A Transtheoretical Model-Based Online Intervention to Improve Medication Adherence for Chinese Adults Newly Diagnosed With Type 2 Diabetes: A Mixed-Method Study

Abstract

Background:

Type 2 diabetes mellitus (T2DM) is increasing in China, with medication non-adherence being a significant contributor to uncontrolled T2DM. The Transtheoretical Model (TTM) has shown effectiveness in chronic disease management, but few studies have applied it in online interventions for T2DM medication adherence.

Aim:

The study aimed to develop and investigate the effects of a TTM-based online health education program on promoting positive stage of change (SOC) movement, improving self-efficacy and medication adherence, as well as reducing HbA1c levels in newly diagnosed patients with T2DM.

Methods:

This sequential mixed-method study was conducted from April 2023 to March 2024. Using the TTM framework, the study initially explored 32 participants’ experiences with hypoglycemic medications, health information acquisition, and perspectives on online programs. Then, a quasi-experimental study design was conducted. Two communities were randomly assigned as the intervention (n = 91) and comparison (n = 98) groups, with 189 newly diagnosed middle-aged T2DM patients from various SOC. The intervention group received short videos health education and participated in WeChat group discussions, compared with usual care in the comparison group. Data were collected at baseline, 3-month, and 6-month follow-ups.

Results:

The intervention group was more likely to achieve positive SOC movement (P < .001, Adj OR = 13.69 95% CI = 6.76-27.71) compared to the comparison group. The intervention group also had significantly higher mean CDMSS-11 and MMAS-8 scores at 6 months (P = .03 and <.001, respectively) and more likely to achieve clinically significant glycated Hemoglobin (HbA1c) change at 3 months (P < .001, Adj OR = 3.91, 95% CI = 1.77-8.63) and at 6 months (P < .001, Adj OR = 5.62, 95% CI = 2.70-11.69) compared to the comparison group.

Conclusion:

These findings support that applying the TTM to develop an online program could promote behavior change, improve self-efficacy and medication adherence, and could lead to better glycemic control in newly diagnosed T2DM patients.

Introduction

T2DM has an increasing incidence globally, with China having the highest number of diabetic patients worldwide. ¹ Diabetes can lead to severe complications impose substantial economic burdens. ¹ Glycemic control is crucial, particularly during the first year after diagnosis, as optimal glycemic control during this critical period can reduce the risk of future complications. ² Although lifestyle modifications are essential, oral antidiabetic medications remain the primary means of achieving glycemic control. ³ In China, the prevention and treatment of diabetes have shifted from large general hospitals to community health service centers (CHSCs). ⁴ However, only 23.5% of T2DM patients in Chinese communities exhibit good medication adherence. ⁵

As a prominent theory of behavior change, the TTM has been extensively employed as the theoretical framework for research on health behaviors including medication adherence. ⁶ TTM focuses on the specific SOC an individual is in during the process of adopting a healthy behavior, which involves pre-contemplation (PC), contemplation (C), preparation (PR), action (A), and maintenance (M). ⁷ By assessing the patient’s SOC, TTM guides the development of stage-matched, tailored interventions to promote gradual transition toward improved adherence levels.

In previous studies, TTM has been widely applied to improve medication adherence among patients with chronic diseases. The main intervention measures include telephone, printed materials, and face-to-face interventions. ⁶ However, while TTM has shown certain effects in enhancing medication adherence among other chronic disease patients, studies applying TTM to enhance medication adherence in T2DM patients are relatively scarce. One study from China employed traditional face-to-face interventions based on TTM, including interviews, community lectures, and home visits, which improved medication adherence and glycemic control in T2DM patients. ⁸

In China, CHSCs are crucial in providing primary healthcare services to community residents, including chronic disease management and health education. ⁹ Nevertheless, due to limited community healthcare resources and staff shortage, each community nurse needs to serve approximately 4000 residents on average, leading to a lower level of patient management in communities than in tertiary hospitals.^4,10 Community nurses are the primary providers of diabetes management and health education. However, with the large population served by the community and the limited healthcare resources, traditional on-site intervention methods are difficult to implement on a large scale in the community setting. To better meet the health management needs of T2DM patients in the community, it is imperative to explore a health education approach that can fully utilize existing resources while being widely accessible.

