Correlates of physical activity behavior in adults: a data mining approach

Using interpretable machine learning methods to identify the relative importance of lifestyle factors for overweight and obesity in adults: pooled evidence from CHNS and NHANES

BACKGROUND

Overweight and obesity pose a huge burden on individuals and society. While the relationship between lifestyle factors and overweight and obesity is well-established, the relative contribution of specific lifestyle factors remains unclear. To address this gap in the literature, this study utilizes interpretable machine learning methods to identify the relative importance of specific lifestyle factors as predictors of overweight and obesity in adults.

METHODS

Data were obtained from 46,057 adults in the China Health and Nutrition Survey (2004-2011) and the National Health and Nutrition Examination Survey (2007-2014). Basic demographic information, self-reported lifestyle factors, including physical activity, macronutrient intake, tobacco and alcohol consumption, and body weight status were collected. Three machine learning models, namely decision tree, random forest, and gradient-boosting decision tree, were employed to predict body weight status from lifestyle factors. The SHapley Additive exPlanation (SHAP) method was used to interpret the prediction results of the best-performing model by determining the contributions of specific lifestyle factors to the development of overweight and obesity in adults.

RESULTS

The performance of the gradient-boosting decision tree model outperformed the decision tree and random forest models. Analysis based on the SHAP method indicates that sedentary behavior, alcohol consumption, and protein intake were important lifestyle factors predicting the development of overweight and obesity in adults. The amount of alcohol consumption and time spent sedentary were the strongest predictors of overweight and obesity, respectively. Specifically, sedentary behavior exceeding 28-35 h/week, alcohol consumption of more than 7 cups/week, and protein intake exceeding 80 g/day increased the risk of being predicted as overweight and obese.

CONCLUSION

Pooled evidence from two nationally representative studies suggests that recognizing demographic differences and emphasizing the relative importance of sedentary behavior, alcohol consumption, and protein intake are beneficial for managing body weight status in adults. The specific risk thresholds for lifestyle factors observed in this study can help inform and guide future research and public health actions.

Analytic Methods for Understanding the Temporal Patterning of Dietary and 24-H Movement Behaviors: A Scoping Review

Dietary and movement behaviors [physical activity (PA), sedentary behavior (SED), and sleep] occur throughout a 24-h day and involve multiple contexts. Understanding the temporal patterning of these 24-h behaviors and their contextual determinants is key to determining their combined effect on health. A scoping review was conducted to identify novel analytic methods for determining temporal behavior patterns and their contextual correlates. We searched Embase, ProQuest, and EBSCOhost databases in July 2022 to identify studies published between 1997 and 2022 on temporal patterns and their contextual correlates (e.g., locational, social, environmental, personal). We included 14 studies after title and abstract (n = 33,292) and full-text (n = 135) screening, of which 11 were published after 2018. Most studies (n = 4 in adults; n = 5 in children and adolescents), examined waking behavior patterns (i.e., both PA and SED) of which 3 also included sleep and 6 included contextual correlates. PA and diet were examined together in only 1 study of adults. Contextual correlates of dietary, PA, and sleep temporal behavior patterns were also examined. Machine learning with various clustering algorithms and model-based clustering techniques were most used to determine 24-h temporal behavior patterns. Although the included studies used a diverse range of methods, behavioral variables, and assessment periods, results showed that temporal patterns characterized by high SED and low PA were linked to poorer health outcomes, than those with low SED and high PA. This review identified temporal behavior patterns, and their contextual correlates, which were associated with adiposity and cardiometabolic disease risk, suggesting these methods hold promise for the discovery of holistic lifestyle exposures important to health. Standardized reporting of methods and patterns and multidisciplinary collaboration among nutrition, PA, and sleep researchers; statisticians; and computer scientists were identified as key pathways to advance future research on temporal behavior patterns in relation to health.

Associations of accelerometer-estimated free-living daily activity impact intensities with 10-year probability of osteoporotic fractures in adults

PURPOSE

Accelerometers are used to objectively measure physical activity; however, the relationship between accelerometer-based activity parameters and bone health is not well understood. This study examines the association between accelerometer-estimated daily activity impact intensities and future risk estimates of major osteoporotic fractures in a large population-based cohort.

