The number of repeated observations needed to estimate the habitual physical activity of an individual to a given level of precision

Fitness-related physical activity intensity explains most of the association between accelerometer data and cardiometabolic health in persons 50-64 years old

OBJECTIVES

To investigate the physical activity (PA) intensity associated with cardiometabolic health when considering the mediating role of cardiorespiratory fitness (CRF).

METHODS

A subsample of males and females aged 50-64 years from the cross-sectional Swedish CArdioPulmonary bioImage Study was investigated. PA was measured by accelerometry and CRF by a submaximal cycle test. Cardiometabolic risk factors, including waist circumference, systolic blood pressure, high-density lipoprotein, triglycerides and glycated haemoglobin, were combined to a composite score. A mediation model by partial least squares structural equation modelling was used to analyse the role of CRF in the association between PA and the composite score.

RESULTS

The cohort included 4185 persons (51.9% female) with a mean age of 57.2 years. CRF mediated 82% of the association between PA and the composite score. The analysis of PA patterns revealed that moderate intensity PA explained most of the variation in the composite score, while vigorous intensity PA explained most of the variation in CRF. When including both PA and CRF as predictors of the composite score, the importance of vigorous intensity increased.

CONCLUSION

The highly interconnected role of CRF in the association between PA and cardiometabolic health suggests limited direct effects of PA on cardiometabolic health beyond its impact on CRF. The findings highlight the importance of sufficient PA intensity for the association with CRF, which in turn is linked to better cardiometabolic health.

The interrelationship between physical activity intensity, cardiorespiratory fitness, and executive function in middle-aged adults: An observational study of office workers

Background

Previous evidence supports a beneficial effect of physical activity on executive function across the whole lifespan. Yet, the interrelationships of the intensities of physical activity, cardiorespiratory fitness, and executive function require further investigation in adults.

Aim

Using unfiltered accelerometry data and high-resolution intensity classification, we sought to estimate the associations of physical activity with cardiorespiratory fitness and executive function in adult office workers.

Methods

We included 343 full-time office workers (mean age: 42.41 years, range of age: 36-49 years). Executive function was assessed using Stroop, Trail making tests (part-B), and 2-back tests, and a composite score was produced to reflect the general executive function performance. Physical activity was assessed using the Actigraph GT3X+-monitor, worn by each participant for seven days at the hip. Raw accelerometry data were processed by the 10 Hz frequency extended method and divided into 22 intensity bins and sleep time. Cardiorespiratory fitness was estimated using the submaximal Ekblom-Bak cycle ergometer test. Data were analyzed using partial least squares regressions.

Results

In adults, cardiorespiratory fitness was closely correlated with a wide range of absolute physical activity intensity patterns. A higher level of executive function in adults was associated with both higher absolute physical activity intensities and cardiorespiratory fitness, which was independent of age, sex, and education levels. A very weak association between intensities, fitness, and executive function was observed in high-fit adults. Among low-fit adults, although a positive association started already toward the upper end of moderate intensity, there still appeared to be an association between intensities, cardiorespiratory fitness, and executive function. That is, cardiorespiratory fitness may mediate the association between absolute physical activity intensities and executive function up to a certain level.

Conclusion

The maintenance of executive function in adulthood was related to both physical activity intensities and cardiorespiratory fitness, while their interrelationship was not equal across fitness levels. It is highly recommended to consider the cardiorespiratory fitness level in future studies that focus on executive functions in aging as well when designing individualized physical activity training programs.

Objectively measuring the association between the built environment and physical activity: a systematic review and reporting framework

Background

Objective measures of built environment and physical activity provide the opportunity to directly compare their relationship across different populations and spatial contexts. This systematic review synthesises the current body of knowledge and knowledge gaps around the impact of objectively measured built environment metrics on physical activity levels in adults (≥ 18 years). Additionally, this review aims to address the need for improved quality of methodological reporting to evaluate studies and improve inter-study comparability though the creation of a reporting framework.

Methods

A systematic search of the literature was conducted following the PRISMA guidelines. After abstract and full-text screening, 94 studies were included in the final review. Results were synthesised using an association matrix to show overall association between built environment and physical activity variables. Finally, the new PERFORM (’Physical and Environmental Reporting Framework for Objectively Recorded Measures’) checklist was created and applied to the included studies rating them on their reporting quality across four key areas: study design and characteristics, built environment exposures, physical activity metrics, and the association between built environment and physical activity.

