A randomized controlled trial was conducted using parallel treatment groups for 6 months. Participants were randomly allocated to either the self-management app for repeat tutoring (SMART-Liver, intervention group) or the standard care group (control group) in a 1:1 ratio.

The participants were recruited from the outpatient department of gastroenterology at a tertiary university hospital in Seoul, South Korea. In total, 138 patients with NAFLD were assessed for eligibility and participated in a study between November 2020 and January 2022. Inclusion criteria were patients who were aged 19 years or older, diagnosed with NAFLD without secondary causes for liver fat accumulation, able to read and write in English, and owned a smartphone using an Android system. The exclusion criteria were a history of alcohol intake of more than 20 g per day for women and 30 g per day for men based on the guidelines of the Korean Association for the Study of the Liver [20], those with hepatitis B or C virus infections, pregnant women who received hepatotoxic medication, patients who had cirrhosis, those with end-stage disease, those with concomitant liver disease, those diagnosed with psychiatric conditions, and those with clinically or biochemically recognized systemic diseases. The patients were screened for eligibility by clinicians and researchers.

Power analysis was performed using G*Power software (version 3.1; Heinrich Heine University) to determine the sample size. The sample size for each group was calculated to be at least 64 participants to obtain a 0.5 effect size, α=.05, and 80% statistical power. Accounting for a 10% dropout rate, the total sample size was planned to be 140 participants.

The recruited study participants were patients whose physicians recommended screening for eligibility when they visited the clinic. Participants who agreed to participate in the study and provided written informed consent were randomly assigned to either the control or intervention group in a 1:1 ratio. The allocation sequence was concealed using lots marked with the letter I (intervention group) or C (control group) in sealed, impenetrable envelopes prepared by a research assistant who was not involved in the intervention. The lots were stored in cabinets with locks before and after recruitment.

Participants in the intervention group received a unique code number for logging in to the SMART-Liver app after learning how to use the app and registering it on the spot. We provided a written manual for the SMART-Liver app to each participant in the intervention group. The research assistant surveyed participants in the intervention and control group using a questionnaire to obtain their general characteristics, self-management levels, and other psychosocial behaviors. The study period of 24 weeks began after a week in both groups. Participants in the control group received routine care based on treatment protocol in the hospital.

Data analyses were performed using SPSS Statistics for Windows (version 25.0; IBM). A per-protocol analysis was conducted to evaluate the results of the intervention among 102 participants. General characteristic data are described using mean and SD or frequencies. Differences between the intervention and control groups in sociodemographic, physiological, and psychosocial variables at baseline were analyzed using a 2-tailed t test. This process allowed us to confirm the homogeneity between the intervention and control groups and ensured that we did not require further adjustments for any imbalance.

We conducted a paired 2-tailed t test for continuous variables within each group and an independent 2-tailed t test between groups to compare the baseline and 3- and 6-month changes. Subsequently, a linear mixed model was used to evaluate any statistically significant differences over time in repeated measurements between the study groups. This approach assessed the intervention effect by simultaneously adjusting for the correlation between the intervention and control groups and between repeated measurements of the same participants.

This study was approved by the institutional review board of Yonsei University Health System (IRB 4-2020-0955). All participants were provided written informed consent prior to participating. This study only recruited participants who were willing to participate, and they were fully informed of the rights and obligations associated with the intervention. They retained the freedom to withdraw from the study at any time. All study-related and participant data were securely stored in a password-protected file cabinet with restricted access. To ensure the confidentiality of participants, all identifiable data were de-identified using coding. Participants were compensated for their participation and completion of the study survey on three occasions (Baseline, 3 months, and 6 months). At the 3-month survey, participants received US $15 upon completion of both the baseline and 3-month surveys. Additionally, participants who completed the 6-month intervention were provided with a free transient elastography test for liver fat valued at US $150.

The trial was prospectively registered in the World Health Organization–accredited Clinical Research Information Service on October 29, 2020 (registration number KCT0005549). We performed this trial based on the CONSORT-EHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Application and online TeleHealth) guidelines (Multimedia Appendix 1).