METHODS

Participants were 3165 adults 46 years of age from the Northern Finland Birth Cohort 1966 who agreed to wear a hip-worn accelerometer during all waking hours for 14 consecutive days. Raw accelerometer data were converted to resultant acceleration. Impact magnitude peaks were extracted and divided into 32 intensity bands, and the osteogenic index (OI) was calculated to assess the osteogenic effectiveness of various activities. Additionally, the impact peaks were categorized into three separate impact intensity categories (low, medium, and high). The 10-year probabilities of hip and all major osteoporotic fractures were estimated with FRAX-tool using clinical and questionnaire data in combination with body mass index collected at the age of 46 years. The associations of daily activity impact intensities with 10-year fracture probabilities were examined using three statistical approaches: multiple linear regression, partial correlation, and partial least squares (PLS) regression.

RESULTS

On average, participants' various levels of impact were 8331 (SD = 3478) low; 2032 (1248) medium; and 1295 (1468) high impacts per day. All three statistical approaches found a significant positive association between the daily number of low-intensity impacts and 10-year probability of hip and all major osteoporotic fractures. In contrast, increased number of moderate to very high daily activity impacts was associated with a lower probability of future osteoporotic fractures. A higher OI was also associated with a lower probability of future major osteoporotic fractures.

CONCLUSION

Low-intensity impacts might not be sufficient for reducing fracture risk in middle-aged adults, while high-intensity impacts could be beneficial for preventing major osteoporotic fractures.

Deep learning of movement behavior profiles and their association with markers of cardiometabolic health

BACKGROUND

Traditionally, existing studies assessing the health associations of accelerometer-measured movement behaviors have been performed with few averaged values, mainly representing the duration of physical activities and sedentary behaviors. Such averaged values cannot naturally capture the complex interplay between the duration, timing, and patterns of accumulation of movement behaviors, that altogether may be codependently related to health outcomes in adults. In this study, we introduce a novel approach to visually represent recorded movement behaviors as images using original accelerometer outputs. Subsequently, we utilize these images for cluster analysis employing deep convolutional autoencoders.

METHODS

Our method involves converting minute-by-minute accelerometer outputs (activity counts) into a 2D image format, capturing the entire spectrum of movement behaviors performed by each participant. By utilizing convolutional autoencoders, we enable the learning of these image-based representations. Subsequently, we apply the K-means algorithm to cluster these learned representations. We used data from 1812 adult (20-65 years) participants in the National Health and Nutrition Examination Survey (NHANES, 2003-2006 cycles) study who worn a hip-worn accelerometer for 7 seven consecutive days and provided valid accelerometer data.

RESULTS

Deep convolutional autoencoders were able to learn the image representation, encompassing the entire spectrum of movement behaviors. The images were encoded into 32 latent variables, and cluster analysis based on these learned representations for the movement behavior images resulted in the identification of four distinct movement behavior profiles characterized by varying levels, timing, and patterns of accumulation of movement behaviors. After adjusting for potential covariates, the movement behavior profile characterized as "Early-morning movers" and the profile characterized as "Highest activity" both had lower levels of insulin (P < 0.01 for both), triglycerides (P < 0.05 and P < 0.01, respectively), HOMA-IR (P < 0.01 for both), and plasma glucose (P < 0.05 and P < 0.1, respectively) compared to the "Lowest activity" profile. No significant differences were observed for the "Least sedentary movers" profile compared to the "Lowest activity" profile.

CONCLUSIONS

Deep learning of movement behavior profiles revealed that, in addition to duration and patterns of movement behaviors, the timing of physical activity may also be crucial for gaining additional health benefits.