With the advancement of information technology, healthcare providers should optimally utilize the necessary tools and conditions. The rise of short videos has provided a new form of health information dissemination. Delivering information through audiovisual media in an easily understandable format has become a common practice and has been proven to enhance patients’ health knowledge.^11,12 However, research on TTM-based online short video interventions to improve medication adherence in T2DM patients is still limited. Therefore, a 2-phase study design was used to develop and test a TTM-based online short video health education intervention. The first objective was to develop an online short video health education intervention based on the perspectives and suggestions of interviewees at different SOC regarding oral hypoglycemic medication and online health education content. The second objective was to evaluate whether this online program can promote positive SOC movement, improve self-efficacy and medication adherence, and reduce glycated Hemoglobin (HbA1c) levels in newly diagnosed T2DM patients.

Methods

This study was carried out between April 2023 and March 2024, employing a 2-phase, exploratory sequential mixed-method study design. Phase 1 focused on program development, utilizing qualitative methods to explore patients’ perspectives toward the online program. Phase 2 involved the assessment of the program.

Phase 1: Program Development

This phase comprised program development based on findings from semi-structured interviews.

In phase 1, the researcher conducted semi-structured interviews with participants in the PC, C, PR, and A stages at CHSC, participants’ homes, or their local communities based on their willingness. The interviews primarily focused on their experiences with hypoglycemic medications, health information acquisition, and perspectives on online health education programs. With the consent of the participants, the interviews were audio recorded and transcribed verbatim. The main interview questions were: (1) What concerns do you have while taking hypoglycemic medications? Or what areas do you need support and help with? (2) What channels do you use to obtain health knowledge about diabetes? Do you think the health education provided by these channels has some shortcomings in helping you manage diabetes? (3) If health education is provided through short videos, what aspects of information would you like to learn? Do you have any suggestions, such as the duration, filming format, content, etc.? Each participant was interviewed for an average of 35 min. After conducting the interviews, 2 researchers analyzed and coded the interview content using the thematic analysis proposed by Braun and Clarke. ¹³ Key findings and perspectives were identified to inform the development of the online health education program.

Phase 2: Program Implementation and Evaluation

This phase constituted a 2-armed, quasi-experimental study with a 3-month intervention and a 3-month follow-up from September 2023 to March 2024, with all participants starting the intervention simultaneously. The study was conducted at 2 CHSCs purposively selected within the same district of Sichuan Province, which provided standardized basic management services for chronic diseases. After obtaining consent from the 2 CHSCs, they were then randomly assigned to either the intervention or the comparison group. The intervention design was based on preliminary qualitative findings and the empirically supported TTM. The familiar WeChat platform was used to enhance the intervention’s acceptability and feasibility within the community setting.

Participants and recruitment

Inclusion criteria were: aged 40 to under 65 years; diagnosed with T2DM with HbA1c 7.0% to 9.0%; disease course within 12 months; taking oral hypoglycemic medication; in PC, C, PR, or A; literate and able to use WeChat; agreed to participate and provide informed consent. Exclusion criteria were: M stage, severe comorbidities, mental disorder, and participating in other diabetes education programs. Dropout criteria were failure to engage in the intervention and loss to follow-up.

The sample size was calculated using G*Power 3.1 analysis software, a widely utilized program for conducting statistical power analyses and determining appropriate sample sizes.

Based on prior research,^8,14,15 with power set at 0.68 and considering the potential influence of various factors on participants’ engagement in this online intervention study, a total sample of 172 participants was determined to be adequate to provide 90% power to detect significance with a medium effect size (d = 0.50) for a 2-tailed t-test at the .05 α level. Considering a potential dropout rate of 25% according to previous WeChat-based online studies,^16,17 a total sample size of at least 216 participants (108 per group) was recruited for this study.

