Experience Sampling Method Data Research Articles

Preprocessing Experience-Sampling-Method Data: A Step-by-Step Framework, Tutorial Website, R Package, and Reporting Templates

Experience-sampling-method (ESM) studies have become a very popular tool to gain insight into the dynamics of psychological processes. Although the statistical modeling of ESM data has been widely studied, the preprocessing steps that precede such modeling have received relatively limited attention despite being a challenging phase. At the same time, adequate preprocessing of ESM data is crucial: It provides valuable information about the quality of the data and, importantly, helps to resolve issues in the data that may compromise the validity of statistical analyses. To support researchers in properly preprocessing ESM data, we have developed a step-by-step framework, a tutorial website that provides a gallery of R code, an R package, and templates to report the preprocessing steps. Particular attention is given to three different aspects in preprocessing: checking adherence to the study design (e.g., whether the momentary questionnaires were delivered according to the sampling scheme), examining participants’ response behaviors (e.g., compliance, careless responding), and describing and visualizing the data (e.g., examining distributions of variables).

Forecasting the onset of depression with limited baseline data only: A comparison of a person-specific and a multilevel modeling based exponentially weighted moving average approach.

The onset of depressive episodes is preceded by changes in mean levels of affective experiences, which can be detected using the exponentially weighted moving average procedure on experience sampling method (ESM) data. Applying the exponentially weighted moving average procedure requires sufficient baseline data from the person under study in healthy times, which is needed to calculate a control limit for monitoring incoming ESM data. It is, however, not trivial to obtain sufficient baseline data from a single person. We therefore investigate whether historical ESM data from healthy individuals can help establish an adequate control limit for the person under study via multilevel modeling. Specifically, we focus on the case in which there is very little baseline data available of the person under study (i.e., up to 7 days). This multilevel approach is compared with the traditional, person-specific approach, where estimates are obtained using the person's available baseline data. Predictive performance in terms of Matthews correlation coefficient did not differ much between the approaches; however, the multilevel approach was more sensitive at detecting mean changes. This implies that for low-cost and nonharmful interventions, the multilevel approach may prove particularly beneficial. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Experience sampling of suicidality, religiosity and spirituality in depression: Network analyses using dynamic time warping

BackgroundSuicidality is a clinically important and multifaceted phenomenon, frequently present in depressed subjects. Religiosity and spirituality (R/S) can have an attenuating as well as a reinforcing effect on suicidality. MethodsFrom two Dutch mental health care settings, a sample of 31 depressed and in- and outpatients with suicidal ideation, self-identifying as being religious or spiritual, was selected by convenience sampling. Using an experience sampling method (ESM) mobile application, during six days (mean of 42 assessments per subject), the association between symptoms of depression, suicidality, and specific positive-supportive affective R/S and positive psychology variables. For 28 participants symptom network plots on a group level, and on an individual level, were analyzed using dynamic time warping (DTW). ResultsParticipants were on average 35.7 years old, and 65 % were women. In the group-level undirected network, R/S variables were linked to positive psychology variables via a bridge function of inner peace. Changes in the experience of inner peace and enjoying a physical activity preceded changes of several other symptoms. A network dynamic appeared with a dense cluster of ‘positive psychology’ items. LimitationsOnly a limited number of R/S variables were included. ConclusionThe results of this study suggest that religiosity and spirituality function as meaningful factors in depression and suicidality in religiously or spiritually engaged persons. Experienced inner peace has a positive association with reasons to live. Experience sampling method data can be effectively analyzed using dynamic time warping. Exploring individual religious or spiritual engagement can prove important in treating suicidality and depression.

A theory-informed emotion regulation variability index: Bray-Curtis dissimilarity.

