One aim of this study was to describe the development process of a mobile app linked with an accelerometer for a clinic-based weight loss program. This study was part of a project to develop a personalized media service using a device-cloud interconnection for improving patient-doctor communication, which was funded by a grant from the Korean Evaluation Institute of Industrial Technology Research Fund. The development of the mobile app was planned and organized by a multidisciplinary team, including a team from Xeron Healthcare Corporation (Seoul, Republic of Korea), app designers, and teams from the hospital staff (family physicians and dietitians with a research coordinator; Seoul National University Bundang Hospital, Seongnam, Republic of Korea) between September 2013 and February 2014.

During the period of March 2014 to May 2014, we applied a pre-post single-group design to evaluate the usability and acceptability, as well as the early effects on patient-doctor relationship, of our newly developed mobile app, which was linked with an accelerometer. Usability assessments measured the technical effectiveness of the mobile app by asking whether the users could easily follow the steps without making any errors. We assessed acceptability by the users’ overall experience in using the mobile app, their perceived confidence in the information, and their satisfaction with the experience. Our aim was to implement a mobile app that runs on mobile phones as a clinical program. We did not provide telephone numbers, emails, or Web-based portal services to patients. A Web-based portal was only used for researchers to summarize participant’s anthropometric measurements, laboratory measurements, and corresponding obesity-related problems.

We briefly outline the clinic-based weight loss program as follows.

If a patient came to the clinic, either on their own or referred from other departments, a family physician obtained the patient’s medical history, health behaviors, previous attempts at losing weight, preferences, and goals. Then, the patient was scheduled for a meeting with a dietitian and was evaluated regarding obesity-related problems, including screening questionnaires for depression and eating disorders, body composition measurements, and laboratory examinations, such as tests for glucose, cholesterol, liver function, and kidney function. The patient could be referred to specialists in accordance with the problems identified in their initial assessment. The patient then received a treatment program consisting of lifestyle-focused interventions, including a nutritional plan, targeted behavioral changes, and an exercise program. The patient was also provided with pharmacotherapy or was referred to a bariatric surgical team if their body mass index (BMI) was >35 kg/m² with comorbidities, they had previously attempted lifestyle modification and pharmacotherapy without success, and they were interested in surgical treatment. A follow-up meeting with the physician was typically scheduled at 2- to 4-week intervals for the first 3 months.

We recruited 30 participants from patients who visited our clinic for weight loss programs. The World Health Organization’s Regional Office for the Western Pacific Region defines obesity in Asians as those with a BMI ≥25 kg/m² [39]. The Korean government officially uses this definition when defining and implementing health policies regarding obesity in Korea.

If patients met the following criteria, they were eligible for participation: BMI of ≥25 kg/m², possession of an Android mobile phone, and between 20 and 70 years of age. Patients were excluded if they had been admitted to a hospital due to a cardiovascular disease in the previous 6 months, being immobile due to surgical procedures in the previous 3 months, being pregnant, having experienced weight fluctuations of >10% in the previous year, or having a history of alcohol use disorder, an eating disorder, or of using weight loss pills in the previous 3 months.

Patients who were eligible were identified and informed about the study. On their second visit to the clinic, patients were invited to participate and discuss their results from the initial assessment. If a patient agreed to participate and provided written informed consent, he or she was scheduled to contact a research staff member for instructions regarding the process of downloading, authenticating, and using the app, as well as other pertinent information, including receiving manuals and explanations for the accelerometer. Moreover, patients were followed up 4 weeks after downloading the study app.

We also explained the delay of pharmacotherapy or surgical treatment for the 1 month of the study period so that we could perform pilot testing of the app’s effect on weight changes. Patients received the equivalent of US $20 per visit as reimbursement. The study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (B-1402-238-005).

The primary outcomes measured were the acceptability of our mobile app linked with an accelerometer and changes in the patient-doctor relationship after the intervention. We measured the usability of our mobile app linked with an accelerometer via a 5-point Likert scale by asking whether the users could easily follow every step without making errors and successfully complete all steps. We analyzed usage information of each module in our mobile app as follows. (1) mean number of log-ins per day: the mean number of log-ins to the mobile app per day during the study period, which represented how often a participant used the app in a day, (2) use of each module: the number of clicks per specific module divided by the total number of clicks in the mobile app during the study period, which represented how often a participant used a specific module in this app, (3) mean exercise calories per week: the mean exercise calories per week was recorded by an accelerometer during the study period, which represented the actual exercise activities recorded by an accelerometer in a week, (4) mean percentage of weekly exercise goals achieved: the mean percentage of actual exercise calories divided by suggested exercise goals per week, which represented the amount of exercise a participant achieved in a week, (5) total numbers of messages posted on the social network services: the numbers of messages, including replies participants posted on the social network services, which represented how frequently a participant used the social network services.