Results

Studies came from 21 countries and ranged from two days to six years in duration. Accelerometers and using geographic information system (GIS) to define the spatial extent of exposure around a pre-defined geocoded location were the most popular tools to capture physical activity and built environment respectively. Ethnicity and socio-economic status of participants were generally poorly reported. Moderate-to-vigorous physical activity (MVPA) was the most common metric of physical activity used followed by walking. Commonly investigated elements of the built environment included walkability, access to parks and green space. Areas where there was a strong body of evidence for a positive or negative association between the built environment and physical activity were identified. The new PERFORM checklist was devised and poorly reported areas identified, included poor reporting of built environment data sources and poor justification of method choice.

Conclusions

This systematic review highlights key gaps in studies objectively measuring the built environment and physical activity both in terms of the breadth and quality of reporting. Broadening the variety measures of the built environment and physical activity across different demographic groups and spatial areas will grow the body and quality of evidence around built environment effect on activity behaviour. Whilst following the PERFORM reporting guidance will ensure the high quality, reproducibility, and comparability of future research.

Supporting Behavior Change in Sedentary Adults via Real-time Multidimensional Physical Activity Feedback: Mixed Methods Randomized Controlled Trial

BACKGROUND

Increasing physical activity (PA) behavior remains a public health priority, and wearable technology is increasingly being used to support behavior change efforts. Using wearables to capture and provide comprehensive, visually persuasive, multidimensional feedback with real-time support may be a promising way of increasing PA in inactive individuals.

OBJECTIVE

This study aims to explore whether a 6-week self-monitoring intervention using composite web-based multidimensional PA feedback with real-time daily feedback supports increased PA in adults.

METHODS

A 6-week, mixed methods, 2-armed exploratory randomized controlled trial with 6-week follow-up was used, whereby low to moderately active (PA level [PAL] <2.0) adults (mean age 51.3 years, SD 8.4 years; women 28/51, 55%) were randomly assigned to receive the self-monitoring intervention (36/51, 71%) or waiting list control (15/51, 29%). Assessment of PA across multiple health-harnessing PA dimensions (eg, PAL, weekly moderate to vigorous intensity PA, sedentary time, and steps), psychosocial cognitions (eg, behavioral regulation, barrier self-efficacy, and habit strength), and health were made at the prerandomization baseline at 6 and 12 weeks. An exploratory analysis of the mean difference and CIs was conducted using the analysis of covariance model. After the 12-week assessment, intervention participants were interviewed to explore their views on the program.

RESULTS

There were no notable differences in any PA outcome immediately after the intervention; however, at 12 weeks, moderate-to-large effects were observed with a mean difference in PAL of 0.09 (95% CI 0.02-0.15; effect size [Hedges g] 0.8), daily moderate-intensity PA of 24 (95% CI 0-45; Hedges g=0.6) minutes, weekly moderate-to-vigorous intensity PA of 195 (95% CI 58-331; Hedges g=0.8) minutes, and steps of 1545 (95% CI 581-2553; Hedges g=0.7). Descriptive analyses suggested that the differences in PA at 12 weeks were more pronounced in women and participants with lower baseline PA levels. Immediately after the intervention, there were favorable differences in autonomous motivation, controlled motivation, perceived competence for PA, and barrier self-efficacy, with the latter sustained at follow-up. Qualitative data implied that the intervention was highly informative for participants and that the real-time feedback element was particularly useful in providing tangible, day-to-day behavioral support.

CONCLUSIONS

Using wearable trackers to capture and present sophisticated multidimensional PA feedback combined with discrete real-time support may be a useful way of facilitating changes in behavior. Further investigation into the ways of optimizing the use of wearables in inactive participants and testing the efficacy of this approach via a robust study design is warranted.

TRIAL REGISTRATION

ClinicalTrials.gov NCT02432924; https://clinicaltrials.gov/ct2/show/NCT02432924.

Characterisation of Temporal Patterns in Step Count Behaviour from Smartphone App Data: An Unsupervised Machine Learning Approach

The increasing ubiquity of smartphone data, with greater spatial and temporal coverage than achieved by traditional study designs, have the potential to provide insight into habitual physical activity patterns. This study implements and evaluates the utility of both K-means clustering and agglomerative hierarchical clustering methods in identifying weekly and yearlong physical activity behaviour trends. Characterising the demographics and choice of activity type within the identified clusters of behaviour. Across all seven clusters of seasonal activity behaviour identified, daylight saving was shown to play a key role in influencing behaviour, with increased activity in summer months. Investigation into weekly behaviours identified six clusters with varied roles, of weekday versus weekend, on the likelihood of meeting physical activity guidelines. Preferred type of physical activity likewise varied between clusters, with gender and age strongly associated with cluster membership. Key relationships are identified between weekly clusters and seasonal activity behaviour clusters, demonstrating how short-term behaviours contribute to longer-term activity patterns. Utilising unsupervised machine learning, this study demonstrates how the volume and richness of secondary app data can allow us to move away from aggregate measures of physical activity to better understand temporal variations in habitual physical activity behaviour.

