Regression tree analysis of ecological momentary assessment data

Using Machine Learning for the Risk Factors Classification of Glycemic Control in Type 2 Diabetes Mellitus

Several risk factors are related to glycemic control in patients with type 2 diabetes mellitus (T2DM), including demographics, medical conditions, negative emotions, lipid profiles, and heart rate variability (HRV; to present cardiac autonomic activity). The interactions between these risk factors remain unclear. This study aimed to use machine learning methods of artificial intelligence to explore the relationships between various risk factors and glycemic control in T2DM patients. The study utilized a database from Lin et al. (2022) that included 647 T2DM patients. Regression tree analysis was conducted to identify the interactions among risk factors that contribute to glycated hemoglobin (HbA1c) values, and various machine learning methods were compared for their accuracy in classifying T2DM patients. The results of the regression tree analysis revealed that high depression scores may be a risk factor in one subgroup but not in others. When comparing different machine learning classification methods, the random forest algorithm emerged as the best-performing method with a small set of features. Specifically, the random forest algorithm achieved 84% accuracy, 95% area under the curve (AUC), 77% sensitivity, and 91% specificity. Using machine learning methods can provide significant value in accurately classifying patients with T2DM when considering depression as a risk factor.

Warming of surface water in the large and shallow lakes across the Yangtze River Basin, China, and its driver analysis

A variety of physical, chemical, and biological processes within the lakes relies on the surface water temperature while the spatial pattern of large lakes of different warming trends and their connections with climate change remain unclear. Using correlation analysis, regression tree analysis (RTA), and general linear models (GLMs), we have estimated the warming trends of 192 lakes since 2000 in the populated Yangtze River Basin, China, to identify dominant climate drivers and quantify their contributions. The results show that surface water temperature has increased substantially in the majority of the investigated lakes (179 from a total of 192 lakes) at a rate of 0.29 (- 0.12 to 0.62) °C/decade (median and 95% confidence interval). The shallower lakes (< 13.1 m in depth) usually have the faster median warming rates than the deeper lakes (i.e., 0.37 °C/decade versus 0.16 °C/decade). We find that in the shallow lakes, rising air temperatures and declining wind speeds can explain the majority of variation in surface water temperature (i.e., 31.4‒80.3% and 13.0‒21.0%, respectively). In contrast, in deeper lakes, change of air temperatures plays a dominant role in water warming (75.4‒91.2%). This study has emphasized the importance of declining wind speed in water warming in large and shallow lakes and illustrated a difference of dominant climatic drivers in water warming between the shallow and deep lakes.

Low Subjective Wellbeing Is Associated with Psychological Distress in People Living with Inflammatory Bowel Disease

BACKGROUND

Inflammatory bowel disease (IBD) is a common and debilitating disease of the gastrointestinal tract. Psychological distress is highly comorbid to IBD, especially during periods of active disease. However, a controversy exists on how to best manage its symptoms in the IBD population.

AIMS

This study aimed to explore protective and risk factors of psychological distress in IBD.

METHODS

A cross-sectional online survey was conducted via social media and online patient forums. Respondents (N = 235) filled out questionnaires on demographics, health characteristics and a range of psychological variables. Measures of pain, disease activity, comorbid functional symptom severity, social support, subjective wellbeing, sleep quality, fatigue, stress, age, BMI and gender were entered into the Classification and Regression Tree Analysis model.

RESULTS

Overall, 87 participants (37%) reported distress. Self-reported stress significantly discriminated between cases of probable psychological distress. In those with high stress, patients with and without probable psychological distress were separated by subjective wellbeing. Among patients with low stress, fatigue was the primary discriminator.

CONCLUSIONS

Monitoring patients for low subjective wellbeing and high stress in clinical settings could offer an opportunity to engage in early intervention to limit psychological distress development. Monitoring for fatigue in patients who seem otherwise psychologically well could offer preventative benefits.

Measuring Engagement in eHealth and mHealth Behavior Change Interventions: Viewpoint of Methodologies

Engagement in electronic health (eHealth) and mobile health (mHealth) behavior change interventions is thought to be important for intervention effectiveness, though what constitutes engagement and how it enhances efficacy has been somewhat unclear in the literature. Recently published detailed definitions and conceptual models of engagement have helped to build consensus around a definition of engagement and improve our understanding of how engagement may influence effectiveness. This work has helped to establish a clearer research agenda. However, to test the hypotheses generated by the conceptual modules, we need to know how to measure engagement in a valid and reliable way. The aim of this viewpoint is to provide an overview of engagement measurement options that can be employed in eHealth and mHealth behavior change intervention evaluations, discuss methodological considerations, and provide direction for future research. To identify measures, we used snowball sampling, starting from systematic reviews of engagement research as well as those utilized in studies known to the authors. A wide range of methods to measure engagement were identified, including qualitative measures, self-report questionnaires, ecological momentary assessments, system usage data, sensor data, social media data, and psychophysiological measures. Each measurement method is appraised and examples are provided to illustrate possible use in eHealth and mHealth behavior change research. Recommendations for future research are provided, based on the limitations of current methods and the heavy reliance on system usage data as the sole assessment of engagement. The validation and adoption of a wider range of engagement measurements and their thoughtful application to the study of engagement are encouraged.