Emotion regulation (ER) variability refers to how individuals vary their use of ER strategies across time. It helps individuals to meet contextual needs, underscoring its importance in well-being. The theoretical foundation of ER variability recognizes two constituent processes: strategy switching (e.g., moving from distraction to social sharing) and endorsement change (e.g., decreasing the intensity of both distraction and social sharing). ER variability is commonly operationalized as the SD between strategies per observation (between-strategy SD) or within a strategy across time (within-strategy SD). In this article, we show that these SD-based approaches cannot sufficiently capture strategy switching and endorsement change, leading to ER variability indices with poor validity. We propose Bray-Curtis dissimilarity, a measure used in ecology to quantify biodiversity variability, as a theory-informed ER variability index. First, we demonstrate how Bray-Curtis dissimilarity is more sensitive than SD-based approaches in detecting ER variability through two simulation studies. Second, assuming that higher ER variability is adaptive in daily life, we test the relation between ER variability and negative affect in three experience sampling method data sets (total N = [70, 95, 200], number of moment-level observations = [5,040, 6,329, 14,098]). At both the moment level and person level, higher Bray-Curtis dissimilarity predicted lower negative affect more consistently than SD-based indices. We conclude that Bray-Curtis dissimilarity may better capture moment-level within-person ER variability and could have implications for studying variability in other multivariate dynamic processes. The article is accompanied by an R tutorial and practical recommendations for using Bray-Curtis dissimilarity with experience sampling method data. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Assessing psychological variables on mobile devices: an introduction to the experience sampling app ESM-Quest.

The use of the Experience Sampling Method (ESM), which involves repeated assessments in people's daily lives, has increased in popularity in psychology and associated disciplines in recent years. A rather challenging aspect of ESM is its technical implementation. In this paper, after briefly introducing the history of ESM and the main reasons for its current popularity, we outline the ESM-Quest experience sampling app which is currently being developed at the University of Vienna. ESM-Quest runs on different operating systems, specifically on mobile devices such as smartphones and tablets running either iOS or Android. An internet connection is not necessary during the assessment. Compared to most other ESM apps, ESM-Quest allows event-based random sampling, which is very helpful when assessments need to be collected within specific situations. Currently, ESM-Quest is being utilized at the University of Vienna and will be made available for research groups worldwide upon request. We introduce the technical aspects of ESM-Quest and provide examples of analyses on ESM data collected through this app, such as examining fluctuations in constructs within individuals. Finally, we outline potential next steps in ESM research.

A smartphone-based intervention for young people who self-harm (‘PRIMARY’): study protocol for a multicenter randomized controlled trial

BackgroundSelf-harm in young people is a public health concern connected with severe mental health problems, such as personality pathology. Currently, there are no specific evidence-based interventions available for young people who self-harm. Therefore, we developed PRe-Intervention Monitoring of Affect and Relationships in Youth (PRIMARY), a smartphone-based intervention, co-designed by clinicians and young people with lived experience of mental ill-health. PRIMARY combines the Experience Sampling Method (ESM) with weekly report sessions. The study aims to examine the effectiveness of PRIMARY with regard to reducing self-harm, and improving emotion regulation and quality of relationships.MethodsThis study is a multicenter, parallel groups, randomized controlled trial (RCT) comparing the PRIMARY intervention to a waiting list control group. PRIMARY comprises 28 consecutive days of questionnaires five times each day (i.e., ESM) and four weekly report sessions. Participants will comprise 180 young people referred for treatment to the participating Dutch mental healthcare institutions and (1) are aged 12 to 25 years, and (2) engaged in ≥ 1 act of self-harm in the past year. Participants are randomly allocated to a study group after screening in a 1:1 ratio by an independent researcher using computer-generated randomization sequences with stratified block randomization by age (12 to 15 years / 16 to 25 years). Staff will conduct assessments with all participants at baseline (Wave 1), after 28 days (Wave 2), and in a subsample after 10 weeks of subsequent specialized treatment (Wave 3). The primary outcomes are self-harm, emotion regulation, and quality of relationships. Secondary outcomes include patient and clinician satisfaction. Exploratory analyses of ESM data will examine the relationship between emotions, social relationships, and self-harm.DiscussionThe results of this trial will clarify whether an innovative smartphone-based intervention is effective for reducing self harm and improving emotion regulation and the quality of social relationships. It has the potential to fill a treatment gap of interventions specifically targeting self-harm. If proven effective, it would provide an accessible, easy-to-implement, low-cost intervention for young people. Furthermore, the ESM-data will allow detailed analyses into the processes underlying self-harm, which will contribute to theoretical knowledge regarding the behavior.Trial registrationISRCTN42088538 (https://doi.org/10.1186/ISRCTN42088538), retrospectively registered on the 26th of October 2022.

