Ankylosing spondylitis (AS) is a chronic rheumatic disease characterized by inflammatory back pain, morning stiffness, and reduced physical function that requires long-term management [1]. Regular disease management, including pharmacological therapy and exercise, is essential to control disease and prevent progression of the disease [1,2]. The inflammatory back pain and morning stiffness of AS patients improves with exercise but not with rest or inactivity [3]. Evidence reveals that exercise (such as stretching or aerobic exercise) can improve health outcomes, including pain, morning stiffness, and physical function, and the effects of exercise depend on patient adherence [4,5]. Exercise adherence refers to the extent to which people undertake the prescribed exercise from health care providers [5]. However, AS patients face severe challenges during disease management, such as lack of knowledge about the disease and nonadherence to medication and exercise [6-8]. Lack of knowledge about AS is a barrier to exercise and medication adherence [8]. Lack of exercise adherence negatively influences health outcomes among AS patients [8].

Patient education comprises educational activities designed to influence patient knowledge and health behaviors, enable patients to manage their disease, and optimize health outcomes [2,9]. Ndosi et al [10] revealed that patient education should aim at improving patient self-efficacy since self-efficacy is a predictor of health behaviors and health outcomes. Self-efficacy is defined as an individual’s confidence in performing a specific behavior [11]. Previous studies indicated that patient education can increase disease knowledge, self-efficacy, and adherence in arthritis patients [10,12,13]. However, only several published studies explored the effects of patient education among AS patients [14-20]. These interventions mostly reported small sample sizes, the lack of theoretical basis, and found limited effects on self-efficacy and adherence, and inconsistent results on health outcomes (eg, disease activity, physical function) [15-20]. The intervention delivery methods of many previous studies relied heavily on face-to-face interactions, which can be difficult due to travel restrictions, time constraints due to any number of factors, and costs of missing work [18,21].

Mobile health (mHealth) interventions can deliver timely health service and overcome the obstacles of time, distance, and cost [22]. The wide use of mobile phones has increased the possibility of delivering through health-focused interventions via apps [23]. In recent years, WeChat has been the most popular social networking app in China, with over 1 billion monthly active users [24]. WeChat can offer free message, voice/video calls, and enhance effective communication and information sharing [23,25]. WeChat has been used as a tool for educational interventions in patients with cancer, hypertension, coronary heart disease, and these studies reported positive effects [23,26-29]. Evidence revealed that AS patients need available education through phones and apps [30]. However, there is little evidence of an educational intervention through WeChat for AS patients.

AS patients may benefit from effective theory-based interventions to improve health behaviors and health outcomes (eg, disease activity, physical function). The health belief model (HBM) is developed to explain how to change health behaviors and focuses on an individual’s likelihood of engaging in healthy behaviors [11,31]. Previous studies revealed that the HBM is effective in developing interventions to change an individual’s beliefs and healthy behaviors in cancer, pulmonary tuberculosis patients [32,33]. However, this theory has not been used to develop an educational intervention for AS patients.

Based on these findings, a randomized controlled trial for AS patients was conducted and aimed to compare patient outcomes in a theory-based mHealth intervention via WeChat with standard care. Results regarding quality of life, depression, and selected clinical outcomes have been published elsewhere [34]. This paper describes the primary outcomes of this intervention, including disease knowledge, self-efficacy, exercise adherence, disease activity, and physical function. We hypothesized that the HBM-based mHealth intervention would improve the disease knowledge, self-efficacy, exercise adherence, physical function, and control disease activity of AS patients.

The key constructs of HBM consist of perceived susceptibility, perceived severity, perceived benefits, perceived barriers, cues to action and self-efficacy [11,35]. In HBM, health behaviors are based on people’s perceptions of susceptibility to and severity of health problems, barriers and benefits to enacting health behaviors and cues to action [31]. Self-efficacy can improve the efficacy of the model, and change subsequent health behaviors [11]. This intervention helped patients understand the severity of AS and provided strategies for managing their disease in order to improve self-efficacy, healthy behaviors, and achieve better health outcomes. We applied the theory to the AS intervention by using the constructs to guide the design of the intervention and match the intervention elements to the hypothesized outcomes. The primary goals of patient education are to transfer knowledge about disease and improve health behaviors [9]. Exercise adherence is a crucial health behavior related to AS patients, and self-efficacy is a predictor of health behaviors. This educational intervention was designed to explore short-term effects. Thus, we selected disease knowledge, self-efficacy, and exercise adherence as primary outcomes while health outcomes (ie, disease activity and physical function) were secondary outcomes.

