Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Class |
|---|---|
| ETH Zurich | OTHER |
| National University of Singapore | OTHER |
| Nanyang Technological University | OTHER |
Not provided
Not provided
Not provided
Non-communicable diseases (NCDs), such as cardiovascular disease, diabetes, or cancer, and common mental disorders (CMDs), such as depression or anxiety, represent the primary causes of death and disability worldwide, causing major health and financial burdens. Lifestyle behaviours, including physical activity, diet, stress and emotional regulation, tobacco smoking, alcohol consumption, and sleep are important modifiable risk factors associated with the prevention and management of both NCDs and CMDs.
LvL UP is a mHealth intervention aimed at preventing NCDs and CMDs in adults from multi-ethnic Southeast Asian populations (Castro et al., 2023). Building upon leading evidence- and theory-based frameworks in the areas of mental health and behaviour change, a multidisciplinary team of researchers developed LvL UP as a holistic intervention centred around three core pillars: Move More, Eat Well, Stress Less.
The goal of this pilot study is to assess the feasibility of a Sequential, Multiple Assignment, Randomized Trial (SMART) aimed at (i) evaluating the effectiveness and cost-effectiveness of LvL UP and (ii) establishing the optimal blended approach in LvL UP that balances effective personalized lifestyle support with scalability. The main questions it aims to answer are:
Participants will:
Participants will be randomly allocated to groups 1 or 2 (LvL UP or comparison) following a 2:1 ratio favoring the LvL UP group. At week 4 (decision point), participants from the LvL UP group will be classified as responders or non-responders based on pre-specified criteria. Non-responder participants will be re-randomized with equal probability (1:1) to one of the two second-line conditions: (i) continuing with the initial intervention (LvL UP) or (ii) additional MI support sessions (LvL UP + MI). Response / non-response categorization will involve assessing preliminary intervention effects (positive change in one of the following: physical activity, diet or mental well-being) as well as satisfaction with the intervention (net promoter score).
Participants will also take part in a process evaluation informed by the UK Medical Research Council's guidelines (Moore et al., 2015) to explore implementation (process, fidelity, dose, adaptations, reach), mechanisms of action (participant experience and response to intervention, mediators, unexpected pathways and consequences), and contextual factors that may affect implementation and intervention outcomes. Methods will entail qualitative and quantitative approaches, including surveys, interviews, web-based and app-based analytic data, and direct observation. In addition, other measures have been added to the research protocol to address the following exploratory aims: (i) explore time-varying and baseline moderators on intervention outcomes (e.g., sociodemographic variables, personality), (ii) identify the most cost-effective intervention condition from the societal perspective (e.g., self-reported sickness absence), and (iii) investigate which LvL UP app and smartphone sensor data (e.g., usage patterns, location) might be effective in predicting participant's engagement, state of receptivity to LvL UP notifications, or state of vulnerability (mental health conditions). All measures are described in the outcomes section.
To formalize the sample size calculations for the pilot study, the precision-based approach by Yan et al. (2020) was used to ensure the estimated outcomes for LvL UP and LvL UP + MI are controlled within a certain precision (i.e., the margin of error as a proportion of the outcome's standard deviation). Assuming a 50% responder rate, a two-tailed Type I error of 5% and a precision of 30%, a total of 97 participants will be required. Taking an 82% retention estimate for digital health interventions lasting ≤8 weeks (Jabir et al., 2023) and rounding up the number, 120 participants will be recruited. Therefore, 40 and 80 participants will be randomly assigned to the comparison condition (group 2) and LvL UP (group 1), respectively. In addition, an estimated 40 participants from group intervention will be re-randomized to continue with LvL UP or have LvL UP + MI. Considering participants assigned to group 1 are asked to nominate a LvL UP Buddy, which will be part of the process evaluation of the trial and thus technically considered participants, the total estimated sample size is 200 (120 trial participants and 60 LvL UP Buddies).