Study participants were screened and recruited between April and June 2023. The initial screening and recruitment process was conducted by the staff at the CHSCs, who have authorized access to the chronic disease management systems. They queried information based on the inclusion criteria, including disease duration, medical history, medication adherence status, etc. Potential participants who met the criteria were assigned a unique identifier and randomly sorted using the random number generation method in SPSS. Starting from the top of the sorted list, the researcher contacted potential participants by telephone.

The researcher explained the study’s purpose, procedure, associated risks and benefits, and the need for permission to access participants’ electronic medical records for research purposes. Verbal informed consent was then obtained from the participants. After that, the researcher further confirmed the participants’ SOC for medication adherence, literacy, and ability to use WeChat. Those who self-reported being in the M stage or participating in other education programs were excluded. The researcher invited eligible participants to undergo HbA1c testing at the CHSC. If participants did not meet the HbA1c 7.0% to 9.0% criteria, the researcher repeated the second step until both groups’ predetermined sample size requirements were satisfied.

Intervention

The qualitative data was used to inform intervention content and mechanism, including format, duration, and scheduling (Supplemental Material 1). The intervention strategies included sending short videos health education and conducting WeChat group discussions for all participants. Based on participants’ needs, videos were scheduled to be sent every Wednesday and Saturday from 7 to 8 pm, following a logical sequence presented in Supplemental Material 2, without specifying a viewing order for participants. WeChat discussions were held on the last Sunday of each month from 7 to 8 pm. We established 4 WeChat groups corresponding to the 4 stages (PC group, C group, PR group, and A group). Before the intervention, participants were provided with instructions to help them participate optimally in the program. The monthly intervention strategies for each stage-specific WeChat group remained constant, while the patients entering or leaving the WeChat groups varied each month.

Short videos were developed based on findings from phase 1 and the change processes of each SOC in the TTM, ⁷ and referred to the “Popular Science Version of the Chinese Guidelines for the Prevention and Treatment of T2DM (2022 Edition).” ¹⁸ Researchers confirmed the accuracy and scientific validity of the video scripts. The researcher, as the on-camera health educator, produced 37 short videos with an average duration of 139 s. The videos included 2 categories: 28 stage-targeted videos (5 for PC, 8 for C, 8 for PR, and 7 for A) and 9 general videos (3 each for diet, exercise, and monitoring). The general videos were created to address the comprehensive nature of diabetes management, encompassing medication-related knowledge, including how diet, exercise, and blood glucose monitoring should be managed concerning medication use, as informed by the needs expressed by phase 1 interviewees.

Health education short videos were shared via the WeChat groups. In the first month, for the PC, C, and PR groups, in addition to sending stage-targeted videos specific to their current stage of behavior change, all general videos were also sent simultaneously to avoid participants in later stages having to watch general videos repeatedly. For the A group, general videos were sent in batches over the first and second months. This prevents participants from remaining in the same stage group in the second month without new video content. Compared to the pre-action stages, the A stage requires a longer time to progress to the next stage, as there is a need to maintain medication adherence behavior.

WeChat group discussions combined voice messages and text discussions, allowing asynchronous participation where participants could conveniently share experiences, thoughts, and questions at their own pace and through their preferred mode of expression. Each stage-specific WeChat group included patients and a health provider team (a researcher, a community family doctor, a community nurse, and a nursing undergraduate volunteer). The researcher initiated the discussion topics, the family doctor summarized and supplemented the content, and the community nurse and nursing undergraduate volunteer assisted in organizing the discussion, ensuring smooth facilitation and participant engagement. The specific themes of health education short videos and WeChat group discussions across stages are detailed in Supplemental Material 2. Additionally, The researcher provided one-on-one online health guidance for participants whose SOC did not change or had negative changes at the end of the first and second-month measurements.

Routine care

In the comparison group, participants received routine health education following the Guidelines for National Basic Public Health Services (third edition). ⁹ All Patients received brief health education from family doctor teams during community outpatient or telephone follow-ups. However, the extent to which each patient actually engaged in or participated in health education posters displayed at the CHSC and self-collected health education brochures was not determined. Community follow-ups were conducted every 3 months, which included measuring fasting blood glucose levels, scheduling patients for outpatient visits, and telephone tracking. For patients with well-controlled blood glucose, the next follow-up appointment was scheduled. However, for those with unsatisfactory blood glucose control, the community family doctor adjusted their medication, and a follow-up visit was arranged within 2 weeks.

