Health@NUS is a prospective cohort study conducted by researchers at the National University of Singapore (NUS) in partnership with the Singapore Government’s Health Promotion Board (HPB). Reporting of this study protocol follows the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement [37].

Health@NUS follows participants longitudinally over a 2-year period during university and transition to early working life. The study uses mHealth data collection methods, including a Fitbit smartwatch and the Health Insights SG (hiSG) mobile app developed by HPB, which collect data continuously over the 2-year period. Objective biometric measurements and questionnaires are collected at the 3 time points (baseline, 12-, and 24-month follow-up visits) with additional app-based ecological momentary surveys pushed out between the time points. Upon enrollment, participants also indicate whether they consent to be contacted for further follow-up assessments beyond the 2-year study period.

In addition to providing informed consent to join the study, participants were required to pay an SGD 20 (US $14.50) deposit to receive the Fitbit smartwatch (Versa Lite), which is refundable upon completion of the study. Participants were also required to setup a Direct Debit Authorization prior to approval to make a deduction from their bank account with HPB for SGD 248 (US $183) to recover the cost of the smartwatch should they lose it or withdraw from the study partway through.

Participants were recruited between August 2020 and June 2022 via emails, internal student communication systems, and word-of-mouth. Data collection was expected to be completed by June 2024.

To be eligible, participants had to be (1) a full-time undergraduate or graduate university student, (2) a resident of Singapore, (3) aged between 18 and 26 years, and (4) own a smartphone with a Singapore registered mobile number and an iOS or Android operating system compatible with both the Fitbit app and hiSG mobile app. Students were excluded from participating if they (1) were pregnant at the point of recruitment, (2) traveled overseas for more than 6 weeks at a time, (3) were not willing to make an upfront deposit of SGD 20 (US $14), or (4) were not willing to set up a Direct Debit Authorization with HPB.

Initially, between October 2020 and April 2021, only students who were between 1 and 3 semesters away from graduating were recruited. However, recruitment challenges were faced due to Singapore’s strict safety management measures during the COVID-19 pandemic; thus, this selection criterion was removed after April 2021. Students completed their digital registration, which included initial eligibility screening. Participants who were screened as eligible were invited to visit the study site to provide written informed consent and baseline biometric measurements (Figure 1). All study visits at baseline and 12- and 24-month follow-ups are conducted at the health screening site at NUS.

Biometric measurements, including height, weight, blood pressure, waist circumference, and resting heart rate, were conducted by trained staff using standardized protocols [38]. In addition, participants were issued a Fitbit smartwatch at baseline and were required to self-administer a web-based questionnaire on diet, physical activity, and other lifestyle factors and their underlying determinants at baseline, and the 12- and 24-month follow-up visits (Figure 2). Full details of the baseline questionnaire are presented in Table S1 in Multimedia Appendix 1 [38-79].

Repeated 2-week bursts of ecological momentary assessment (EMA) surveys were sent between the 3 main data collection time points (baseline, 12 months, and 24 months) via the hiSG app. Adapted from another ongoing study in Singapore [80], the EMA surveys contain questions relating to physical activity, diet, sleep, well-being, affect and physical feeling, behavioral cognitions (intentions and self-efficacy), cigarette smoking, alcohol consumption, and screen use (Table S2 and Figure S1 in Multimedia Appendix 1). Participants received an electronic voucher for completing the baseline assessments (SGD 5, US $3.50), each follow-up visit (worth SGD 50, US $36), and each of the EMA surveys was rewarded through Healthpoints (SGD 7, US $5).

During the 2-year period, participants were asked to fulfill minimum data logging requirements relating to (1) wear time (awake and sleep) and (2) food logging (Table 1). Participants receive Healthpoints (150 Healthpoints=SGD 1, US $0.7) for completing the minimum study requirements across the 2-year study duration. The rewards would commensurate the level and duration of participation in the study and participants will receive milestone rewards of approximately SGD 80 (US $58) worth of Healthpoints (Figure S2A in Multimedia Appendix 1). Accumulated Healthpoints can be exchanged for electronic vouchers within the hiSG mobile app. Monthly reminders are sent via SMS text messages and emails to participants who do not fulfill the minimum logging requirements. Additionally, study participants are allowed to keep their Fitbit smartwatch at the end of the 2-year study period.