We evaluated the patient-doctor relationship by the scores of the Patient-Doctor Depth-of-Relationship Scale [40]. This scale is an 8-item self-completed questionnaire for measuring 4 elements: knowledge, trust, loyalty, and respect. The survey instrument has good reported reliability (Cronbach alpha=.93).

The secondary outcomes were changes in weight and abdominal circumferences from before to after the study. A research nurse measured these when patients visited the clinic.

According to the Personal Information Protection Act in Korea, unique identifiers, social security numbers, and bioinformation should be safely encrypted when they are intended to be transferred via a network or subsidiary storage devices. In our mobile app, we stored all individual information on the internal server, where a firewall was installed to block unauthorized access. We also incorporated an additional layer of authentication using an email address and a password.

We used descriptive statistics to summarize the baseline characteristics of the study participants. We analyzed the information on use for each functional category in our mobile app along with the proportion of patients who achieved their first month’s target goal. For the acceptability tests, we present the percentages for each response. Changes in weight and abdominal circumferences were calculated using a paired t test, and changes in the patient-doctor relationship scales were tested using a Wilcoxon signed rank test. We performed all statistical analyses using Stata version 12.1 (StataCorp LP) and considered P<.05 to be statistically significant.

A total of 30 individuals agreed to participate in the study. At 1 month, 93% (28/30) of participants completed their follow-up surveys. Although 2 participants used the mobile app for 1 month, they were unable to attend their 1-month follow-up visit. Among the 30 participants, 60% (18/30) were male. Participants had a mean age of 49.07 (SD 8.84) years and a mean BMI of 27.64 (SD 2.14) kg/m²; 60% (18/30) of participants had dyslipidemia, 27% (8/30) had hypertension, and 10% (3/30) had diabetes mellitus. (Table 1)

All study participants used the mobile app at least once during the study period even if they did not complete the follow-up visits. Table 2 presents usage information.

Participants used the study app for a median of 1.21 log-ins per day with a maximum of 6.38 log-ins per day. The most frequently used module was goal setting (median 199 times, 74%), followed by tracking of physical activities (median 30 times, 13.5%) and viewing of personal health status (median 34.5 times, 10.8%).

Most study participants achieved more than their recommended exercise; the median percentage of target exercise goals achieved per week was 125.9%. The mean number of messages posted on the social network service was 1 during the entire study period, and only 6 patients used the social network service.

With regard to the app’s ease of use, participants reported a high rate of positive responses for the log-in process (25/28, 89%) and goal setting process (23/28, 82%), as presented in Table 3. However, the rate of positive responses for the message board was low, including posting (11/28, 39%) and replying (10/28, 36%) on the social network service. Additionally, the positive response rate for the educational videos was 46% (13/28).

As Table 4 shows, regarding participants’ satisfaction with each module of the mobile app, setting personal goals had the highest positive response rate, at 75% (21/28). Ratings of satisfaction with the monitoring and feedback on physical activities were also highly positive. However, satisfaction with the social network service, educational videos, problem solving in case of errors, and reliability of the contents was approximately 50% (14/28, 14/28, 15/28, and 15/28), which was low relative to other modules. Nonetheless, most users (26/28, 93%) responded positively to receiving help with managing weight from the mobile app. Specifically, 75% (21/28) of users were willing to recommend this app to their family or friends, and 79% (22/28) (expressed interest in continuing to use this app for weight management.

Scales measuring the depth of the patient-doctor relationship decreased during the 1-month period from 27.6 (SD 4.8) to 25.1 (SD 4.5) by a Wilcoxon signed rank test (P=.02).

After the 1-month study period, the mean weight change was –0.1 kg (95% CI –0.6 to 0.8) and we observed no significant changes. However, abdominal circumference was significantly reduced by a mean of –1.84 cm (95% CI –3.3 to –0.4) (see Multimedia Appendix 2).