Exercise makes better mind: a data mining study on effect of physical activity on academic achievement of college students

The effect of physical activity (PA) on academic achievement has long been a hot research issue in physical education, but few studies have been conducted using machine learning methods for analyzing activity behavior. In this paper, we collected the data on both physical activity and academic performance from 2,219 undergraduate students (Mean = 19 years) over a continuous period of 12 weeks within one academic semester. Based on students' behavioral indicators transformed from a running APP interface and the average academic course scores, two models were constructed and processed by CHAID decision tree for regression analysis and significance detection. It was found that first, to attain higher academic performance, it is imperative for students to not only exhibit exceptional activity regularity, but also sustain a reduced average step frequency; second, the students completing running exercise with an average frequency of 1 time/week and the duration of 16-25 min excelled over approximately 88 percentage of other students on academic performance; third, the processing validity and reliability of physical observation data in complex systems can be improved by utilizing decision tree as a leveraging machine learning tool and statistical method. These findings provide insights for educational practitioners and policymakers who will seek to enhance college students' academic performance through physical education programs, combined with data mining methods.

Identifying the sociodemographic and work-related factors related to workers' daily physical activity using a decision tree approach

BACKGROUND

The social and behavioural factors related to physical activity among adults are well known. Despite the overlapping nature of these factors, few studies have examined how multiple predictors of physical activity interact. This study aimed to identify the relative importance of multiple interacting sociodemographic and work-related factors associated with the daily physical activity patterns of a population-based sample of workers.

METHODS

Sociodemographic, work, screen time, and health variables were obtained from five, repeated cross-sectional cohorts of workers from the Canadian Health Measures Survey (2007 to 2017). Classification and Regression Tree (CART) modelling was used to identify the discriminators associated with six daily physical activity patterns. The performance of the CART approach was compared to a stepwise multinomial logistic regression model.

RESULTS

Among the 8,909 workers analysed, the most important CART discriminators of daily physical activity patterns were age, job skill, and physical strength requirements of the job. Other important factors included participants' sex, educational attainment, fruit/vegetable intake, industry, work hours, marital status, having a child living at home, computer time, and household income. The CART tree had moderate classification accuracy and performed marginally better than the stepwise multinomial logistic regression model.

CONCLUSION

Age and work-related factors-particularly job skill, and physical strength requirements at work-appeared as the most important factors related to physical activity attainment, and differed based on sex, work hours, and industry. Delineating the hierarchy of factors associated with daily physical activity may assist in targeting preventive strategies aimed at promoting physical activity in workers.

Replacing sedentary time with physical activity and sleep: Associations with cardiometabolic health markers in adults

This study aimed to examine the associations of sedentary time, and substituting sedentary time with physical activity and sleep, with cardiometabolic health markers while accounting for a full 24 h of movement and non-movement behaviors, cardiorespiratory fitness (CRF), and other potential confounders. The participants were 4585 members of the Northern Finland Birth Cohort 1966, who wore a hip-worn accelerometer at the age of 46 years for 14 consecutive days. Time spent in sedentary behaviors, light-intensity physical activity (LPA), and moderate-to-vigorous-intensity physical activity (MVPA) were determined from the accelerometer and combined with self-reported sleep duration to obtain the 24-h time use. CRF was estimated from the peak heart rate in a submaximal step test. An isotemporal substitution paradigm was used to examine how sedentary time and substituting sedentary time with an equal amount of LPA, MVPA, or sleep were associated with adiposity markers, blood lipid levels, and fasting glucose and insulin. Sedentary time was independently and adversely associated with the markers of cardiometabolic health, even after adjustment for CRF, but not in partition models including LPA, MVPA, sleep, and CRF. Substituting 60, 45, 30, and 15 min/day of sedentary time with LPA or MVPA was associated with 0.2%-13.7% favorable differences in the cardiometabolic health markers after accounting for LPA, MVPA, sleep, CRF, and other confounders. After adjustment for movement and non-movement behaviors within the 24-h cycle, reallocating additional time to both LPA and MVPA was beneficially associated with markers of cardiometabolic health in middle-aged adults regardless of their CRF level.

Device-measured physical activity and sedentary time in a national sample of Luxembourg residents: the ORISCAV-LUX 2 study

BACKGROUND

Existing information about population physical activity (PA) levels and sedentary time in Luxembourg are based on self-reported data.

METHODS

This observational study included Luxembourg residents aged 18-79y who each provided ≥4 valid days of triaxial accelerometry in 2016-18 (n=1122). Compliance with the current international PA guideline (≥150 min moderate-to-vigorous PA (MVPA) per week, irrespective of bout length) was quantified and variability in average 24h acceleration (indicative of PA volume), awake-time PA levels, sedentary time and accumulation pattern were analysed by linear regression. Data were weighted to be nationally representative.