Time Effect on Acute Postoperative Pain After Total Knee Replacement Surgery: An Exploratory Study Using the Experience Sampling Method.

Acute postoperative pain (APP) is the main cause of postoperative dissatisfaction; however, traditional methods of pain assessment provide limited insights into the dynamics and development of APP. This study used the experience sampling method (ESM) to understand the dynamics of APP over time in relation to various patient factors. Forty patients scheduled to undergo total knee replacement surgery were recruited in this study. Following an initial assessment, a short report questionnaire was sent to the patients via 10 digital alerts per day to assess the pain levels during two preoperative and the first 6 postoperative days. The data were analyzed using multilevel regression including random intercept and slope. Thirty-two patients submitted the pre-specified minimum of 30% of their short reports, yielding 1217 records. The analysis revealed significant (P<0.001) linear and quadratic decreases in APP and a quadratic time effect. The lowest between-days and within-day pain levels were observed on postoperative day 4.8 and during the time slot 3.8 or approximately 19:15, respectively. Significant random intercepts and slopes were noted, indicating variations in the mean pain level between patients and a decrease in pain. None of the 10 patient factors had any confounding effect. Using ESM data combined with multilevel analysis, we were able to evaluate the postoperative pain course while considering inter-individual differences in the baseline pain level and non-linear pain course over time. The findings of this study could aid clinicians in personalizing the treatment for APP.

Using the experience sampling method to support clinical practice: An illustration with problematic cannabis use

Objective The experience sampling method (ESM) has been frequently used in clinical research; however, there is low translational uptake in clinical practice. This may be due to challenges with interpreting individual-level data at granular intervals. We provide an illustrative example of how ESM can be leveraged to generate personalized cognitive–behavioral strategies for problematic cannabis use. Method We conducted a descriptive case series analysis using ESM data from 30 individuals reporting on problematic cannabis use, craving, affect, and coping four times daily for 16-days (t = 64, T = 1,920). Results Analyzing ESM data using descriptive statistics and visualizations from individuals with similar clinical and demographic profiles supported a diverse array of personalized clinical insights and recommendations for each case. These recommendations included psychoeducation regarding affect- and boredom-regulation strategies, functional analyses of occasions during which cannabis was not used, and discussions on how cannabis use intersects with one’s personal values. Conclusion While many clinicians utilize measurement-based care, barriers have limited the incorporation of ESM towards personalized, data-informed approaches to treatment. We provide an illustrative example of how ESM data can be used to generate actionable treatment strategies for problematic cannabis use and highlight continued challenges with interpreting time-series data.

Practitioner perspectives on the use of the experience sampling software in counseling and clinical psychology