In contrast with many other weight loss apps that target the general population as a health and fitness tool, our mobile app was targeted to the population that participates in a clinic-based weight management program. Our choice of target population was based on several factors that are particular to hospital settings.

Numerous different phenotypic presentations [41,42] and risks that affect obesity-related mortality have been noted [43], even within the same BMI. Therefore, recognizing obesity-related comorbidities and functional status is essential for treatment planning, as well as for prescribing exercise or nutrition. These evaluations are typically possible only in a clinic or hospital setting with a multidisciplinary health care team. Hence, management of weight loss should not only target weight loss itself but also aim at controlling obesity-related health problems. Consequently, a commercially available weight loss app cannot deliver tailored information to each patient. We attempted to incorporate medical information from EHRs to deliver the appropriate information to each patient in need via tailored education videos and nutritional information. Moreover, bariatric surgery continues to be the most effective method for initiating significant improvement in obesity-related comorbidities [44,45], and the need for multidisciplinary weight management is increasing in hospital settings for both pre- and postoperational care. Therefore, mobile technology can contribute to a precise assessment of obesity-related comorbidities, thereby enhancing the medical and surgical treatments.

Given that informed, active patients, or patients who have the motivation, knowledge, skill, and confidence, play a large role in improving their own health outcomes [46], as well as reducing health care costs [47], our mobile app could have potential use as a tool for supporting patient activation in hospitals by providing tailored education along with self-management skills.

A mobile app linked with an accelerometer could be a good resource for clinicians for monitoring patient behaviors or physiological signals. In addition, if such information is recognized and used during a clinic or hospital visit, high-quality doctor-patient exchanges both inside and outside of the office visit can be achieved, which could favorably influence patients’ behavior and attitude [24].

Although our study did not use EHR-linked personal health records, our methods can be applied to incorporating them in chronic disease management, where patients require constant care and regular follow-ups.

Based on app usage data, the most frequently used module was goal setting, as well as the monitoring of and feedback on physical activities, which was concordant with the subjective satisfaction measures reported in the questionnaires. The median number of log-ins per day was 1.2, and all participants used our app over the course of the 1-month intervention period. Furthermore, participants achieved more than their suggested activity amount. We linked an accelerometer to our mobile app automatically, which potentially contributed to its frequent use and the successful achievements within our app. The self-regulation function was the main feature of our program, and participants evaluated this feature as the most comfortable and helpful through both objective and subjective measurements. These results were consistent with other studies [48,49], which further validates the benefit of mobile health apps in increasing physical activity. This increase may also result in a reduction in the abdominal circumferences of study participants by the end of the study period.

Conversely, we only provided nutritional guidelines that matched with the metabolic profiles of each participant, and we did not track nutrition, which may have affected the weight loss outcomes. Although abdominal circumferences were reduced by a mean of -1.84 cm during the study period, we noted no weight change. Research has demonstrated that frequent recording of dietary intake can contribute to weight loss [50]. However, recording food intake has been the most laborious and challenging work in mobile phone apps or websites for weight loss [51]. Healthy diet education via a mobile app did not affect weight loss. However, eliminating the food diary, which is time consuming, might also lead to a low attrition rate for mobile apps. Furthermore, a food recording function that is user friendly and less time consuming is necessary for achieving desirable outcome in a clinic-based weight loss program.

The satisfaction rate was low for the personalized mobile education module. We attempted to provide the study participants with tailored educational videos. Most participants did not download the educational videos from an Android operating system even if they received instructions on how to download the videos. The size of each educational video was between 31.8 and 52.4 MB. Therefore, the process of downloading and viewing educational videos may have been too difficult or too time consuming, or patients may have simply been in an area with no Internet or network connection.

Moreover, only a few participants used the social network service. Although social networking has been used extensively in intervention approaches, including those targeted at weight loss, no clear effect of social networking on weight change has been reported in randomized controlled trials [52]. However, participants received more encouragement from the community in many weight loss intervention programs that used social network services [52]. The low usage of this module combined with low satisfaction among our study participants may be attributed to several factors. The patients were all new users to our social network service, and no one had adequate experience to act as a facilitator. They study intervention period was only 1 month long, hindering the formation of a strong support group. Given the age and sex of our study participants, patients might not have been exposed to online social networking in general. The contents of the online networking in this study app might not have met participants’ needs, or the usability might not have been optimal, which is important for both user engagement and behavioral change [53]. However, research into the theory and function of online social networking and its impact on health outcomes is still in early development, and more studies regarding the actual usability combined with the contextual factors of social networking are needed. According to a systematic review [54], most studies reported that online social networking in attempts to change health behavior had poor retention and engagement.