Outcomes

The primary outcome was positive SOC movement, and secondary outcomes were medication self-efficacy, medication adherence, and HbA1c levels. These outcomes are aligned with theoretical constructs, key health behaviors, and clinical indicator. Three experts initially evaluated all questionnaires for accuracy and appropriateness, with a content validity index (CVI) ranging from .93 to 1. All participants completed the HbA1c measurement at the CHSCs and self-report questionnaires via telephone inquiries at baseline, 3 months, and 6 months. Additionally, the intervention group completed SOC assessment in month 1 and 2 to determine stage transitions for tailored interventions.

Questionnaire for Stages of Change in Medication Adherence of Diabetic Patients (QSCMC) was employed to assess SOC in medication adherence of T2DM patients. ⁸ Patient categorization was based on non-adherence occasions in the preceding month, as well as their intention and actual adherence behavior: PC stage denoted ≥4 occasions with no intention to change within 6 months; C stage involved ≥4 occasions while planning change in the 6months; PR stage entailed ≥4 occasions with a planned change in the next month; A stage comprised <4 occasions and adherence behavior sustained for <6 months; M stage constituted <4 occasions and adherence behavior sustained for ≥6 months. The retest reliability was .80.

The Chinese version of the Morisky Medication Adherence Scale-8 (MMAS-8) was used to assess medication adherence. ¹⁹ It comprises 8 items: Items 1 to 7 are dichotomously scored (yes = 0, no = 1), with item 5 reverse-scored; item 8 is scored as 1, 0.75, 0.50, 0.25, or 0 for responses “never,” “rarely,” “sometimes,” “often,” and “always,” respectively. Higher scores (a maximum of 8) indicate better adherence. Cronbach’s alpha was .65, and the intra-class correlation coefficient was .80.

Chinese Diabetes Medication Self-efficacy Scale (CDMSS-11) was used to evaluate medication self-efficacy. ²⁰ with 11 items scored on a 3-point Likert scale (not at all = 1, somewhat sure = 2, and very sure = 3). Higher scores indicate greater self-efficacy. Cronbach’s α was .94.

Statistical analyses

Statistical analyses were conducted using IBM SPSS Statistics 18.0 software. SOC and clinically significant changes in HbA1C were described using frequency and percentage. MMAS-8 scores, CDMSS scores, and baseline HbA1c levels were described using mean and standard deviation. Chi-square tests examined between-group differences in SOC distribution at baseline.

For the analysis of positive SOC movement and clinically significant change in HbA1C at different time points, we conducted multiple binary logistic regression analyses. We considered other potential confounding variables (age, gender, marital status, and education level) and performed univariate analyses and collinearity diagnostics to select independent variables. Variables significantly associated with the outcome and free from collinearity were ultimately included in the models for each time point. MMAS-8 scores and CDMSS scores before and after the intervention were compared between the 2 groups using Mann-Whitney U tests. Friedman’s ANOVA was used for each group to test for differences among the 3 time points. If the overall test was significant, Bonferroni-adjusted pairwise comparisons were performed to identify differences between any 2 time points. The significance level was set at P < .05 for all analyses.

Ethical consideration

This study received approval from the Ethical Review Committee for Human Research (MUPH 2022-154) and the Institutional Review Board of Biomedical Ethics Committee (SWMUIRBKS-202312-0014). The study procedures were conducted following the Declaration of Helsinki. To minimize potential discomfort, we consulted with patients to select locations to create a comfortable environment during interviews. Additionally, preventive measures during blood sampling, including infection control strategies were implemented to prevent adverse reactions or infection from blood sampling.

Results

Phase 1: Program Development

The study interviewed 32 T2DM patients from the intervention group, with 8 patients in each of the PC, C, PR, and A stages. The average age was 54.3 years, 71.9% (23/32) were female, 96.9% (31/32) were married, and 81.3% (26/32) had an education level below high school.