ABSTRACT Facilitating the uptake and making better use of technological advances will be pivotal for counseling and clinical psychology to respond to the rising call for more community-based and person-centred care. While the Experience Sampling Method (ESM), a structured self-report digital diary, could help facilitate this transition, it is currently unclear how practitioners envision using ESM in clinical practice. Therefore, we organised focus groups with 36 mental health practitioners (Mage = 39.37, SDage = 12.18, 58.33% female) across Flanders (Belgium). Four broad topics were discussed: (1) how to use ESM, (2) how to visualise clinically relevant information, (3) the software requirements thereof, and (4) barriers and facilitators for implementing ESM in clinical practice. Thematic analysis was conducted and Cohen's Kappa was calculated (κ = .79). Different clinical applications emerged, such as screening, evaluating the effectiveness of therapy, and elucidating determinants of mental health problems in daily life. However, practitioners also expressed difficulty determining how to visualise ESM data, and novel features for use emerged (e.g. integration with electronic health records). Various barriers (e.g. lack of best-practice guidelines) and facilitators (e.g. simplicity) were identified, with the implications of these findings for future clinical implementation studies discussed.

A mixed-method investigation into measurement reactivity to the experience sampling method: The role of sampling protocol and individual characteristics.

Since the introduction of the experience sampling method (ESM), there have been concerns that the repeated assessments typically related to this method may alter the behavior, thoughts, or feelings of participants. Previous studies have offered mixed results with some studies reporting reactive changes, while others failed to find such effects. Our aim was to investigate under which circumstances ESM induces reactive effects. Students (N = 151) were randomly assigned to receive a questionnaire containing 30 or 60 items three, six, or nine times per day for 14 days. A random sample of 50 participants took part in qualitative interviews after the end of the data collection. We investigated changes over time in the data, while taking into account the sampling protocol and characteristics of participants, and analyzed qualitative reports of measurement reactivity. Decreases in completion time, within-person variance of ratings and subjective reports of habituation point toward the existence of a habituation period. While participants reported increases in emotional awareness in interviews, ESM measures indicated a decrease in emotional awareness over time. Changes in behavior were rare in quantitative and qualitative reports. Positive affect was decreasing over time in the ESM data, and various changes in affect, emotion regulation, and thoughts were reported in interviews. Individual characteristics and sampling protocol had inconsistent effects on changes over time. The results suggest that ESM induces changes in within-person variability, completion times, affect, or emotional awareness over time. Further research is needed to explore whether observed changes affect the validity of ESM data. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

The Lang‐Track‐App: Open‐Source Tools for Implementing the Experience Sampling Method in Second Language Acquisition Research

AbstractThis paper introduces the Lang‐Track‐App, a smartphone application and backend system to aid second language acquisition researchers in implementing the experience sampling method (ESM). Surveying research participants with the Lang‐Track‐App multiple times per day can yield exceptionally rich data that can be analyzed in multiple ways. This article explains the different Lang‐Track‐App components and describes workflows for both researchers and participants. To demonstrate the step‐by‐step implementation, we present a recent study of language use in study abroad as a test case in which we used the Lang‐Track‐App to collect ESM data about frequency and qualitative aspects of language use. A selection of findings from this study highlights the richness of ESM data and how they can be analyzed at various levels. Researchers wishing to use the Lang‐Track‐App to implement the ESM in their own research can access the full open‐source code and documentation via GitHub (https://github.com/HumlabLu).

Applying multiverse analysis to experience sampling data: Investigating whether preprocessing choices affect robustness of conclusions.

The experience sampling method (ESM) has revolutionized our ability to conduct psychological research in the natural environment. However, researchers have a large degree of freedom when preprocessing ESM data, which may hinder scientific progress. This study illustrates the use of multiverse analyses regarding preprocessing choices related to data exclusion (i.e., based on various levels of compliance and exclusion of the first assessment day) and the calculation of constructs (i.e., composite scores calculated as the mean, median, or mode) by reanalyzing established group differences in negative affect, stress reactivity, and emotional inertia between individuals with and without psychosis. Data came from five studies and included 233 individuals with psychosis and 223 healthy individuals (in total, 26,892 longitudinal assessments). Preprocessing choices related to data exclusion did not affect conclusions. For both stress reactivity and emotional inertia of negative affect, group differences were affected when negative affect was calculated as the mean compared to the median or mode. Further analyses revealed that this could be attributed to considerable differences in the within- and between-factor structure of negative affect. While these findings show that observed differences in affective processes between individuals with and without psychosis are robust to preprocessing choices related to data exclusion, we found disagreement in conclusions between different central tendency measures. Safeguarding the validity of future experience sampling research, scholars are advised to use multiverse analysis to evaluate the robustness of their conclusions across different preprocessing scenarios.