Our mobile app had a unique feature: an interactive message board that was used to help patients share information with their doctors. However, according to the usage analysis, only a few participants used the interactive message board for communication. Family physicians in charge of patients left at least 2 messages for all participants under their care: 1 at the beginning of the study period to encourage behavioral change, and the other 2 weeks later to address questions about the weight loss program or process. Only 2 patients sent messages to their doctors: 1 pertained to how to take care of high blood pressure measured at home, and the other pertained to medication for controlling diabetes. Doctors replied to the participant’s question via the message board, but the participants were informed that the doctor’s reply may be instant or provided at the time of their scheduled session. Such misinformation may have inhibited the participants from using the interactive message board.

In our study, scales measuring the depth of the patient-doctor relationship decreased during the 1-month study period. This effect could reflect several aspects related to using the mobile app in a clinical setting. First, the doctor’s passive participation in the app, particularly in the interaction with a patient, might have been a factor in the decreased depth in the patient-doctor relationship. Second, the doctor’s encouragement or feedback regarding app usage might not have been recognized during the clinic visit, or the doctor’s feedback might not have been sufficiently positive. According to one survey regarding information technology and its impact on doctor-patient communication [55], panelists, including both medical professionals and patient advocates, expected improvement in the relationship. However, the lack of acceptance by doctors and issues of data security and monetary costs were regarded as barriers. Third, patient-doctor relationship scales might not adequately reflect short-term changes. Patient-doctor relationship scales were designed to measure the quality of the relationship in the context of continuity of care, whereas our study period was only 1 month. Fourth, empowering patients with self-management skills could make the patient-doctor relationship more complicated, and this phenomenon could have a negative effect on the short-term patient-doctor relationship if a doctor did not exhibit a supportive attitude toward using the apps. Moreover, cultural aspects of the relationship between a patient and his or her doctor in Korea may have played a role in preventing any improvement [56].

This study had several limitations. It was a pilot study of a single group that used a pre-post intervention design, which only demonstrated the possibilities for further improvement and clarification of the app’s design and treatment functions. The research was performed in 1 clinic for a limited time using a small number of participants, which may not reflect the diverse population of patients with weight problems. We attempted to implement the function of self-monitoring of physical activity by synchronizing the accelerometer with the app, but LG LifeGram has not published its validity test, which might affect the usage pattern for this study app.

However, to our knowledge, this study is one of the first to demonstrate the usability and acceptability of a mobile app linked with an accelerometer for a clinic-based weight loss program and its potential impact on the doctor-patient relationship. The mean age of the study participants was 49.07 years, which is generally older than other mobile app studies [30,57] with a greater proportion being male participants. This finding could potentially indicate that these were less-active users for a mobile weight loss app. However, all participants used the mobile app during the study period, achieved their suggested physical activity goals, and reported a high rate of satisfaction with the self-regulation function. One study [58] suggested that men would be willing to use Web-based self-monitoring tools delivered by mobile phones if interventions were easy, quick, and simple to use but showed little interest in using the social networking service. Our study results confirm these observations to some degree. Further studies identifying a mediating effect of age and sex will be needed to deliver a tailored intervention through a mobile app.

This mobile app has the potential for furthering improvements in achieving weight loss if it includes more strategies for reaching dietary goals, improvements in pattern monitoring and stress management techniques [59], and improvements in its structure, ease of use, personalized features, and accessibility in technology and design [60]. Further studies with a larger representative population and longer duration are needed.

In this study, we evaluated a mobile app linked with an accelerometer for a clinic-based weight loss program. Mobile health apps have the potential to be integrated into a clinic-based weight management program by empowering patients with self-regulation tools that enhance physical activity.

An easier method for recording daily food intake should be incorporated into the mobile app for better weight loss outcomes. Further reflection of the user’s needs with regard to the online social network and the patient-doctor relationship should also be considered.