We found that patients’ concerns and needs regarding hypoglycemic medications varied across different SOC. In the PC stage, patients mainly expressed doubts about the necessity of taking the medications prescribed by doctors and what benefits the medications could bring: “I don’t think taking medications is that important. Controlling my diet should be enough.” “Medications are like poisons. If I take them for too long, I’ll get addicted and can never stop.” The C stage patients expressed skepticism toward medication adherence and were puzzled over whether traditional Chinese medicine or Western medicine was more effective: “My blood sugar levels have been going up and down. It doesn’t seem to make any difference whether I take the medications or not. Why do I have to take pills every day?” “I see some people can control their levels quite well with Chinese herbs, but others say herbs are too slow, while Western medicines act faster. Which one is better?” The PR stage patients sought information on medication side effects and medication regimen, as well as government subsidies for T2DM: “I want to know if these medications have any harmful effects on other parts of the body, or if they may cause any side effects or adverse reactions.” “Why is my medication different from my friend’s? . . . you can explain.” “I’ve heard the government provides subsidies for patients with chronic diseases. Does diabetes count? What kind of discounts or benefits can we get?” In the A stage, patients expressed needing to know the precautions for taking medications and what to do when experiencing hypoglycemia: “Sometimes I forget to take my medications, but I don’t know if I should take a make-up dose or not. What should I be mindful of?” “A few times, I woke up in the middle of the night feeling powerless and sweating, which I think were signs of low blood sugar. How should I handle that?”

Regarding patients’ previous experiences obtaining health information, we found that community follow-up visits are the primary source of health knowledge for patients: “I get most of it (health information) from my family doctor. Every time I go for a follow-up, they talk to me and share some precautions and things to keep in mind.” However, the interviewees identified 3 aspects of inadequacy in the health education provided during follow-up visits. Firstly, the time allocated for health education during follow-ups is limited, and the content is incomplete: “Every time I go to the community, there are so many people that the doctors don’t have time to say a few more words to me.” “Every time doctors talk in general, they always say to eat less, take medication on time, and exercise more. Who doesn’t understand these, but I still don’t know how to do them.” Secondly, the health information obtained during follow-ups is quickly forgotten and difficult to review: “I remember the doctor said these things to me, but every time I have to do them, I forget. My memory is not good.” Lastly, participation in offline health education activities is constrained by time and space: “The community called me and asked me to attend the lecture. I don’t have time and must work to earn money.” “I moved to another district, and it’s too far from the community to attend.”

Regarding the suggestions for an online program, participants expressed a welcoming and accepting attitude, recognizing the advantages of video-based health education, such as its intuitiveness, flexibility, and the ability to facilitate repeated learning: “I think the video is quite good, much better than reading those materials (health education manuals), it’s clear at a glance.” “With the video, I can watch it whenever I have free time, without having to go to the community to attend lectures (on-site health education activities).” “My memory is bad, and I often forget the (health knowledge) that my family doctor tells me during clinic visits. But the videos are different, and I can still watch them repeatedly.” Furthermore, the interviewees hoped that the video content would provide comprehensive guidance on various aspects of diabetes management, including medication, diet, exercise, and blood glucose monitoring: “Taking medicine alone is not enough, and diet is also important. How can we ordinary people get away with three meals a day?” “They always say that medicine is three parts poison. What are the side effects of this medicine? I heard that there is a new medication that can cure my disease (diabetes), you know what?” “They all said I need to exercise, but when I exercise, I feel dizzy. One time, I ran and almost fainted, which scared my family. You can talk about these aspects (exercise).” “I have this (blood glucose meter), but I found it inaccurate. It’s different from when I went to the community for a check-up. I’m wondering if there’s a problem with the instrument or if there’s a problem with what I measured.”