Risk and Protective Factors of College Students’ Psychological Well-Being During the COVID-19 Pandemic: Emotional Stability, Mental Health, and Household Resources

Colleges and universities have increasingly worried in recent decades about college students’ well-being, with the COVID-19 pandemic aggravating these concerns. Our study examines changes to undergraduate emotional sentiments and psychological well-being from before to after the onset of the pandemic. In addition, we explore whether certain risk factors (i.e., prior mental health impairments, trait emotional stability) and protective factors (i.e., subjective socioeconomic status, parental education, household resources) predicted students’ emotions and their intraindividual changes due to the pandemic onset. We compared experience sampling method data from 120 students from before and after the pandemic onset, examining intraindividual trajectories. There was only little change in students’ emotions. Prior mental health impairment and trait emotional stability predicted students’ emotions, averaged across time points, but not emotion changes. Few associations with emotions were found for subjective socioeconomic status and parental education, but study-related household resources predicted levels and changes in emotions.

Daily Associations Among Alcohol Intoxication, Partner Familiarity, Participant Effortful Control, Urgency, and PrEP Uptake on Sexual Behavior in Men Who Have Sex with Men.

The purpose of this study was to examine the effects of alcohol intoxication and its interaction with contextual or situation (partner familiarity) and individual differences variables (effortful control, urgency, and whether taking pre-exposure prophylaxis (PrEP) medication) on sexual behaviors in men who have sex with men (MSM), a subgroup for whom HIV continues to be a major public health problem in the U.S. The participants were 236 men recruited from two northeastern U.S. cities and aged 21-50 years, M = 27.8). These men participated in a 6-week (two 3-week sampling bursts) experience sampling method (ESM) study. The ESM data were collected via use of software installed on the participant's own or study-provided mobile phone. Individual differences variables were measured by participants' completing questionnaires measuring effortful control and urgency, and the participant's self-report of whether he was currently taking PrEP. The ESM data pertained to sexual behavior as well as situation variables of familiarity of relevant sexual partners and number of standard alcohol drinks consumed. The results generally were consistent with hypotheses, as alcohol intoxication showed a curvilinear relation to the occurrence of condomless anal intercourse. Furthermore, the likelihood of occurrence of condomless anal sex increased with increased familiarity of the sexual partner. Similarly, taking PrEP increased the likelihood of occurrence of condomless anal sex. At the same time, alcohol's effects were moderated by all three individual differences variables as expected, but the prediction that partner familiarity would moderate alcohol's effects on the occurrence of condomless sex was not supported. Clinical implications of the findings center on the application of the data to HIV prevention programs toward inclusion of more empirically supported, nuanced information on the relation between acute alcohol intoxication and sexual behavior. Directions for further research address the need for additional testing and refinement of a person × situation approach to alcohol and sexual behavior. Furthermore, it is argued that it is important to refine further the concept of sexual risk in the context of taking PrEP and to conduct more detailed, multivariate studies of the relation between taking PrEP and patterns of sexual behavior.