Test-retest reliability of physical activity questionnaires in Parkinson's disease

BACKGROUND

People with Parkinson's disease are less physically active than controls. It is important to promote physical activity, which can be assessed using different methods. Subjective measures include physical activity questionnaires, which are easy and cheap to administer in clinical practice. Knowledge of the psychometric properties of physical activity questionnaires for people with Parkinson's disease is limited. The aim of this study was to evaluate the test-retest reliability of physical activity questionnaires in individuals with Parkinson's disease without cognitive impairment.

METHODS

Forty-nine individuals with Parkinson's disease without cognitive impairment participated in a test-retest reliability study. At two outpatient visits 8 days apart, the participants completed comprehensive questionnaires and single-item questions: International Physical Activity Questionnaire-Short Form (IPAQ-SF), Physical Activity Scale for the Elderly (PASE), Saltin-Grimby Physical Activity Level Scale (SGPALS) and Health on Equal Terms (HOET). Test-retest reliability was evaluated using the intraclass correlation coefficient (ICC), standard error of measurement (SEM), limits of agreement, weighted kappa or the Svensson method.

RESULTS

Several of the physical activity questionnaires had relatively low test-retest reliability, including the comprehensive questionnaires (IPAQ-SF and PASE). Total physical activity according to IPAQ-SF had an ICC value of 0.46 (95% confidence interval [CI], 0.21-0.66) and SEM was 2891 MET-min/week. The PASE total score had an ICC value of 0.66 (95% CI, 0.46-0.79), whereas the SEM was 30 points. The single-item scales of SGPALS-past six months (SGPALS-6 m) and HOET question 1 (HOET-q1) with longer time frames (6 or 12 months, respectively) showed better results. Weighted kappa values were 0.64 (95% CI, 0.45-0.83) for SGPALS-6 m and 0.60 (95% CI, 0.39-0.80) for HOET-q1, whereas the single-item questions with a shorter recall period had kappa values < 0.40.

CONCLUSIONS

Single-item questions with a longer time frame (6 or 12 months) for physical activity were shown to be more reliable than multi-item questionnaires such as the IPAQ-SF and PASE in individuals with Parkinson's disease without cognitive impairments. There is a need to develop a core outcome set to measure physical activity in people with Parkinson's disease, and there might be a need to develop new physical activity questionnaires.

Socio-demographic determinants of physical activity and app usage from smartphone data

The increasing ubiquity of smartphones provides a potential new data source to capture physical activity behaviours. Though not designed as a research tool, these secondary data have the potential to capture a large population over a more extensive spatial area and with longer temporality than current methods afford. This paper uses one such secondary data source from a commercial app designed to incentivise activity. We explore the new insights these data provide, alongside the sociodemographic profile of those using physical activity apps, to gain insight into both physical activity behaviour and determinants of app usage in order to evaluate the suitability of the app in providing insights into the physical activity of the population. We find app usage to be higher in females, those aged 25-50, and users more likely to live in areas where a higher proportion of the population are of a lower socioeconomic status. We ascertain longer-term patterns of app usage with increasing age and more male users reaching physical activity guideline recommendations despite longer daily activity duration recorded by female users. Additionally, we identify key weekly and seasonal trends in physical activity. This is one of the first studies to utilise a large volume of secondary physical activity app data to co-investigate usage alongside activity behaviour captured.