This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the National Health Care Group Domain Specific Review Board (Ref: 2019/00285, Singapore) and registered at ClinicalTrials.gov (NCT05154227). Written informed consent to participate in the study was obtained from all participants prior to data collection. Participants were offered SGD 120 (US $87) worth of Healthpoints for fulfilling the minimum data logging requirements, and up to SGD 425 (US $312) worth of Healthpoints were offered for additional wear time, sleep time, food logging, and survey data provided.

The study initially aimed to recruit 1300 students to account for attrition over the 2-year follow-up period and to achieve a target sample size of 1000. However, due to the challenges associated with recruitment during the COVID-19 pandemic, the original recruitment target was not met and 776 participants were recruited.

The sample size of 776 allows the detection of a small exposure effect (f²=0.02) on a continuous outcome when performing a multiple linear regression that accounts for 9 confounders at a significance level of 5% with more than 80% power [85,86]. The sample size of 776 also allows the detection of an odds ratio of 1.85 when performing a multiple logistic regression where 25% of the binary exposure’s variability is explained by the other predictors in the model, the probability of an outcome is between 0.2 and 0.7 among the unexposed, and the prevalence of being exposed is between 0.3 and 0.7 at a significance level of 5% with power at 80% or more [87]. The sample size (N=776) is larger than the number of participants recruited in previous EMA studies on youth [88-90].

Demographic data were analyzed descriptively. As the health and lifestyle behavior of university students are influenced by gender [91-93], comparisons of means between male and female participants were made via 2-tailed t test, and comparisons of medians between males and female participants were made via the Mann-Whitney U test. For categorical variables, differences between males and female participants were assessed by the chi-square test or the Fisher exact test for variables with small cell counts.

Interested participants (n=1556) were screened for eligibility prior to joining the study. We excluded 780 participants due to incomplete registration forms (n=460), ineligibility (n=240), and refusal to participate (n=80), as shown in Figure 2. A total of 776 participants were enrolled in the study.

Of the 776 participants, 4 participants had incomplete baseline questionnaire data. The mean age of participants is 22.70 (SD 1.7) years. Participants are mostly female (441/776, 56.8%), Chinese ethnicity (714/776, 92%), undergraduate students (759/776, 97.8%), and not married (773/776, 99.6%). Half of the students (403/776, 51.9%) were in year 4 or year 5 of their undergraduate degree and were expected to transition into their working life during the study period. More details on participant demographics are presented in Table 2.

Table 3 shows anthropometric and lifestyle characteristics stratified by sex. Participants had a mean BMI of 21.9 (SD 3.3) kg/m². Based on Asian classifications for BMI [84], 26.1% (202/773) were classified as being overweight (BMI ≥23 and <27.5 kg/m²) and 5.43% (42/773) were classified as being obese (BMI≥27.5 kg/m²). Participants had a mean systolic blood pressure of 113.7 (SD 11.42) mm Hg and a mean diastolic blood pressure of 64.76 (SD 7.94) mm Hg.

Most participants were nonsmokers (746/772, 96.6%), and more male participants (16/334, 4%) reported as being current or ex-smoker as compared to female participants (4/438, 1%). For dietary intake, the median intake of standard servings per day is reported in Table 3. Participants consumed 2.86 (IQR 1.71-4.64) servings of fruits and vegetables per day. Furthermore, participants reported consuming 1.53 (IQR 0.89-2.75) servings of refined grains as compared to 0.39 (IQR 0.08-1.00) servings of whole grains per day. For seafood, red meat, and poultry, participants consumed 0.71 (IQR 0.36-1.00) servings of poultry, 0.44 (IQR 0.23-0.86) servings of red meat, and 0.37 (IQR 0.20-0.64) servings of seafood per day. Only a small percentage of participants reported binge drinking (54/759, 7%).

For movement behaviors, participants engaged in an average of 0.78 (SD 0.76) hours per day in moderate to vigorous physical activity (MVPA), 10.87 (SD 3.60) hours of sedentary time, 13.39 (SD 4.22) hours of screen time, and 6.84 (SD 1.08) hours of sleep. For mental well-being, 37.7% (291/772) of the participants had poor mental well-being (WHO-5 score ≤50), and 14.1% (109/772) of the participants were at higher risk for psychological distress (K6≥13).

