Maya Stemmer

Background: Fully remote online health research is vulnerable to fraudulent participation that can threaten data validity and divert study resources. Although fraud mitigation is increasingly necessary in online research, mitigation efforts are rarely described systematically or evaluated transparently across studies. Objective: This study demonstrates the application of the configure, assess, triage, corroborate, and hone (CATCH) framework to an ongoing remote mental health study and evaluates how structured mitigation procedures have affected participant progression and operational costs. Methods: This study was conducted within an ongoing, fully remote mental health study evaluating hallucinations through longitudinal digital data collection. Following detection of fraudulent activity, the CATCH framework was applied to organize mitigation procedures into stages and refine detection and verification workflows. New screening safeguards were configured, classification thresholds were recalibrated, enrolled participants were reassessed and triaged accordingly, and secondary corroboration through synchronous video verification was introduced when warranted. Procedures were honed through ongoing monitoring in response to emerging fraud patterns. We compared 2 temporally defined cohorts before (March-August 2025) and after (September 2025-January 2026) the implementation of the refined procedures. Primary outcomes included the stage at which fraudulent participants were identified and cost per eligible participant, decomposed into costs attributable to fraudulent participants, nonfraudulent participants, and fraud-mitigation technology. Participant-level costs were estimated using a stage-based model incorporating participant compensation and research staff effort. Sensitivity analyses examined alternative compensation and staffing scenarios to assess the generalizability of the proposed cost-accounting mechanism. Results: Following CATCH implementation, fraudulent participants were identified earlier in the study workflow, reducing progression to later study stages. Overall, 41.5% (161/388) of reassessed participants were discharged for suspected fraud, while 58.5% (227/388) were retained following adjudication. Among those flagged as high risk and invited to video verification, 1.8% (3/164) successfully confirmed their identity, providing a lower-bound estimate of the false-positive classification rate. The mean sunk cost per fraudulent participant decreased from US $12.75 (SD US $40.76) in the pre-CATCH cohort to US $3.29 (SD US $10.91) in the post-CATCH cohort. Total cost per eligible participant decreased from US $170.48 to US $134.90, driven by a substantial reduction in the fraud-related portion from US $80.46 to US $19.73. Sensitivity analyses showed reductions in fraud-related cost per eligible participant across all alternative costing scenarios. To our knowledge, this is the first empirical demonstration of a structured fraud-mitigation framework implemented and evaluated within an active health study. Conclusions: Structuring fraud mitigation using an explicit framework enabled systematic documentation of procedural adaptations and their operational consequences. Beyond study-specific findings, this work illustrates how organizing fraud mitigation into defined, reportable stages can promote transparency, reproducibility, and cumulative learning across remote health studies. Treating fraud mitigation as an integral methodological component is essential for safeguarding scientific integrity and legitimate resource expenditure in online research.

BACKGROUND Fully remote online health research is vulnerable to fraudulent participation that can threaten data validity and divert study resources. Although fraud mitigation is increasingly necessary in online research, mitigation efforts are rarely described systematically or evaluated transparently across studies. OBJECTIVE This study demonstrates the application of the configure, assess, triage, corroborate, and hone (CATCH) framework to an ongoing remote mental health study and evaluates how structured mitigation procedures have affected participant progression and operational costs. METHODS This study was conducted within an ongoing, fully remote mental health study evaluating hallucinations through longitudinal digital data collection. Following detection of fraudulent activity, the CATCH framework was applied to organize mitigation procedures into stages and refine detection and verification workflows. New screening safeguards were configured, classification thresholds were recalibrated, enrolled participants were reassessed and triaged accordingly, and secondary corroboration through synchronous video verification was introduced when warranted. Procedures were honed through ongoing monitoring in response to emerging fraud patterns. We compared 2 temporally defined cohorts before (March-August 2025) and after (September 2025-January 2026) the implementation of the refined procedures. Primary outcomes included the stage at which fraudulent participants were identified and cost per eligible participant, decomposed into costs attributable to fraudulent participants, nonfraudulent participants, and fraud-mitigation technology. Participant-level costs were estimated using a stage-based model incorporating participant compensation and research staff effort. Sensitivity analyses examined alternative compensation and staffing scenarios to assess the generalizability of the proposed cost-accounting mechanism. RESULTS Following CATCH implementation, fraudulent participants were identified earlier in the study workflow, reducing progression to later study stages. Overall, 41.5% (161/388) of reassessed participants were discharged for suspected fraud, while 58.5% (227/388) were retained following adjudication. Among those flagged as high risk and invited to video verification, 1.8% (3/164) successfully confirmed their identity, providing a lower-bound estimate of the false-positive classification rate. The mean sunk cost per fraudulent participant decreased from US $12.75 (SD US $40.76) in the pre-CATCH cohort to US $3.29 (SD US $10.91) in the post-CATCH cohort. Total cost per eligible participant decreased from US $170.48 to US $134.90, driven by a substantial reduction in the fraud-related portion from US $80.46 to US $19.73. Sensitivity analyses showed reductions in fraud-related cost per eligible participant across all alternative costing scenarios. To our knowledge, this is the first empirical demonstration of a structured fraud-mitigation framework implemented and evaluated within an active health study. CONCLUSIONS Structuring fraud mitigation using an explicit framework enabled systematic documentation of procedural adaptations and their operational consequences. Beyond study-specific findings, this work illustrates how organizing fraud mitigation into defined, reportable stages can promote transparency, reproducibility, and cumulative learning across remote health studies. Treating fraud mitigation as an integral methodological component is essential for safeguarding scientific integrity and legitimate resource expenditure in online research.