Sociodemographic Characteristics and Physical Activity in Patients with COPD: A 3-Month Cohort Study

Decreased physical activity (PA) is associated with morbidity and mortality in COPD patients. In this secondary analysis of data from a 12-week longitudinal study, we describe factors associated with PA in COPD. Participants completed the Physical Activity Checklist (PAC) daily for a 7- to 8-day period. PA was measured monthly using the Physical Activity Scale for the Elderly (PASE). At three different time points, daily step count was measured for one week with an Omron HJ-720ITC pedometer. The 35 participants were primarily male (94%) and White (91%), with an average age of 66.5 years and FEV₁ 44.9% predicted. Common activities reported on the PAC were walking (93%), preparing a meal (89%), and traveling by vehicle (96%). PA measured by both PASE score (p = 0.01) and average daily step count (p = 0.04) decreased during follow-up. In repeated measures multivariable modeling, participants living with others had a higher daily step count (ß = 942 steps, p = 0.01) and better PASE scores (ß = 46.4, p < 0.001). Older age was associated with decreased step count (ß = -77 steps, p < 0.001) whereas White race was associated with lower PASE scores (ß = -55.4, p < 0.001) compared to non-White race. Other demographic factors, quality of life, and medications were not associated with PA. A better understanding of the role of social networks and social support may help develop interventions to improve PA in COPD.

Influence of the Duration and Timing of Data Collection on Accelerometer-Measured Physical Activity, Sedentary Time and Associated Insulin Resistance

Accelerometry is a commonly used method to determine physical activity in clinical studies, but the duration and timing of measurement have seldom been addressed. We aimed to evaluate possible changes in the measured outcomes and associations with insulin resistance during four weeks of accelerometry data collection. This study included 143 participants (median age of 59 (IQR9) years; mean BMI of 30.7 (SD4) kg/m²; 41 men). Sedentary and standing time, breaks in sedentary time, and different intensities of physical activity were measured with hip-worn accelerometers. Differences in the accelerometer-based results between weeks 1, 2, 3 and 4 were analyzed by mixed models, differences during winter and summer by two-way ANOVA, and the associations between insulin resistance and cumulative means of accelerometer results during weeks 1 to 4 by linear models. Mean accelerometry duration was 24 (SD3) days. Sedentary time decreased after three weeks of measurement. More physical activity was measured during summer compared to winter. The associations between insulin resistance and sedentary behavior and light physical activity were non-significant after the first week of measurement, but the associations turned significant in two to three weeks. If the purpose of data collection is to reveal associations between accelerometer-measured outcomes and tenuous health outcomes, such as insulin sensitivity, data collection for at least three weeks may be needed.

Utility of wearable physical activity monitors in cardiovascular disease: a systematic review of 11 464 patients and recommendations for optimal use

Aims

Physical activity (PA) plays an important role in primary and secondary prevention of cardiovascular disease (CVD), functioning as a marker of disease progression and response to therapy. Real-world measurement of habitual PA is now possible through wearable activity monitors, however, their use in cardiovascular patients is not well described.

Methods and results

We performed a systematic review to summarize how wearable activity monitors have been used to measure PA in patients with CVD, with 11 464 patients included across 108 studies. Activity monitors were primarily used in the setting of cardiac rehabilitation (46, 43%). Most often, triaxial accelerometers (70, 65%) were instructed to be worn at the hip (58, 54%) for 7 days (n = 54, 50%). Thirty-nine different activity monitors were used, with a range of accelerometer specific settings for collection and reporting of activity data. Activity was reported most commonly as time spent in metabolic equivalent-defined activity levels (49, 45%), while non-wear time was defined in just 16 (15%) studies.

Conclusion

The collecting, processing, and reporting of accelerometer-related outcomes were highly heterogeneous. Most validation studies are limited to healthy young adults, while the paucity of methodological information disclosed renders interpretation of results and cross-study comparison challenging. While accelerometers are promising tools to measure real-world PA, we highlight current challenges facing their use in elderly multimorbid cardiology patients. We suggest recommendations to guide investigators using these devices in cardiovascular research. Future work is required to determine optimal methodology and consensus-based development of meaningful outcomes using raw acceleration data.

Measurement Reactivity of Accelerometer-Based Sedentary Behavior and Physical Activity in 2 Assessment Periods

BACKGROUND

The purposes of this study were to examine accelerometer measurement reactivity (AMR) in sedentary behavior (SB), physical activity (PA), and accelerometer wear time in 2 measurement periods and to quantify AMR as a human-related source of bias for the reproducibility of SB and PA estimates.

METHODS

In total, 136 participants (65% women, mean age = 54.6 y) received 7-day accelerometry at the baseline and after 12 months. Latent growth models were used to identify AMR. Intraclass correlations were calculated to examine the reproducibility using 2-level mixed-effects linear regression analyses.