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| LvL UP | Experimental | Downloading and using the LvL UP app |
|
| LvL UP + MI | Experimental | Downloading and using the LvL UP app + taking part in motivational interviewing support sessions |
|
| Comparison | Active Comparator | Receiving healthy lifestyle and mental well-being resources from Singapore's Health Promotion Board |
|
| Name | Type | Description | Arm Group Labels | Other Names |
|---|---|---|---|---|
| LvL UP | Behavioral | The LvL UP app includes four lifestyle intervention components centred around three core pillars, Move More (physical activity), Eat Well (healthy nutrition), and Stress Less (mental well-being), as follows: (i) conversational agent-delivered health literacy coaching sessions, (ii) daily "Life Hacks" (healthy habit suggestions), (iii) therapeutic Tools including step-based activity tracker, food diary, and journal and (iv) gamified slow-paced breathing training (Breeze). These components are delivered using an innovative engagement approach that combines storytelling, MI, feedback on progress, just-in-time adaptive notifications and gamification. As part of the LvL UP app onboarding, participants are asked to nominate a 'LvL UP Buddy' (e.g., a friend, family member, or spouse) to provide additional support. Buddies are expected to complete different tasks, such as sending messages of encouragement or engaging in intervention-related activities together with the participant. |
| Measure | Description | Time Frame |
|---|---|---|
| Mental well-being (Warwick-Edinburgh Mental Well-being Scale) | Assessed using the Warwick-Edinburgh Mental Well-being Scale (WEMWBS-14). The total score ranges from 14 to 70, with higher values indicating higher mental well-being. | Baseline, week 4, week 8 |
| Measure | Description | Time Frame |
|---|---|---|
| Subjective well-being (World Health Organization Well-Being Index) | Assessed using the World Health Organization Well-Being Index (WHO-5). The total score ranges from 0 to 25, with higher values indicating higher well-being. | Baseline, week 4, week 8 |
| Mental health (depression - Patient Health Questionnaire-9) |
| Measure | Description | Time Frame |
|---|---|---|
| LvL UP app: Response time to app notifications | Timestamp when a message was sent by the app to the participant and the timestamp when an answer was provided by the participant | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Duration of app usage in minutes |
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Affiliation | Role |
|---|---|---|
| Falk Mueller-Riemenschneider, Professor | National University of Singapore | Principal Investigator |
| Tobias Kowatsch, Professor | ETH Zurich | Principal Investigator |
| Konstantina Griva, Professor | Nanyang Technological University | Principal Investigator |
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| Saw Swee Hock School of Public Health | Singapore | 117549 | Singapore | |||
| Singapore ETH Center |
| PubMed Identifier | Type | Citation | Retractions |
|---|---|---|---|
| 37234382 | Background | Castro O, Mair JL, Salamanca-Sanabria A, Alattas A, Keller R, Zheng S, Jabir A, Lin X, Frese BF, Lim CS, Santhanam P, van Dam RM, Car J, Lee J, Tai ES, Fleisch E, von Wangenheim F, Tudor Car L, Muller-Riemenschneider F, Kowatsch T. Development of "LvL UP 1.0": a smartphone-based, conversational agent-delivered holistic lifestyle intervention for the prevention of non-communicable diseases and common mental disorders. Front Digit Health. 2023 May 10;5:1039171. doi: 10.3389/fdgth.2023.1039171. eCollection 2023. | |
| 38412023 |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D015438 | Health Behavior |
| D000073296 | Noncommunicable Diseases |
| D040242 | Risk Reduction Behavior |
| D009043 | Motor Activity |
| D000092862 | Psychological Well-Being |
| ID | Term |
|---|---|
| D001519 | Behavior |
| D020969 | Disease Attributes |
| D010335 | Pathologic Processes |
| D013568 | Pathological Conditions, Signs and Symptoms |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
|
| LvL UP + MI | Behavioral | After 4 weeks, participants categorized as non-responders from the LvL UP group will be re-randomized into second-stage conditions: (i) continuing with the initial intervention (LvL UP) or (ii) additional human-delivered motivational interviewing (MI) support sessions (LvL UP + MI). The MI-informed sessions for non-responders will consist of four sessions delivered via WhatsApp, lasting between 30 and 40 minutes. The content of support will include various MI-based strategies such as use of ask-offer-ask framework and strategic use of communication skills (open-ended questions, reflections, affirmations and summaries) as per four MI processes:
