UCR SOM - TriNetX Summer Bootcamp 2025, Week 2 – Defining and Exploring Cohorts

(Summarized by the Zoom AI Meeting Agent; may contain errors or inaccuracies) TriNetX Advanced Cohort Definition Bootcamp The second session of the TriNetX Summer Bootcamp focused on advanced cohort definition and exploration, which are central to using the platform effectively. Daniel Novak, the Director of Student Scholarly Activities at the School of Medicine, emphasized the importance of understanding cohorts for developing studies and ensuring data quality. He introduced tools like the cohort explorer and discussed topics such as confounding factors, inclusion and exclusion criteria, and the role of time in studies. Daniel also mentioned ongoing developments of additional tools to aid in research, including a bar chart maker and a power set and sample size checker. The session aimed to build a strong foundation for participants to conduct effective research using the TriNetX platform. Epidemiological Study Methodologies and TriNetX Daniel discussed the methodology and design of epidemiological studies, focusing on cohort studies, case-control studies, and cross-sectional studies. He emphasized the importance of accurately defining cohorts and using index events to anchor timelines for measuring outcomes. Daniel also introduced the concept of TriNetX, a platform for conducting health research, and explained its query builder features, including term groups and cohort definitions. He highlighted the dynamic nature of the network, which constantly evolves with new healthcare organizations joining, affecting the number of patients available for study. Analyzing Cohorts With TriNetX Daniel demonstrated how to use the related group function in TriNetX to define and analyze cohorts based on specific criteria and temporal relationships. He explained how to create groups of terms, set time constraints, and use Boolean logic to focus on specific populations. Daniel emphasized the importance of properly defining temporal relationships between events to avoid confounding factors and ensure accurate study results. He also advised naming groups clearly and using stepwise approaches to create complex groupings. Cohort Management in TriNetX Daniel explained the process of creating and managing cohort groups in TriNetX, emphasizing the distinction between groups of terms and cohorts of people. He demonstrated how to copy, paste, and modify groups, as well as use logical operators to define complex cohorts. Daniel also highlighted the importance of creating separate cohorts for comparison and warned against combining all groups into one cohort. He introduced the Explore Cohort function, which provides demographic information about the cohort, and discussed the limitations of using unknown categories for race and ethnicity. Cohort Analysis Best Practices Daniel discussed the limitations of using certain medical categories and emphasized the importance of being cautious when comparing outcomes for different patient groups. He introduced the "Doctors Don't Pay Money Late" mnemonic to remember the key features available for cohort analysis: demographics, diagnoses, procedures, medications, and labs. Daniel explained how to use these features to build cohorts and highlighted the importance of not using top-level diagnostic codes for inclusion or exclusion criteria. He also demonstrated how to export data to Excel for further analysis and advised checking both CPT and ICD-10 codes when analyzing procedures. Cohort Selection and Bias Mitigation Daniel discussed methods for building and comparing cohorts in research studies, emphasizing the importance of selecting appropriate comparison groups to avoid bias. He advised against comparing small groups to large populations and recommended using AI tools like Perplexity to assist in identifying relevant comorbidities and exclusion criteria. Daniel also shared his experience with a study on statins and Alzheimer's disease, highlighting the complexity of the issue and the various exclusion criteria used to control for confounding factors. Cohort Diagrams for Study Design Daniel discussed the importance of using cohort diagrams to clearly represent inclusion and exclusion criteria in studies, emphasizing their value in communicating study designs effectively. He highlighted the need to document the TriNetX network's dynamic nature and the importance of using current network data for analysis. Daniel also addressed questions about confounders, readmission frequency, and AI tool subscriptions, providing guidance on how to handle these issues in research. The session concluded with a reminder about upcoming topics on comparing cohorts and propensity score matching. #TriNetX #ClinicalInformatics #MedicalStudentResearch #RealWorldEvidence #MedicalEducation #HealthDataAnalytics #ClinicalTrials #PropensityScoreMatching #EHRData #ICD10Codes #DigitalHealth #HealthTech #PopulationHealth