In total, 111 participants were enrolled in this study (Figure 2). During the first month, 2 patients withdrew from the study for personal reasons. Five patients in the control group dropped out of the study because they did not respond to the several contact attempts made by the researchers. Two patients in the intervention group dropped out due to poor compliance and reluctance to use the app. Thus, a total of 102 (91.9%) participants were included in the final analysis (48 in the intervention group and 54 in the control group). Overall, the mean age of the participants was 48.9 (SD 13.74) years.

Table 1 shows the demographic data and baseline characteristics of the participants in each group. There were no significant differences between the groups in terms of age, sex, socioeconomic status, and the presence of relevant comorbidities (all P>.05). The physiological and psychosocial outcome variables showed no differences between the participants in either group at the beginning of the intervention (Table 2).

The purposes of this study were to develop and evaluate the physiological and psychosocial effects of the lifestyle coaching intervention for patients with NAFLD. The results showed that there were significant changes in physiological outcomes (body weight, BMI, liver fat score, AST, ALT, and ɣ-GT) and psychosocial outcomes (self-management, fatigue, depression, and quality of life) in the intervention group for 6 months, and the intervention-by-time interaction was effective for self-management and fatigue improvement.

There was a significant reduction in weight within the intervention group from baseline at 6 months in this study. Similarly, there was a significant change in BMI in the intervention group than in the control group, which meant a more accurate weight loss representation. This result is consistent with previous studies that reported effects on weight loss in patients with NAFLD [17,29]. Weight loss is recommended for the treatment of NAFLD in the absence of pharmacological treatment. Previous intervention studies on patients with NAFLD have mainly focused on improving physiological outcomes, including body weight and clinical aspects [17,29,30].

In this study, other physiological outcomes, such as liver fat score, AST, ALT, and ɣ-GT levels, significantly changed in the intervention group at 6 months. The previous studies presented that weight loss through lifestyle modification is known to be associated with a reduction in liver fat and recommend a 5% minimum weight loss [5,31,32]. Lifestyle modifications, such as diet changes and exercise routines, have been reported to be effective in reducing body weight [33-35] but did not directly result in the reduction of participants’ liver fat scores in this study. The reason for this ineffectiveness in improving liver fat scores in this intervention may be related to the fact that participants did not present a significant body weight loss that could result in the improvement of NAFLD-related outcomes. Despite the lack of liver fat reduction at 6 months, there were significant changes in liver enzymes (AST, ALT, and ɣ-GT). Generally, liver enzyme levels in simple steatosis cases are mostly not high because their abnormalities are related to inflammatory processes [36]. Nevertheless, our results showed a reduction in the mean levels of liver enzymes at 6 months. Furthermore, we found that changes in liver enzymes were significantly different only among participants with a high level of intervention compliance. Although AST and ALT have a limitation of moderate accuracy in evaluating the fatty liver [37], this result suggests that further intervention studies need to be developed to identify the effects of changes in liver enzymes and to determine the significant level of weight loss and adherence on these changes.

In the interest of improving the health conditions of patients with NAFLD, psychological factors should also be taken into consideration, as many studies have reported a relationship between physiological biomarkers and psychosocial factors [38,39]. In this respect, our study identified that participants’ level of self-management improved in the intervention group compared to the control group and that the SMART-Liver app was significantly effective in improving self-management among patients with NAFLD at 6 months. Self-management behaviors that modify unhealthy lifestyles related to diet and lack of physical activity are a cornerstone of NAFLD treatment [35]. Several studies have conducted lifestyle modification interventions using mobile phones or other electronic devices and have identified the effects of physiological factors in patients with NAFLD [17,29,40,41]. However, other interventions targeting patients with chronic illnesses have evaluated their effects on psychosocial variables, such as self-management, self-efficacy, and quality of life, as well as on physical factors [40-43]. Our findings support the idea that lifestyle modification interventions can significantly improve psychosocial and physical parameters among patients with NAFLD.

