Effects of a Video-based mHealth Program

Methods/Design The SPIRIT statement was used as a guideline for this protocol paper. Study design and setting This is a single-blinded, two-armed randomized controlled trial. The research assistant who collects the data is blinded, but the subjects and the health care providers who are involved in the intervention are not. This study is supported by five community centers. The centers are widely scattered across Hong Kong, which helps to maximize the generalizability effect of this program.

Participants and recruitment strategy, and randomization The leaders of the community centers will help draw up lists of potential subjects. Simple random sampling with a fixed sample size will be employed to recruit subjects. The inclusion criteria of this study are (1) being aged 60 or over, (2) going outdoors less than once a week in the last six months, (3) living within the service area of a participating community center, and (4) ability to use a smartphone. The exclusion criteria are (1) being diagnosed with dementia, (2) inability to hear, see or communicate, (3) being bed-bound, (4) having an active psychiatric illness with recent hospital admission within the last six months, (5) living within an area with no internet coverage, and (6) being already engaged in other mHealth programs.

Potential subjects will be approached and invited to participate in the study. A consent form will be signed once the subject agrees to participate. The random assignment schedule, generated using the computer software Research Randomizer, will be compiled by a research team member who is not involved in subject recruitment. The group assignments will be put in a sealed envelope and revealed sequentially at the time of randomization. The research assistant, after successfully recruiting a subject, will call the research team member for the random assignment. The research team member who has no knowledge of the identity of the subject will make the assignment based on computer numbers ('1' = video-based mHealth program; '2' = control group).

Interventions A three-month program will be designed in this study. There will be two groups involved in this study, described as follows.

Intervention group This group of participants will receive a video-based mHealth program that includes two main elements: 1) nurse case management supported by a social service team, and 2) individual-specific video messages covering self-care topics delivered via smartphone.

Upon enrolment, each client in the intervention group will be assigned to a nurse case manager. The nurse case manager will conduct a structured assessment via telephone call using the Omaha System. The Omaha System is a comprehensive assessment-intervention-evaluation tool that identifies 42 problems under the domains of environmental, psychological, physiological, and health-related behaviors. The Omaha System has been found to be applicable for older adults in the community and proven valid for use in Hong Kong. Following assessment, the nurse case manager will guide and empower clients to set their own goals and co-decide their own plans to manage their health problems.

Apart from monitoring clients' progress in accordance with their contract goals in a weekly telephone call, the nurse case manager will also send weekly, individual-specific videos with tips and reminders via smartphone. These videos cover topics of importance to clients according to the results of the Omaha System. The content of videos includes but is not limited to chronic disease self-management, healthy behaviors, simple self-care practices, and medication knowledge. The selected videos are limited to 10 minutes to accommodate the short attention span of older adults. All videos provided to the intervention group originate from reliable sources such as the Hospital Authority, the Department of Health, or credible health or charity organizations. The nurse case manager will first send a testing video to the client in the first home visit to ensure that the client can receive the video on his/her mobile. The client can either download the videos or view them online. The videos will run continuously unless the client presses the "pause" button. Videos can be replayed once by clicking the "play" button again.

A protocol for a referral system will be established with the full support of the health-social team in accordance with the conditions and needs of the client. The nurse can initiate a client referral according to the set guidelines. Health referrals include primary care consultation and hospital service if indicated. Social referrals include home meal service, psychological counselling and the like. In this team, the nurse case manager is responsible for providing health-related services such as health education and drug compliance monitoring, while social workers can provide services including arranging home-delivered meals, providing psychosocial support, and mobilizing community resources available in the district. A monthly case conference will be held in the community center between the nurse case manager and the social worker. Issues such as the progress and any concerns of the clients, suggestions for interventions, and requirements for modification or adjustment of goals will be discussed during the meeting. The shared responsibilities are based on standardized protocols and agreed referral forms and records.

Control group Both the intervention and control groups will receive their respective community services as usual. The most common services provided by community centers include meal and laundry services, transportation, health checks and education, and social activities, but most of them are episodic in nature. Social calls will be provided to the control group clients by a research team member who will not be involved in data collection, to minimize possible social effects. Social questions such as "Where will you go tomorrow?" and "What is your hobby?" will be set in the protocol. The research team member will follow the protocol strictly and will not provide health advice. The older adult clients will be encouraged to seek medical help whenever they express concern about their health.

Data collection Data will be collected at two time points-pre-intervention (T1) and post-intervention (T2). The baseline and post-intervention data will be collected by the research assistant, who will be blinded to the grouping. A previous study has found that collecting data via telephone call can yield the highest response from homebound older adults, since they seldom go out. The research assistant will be trained and tested on the inter-rater and intra-rater reliability. Five percent of the data will be randomly selected for independent review to ensure data quality.

Quality assurance mechanism This study will adopt a number of measures to ensure that the intervention delivery is valid and reliable. The nurse case manager employed in this study will have extensive experience in community elderly care, with competent level in using a smartphone, good communication skills, and the ability to work with a team. A training session will be given to providers in the areas of case management, the referral system, the Omaha system, telephone counselling skills and guidelines for telephone follow-up. During the training session and prior to actual implementation of the study, providers will be required to provide interventions to a simulated client in front of the research team to ensure that they understand and comply with the set protocols. The health-social team and the research team will meet regularly to review and discuss the progress of cases.

Ethical considerations The present study was granted ethical approval from the Human Subjects Ethics Sub-Committee (HSESC) of the University prior to the commencement of the study. This study will not cause any pain or discomfort to participants. Information about and an explanation of the ethical observations of the study will be provided to all eligible participants, and they will be asked to sign an informed consent. Participants will be reassured that they can withdraw from the study at any time without any adverse consequences. They will remain anonymous and all data will be identified by a case number only. All data collected from questionnaires will be stored in a password-secured cabinet.

Sample size The sample size calculation is based on power analysis. Assuming a two-tailed alpha of 0.05, a probability of 0.2 for beta error (80% power), and an effect size of 0.45 with reference to previous video-based mHealth programs with the same primary outcome measure (self-efficacy), 84 participants per group are required. With reference to the 10% to 15% attrition rate reported in previous programs, we assume a 15% drop-out rate in this study, thus the total sample size needed is 97 participants per group, i.e. a total of 194 participants. However, since this is a pilot study, at least 9% of the sample size of the main planned trial should be used. To ensure that the power is enough to estimate the group difference, this will be set at 35% of the sample size of the main planned trial. The total number of clients required is therefore 68. The flow of this study will follow the Modified CONSORT Statement extension for individual randomized controlled trials of non-pharmacological treatments.

Data processing and analysis The research assistant and the research team member will independently enter the data into the SPSS software (version 25.0, IBM). The inconsistencies between these two files will be resolved by retrieving the raw data from the questionnaires. Descriptive statistics will be used to detect potential outliers.

The participants' baseline characteristics will be compared using the chi-squared test or Fisher's exact test (in case the expected frequencies in any of the cells are lower than 5) for categorical variables and the two samples independent t-test for continuous variables. The p-value is set at less than 0.05 as a significant result for the two-tailed test.

The between-group (group), within-group (time), and the interaction effects (group x time) of outcome variables will be analyzed using the Generalized Estimating Equation (GEE). Linear link function will be used for all outcome measures except health service utilization while Poisson link function will be employed for health service utilization outcomes. Unstructured working correlation matrix will be adopted to indicate the same spacing between measurements for each subject. Adjusted GEE models will be employed to evaluate each outcome measures. All analyses will be conducted according to the intention-to-treat (ITT) method.