ESMvis: a tool for visualizing individual Experience Sampling Method (ESM) data

PurposeThe experience sampling method (ESM) is used for intensive longitudinal time-series data collection during normal daily life. ESM data give information on momentary affect, activities and (social) context of, for example, patients suffering from mental disorders, and allows for person-specific feedback reports. However, current personalized feedback reports only display a selection of measured variables, and typically involve only summary statistics, thus not reflecting the dynamic fluctuations in affect and its influencing factors. To address this shortcoming, we developed a tool for dynamically visualizing ESM data.MethodsWe introduce a new framework, ESMvis, for giving descriptive feedback, focusing on direct visualization of the dynamic nature of raw data. In this ESM feedback approach, raw ESM data are visualized using R software. We applied ESMvis to data collected for over 52 weeks on a patient diagnosed with an obsessive–compulsive disorder with comorbid depression.ResultsWe provided personalized feedback, in which both the overall trajectory and specific time moments were captured in a movie format. Two relapses during the study period could be visually determined, and subsequently confirmed by the therapist. The therapist and patient evaluated ESMvis as an insightful add-on tool to care-as-usual.ConclusionESMvis is a showcase on providing personalized feedback by dynamic visualization of ESM time-series data. Our tool is freely available and adjustable, making it widely applicable. In addition to potential applications in clinical practice, ESMvis can work as an exploratory tool that can lead to new hypotheses and inform more complex statistical techniques.

Usability of the Experience Sampling Method in Parkinson's Disease on a Group and Individual Level.

ABSTRACTBackgroundAround 50% of PD patients experience motor fluctuations, which are often accompanied by mood fluctuations. The nature of the relationship between motor and mood fluctuations remains unknown. It is suggested that the experience sampling method can reveal such associations on both a group and individual level. Revealing group patterns may enhance our understanding of symptom interactions and lead to more general treatment recommendations, whereas analyses in individual patients can be used to establish a personalized treatment plan.ObjectivesTo explore the usability of routinely collected experience sampling method data over a brief period of time to detect associations between motor fluctuations, affective state, and contextual factors in PD patients with motor fluctuations on a group level and on an individual level.MethodsEleven patients with motor fluctuations collected data at 10 semirandom moments over the day for 5 consecutive days.ResultsOn a group level, multilevel analyses showed significant associations between all motor symptoms and positive affect. Being at home was associated with increased balance problems and rigidity. Analyses on an individual level revealed much less significant associations that mostly, but not always, were in line with the results on a group level.ConclusionThis exploratory study showed significant associations between affective state, motor symptoms, and contextual factors in a group of PD patients with motor fluctuations, but less so in individual patients. Given that the ultimate aim is to use the experience sampling method as an aid to personalize treatments, the sensitivity of the approach needs to be increased. © 2020 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.

For whom, and for what, is experience sampling more accurate than retrospective report?

The experience sampling method (ESM) is often used in research, and promoted for clinical use, with the rationale that it avoids problematic inaccuracies and biases that attend retrospective measures of mental phenomena. Research suggests that averaged scores from ESM data are more accurate than retrospective ratings. However, it is not known how well individuals can remember information about momentary (rather than averaged) mental states, nor how accurately they estimate the dynamic covariation of these states. Individual differences in retrospective accuracy are also poorly understood. In two pre-registered studies, we examined differences between retrospective memory for stress and self-esteem and data gathered via experience sampling and examined whether alexithymia predicted accuracy. Results of both studies revealed substantial discrepancies between retrospective ratings and ESM ratings, especially for momentary states and their covariation. Alexithymia was positively related to recognition of stress means and variability but unrelated to recall of either stress or self-esteem, their variability, or their covariation. These findings suggest that experience sampling may be more useful than self-report when precise information is needed about the timing of mental states and dynamics among them.

Applications of the experience sampling method: a research agenda for hospitality management

PurposeThis study aims to provide researchers in hospitality management with a comprehensive understanding of the experience sampling method (ESM) and to engage them in the use of ESM in their future research. With this critical discussion of the advantages and challenges of the method, researchers can apply it appropriately to deepen and broaden their research findings.Design/methodology/approachThis study chooses an empirical example in the context of hotel employees’ surface acting, tiredness and sleep quality to illustrate the application of ESM. Based on the example, this paper conducts a two-level modeling in Mplus, including a cross-level mediation analysis and mean centering.FindingsThis paper demonstrates the applicability and usefulness of ESM for hospitality research and provides a detailed demonstration of how to use the statistical program Mplus to analyze ESM data. With this paper, researchers will be able to consider how to engage ESM in their future studies.Originality/valueTo the best of the authors‘ knowledge, this paper is among the first to provide a hands-on demonstration of ESM to hospitality researchers. The authors call for more research in hospitality management to use ESM to answer complex and pressing research questions.