RESULTS

Participants spent 51% of daily time sedentary (mean (95% confidence interval (CI)): 12.1 (12.0 to 12.2) h/day), 11% in light PA (2.7 (2.6 to 2.8) h/day), 6% in MVPA (1.5 (1.4 to 1.5) h/day), and remaining time asleep (7.7 (7.6 to 7.7) h/day). Adherence to the PA guideline was high (98.1%). Average 24h acceleration and light PA were higher in women than men, but men achieved higher average accelerations across the most active periods of the day. Women performed less sedentary time and shorter sedentary bouts. Older participants (aged ≥55y) registered a lower average 24h acceleration and engaged in less MVPA, more sedentary time and longer sedentary bouts. Average 24h acceleration was higher in participants of lower educational attainment, who also performed less sedentary time, shorter bouts, and fewer bouts of prolonged sedentariness. Average 24h acceleration and levels of PA were higher in participants with standing and manual occupations than a sedentary work type, but manual workers registered lower average accelerations across the most active periods of the day. Standing and manual workers accumulated less sedentary time and fewer bouts of prolonged sedentariness than sedentary workers. Active commuting to work was associated with higher average 24h acceleration and MVPA, both of which were lower in participants of poorer self-rated health and higher weight status. Obesity was associated with less light PA, more sedentary time and longer sedentary bouts.

CONCLUSIONS

Adherence to recommended PA is high in Luxembourg, but half of daily time is spent sedentary. Specific population subgroups will benefit from targeted efforts to replace sedentary time with PA.

Joint Profiles of Sedentary Time and Physical Activity in Adults and Their Associations with Cardiometabolic Health

PURPOSE

This study aimed to identify and characterize joint profiles of sedentary time and physical activity among adults and to investigate how these profiles are associated with markers of cardiometabolic health.

METHODS

The participants included 3702 of the Northern Finland Birth Cohort 1966 at age 46 yr, who wore a hip-worn accelerometer during waking hours and provided seven consecutive days of valid data. Sedentary time, light-intensity physical activity, and moderate- to vigorous-intensity physical activity on each valid day were obtained, and a data-driven clustering approach ("KmL3D") was used to characterize distinct joint profiles of sedentary time and physical activity intensities. Participants self-reported their sleep duration and performed a submaximal step test with continuous heart rate measurement to estimate their cardiorespiratory fitness (peak heart rate). Linear regression was used to determine the association between joint profiles of sedentary time and physical activities with cardiometabolic health markers, including adiposity markers and blood lipid, glucose, and insulin levels.

RESULTS

Four distinct groups were identified: "active couch potatoes" ( n = 1173), "sedentary light movers" ( n = 1199), "sedentary exercisers" ( n = 694), and "movers" ( n = 636). Although sufficiently active, active couch potatoes had the highest daily sedentary time (>10 h) and lowest light-intensity physical activity. Compared with active couch potatoes, sedentary light movers, sedentary exercisers, and movers spent less time in sedentary by performing more physical activity at light-intensity upward and had favorable differences in their cardiometabolic health markers after accounting for potential confounders (1.1%-25.0% lower values depending on the health marker and profile).

CONCLUSIONS

After accounting for sleep duration and cardiorespiratory fitness, waking activity profiles characterized by performing more physical activity at light-intensity upward, resulting in less time spent in sedentary, were associated with better cardiometabolic health.

How Europeans move: a moderate-to-vigorous physical activity and sitting time paradox in the European Union

OBJECTIVES

This study aimed to assess the interactions between physical activity (PA) and sedentary behaviour in a large population taking account of major sociodemographic characteristics.

STUDY DESIGN

Cross-sectional population-based study.

METHODS

Data from 28,031 individuals living in the European Union who were aged ≥15 years were retrieved from a cross-sectional survey, the Eurobarometer 2017. Interactions among the four mobility components (vigorous, moderate, walking activity and sitting time) were assessed at the individual level across age, gender and place of residence, and at the country level by compositional data analysis, hierarchical linear regressions and principal component analysis.