Significant differences between male and female participants (P<.05) were observed for all anthropometric measures (eg, height, weight, BMI, blood pressure, waist circumference, and average resting heart rate). Female participants were more likely to be underweight and had a higher resting heart rate, while male participants were more likely to be overweight. Significant differences between males and female participants were observed only in a few lifestyle behaviors such as consumption of refined grains and red meat, time spent in MVPA, and mental well-being and psychological distress (Table 3). Male participants consumed more refined grains and red meat and spent more time engaging in MVPA than female participants. A larger proportion of females reported poor mental well-being and psychological distress, as compared to male participants.

Health@NUS is the first longitudinal study of its kind to explore the interrelationships and associations among health behaviors, health, and well-being of young adults as they transition from university and into their working lives. In addition to traditional assessments used in cohort studies (self-report surveys and physical examination), Health@NUS uses mHealth technologies for continuous and repeated assessments of health behaviors and related contextual information at a level of granularity that has not previously been possible [36]. These data will provide important insights into the determinants and trajectories of health behaviors, health, well-being and related contextual information, and the dynamic and synergistic interrelationships between them.

With an increasing number of young adults graduating from universities in Singapore [94] and across the world, more young adults are subjected to this transitionary period between university and working life [6]. Understanding young adults’ health behaviors, health, and well-being during this vulnerable period in their lives, how they change, and what influences these changes has important implications for the promotion of population health. The early results from Health@NUS baseline assessments provide insights into indicators of physical health (eg, BMI, waist circumference, and blood pressure) among university students in Singapore. For instance, participants of Health@NUS reported an average BMI of 21.91 kg/m², and the prevalence of obesity (5.43%) was substantially lower than among all Singapore residents aged 18 to 29 years (13.1%) [4]. This may be in part due to the low age of the study population, even though most weight gain occurs between the ages of 18 and 29 years in Singapore [95]. The proportion of underweight female participants (n=91, 17.3%) was greater than Singapore residents (10%) [4], which is an underrecognized public health problem prevalent among young female participants [96]. Further research is warranted to investigate this issue. In addition, the average waist circumference for both male (77.46 cm) and female (69.78 cm) participants was lower than the cutoffs for abdominal obesity in Asian populations [4,97]. Moreover, the prevalence of smoking (11/772, 1.4%) and binge drinking (54/759, 7.1%) among participants of Health@NUS was very low compared to Singapore residents aged 18 to 29 years who reported a prevalence of smoking and binge drinking of 8.3% and 15.6%, respectively [98]. The prevalence of smoking and binge drinking in this study population was also low in comparison to other studies among university students in Singapore and globally [99-102].

In contrast, baseline data on movement behaviors and diet present a less favorable outlook of the student population. While physical activity levels and sleep duration appear low, time spent being sedentary and using screen devices is high. This is in contrast to the physical activity levels observed among Singapore residents aged 18 to 29 years, who were reported to be most likely to engage in regular exercise [4,98]. Compared to Singapore residents aged 18 to 29 years (76.3%), our findings suggest that fewer students (n=555, 72.7%) met the minimum recommendations of at least 150 minutes per week of MVPA as defined by the World Health Organization guidelines [4,103,104]. This was particularly obvious among female students, for whom the proportion meeting recommendations was markedly lower than among their male counterparts. Furthermore, time spent in sedentary behavior among Health@NUS participants was greater than in other studies in Singapore, which reported an average of 5 to 6 hours per day [105,106]. Low levels of physical activity and large amounts of time spent in sedentary behavior by Health@NUS participants may in part be attributable to the fact that university students spent the longest time per week globally studying [107]. With regard to the consumption of whole grains, fruits, and vegetables, the situation also appears suboptimal compared to the “My Healthy Plate” dietary recommendations from the HPB for Singapore residents [108]. On average, students did not meet the recommended daily 5 to 7 servings of whole grains, and daily 4 servings of fruits and vegetables [109]. The daily consumption of sugar-sweetened beverages on the other hand was exceptionally low as compared to Singapore residents [110,111], where sugar-sweetened beverages accounted for 52% of dietary sugar intake [111]. These findings warrant further investigation based on Health@NUS’s extensive diet logging and EMA data to better understand dietary patterns of university students in Singapore.