Even though NAFLD is an asymptomatic disease, the SMART-Liver program was also effective in managing NAFLD’s nonspecific symptoms, such as fatigue. Recently, Golubeva et al [44] reported that fatigue was almost twice more common in patients with NAFLD than in those without it. However, chronic liver disease fatigue management remains a challenge because the pathophysiology of fatigue is complex. From this perspective, health outcomes following lifestyle modifications such as weight loss or reduction in liver fat may positively affect fatigue.

Additionally, participants with high compliance (90% and above) showed higher reductions in body weight and BMI at 6 months. Clinical factors (liver fat, AST, and ɣ-GT) and psychosocial factors (self-management and self-management knowledge) were also significantly improved by higher participant compliance (90% and above) compared to other groups. Similarly, there were significant differences in the level of fatigue and quality of life among compliance groups, with higher changes in the 70% or more compliance groups than in the group of 70% or less. Generally, higher compliance can improve health outcomes [45-47]. However, evidence on the standard level of compliance in lifestyle modification interventions using mobile apps is lacking, and many studies that included patients with various chronic diseases have mainly reported medication adherence [48,49]. Our study was conducted to derive changes in health outcomes according to lifestyle behavior modification compliance. These findings may be helpful in setting goals to improve the outcomes of the intervention in this population.

In this study, the attrition rate in the intervention group was low (9/111, 8.1%). The attrition of mobile-using interventions for chronic diseases was previously reported to be 40% [49]. In prior interventions in patients with NAFLD, attrition rates ranged from 5% to 27%, and interventions using the coaching method had especially lower rates [15,16]. Health coaching is an individual education method that can improve health behaviors and the effectiveness of interventions for chronic disease management [50]. Our study applied the coaching method to guide tailored goals by considering the participants’ circumstances. The participants were provided counseling via a remote meeting system halfway through the intervention period, which reset the goal based on their behavior and compliance and discussed their obstacles in changing unhealthy habits. In addition, the number of male participants with NAFLD was higher than female participants in this study even though we did not control for the sex ratio. Generally, in research, the proportion of female participants is higher than that of male participants [51]. As this study did not analyze the difference between male and female participants, further research is needed to compare this aspect.

In addition, compared to other liver diseases, most patients with NAFLD have less severe physical symptoms, and they can continue to engage in social activities, including work, relatively without constraints. In this respect, patients could participate in this program using a mobile app without time or space constraints, and if they had obstacles in sustaining their healthy lifestyles, they could contact and receive help from our research team. This might contribute to lowering the attrition rate. Further studies in this population may be successful if conducted considering these individual approaches.

We found that this lifestyle coaching intervention using a mobile app was successful in improving the level of self-management, which could facilitate healthy behaviors for weight loss and reduction of liver fat. Therefore, the results of this study can help future studies to develop lifestyle modification interventions and apply them to interventions for patients with other chronic diseases.

This study had some limitations. First, the intervention was conducted in a single hospital with only mobile phone users using the Android system. Although the use of mobile phones has increased, it can be affected by age, educational level, and household income. Therefore, future studies could include various types of mobile phones to diversify the target population. Second, the intervention involved counseling for goal setting via a remote meeting system that was not embedded in the app. The research team separately provided a meeting link to each participant, even though the meeting was informed through a chat channel within the app. Future studies may benefit from the development of an app with all possible functions needed for intervention. Finally, it would be beneficial to identify the effects of self-management on participant groups classified by NAFLD progression or clinical findings. Additionally, we did not collect data on app usage of frequency in this study. However, this frequency may be associated with the intervention effect [52]. Further study may be conducted on the relationship between intervention effect and app usage frequency in this population.

We developed a mobile app–based intervention for patients with NAFLD that provides personalized information, counseling, and interactions to facilitate self-management behaviors. This study showed that mobile app–based lifestyle modification interventions for patients with NAFLD effectively improved their physiological and psychosocial health outcomes. Our results suggest that future studies should include a more diverse and wider population and evaluate the effects of psychosocial aspects on self-management, as well as on clinical health outcomes in lifestyle coaching interventions using a mobile app. To confirm the effectiveness of the only coaching intervention, further studies that consider a mobile app–based lifestyle intervention without coaching as a group should be conducted.