Moreover, participants believed that the focus of video production should prioritize the accuracy, scientific basis, and feasibility of the content rather than emphasizing the format or duration of the videos: “The length of the video doesn’t matter. The key is that the content is useful and can be understood and learned by ordinary people.” “I don’t care how the video is filmed; fancy (overly complex production) is useless. The most important thing is whether the content is correct and reasonable.” The more specific themes, categories, and codes of phase 1 are in Supplemental Material 1. These qualitative findings informed intervention content and mechanism, including format, duration, and scheduling.

Phase 2: Program Implementation and Evaluation

Of 216 participants, 17 participants dropped out: 9 failed to engage, and 8 were lost to follow-up in the intervention group; 10 were lost to follow-up in the comparison group. The final analysis included 189 participants: 91 in the intervention and 98 in the comparison group. Each intervention group participant received a mean of 19.8 (SD = 4.7) videos and discussed a mean of 1.7 (SD = 0.9) topics.

Table 1 shows no significant differences in characteristics between the 2 groups. The intervention group had a mean age of 56.4 years (SD = 3.3), with half being female, mostly married, and having an education level below high school. At baseline, there were no significant differences between the 2 groups in SOC (P = .93), MMAS-8 (P = .45), CDMSS-11 (P = .84), and HbA1c (P = .24).

Table 1.

Baseline T2DM Patients’ Characteristics.

Characteristics	Total (N = 189)	Intervention group (n = 91)	Comparison group (n = 98)	Statistical test	P-value
Female, n (%)	104 (55.0)	49 (47.1)	55 (52.9)	0.099	.75 a
Age, mean (SD)	56.7 (4.2)	56.4 (3.3)	56.9 (4.8)	1.531	.13 b
Marital, n (%)	172 (91.0)	82 (90.1)	90 (91.8)	0.172	.68 a
Education, n (%)
Primary school	61 (32.3)	30 (33.0)	31 (31.6)	0.039	.98 a
Middle school	84 (44.4)	40 (44.0)	44 (44.9)
High school and greater	44 (23.3)	21 (23.1)	23 (23.5)
SOC, n (%)
Precontemplation	22 (11.6)	10 (11.0)	12 (12.2)	0.442	.93 a
Contemplation	38 (20.1)	18 (19.8)	20 (20.4)
Preparation	56 (29.6)	29 (31.9)	27 (27.6)
Action	73 (38.6)	34 (37.4)	39 (39.8)
MMAS-8, mean (SD)		5.2 (1.9)	5.3 (2.0)	0.761	.45 b
CDMSS-11, mean (SD)		23.2 (5.3)	23.4 (5.6)	0.200	.84 b
HbA1c, mean (SD)		7.9 (0.7)	7.8 (0.5)	−1.170	.24 b

Abbreviations: SD, standard deviation; SOC, stage of change.

^aChi-square test.

^bMann-Whitney U test.

As shown in Table 2, the intervention group were more likely to achieve positive SOC movement (P < .001, Adj OR = 13.69 95% CI = 6.76-27.71) compared to the comparison group. The higher number of those participants in the intervention group with positive SOC movement were reported at all time points; at 3 months (P < .001, Adj OR = 17.13, 95% CI = 6.98-42.08), at 6 months (P < .001, Adj OR = 14.23, 95% CI = 6.85-29.53), and at both 3 and 6 months (P < .001, Adj OR = 31.28 95% CI = 9.87-99.20). Among participants who experienced positive SOC movement, 37.3% (28/75) received one-on-one guidance.

Table 2.

Positive SOC Movement and Clinically Significant Change in HbA1C Between Groups.

Outcome	Intervention group (n = 91)	Comparison group (n = 98)	Adjusted OR (95% CI)	Wald χ2	P-value
Positive SOC novement, n (%)
At 3 months	49 (53.8)	7 (7.1)	17.13 (6.98-42.08)	38.41	<.001 a
At 6 months	71 (78.0)	22 (22.4)	14.23 (6.85-29.53)	50.78	<.001 a
At both 3 and 6 months	45 (49.5)	4 (4.1)	31.28 (9.87-99.20)	34.19	<.001 a
Total	75 (82.4)	25 (25.5)	13.69 (6.76-27.71)	52.85	<.001 a
Clinically significant change in HbA1C, n (%)
At 3 months	28 (30.8)	10 (10.2)	3.91 (1.77-8.63)	11.42	.001 a
At 6 months	40 (44.0)	12 (12.2)	5.62 (2.70-11.69)	21.36	<.001 a

Abbreviations: OR, odds ratio.