The high prevalence of poor mental well-being (291/772, 37.7%) and psychological distress (109/772, 14.1%) observed in this study population of young adults is of great concern. Our findings from the WHO-5 scores suggest that university students in Singapore have poorer mental well-being as compared to other countries such as Germany [112], New Zealand [113], Hong Kong [114], the United Kingdom [115], and the United States [116]. According to national studies in Singapore, poor mental well-being is most prevalent in young adults aged 18 to 29 years residing in Singapore [4,8], and the proportion of young adults with poor mental well-being has been gradually rising in recent years [4,8,98]. In comparison to the Singapore Mental Health Study 2016, Singapore residents aged between 18 and 34 years reported the highest prevalence of lifetime- and 12-month mental disorders at 21.6% and 11.9%, respectively [8]. Furthermore, the prevalence of poor mental well-being in Singapore prior to and during the COVID-19 pandemic was found to be 8.4% and 8.7%, which is lower than reported in Health@NUS [117,118]. As most mental illnesses (eg, anxiety and depression) [6] emerge during adolescence or early adulthood [9,119,120], the delay in help-seeking by affected individuals could have long-lasting emotional and social impacts on their adulthood [119,121]. Greater levels of stress experienced by students in university increase their vulnerability toward mental health conditions (eg, depression), substance abuse (eg, alcohol abuse), and chronic health conditions (eg, obesity) [6] in later adulthood without proper coping mechanisms. As modifiable lifestyle behaviors (eg, physical activity, diet, and sleep) are associated with positive mental well-being [122], it is important for Health@NUS to explore the underlying determinants of students’ health behaviors during the transition period from university to working life. To this end, data collected from Health@NUS will be used to inform the design of future interventions to promote healthy behaviors and the health and well-being of students during these formative years and beyond. Given recent attention on the poor mental health and well-being of young people in Singapore [4,5,7,8,119,121], 1 of 4 priorities will also be to use the detailed data from Health@NUS to understand trajectories of well-being and psychological distress, as well as their related contextual factors over time, and to identify potential intervention strategies to promote well-being among the student population in Singapore.

Health@NUS will provide unprecedented information on health behaviors, health, and well-being, and related contextual information. We have recruited a large sample and followed these students for 2 years, over their time at university and as they transition into their postgraduation working life. Combining assessments used in traditional cohort studies (self-reported questionnaires and biometric assessments), as well as repeated and continuous data collection via smartwatches and smartphones will provide detailed longitudinal data. Importantly, much of the data collected required minimal inputs from participants (ie, data from the smartwatch and smartphone), and we captured these data in real time and as participants were going about their daily lives, which enhances the ecological validity of our findings. A potential limitation of presented results relates to the self-reported lifestyle behaviors (eg, dietary intake, cigarette smoking, alcohol consumption, and physical activity). Future analysis of data collected via smartwatches, food logging, and EMA will help to overcome this limitation to some extent. Additionally, there could be a possibility of selection bias as health-conscious and motivated students were more likely to join the study. Disruptions stemming from the COVID-19 pandemic might have affected the job opportunities of university students seeking employment, thus intensifying their stress levels during the challenging transition from study to work. This will have implications on the generalizability of our findings to the wider university student population in Singapore and emphasizes the need for further research in the post–COVID-19 era.

To our knowledge, Health@NUS is a novel prospective cohort study that augments information collected from traditional questionnaires and on-site health screening, with mHealth technologies to capture intensive longitudinal data on individual characteristics, lifestyle behaviors, health, and well-being among young adults during their transition from university to working life in real life [34,35]. Health behaviors shaped in early adulthood often persist into adulthood [11,23] and can have negative implications for health and well-being. A deeper understanding of how health behaviors cluster or interact and of the influence of individual and ecological determinants is essential to developing effective health promotion strategies for young adults in Singapore. Furthermore, the findings from this study will provide insights into the mental health of students during the potentially stressful transition from university to work life.