RESULTS

The most frequently reported PA was walking; however, sitting time represented >95% of the reported weekly times, whereas moderate-to-vigorous PA (MVPA) represented <1%. Women reported less PA and sitting time, age decreased total PA and increased sitting time, and individuals living in large urban areas reported lower PA and higher sitting times. MVPA decreased with age (β = -0.047, P < 0.001) and was lower in women (β = -0.760, P < 0.001) and those living in large urban areas (β = -0.581, P < 0.001), while walking and sitting times increased with age, being higher in women and lower in those living in rural areas. At the country level, sitting time was positively associated with moderate activity (β = 0.389, P = 0.041) and marginally non-significant with MVPA (β = 0.330, P = 0.087).

CONCLUSIONS

Walking was the highest contributor to weekly PA, whereas sitting time was paradoxically associated with higher MVPA. Specific measures to reduce sitting time are required to achieve an active lifestyle.

Conceptualizing and Measuring Appetite Self-Regulation Phenotypes and Trajectories in Childhood: A Review of Person-Centered Strategies

This review uses person-centered research and data analysis strategies to discuss the conceptualization and measurement of appetite self-regulation (ASR) phenotypes and trajectories in childhood (from infancy to about ages 6 or 7 years). Research that is person-centered provides strategies that increase the possibilities for investigating ASR phenotypes. We first examine the utility of examining underlying phenotypes using latent profile/class analysis drawing on cross-sectional data. The use of trajectory analysis to investigate developmental change is then discussed, with attention to phenotypes using trajectories of individual behaviors as well as phenotypes based on multi-trajectory modeling. Data analysis strategies and measurement approaches from recent examples of these person-centered approaches to the conceptualization and investigation of appetite self-regulation and its development in childhood are examined. Where relevant, examples from older children as well as developmental, clinical and educational psychology are drawn on to discuss when and how person-centered approaches can be used. We argue that there is scope to incorporate recent advances in biological and psychoneurological knowledge about appetite self-regulation as well as fundamental processes in the development of general self-regulation to enhance the examination of phenotypes and their trajectories across childhood (and beyond). The discussion and conclusion suggest directions for future research and highlight the potential of person-centered approaches to progress knowledge about the development of appetite self-regulation in childhood.

Accumulation patterns of sedentary time and breaks and their association with cardiometabolic health markers in adults

Breaking up sedentary time with physical activity (PA) could modify the detrimental cardiometabolic health effects of sedentary time. Our aim was to identify profiles according to distinct accumulation patterns of sedentary time and breaks in adults, and to investigate how these profiles are associated with cardiometabolic outcomes. Participants (n = 4439) of the Northern Finland Birth Cohort 1966 at age 46 years wore a hip-worn accelerometer for 7 consecutive days during waking hours. Uninterrupted ≥1-min sedentary bouts were identified, and non-sedentary bouts in between two consecutive sedentary bouts were considered as sedentary breaks. K-means clustering was performed with 65 variables characterizing how sedentary time was accumulated and interrupted. Linear regression was used to determine the association of accumulation patterns with cardiometabolic health markers. Four distinct groups were formed as follows: "Couch potatoes" (n = 1222), "Prolonged sitters" (n = 1179), "Shortened sitters" (n = 1529), and "Breakers" (n = 509). Couch potatoes had the highest level of sedentariness and the shortest sedentary breaks. Prolonged sitters, accumulating sedentary time in bouts of ≥15-30 min, had no differences in cardiometabolic outcomes compared with Couch potatoes. Shortened sitters accumulated sedentary time in bouts lasting <15 min and performed more light-intensity PA in their sedentary breaks, and Breakers performed more light-intensity and moderate-to-vigorous PA. These latter two profiles had lower levels of adiposity, blood lipids, and insulin sensitivity, compared with Couch potatoes (1.1-25.0% lower values depending on the cardiometabolic health outcome, group, and adjustments for potential confounders). Avoiding uninterrupted sedentary time with any active behavior from light-intensity upwards could be beneficial for cardiometabolic health in adults.