^aMultiple binary logistic regression analysis.

Regarding HbA1c, a change of at least 0.5% was required to be accepted as clinically meaningful. ²¹ Finding shows that participants in the intervention group were more likely to achieve clinically significant HbA1c change at 3 months (P = .001, Adj = OR 3.91, 95% CI = 1.77-8.63) and at 6 months (P < .001, Adj OR = 5.62, 95% CI = 2.70-11.69) compared to the comparison group.

As shown in Table 3, the intervention group showed significant improvements in MMAS-8 and CDMSS-11 scores at 3 and 6 months compared to baseline (P < .05). The between-group comparison indicated that the intervention group exhibited significantly higher mean scores for both MMAS-8 and CDMSS-11 than the comparison group at 6 months.

Table 3.

MMAS-8 and CDMSS-11 Scores Within and Between Groups.

Variables	Intervention group (n = 91)	Comparison group (n = 98)	MD (SD) between groups	U-value	P-value
MMAS-8, mean (SD)
Baseline	5.2 (1.9)	5.3 (2.0)	−0.1 (1.9)	0.761	.45 a
3-month	6.1 (1.1) c	5.4 (2.0)	0.7 (1.6)	−1.902	.06 a
6-month	6.3 (1.0) c	5.2 (1.8) d	1.1 (1.6)	−4.629	<.001 a
F-value	44.410	7.887
P	<.001 b	.019 b
CDMSS-11, mean (SD)
Baseline	23.2 (5.3)	23.4 (5.6)	−0.2 (5.4)	0.200	.84 a
3-month	25.8 (4.1) c	25.0 (5.5) c	0.8 (4.9)	−0.550	.58 a
6-month	26.8 (4.2) c	25.0 (5.6) c	1.8 (5.0)	−2.167	.03 a
F-value	48.189	17.687
P	<.001 b	<.001 b

Abbreviations: SD, standard deviation; SOC, stage of change.

^aMann-Whitney U test.

^bFriedman’s ANOVA.

^cBonferroni-adjusted multiple comparisons following a significant Friedman’s ANOVA. Compared to the baseline, the difference was significant (P < .05).

^dBonferroni-adjusted multiple comparisons following a significant Friedman’s ANOVA. Compared to the 3 months, the difference was significant (P < .05).

As shown in Figure 1, the number of participants in the PC, C, and PR stages in the intervention group gradually decreased over time, while those in the A stage progressively increased. Notably, from the 3rd to the 6th month, participants in the A stage kept moving forward to the next stage, resulting in a substantial increase in the number at the M stage. Corresponding to the SOC stage transitions observed, the stage-specific WeChat group participant count ranges within the intervention period were: PC (1-10), C (14-18), PR (24-31), and A (34-45).

Figure 1.

Changes in the distribution of patients across SOC at different time points in the intervention group (n = 91).

Discussion

The study findings demonstrated that the TTM-based online program could effectively promote positive SOC movement, improve medication adherence and self-efficacy, and could lead to better glycemic control in newly diagnosed T2DM patients. The positive changes in medication adherence in this study occurred within a shorter timeframe compared to the previous TTM-based behavior change study. ⁸ The TTM suggests changes in stages occur on a 6-month basis. However, there is no clear sense of how much time is needed for each stage or how long a person can remain in a stage. Prochaska et al ²² stated that moving from 1 stage to the next within 1 month will double one’s chances of acting on changing behavior in the next 6 months. In this study, the 3-month intervention design was used to assist the newly diagnosed patients in increasing their motivation and likelihood of sustaining the medication adherence behavior during the subsequent 3-month follow-up period. Findings on positive SOC movement for a period of fewer than 6 months are also consistent with previous studies on TTM-based interventions.^23,24 Moreover, our study showed that during the 6-month post-intervention follow-up, participants in the A stage continued progressing to the M stage, indicating the long-term effectiveness of our intervention strategy.