RESULTS

Within each 7-day accelerometry assessment, the participants increased their time spent in SB (b = 2.4 min/d; b = 3.8 min/d) and reduced their time spent in light PA (b = -2.0 min/d; b = -3.2 min/d), but did not change moderate to vigorous PA. The participants reduced their wear time (b = -5.2 min/d) only at the baseline. The intraclass correlations ranged from .42 for accelerometer wear time to .74 for SB. The AMR was not identified as a source of bias in any regression model.

CONCLUSIONS

AMR may influence SB and PA estimates differentially. Although 7-day accelerometry seems to be a reproducible measure, our findings highlight accelerometer wear time as a crucial confounder in analyzing SB and PA data.

Test-Retest Reliability of activPAL in Measuring Sedentary Behavior and Physical Activity in People With Type 2 Diabetes

BACKGROUND

To investigate how changes in sedentary behavior relate to health outcomes, it is important to establish the test-retest reliability of activity monitors in measuring habitual sedentary behavior in people with type 2 diabetes (T2D) as a prerequisite for interpreting this information. Thus, the authors' objective was to examine the test-retest reliability of a common activity monitor (activPAL™) in measuring sedentary behavior and physical activity in people with T2D.

METHODS

Sedentary-time, standing-time, stepping-time, step-count, and sit-to-stand transitions were obtained from two 7-day assessment periods separated by at least 1 week. Test-retest reliability was determined with the intraclass correlation coefficient (ICC) to compare sedentary and activity measures between the 2 time points.

RESULTS

A total of 30 participants with self-reported T2D completed the study (age 65 [6] y, 63% women, body mass index 33.3 [5] kg/m2). High test-retest reliability was found for sedentary-time (ICC = .79; 95% confidence interval [CI], .61-.89) and standing-time (ICC = .74; 95% CI, .53-.87). Very high test-retest reliability was found for stepping-time (ICC = .90; 95% CI, .81-.95), step-count (ICC = .91; 95% CI, .83-.96), and sit-to-stand transitions (ICC = .90; 95% CI, .79-.95).

CONCLUSION

The activPAL™ device showed high to very high test-retest reliability in measuring all tested activity categories in people with T2D.

No one accelerometer-based physical activity data collection protocol can fit all research questions

Background

Measuring physical activity and sedentary behavior accurately remains a challenge. When describing the uncertainty of mean values or when making group comparisons, minimising Standard Error of the Mean (SEM) is important. The sample size and the number of repeated observations within each subject influence the size of the SEM. In this study we have investigated how different combinations of sample sizes and repeated observations influence the magnitude of the SEM.

Methods

A convenience sample were asked to wear an accelerometer for 28 consecutive days. Based on the within and between subject variances the SEM for the different combinations of sample sizes and number of monitored days was calculated.

Results

Fifty subjects (67% women, mean ± SD age 41 ± 19 years) were included. The analyses showed, independent of which intensity level of physical activity or how measurement protocol was designed, that the largest reductions in SEM was seen as the sample size were increased. The same magnitude in reductions to SEM was not seen for increasing the number of repeated measurement days within each subject.

Conclusion

The most effective way of reducing the SEM is to have a large sample size rather than a long observation period within each individual. Even though the importance of reducing the SEM to increase the power of detecting differences between groups is well-known it is seldom considered when developing appropriate protocols for accelerometer based research. Therefore the results presented herein serves to highlight this fact and have the potential to stimulate debate and challenge current best practice recommendations of accelerometer based physical activity research.

Stronger Association between High Intensity Physical Activity and Cardiometabolic Health with Improved Assessment of the Full Intensity Range Using Accelerometry

An improved method of physical activity accelerometer data processing, involving a wider frequency filter than the most commonly used ActiGraph filter, has been shown to better capture variations in physical activity intensity in a lab setting. The aim of the study was to investigate how this improved measure of physical activity affected the relationship with markers of cardiometabolic health. Accelerometer data and markers of cardiometabolic health from 725 adults from two samples, LIV 2013 and SCAPIS pilot, were analyzed. The accelerometer data was processed using both the original ActiGraph method with a low-pass cut-off at 1.6 Hz and the improved method with a low-pass cut-off at 10 Hz. The relationship between the physical activity intensity spectrum and a cardiometabolic health composite score was investigated using partial least squares regression. The strongest association between physical activity and cardiometabolic health was shifted towards higher intensities with the 10 Hz output compared to the ActiGraph method. In addition, the total explained variance was higher with the improved method. The 10 Hz output enables correctly measuring and interpreting high intensity physical activity and shows that physical activity at this intensity is stronger related to cardiometabolic health compared to the most commonly used ActiGraph method.