|
|
| Comparison | Behavioral | Participants randomised to the comparison condition will receive a study pack including physical activity, diet and mental well-being content extracted from existing Health Promotion Board (HPB) resources. Established in 2001, the HPB (https://hpb.gov.sg/) is a government organisation under the Ministry of Health committed to promoting healthy living in Singapore. HPB's organises health promotion and disease prevention programmes covering various health domains. HPB-developed resources were selected as the comparator because they are the 'go-to', nation-wide health resources in Singapore which cover LvL UP's domains. |
|
Assessed using the Patient Health Questionnaire-9. The total score ranges from 0 to 27. Scores of 0-4, 5-9, 10-14, 15-19, 20-27 are the ranges for none, mild, moderate, moderately severe and severe depression, respectively. |
| Baseline, week 4, week 8 |
| Mental health (stress - Kessler Psychological Distress Scale) | Assessed using the Kessler Psychological Distress Scale (K6). The total score ranges from 10 to 50, with higher values indicating higher levels of psychological distress. | Baseline, week 4, week 8 |
| Health-related quality of life (European Quality of Life 5 Dimensions 5 Level Version) | Assessed using the European Quality of Life 5 Dimensions 5 Level Version (EQ-5D-5L). The EQ-5D-5L descriptive system of 5 health dimensions (Mobility, Self-care, Usual activities, Pain / discomfort, Anxiety / depression) includes 5 response categories of no problem, slight problems, moderate problems, severe problems, and extreme problems. Health states are scored to give the EQ-5D-5L index using a scoring algorithm from a value set derived from valuation tasks undertaken with general population samples. | Baseline, week 8 |
| Health behaviours (physical activity - International Physical Activity Questionnaire) | Assessed using the International Physical Activity Questionnaire (IPAQ-long). Results can be reported in categories (low activity levels, moderate activity levels or high activity levels) or as a continuous variable (MET minutes a week). | Baseline, week 4, week 8 |
| Health behaviours (sleep - Pittsburgh Sleep Quality Index) | Assessed using the Pittsburgh Sleep Quality Index (PSQI). The total score ranges from 0 to 27, with higher scores indicating more acute sleep disturbances. | Baseline, week 8 |
| Health behaviours (number of steps - smartphone-based) | Smartphone-based steps | Throughout the study period (i.e., assessed every day during 8 weeks) |
| Health behaviours (quantity and frequency of smoking and alcohol consumption - ad hoc questionnaire) | Assessed using a quantity-frequency survey on smoking and alcohol consumption adapted from previous epidemiological studies | Baseline, week 8 |
| Anthropometry (body weight in kilograms) | During study visits research staff will assess participants' body weight in kilograms. Weight and height will be combined to report Body Mass Index (BMI) in kg/m2. | Baseline, week 8 |
| Anthropometry (body height in meters) | During study visits research staff will assess participants' body height in meters. Weight and height will be combined to report Body Mass Index (BMI) in kg/m2. | Baseline, week 8 |
| Anthropometry (waist circumference in meters) | During study visits research staff will assess participants' waist circumference in meters. | Baseline, week 8 |
| Anthropometry (hip circumference in meters) | During study visits research staff will assess participants' hip circumference in meters. | Baseline, week 8 |
| Anthropometry (self-reported height in kilograms and weight in meters) | Self-reported height (kilograms) and weight (meters) via online survey. Weight and height will be combined to report Body Mass Index (BMI) in kg/m2. | week 4 |
| Resting blood pressure in mmHg | During study visits research staff will assess diastolic and systolic blood pressure (mmHg) using an automated blood pressure monitor (Dinamap - Carescape V100, GE Pacific). | Baseline, week 8 |
| Health behaviours (diet quality 1 - Singapore diet screener) | Assessed using a 37-item diet screener developed and validated in Singapore | Baseline, week 4, week 8 |
| Health behaviours (diet quality 2 - ad hoc Food Frequency Questionnaire) | Modified 7-item food frequency questionnaire (ad hoc) assessing diet quality | Baseline, week 4, week 8 |
| Blood metabolic profile (mmol/L concentration of TC, HDL, TG and LDL-calculated) | During study visits research staff will collect a non-fasting venous blood sample to assess the following blood markers: TC, HDL, TG and LDL-calculated. All measured as millimoles per liter (mmol/L). | Baseline, week 8 |
Time measure of how long an individual had the app open |
| Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Actual usage of app components (number of times used and duration) | Timestamps combined with any interactions performed by the participant with the intervention components (coaching sessions, life hacks, Breeze, journaling) | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Coaching topic / sub-topic use | Which coaching topics were used by the participant | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Conversational turns | The number of responses of a participant divided by the number of conversational turns offered by the app's chatbot | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Breeze (slow-paced breathing tool) usage | Audio-recordings of short voice commands and breathing when using "Breeze" (slow-paced breathing tool) | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (location - obfuscated) | Sensor: GPS. Data type: Float. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (location types) | Sensor: Apple SensorKit. Data type: Categorical (work, home, gym). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (raw gyroscope) | Sensor: Gyroscope. Data type: Float. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (raw accelerometer) | Sensor: Accelerometer. Data type: Float. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (physical activity) | Sensor: Accelerometer. Data type: Categorical (sitting, standing, walking, in a vehicle, running or bicycling) . | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (WiFi connection status) | Sensor: WiFi. Data type: Categorical (connected/disconnected) and String (name of network). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (WiFi networks in range) | Sensor: WiFi. Data type: String (name and device IDs of near-by WiFi Access Points). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (bluetooth scan) | Sensor: Bluetooth. Data type: String (name and device IDs of near-by devices). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (battery % level and charging status) | Sensor: Battery Status. Data type: Float representing battery level and Categorical (charging/discharging/none). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (screen on/off) | Sensor: Screen Events. Data type: Binary (on/off). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (call and SMS metadata) | Sensor: Call/SMS logs. Data type: Strings (containing hash of number), time, duration of call / length of text. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (app usage & app name from notifications and action) | Sensor: App Usage. Data type: String/Categorical (social, entertainment, communication, etc). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (step count) | Sensor: Pedometer. Data type: Integer. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (brightness) | Sensor: Screen Brightness. Data type: Integer. | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to smartphone sensors (breeze audio) | Sensor: Audio Mic. Data type: Audio file during Breeze usage (slow-paced breathing tool). | Continuously throughout the study period (i.e., assessed every day during 8 weeks) |
| LvL UP app: Access to Apple Healthkit & Google Fit data | Including the following data types: activity, hearing, vital signs, nutrition, mobility, mindfulness and sleep, workouts | Continuously throughout the study period (i.e., assessed every day during 8 weeks) and up to 1 month before the start of the study |
| Process evaluation: Health literacy questionnaire | Knowledge of physical activity and dietary guidelines (ad hoc) | Baseline, week 8 |
| Process evaluation: Stages of change questionnaire | Stages of change questionnaire for physical activity, eating, and mental wellness habits | Baseline, week 8 |
| Process evaluation: Unintended consequences of participating in the trial (ad hoc) | Unintended consequences of participating in the trial (open-ended text box) | Week 8 |
| Process evaluation (intervention group participants only): Net promoter score | 1-item rating via in-app pop-up: "How likely is it that you would recommend LvL UP to a friend?" rating between 0 (not at all likely) and 10 (extremely likely). | Week 4 |
| Process evaluation (intervention group participants only): System usability scale for LvL UP app | Scale giving a global view of subjective assessments towards the LvL UP app. The total System usability scale score is calculated as a percentage, ranging from 0 to 100. A perfect score of 100% signifies flawless usability and an exceptional user experience. | Week 4 |
| Process evaluation (intervention group participants only): Session alliance inventory for LvL UP app and motivational interviewing coaching sessions | Measures the working alliance between participant and digital (LvL UP app) / MI coach. The total score ranges from 12 to 84, with higher values indicating a more positive rating of working alliance. | Week 4, week 8 |