This study was a single-blinded randomized controlled trial conducted from March to December 2017. Written informed consents were obtained from all participants (legal guardians of participants under 18 years provided written informed consent) after they had received information about the study protocol.

The sample size was determined by self-efficacy score based on our pilot trial, in which standard deviation was 1.47 and mean difference between the two groups was 0.86. With a power of 0.80 and α=.05 (2-sided), each group required 47 participants. The final sample size was 114, allowing a 20% dropout.

The study was registered at the Chinese Clinical Trial Registry [ChiCTR-IPR-16009293], conducted in accordance with the Declaration of Helsinki, and approved by West China Hospital Medical Ethics Committee (ID: 20160364).

The Department of Rheumatology and Immunology from a tertiary hospital in a large city, serving a population of Southwest China, was the site of recruitment. Potential participants were recruited via convenience sampling during their routine care. Participants were included if they (1) were diagnosed with AS according to the Modified New York Classification Criteria for AS, (2) were aged 14 years or older (participants under 18 years need written informed consent signed by legal guardians), (3) could speak and understand Chinese, (4) could use WeChat and had a WeChat account, and (5) were willing to participate in this randomized controlled trial study. We excluded participants if they (1) had severe cognitive or mental problems (comprehension or expression problems, using psychotropic drugs), (2) had other rheumatic diseases, or (3) were participating in other research programs.

To ensure participant assignment was truly randomized without human bias, a well-trained study coordinator generated a random allocation table in Excel (version 2010, IBM Corp). Before recruiting participants, the study coordinator placed randomized numbers in sealed opaque envelopes. Participants selected an envelope at random. Based on the selected number, participants were allocated to the intervention or control group. The researcher who collected the data and study coordinator who generated the random allocation table were blinded to group allocation.

Data were collected at baseline and the 12th week. If participants had difficulty in reading or writing, the researcher would help them complete questionnaires. Baseline data were collected from participants and medical records. The posttest data were collected from participants when they came to the rheumatology clinic for routine care or through an online survey platform [44] or through telephone/WeChat call.

Data were analyzed using SPSS (version 22.0, IBM Corp) software. An intention-to-treat principle was used for analyses, and the last observation carried forward method was used for missing data assessment. Data were described as mean and standard deviation, median and interquartile range, and frequency and percentage. Independent sample t test, Mann-Whitney U test, and chi-square test were used to compare data between the intervention and control groups. Paired sample t tests were used to analyze the changes in outcomes from baseline to the 12th week within each group in continuous variables. P<.05 was considered statistically significant.

A total of 118 participants were included and randomly allocated into the intervention (n=59) or control group (n=59). A total of 89.8% (106/118) of participants completed the study. Additionally, we included 118 participants in data analyses because the intention-to-treat principle was used. Figure 1 shows the study flowchart.

The average age of participants was 29.9 years. The majority of participants were male (93/118, 78.8%). Around half of participants were married (56/118, 47.5%), never smoked (70/118, 59.3%), and had college educational level or above (62/118, 52.5%). There was no statistically significant difference in the variables between the 2 groups (Table 2).

The correct item score, correct answer score, and 4 area scores of knowledge of AS and self-efficacy scores in the intervention group were significantly higher than the control group after the intervention (all P<.001). A larger proportion of participants in the intervention group adhered to regular exercise after the intervention compared with the control group (P=.003, Table 3). The within-group analyses for the intervention group showed significant increases in all scores of AS knowledge and self-efficacy scores (all P<.001). The within-group analyses for the control group detected increases in correct answer score (P=.04) and general knowledge score (P=.002), but no significant difference in self-efficacy score, other scores of knowledge of AS including the correct item score (all P>.05, Table 4).