'Walk this way': results from a pilot randomised controlled trial of a health coaching intervention to reduce sedentary behaviour and increase physical activity in people with serious mental illness

BACKGROUND

Cardiovascular disease (CVD) is the leading cause of premature death among people with serious mental illness (SMI). Sedentary behaviour (SB) is an independent risk factor for CVD and mortality and people with SMI are highly sedentary. We developed a health coaching intervention called 'Walk this Way' to reduce SB and increase physical activity (PA) in people with SMI and conducted a pilot randomised controlled trial (RCT) to test its feasibility and acceptability.

METHODS

We randomised people with SMI from three community mental health teams into either the WTW intervention or treatment as usual. The WTW intervention lasted 17 weeks and included an initial education session, fortnightly coaching, provision of pedometers and access to a weekly walking group. Objective SB and PA were measured with accelerometers. Cardiometabolic risk factors and wellbeing measures were collected.

RESULTS

We recruited 40 people of whom 33 (82.5%) were followed up. 13/20 (65%) of participants allocated to the coaching intervention completed it. In the intervention group SB decreased by 56 min and total PA increased by 32 min per day on average which was sustained 6 months later. There was no change in PA or SB in the control group. When interviewed, participants in the intervention found the intervention helpful and acceptable. No adverse events were reported from the intervention.

CONCLUSIONS

The intervention was feasible and acceptable to participants. Preliminary results were encouraging with improvement seen in both SB and PA. A larger study is needed to assess the effectiveness of the intervention and address any implementation challenges.

TRIAL REGISTRATION

ISRCTN Registry identifier: ISRCTN37724980 , retrospectively registered 25 September 2015.

Measurement of physical activity in clinical practice using accelerometers

Accelerometers are commonly used in clinical and epidemiological research for more detailed measures of physical activity and to target the limitations of self-report methods. Sensors are attached at the hip, wrist and thigh, and the acceleration data are processed and calibrated in different ways to determine activity intensity, body position and/or activity type. Simple linear modelling can be used to assess activity intensity from hip and thigh data, whilst more advanced machine-learning modelling is to prefer for the wrist. The thigh position is most optimal to assess body position and activity type using machine-learning modelling. Frequency filtering and measurement resolution needs to be considered for correct assessment of activity intensity. Simple physical activity measures and statistical methods are mostly used to investigate relationship with health, but do not take advantage of all information provided by accelerometers and do not consider all components of the physical activity behaviour and their interrelationships. More advanced statistical methods are suggested that analyse patterns of multiple measures of physical activity to demonstrate stronger and more specific relationships with health. However, evaluations of accelerometer methods show considerable measurement errors, especially at individual level, which interferes with their use in clinical research and practice. Therefore, better objective methods are needed with improved data processing and calibration techniques, exploring both simple linear and machine-learning alternatives. Development and implementation of accelerometer methods into clinical research and practice requires interdisciplinary collaboration to cover all aspects contributing to useful and accurate measures of physical activity behaviours related to health.

Time Series Visualizations of Mobile Phone-Based Daily Diary Reports of Stress, Physical Activity, and Diet Quality in Mostly Ethnic Minority Mothers: Feasibility Study

Background

Health behavior patterns reported through daily diary data are important to understand and intervene upon at the individual level in N-of-1 trials and related study designs. There is often interest in relationships between multiple outcomes, such as stress and health behavior. However, analyses often utilize regressions that evaluate aggregate effects across individuals, and standard analyses target single outcomes.

Objective

This paper aims to illustrate how individuals’ daily reports of stress and health behavior (time series) can be explored using visualization tools.

Methods

Secondary analysis was conducted on 6 months of daily diary reports of stress and health behavior (physical activity and diet quality) from mostly ethnic minority mothers who pilot-tested a self-monitoring mobile health app. Time series with minimal missing data from 14 of the 44 mothers were analyzed. Correlations between stress and health behavior within each time series were reported as a preliminary step. Stress and health behavior time series patterns were visualized by plotting moving averages and time points where mean shifts in the data occurred (changepoints).

Results

Median correlation was small and negative for associations of stress with physical activity (r=−.14) and diet quality (r=−.08). Moving averages and changepoints for stress and health behavior were aligned for some participants but not for others. A third subset of participants exhibited little variation in stress and health behavior reports.

Conclusions

Median correlations in this study corroborate prior findings. In addition, time series visualizations highlighted variations in stress and health behavior across individuals and time points, which are difficult to capture through correlations and regression-based summary measures.