Furthermore, TTM proposes that the transition from the PC stage to the M stage is a process of shifting from behavior intention to actual behavior, with individuals in the PC and C stage needing to raise consciousness for change, while those in the PR stage and A stage need to develop plans and take action. ²⁵ While previous research focused on providing health information to patients in PC and C stages to improve disease cognition, ⁸ our study found that it is essential to not only focus on the cognition of patients in the PC and C stages but also address the different concerns and needs regarding hypoglycemic medications of patients in the PR and A stages. Therefore, our intervention strategy aimed to enhance participants’ health knowledge about T2DM medications throughout all stages via health education short videos to address their varying concerns and needs. Furthermore, one-on-one online guidance was promptly provided to participants who did not change or experienced negative changes to facilitate behavior change.

Our findings showed that TTM-based interventions can effectively improve patients’ self-efficacy and medication adherence, which is consistent with previous studies.^{6,8,26 -28} However, our study found no significant between-group differences in adherence and self-efficacy at 3 months post-intervention. Significant differences emerged only at the 6-month follow-up. This delay may be attributed to the lengthy process of habit formation, where short-term intervention effects do not immediately translate into stable habits. The impact of the intervention on adherence and self-efficacy may not be immediately apparent, leading to greater variability in early results and wider confidence intervals for the primary outcome of positive SOC movements. Moreover, although personalized online guidance was provided for those who did not change or experienced negative changes during the intervention, individual behavioral and psychological variations, compounded by real-life stressors like work and relationships, could impact self-efficacy and adherence. Therefore, continuous reinforcement and monitoring are crucial for sustaining positive behavioral patterns.

In this study more participants in the intervention group had a reduction in HbA1c of at least 0.5% or more, which is considered a clinically significant change than the comparison group. This study provided evidence for the role of TTM-based interventions in clinical outcome in patients with diabetes, which is consistent with previous research.^29,30

The outcomes of the intervention demonstrated the acceptability and feasibility of the intervention to a considerable extent, as before designing the intervention, qualitative interviews were conducted to understand participants’ needs and perspectives and developed strategies based on the TTM framework. During the intervention, family doctors and community nurses engaged in WeChat groups to provide support, where participants expressed positive feedback such as “very practical” and “content on point.” The intervention results showed that among the initially enrolled 108 participants, only 9 did not participate in the WeChat discussions, and 8 dropped out due to incomplete data collection, considerably lower than the 25% dropout rate reported in previous studies.^16,17

Limitations

The study had some limitations. Although the short duration of the video, averaging 139 s viewing time, made it easy to be fully watched, we could not ascertain the number of views or the impact of repeated viewing on the study outcomes. The measurement of adherence in this study relies only on self-reporting, which might have introduced some uncertainty in determining participants’ SOC.

Future studies should extend the intervention period and follow-up to ensure sustainability, especially for those in the action and maintenance stages. Furthermore, the TTM-based online health education videos should be developed on broader aspects affecting blood sugar control, like diet, exercise, and self-monitoring. Monitoring mechanisms for video-watching in the intervention group, as well as participation in routine care activities in the comparison group should also implemented. Multiple perspectives and multi-modal assessments of acceptability and feasibility should be considered to assure service acceptance and sustainability. Integrating tailored health education content for patients at different behavior change stages and developing videos suitable for diverse populations is also crucial. These resources could be disseminated across communities to enable resource sharing and maximize public health impact.

Conclusion

This study first combined TTM with participants’ perceptions from preliminary interviews, developing a TTM-based online program for newly diagnosed T2DM patients. Findings demonstrated that this program could effectively promote positive SOC movement, improve self-efficacy and medication adherence, and could lead to better glycemic control. More importantly, we explored the application of TTM to health education video development and demonstrated the effectiveness of our intervention approach. This intervention could serve as a model extendable to other topics and populations to enhance intervention effectiveness and reach ultimately improving health outcomes for patients.