Implementing Experience Sampling Technology for Functional Analysis in Family Medicine - A Design Thinking Approach.

BackgroundA paradigm shift in health care from illness to wellbeing requires new assessment technologies and intervention strategies. Self-monitoring tools based on the Experience Sampling Method (ESM) might provide a solution. They enable patients to monitor both vulnerability and resilience in daily life. Although ESM solutions are extensively used in research, a translation from science into daily clinical practice is needed.ObjectiveTo investigate the redesign process of an existing platform for ESM data collection for detailed functional analysis and disease management used by psychological assistants to the general practitioner (PAGPs) in family medicine.MethodsThe experience-sampling platform was reconceptualized according to the design thinking framework in three phases. PAGPs were closely involved in co-creation sessions. In the ‘understand’ phase, knowledge about end-users’ characteristics and current eHealth use was collected (nominal group technique – 2 sessions with N = 15). In the ‘explore’ phase, the key needs concerning the platform content and functionalities were evaluated and prioritized (empathy mapping – 1 session with N = 5, moderated user testing – 1 session with N = 4). In the ‘materialize’ phase, the adjusted version of the platform was tested in daily clinical practice (4 months with N = 4). The whole process was extensively logged, analyzed using content analysis, and discussed with an interprofessional project group.ResultsIn the ‘understand’ phase, PAGPs emphasized the variability in symptoms reported by patients. Therefore, moment-to-moment assessment of mood and behavior in a daily life context could be valuable. In the ‘explore’ phase, (motivational) functionalities, technological performance and instructions turned out to be important user requirements and could be improved. In the ‘materialize’ phase, PAGPs encountered barriers to implement the experience-sampling platform. They were insufficiently facilitated by the regional primary care group and general practitioners.ConclusionThe redesign process in co-creation yielded meaningful insights into the needs, desires and daily routines in family medicine. Severe barriers were encountered related to the use and uptake of the experience-sampling platform in settings where health care professionals lack the time, knowledge and skills. Future research should focus on the applicability of this platform in family medicine and incorporate patient experiences.

Identifying psychosis spectrum disorder from experience sampling data using machine learning approaches

The ubiquity of smartphones opened up the possibility of widespread use of the Experience Sampling Method (ESM). The method is used to collect longitudinal data of participants' daily life experiences and is ideal to capture fluctuations in emotions (momentary mental states) as an indicator for later mental ill-health. In this study, ESM data of patients with psychosis spectrum disorder and controls were used to examine daily life emotions and higher order patterns thereof. We attempted to determine whether aggregated ESM data, in which statistical measures represent the distribution and dynamics of the original data, were able to distinguish patients from controls in a predictive modeling framework. Variable importance, recursive feature elimination, and ReliefF methods were used for feature selection. Model training, tuning, and testing were performed in nested cross-validation, based on algorithms such as Random Forests, Support Vector Machines, Gaussian Processes, Logistic Regression, and Neural Networks. ROC analysis was used to post-process these models. Stability of model performance was studied using Monte Carlo simulations. The results provide evidence that patterns in emotion changes can be captured by applying a combination of these techniques. Acceleration in the variables anxious and insecure was particularly successful in adding further predictive power to the models. The best results were achieved by Support Vector Machines with radial kernel (accuracy = 82% and sensitivity = 82%). This proof-of-concept work demonstrates that synergistic machine learning and statistical modeling may be used to harness the power of ESM data in the future.