At posttest, there was no difference in disease activity or physical function between the intervention and control groups (P>.05). The within-group analyses for the intervention group showed no significant change in disease activity or physical function (P>.05). The within-group analyses for the control group detected a decline in physical function (P=.002), but no significant change in disease activity (P>.05, Table 3 and Table 4).

This study explored the effects of the theory-based educational intervention through WeChat among Chinese patients with AS. Our findings demonstrated that this intervention was feasible and beneficial for improving patient disease knowledge, self-efficacy, and exercise adherence, which was in line with previous studies [12,16,41]. The educational intervention delivered by WeChat can increase access to health care providers for participants, teach knowledge and skill of disease management, and have positive effects in AS patients.

We found that the theory-based educational intervention can increase patient knowledge about AS, which was in line with prior studies [16,41], and corresponds with educating the patients on perceived severity of the disease and perceived susceptibility to increased limited mobility without action. Haglund et al [30] revealed that 43% of spondyloarthritis patients had educational needs. Moreover, patient knowledge levels of AS in this study were relatively low compared with previous studies [36]. In our study, the research nurses provided knowledge of managing AS, which may increase patient knowledge levels of AS.

The educational intervention via WeChat can effectively improve self-efficacy of AS patients, a finding similar to prior studies [18,45]. Self-efficacy can be enhanced through direct experience, alternative experience, and verbal persuasion [46]. In this study, participants gained knowledge and peer experience about disease management through educational information and nurse coaching by praising small accomplishments, which increased their perceived self-efficacy. Learning about useful experiences of others can inspire patients to try strategies to manage disease [47]. Seeing the adaptations of others to AS helps patients manage their disease better and improves confidence in coping with disease, which increases their self-efficacy to manage the disease. Persuasion from research nurses can help patients successfully manage their conditions [48]. That these skills may enhance patient confidence in managing disease and improving their self-efficacy was shown in our intervention.

This intervention effectively improved patient self-efficacy which, in turn, may have contributed to higher adherence to exercise. In this study, a larger proportion of participants in the intervention group adhered to regular exercise compared with the control group after the intervention. The finding was in line with earlier studies [16,41]. Self-efficacy is an important factor influencing exercise behavior in AS patients [49]. Our intervention helped patients perceive the severe consequences of AS, educated them on the importance of disease management, and taught them skills to manage their condition, which may have prompted regular exercise and helped them develop cues to action in their daily lives. In addition, the intervention delivered through WeChat may make it easier to exercise at home. These issues may enhance patient exercise adherence.

The results of this study did not detect significant differences in disease activity and physical function except for a decline in physical function in the control group. Previous reports on the efficacy of patient education on disease activity and physical function are inconsistent [14-18,20]. In our study, patients had relatively low disease activity and functional limitation, and these variables may be difficult to modify. Our 12-week intervention period may not be long enough to detect significant changes in biomarkers, such as disease activity, function, etc. Educational intervention may not produce a direct effect on health outcomes [2]. Thus, future studies should explore the long-term effects of educational intervention on health outcomes.

This study had several strengths. An assessor blinded to group assignment collected pretest and posttest data to reduce biased responses. Furthermore, using HBM might increase the efficacy of this intervention. Finally, we used an intention-to-treat analysis with multiple imputations for missing data to reduce bias in assessment of treatment effects.

This study had several potential limitations. First, we only recruited patients from a tertiary hospital who were able to use WeChat. Although the use of smartphone and internet access are relatively ubiquitous, the use of WeChat limits the generalization of findings to all Chinese AS patients. Second, patient views and cost-effectiveness analysis are important to evaluate and improve this educational intervention, but we did not collect these data because of limited time and financial support. Third, we collected outcomes at 12 weeks, but the effects of WeChat-based education on health outcomes may only become apparent in a long-term.

We demonstrated that the theory-based educational intervention delivered through WeChat, led by experienced nurses, was feasible and effective to improve AS patient disease knowledge, self-efficacy, and exercise adherence in a short-term. WeChat can deliver timely health service for patients with no available time or living in rural communities. During the COVID-19 pandemic period, the intervention approach may help health care providers provide continuous rheumatology care. We suggest that this intervention can be integrated into routine rheumatology care. Future studies should explore long-term effects of this intervention.