Objectives: Mobile health (mHealth) interventions are increasingly used to support illness self-management for individuals with schizophrenia spectrum disorders (SSD). Yet limited research explores how symptom patterns that vary throughout an intervention period can be used to identify individuals who may benefit from tailored support. Leveraging daily self-ratings from FOCUS, an mHealth intervention for SSD, we identified and characterized distinct self-reported symptom patterns. Methods: We conducted a prospective cohort study as a secondary analysis using data from a hybrid type 3 effectiveness–implementation trial of FOCUS across 23 community mental health programs. Participants used FOCUS for up to 90 days and provided daily self-ratings on mood, sleep, and/or auditory hallucinations. Group-Based Trajectory Modeling (GBTM) was used to identify longitudinal trajectories for each domain. We examined demographic, psychiatric, and functioning correlates of trajectory group membership. Results: Overall, 190 participants provided daily self-ratings for mood, sleep, or hallucinations. Participants provided a median of 33 ratings (IQR: 10, 84) across a median 82 days (IQR: 50, 88). For each domain, five distinct trajectory groups emerged. Mood ratings were largely stable over time, with one improving group. Sleep domain showed a unique worsening trajectory. Multiple groups improved in hallucinations self-ratings. Higher early self-rating response rates were associated with improving mood trajectories. Conclusions: Individuals with SSD exhibited distinct symptom expression trajectories over the course of treatment with an mHealth intervention. Trajectory-based approaches may inform adaptive mHealth designs and more targeted allocation of mHealth and clinical support resources.

Fraudulent participation is a growing challenge in digital health research, particularly in online studies where duplicate identities, automated responses, and coordinated sign-ups can distort recruitment, compromise validity, and divert resources. Safeguards intended to prevent fraud might also risk excluding legitimate participants, raising concerns about sample representativeness and study generalizability. Although a wide range of technical and behavioral strategies exists, guidance is lacking on how to organize these methods and report outcomes consistently across studies. To address this gap, we introduce the Configure, Assess, Triage, Corroborate, and Hone (CATCH) framework, a hybrid fraud detection–mitigation model with actionable recommendations for investigators. CATCH begins with pre-study configuration to prepare for fraud mitigation and proceeds through systematic assessment of fraud risk, triage of candidates into risk categories, and corroboration of inconclusive cases, while honing strategies through ongoing monitoring. The framework emphasizes transparent documentation and reporting of actions and outcomes to facilitate comparability across studies and continuous methodological refinement. As fraudulent participation grows and emerging technologies act as both risks and solutions, CATCH can help guide investigators’ efforts to maximize data integrity in digital health research. By synthesizing existing fraud-mitigation strategies into a unified, staged framework, CATCH offers practical guidance for structuring decisions, documenting actions, and balancing data integrity with inclusivity.

This research aims to investigate what patients with inflammatory bowel disease (IBD) are talking about on Twitter and learn from the experimental knowledge they share online. The study presents a framework for analyzing patients’ tweets and comparing their content to tweets of the general population. We started by constructing two datasets of tweets—a dataset of patients’ tweets and a control dataset for comparison. Then, we thematically classified the tweets and obtained a subset of tweets related to health and nutrition. We used a Dirichlet regression to compare the thematic segmentations of the two groups. We continued by extracting keywords from the filtered tweets and applying entity sentiment analysis to determine the patients’ sentiments towards the extracted keywords. Finally, we detected emotions within the tweets and used a Wilcoxon test to compare the emotions conveyed in each group. We found statistically significant differences between the patients’ thematic segmentations and those of the control group and observed significant differences in the emotions each group expressed while talking about health. Not only do patients talk more about health in comparison to the general Twitter population, but they also address the subject with negative sentiments and express more negative emotions. The personal information IBD patients share on Twitter can be used to derive complementary knowledge about the disease and provide an additional foundation to existing medical research on IBD. The four stages of the study are also feasible to extend to other chronic conditions.

