANOVA test.