| Process evaluation (human coaches): Behaviour Change Counselling Index | Behavioral coding system designed to measure treatment fidelity for motivational interviewing | Week 8 |
| Process evaluation (intervention group participants only): User Version of the Mobile Application Rating Scale (uMARS) for LvL UP app | User's assessment of quality of mHealth apps, including the following domains: Engagement, Functionality, Aesthetics, Information, App quality, App subjective quality. All items are assessed on a 5-point scale (1-inadequate, 2-poor, 3-acceptable, 4-good, and 5-excellent). | Week 8 |
| Process evaluation (intervention group participants only): Interview (ad hoc) | Process evaluation interview for a subset of participants exploring mechanisms of action, context, and implementation | Week 8 |
| Process evaluation (Research team): Field Notes | To record any notable information that may impact the trial conduct or outcomes | Continuously throughout the study period (i.e., every day during 8 weeks) |
| Exploratory aims: Sociodemographic (ad hoc) | Age, sex, ethnicity, marital status, current employment status, occupation, completed years of education, accommodation type, and household income | Baseline |
| Exploratory aims: Use of digital health technologies (ad hoc) | Assessed using the a questionnaire on past/current use of wearables and participation in digital health programes (ad hoc) | Baseline |
| Exploratory aims: Ecologic Momentary Assessment of mood state (ad hoc) | Assessed using a graphical question based on the circumplex model of affect. | Every two days throughout the study period (i.e., during 8 weeks) |
| Exploratory aims: Personality (Big 5 Personality Questionnaire) | Assessed using the Big 5 Personality Questionnaire - short. The questionnaire measures five traits (extraversion, agreeableness, openness, conscientiousness, and neuroticism), and consists of 20 items structured as simple sentences rated in the Likert scale of 5 points, ranging from 1 (totally disagree) to 5 (totally agree). | Baseline |
| Exploratory aims: Work Engagement (Work Productivity and Activity Impairment Questionnaire) | Assessed using the Work Productivity and Activity Impairment Questionnaire (WPAI), a five-item scale. For the three first item the answer is number of hours (e.g., During the last 3 months, how many hours did you miss from work because of problems associated with your health?). For the last two items are scored on a scale ranging from 1 to 10, with 1 being 'my health did not affect my work productivity and ability to do regular daily activities at all', and 10 being 'my health affected my work productivity and ability to do regular daily activities completely'. | Baseline, week 8 |
| Exploratory aims: Healthcare utilization (ad hoc) | Healthcare utilization in the past three months | Baseline, week 8 |
| Singapore |
| 138602 |
| Singapore |
| Background |
| Jabir AI, Lin X, Martinengo L, Sharp G, Theng YL, Tudor Car L. Attrition in Conversational Agent-Delivered Mental Health Interventions: Systematic Review and Meta-Analysis. J Med Internet Res. 2024 Feb 27;26:e48168. doi: 10.2196/48168. |
| 25791983 | Background | Moore GF, Audrey S, Barker M, Bond L, Bonell C, Hardeman W, Moore L, O'Cathain A, Tinati T, Wight D, Baird J. Process evaluation of complex interventions: Medical Research Council guidance. BMJ. 2015 Mar 19;350:h1258. doi: 10.1136/bmj.h1258. |
| 32529788 | Background | Yan X, Ghosh P, Chakraborty B. Sample size calculation based on precision for pilot sequential multiple assignment randomized trial (SMART). Biom J. 2021 Feb;63(2):247-271. doi: 10.1002/bimj.201900364. Epub 2020 Jun 11. |
| 41485014 | Derived | Zheng S, Castro O, Mair JL, Jabir AI, Tan SYX, Shenoi A, Negi S, Mathews RR, Koh RWS, Yan X, Chakraborty B, Tai ES, van Dam RM, von Wangenheim F, Fleisch E, Griva K, Kowatsch T, Muller-Riemenschneider F. Feasibility of the "LvL UP" trial: a pilot sequential multiple assignment randomised trial of an adaptive, holistic mHealth lifestyle coaching intervention. Int J Behav Nutr Phys Act. 2026 Jan 3;23(1):16. doi: 10.1186/s12966-025-01869-7. |
| 39900289 | Derived | Castro O, Mair JL, Zheng S, Tan SYX, Jabir AI, Yan X, Chakraborty B, Tai ES, van Dam RM, von Wangenheim F, Fleisch E, Griva K, Kowatsch T, Muller-Riemenschneider F. The LvL UP trial: Protocol for a sequential, multiple assignment, randomised controlled trial to assess the effectiveness of a blended mobile lifestyle intervention. Contemp Clin Trials. 2025 Mar;150:107833. doi: 10.1016/j.cct.2025.107833. Epub 2025 Feb 1. |
| D010549 | Personal Satisfaction |