Inflammatory bowel disease (IBD) is usually classified into Crohn's disease (CD) or ulcerative colitis (UC). Inconclusive cases are diagnosed with IBD unclassified (IBD-U). In 2018, IBD patients shared their disease history on Twitter and signed their tweets with #MyIBDHistory. In this research, we analyzed those tweets and built a logistic regression classifier that predicts patients' IBD type. We constructed tabular classification features and assessed their importance using the regression coefficients and association rules. We identified key features that distinguished CD from UC and used the classifier to predict the disease type of IBD-U patients. Our results correlated with IBD-related research. The two most prominent features that tilted the classification towards CD were suffering from fistulas or nutrient deficiencies. We identified gender differences in disease perspective prior to diagnosis. The research shows that the personal information shared by patients on Twitter can enhance existing medical knowledge regarding their disease.

Background: Patients use social media as an alternative information source, where they share information and provide social support. Although large amounts of health-related data are posted on Twitter and other social networking platforms each day, research using social media data to understand chronic conditions and patients’ lifestyles is limited.Objective: In this study, we contributed to closing this gap by providing a framework for identifying patients with inflammatory bowel disease (IBD) on Twitter and learning from their personal experiences. We enabled the analysis of patients’ tweets by building a classifier of Twitter users that distinguishes patients from other entities. This study aimed to uncover the potential of using Twitter data to promote the well-being of patients with IBD by relying on the wisdom of the crowd to identify healthy lifestyles. We sought to leverage posts describing patients’ daily activities and their influence on their well-being to characterize lifestyle-related treatments.Methods: In the first stage of the study, a machine learning method combining social network analysis and natural language processing was used to automatically classify users as patients or not. We considered 3 types of features: the user’s behavior on Twitter, the content of the user’s tweets, and the social structure of the user’s network. We compared the performances of several classification algorithms within 2 classification approaches. One classified each tweet and deduced the user’s class from their tweet-level classification. The other aggregated tweet-level features to user-level features and classified the users themselves. Different classification algorithms were examined and compared using 4 measures: precision, recall, F1 score, and the area under the receiver operating characteristic curve. In the second stage, a classifier from the first stage was used to collect patients' tweets describing the different lifestyles patients adopt to deal with their disease. Using IBM Watson Service for entity sentiment analysis, we calculated the average sentiment of 420 lifestyle-related words that patients with IBD use when describing their daily routine.Results: Both classification approaches showed promising results. Although the precision rates were slightly higher for the tweet-level approach, the recall and area under the receiver operating characteristic curve of the user-level approach were significantly better. Sentiment analysis of tweets written by patients with IBD identified frequently mentioned lifestyles and their influence on patients’ well-being. The findings reinforced what is known about suitable nutrition for IBD as several foods known to cause inflammation were pointed out in negative sentiment, whereas relaxing activities and anti-inflammatory foods surfaced in a positive context.Conclusions: This study suggests a pipeline for identifying patients with IBD on Twitter and collecting their tweets to analyze the experimental knowledge they share. These methods can be adapted to other diseases and enhance medical research on chronic conditions.

In any research field, data access and data integration are major challenges that even large, well-established consortia face. Although data sharing initiatives are increasing, joint data analyses on nutrition and microbiomics in health and disease are still scarce. We aimed to identify observational studies with data on nutrition and gut microbiome composition from the Intestinal Microbiomics (INTIMIC) Knowledge Platform following the findable, accessible, interoperable, and reusable (FAIR) principles. An adapted template from the European Nutritional Phenotype Assessment and Data Sharing Initiative (ENPADASI) consortium was used to collect microbiome-specific information and other related factors. In total, 23 studies (17 longitudinal and 6 cross-sectional) were identified from Italy (7), Germany (6), Netherlands (3), Spain (2), Belgium (1), and France (1) or multiple countries (3). Of these, 21 studies collected information on both dietary intake (24 h dietary recall, food frequency questionnaire (FFQ), or Food Records) and gut microbiome. All studies collected stool samples. The most often used sequencing platform was Illumina MiSeq, and the preferred hypervariable regions of the 16S rRNA gene were V3–V4 or V4. The combination of datasets will allow for sufficiently powered investigations to increase the knowledge and understanding of the relationship between food and gut microbiome in health and disease.