Supplementary Materials for "Concordance Between Dermatology Electronic Health Record–Derived Studies and Higher-Level Clinical Evidence"
Description
Supplementary materials accompanying the study “Concordance Between Dermatology Electronic Health Record–Derived Studies and Higher-Level Clinical Evidence.” These materials provide additional methodological detail and question-level documentation supporting a methodological concordance study comparing clinical conclusions from dermatology electronic health record (EHR)-derived observational studies with higher-level clinical evidence addressing comparable clinical questions. The repository includes: • Supplemental Methods, detailing study identification, eligibility criteria, study selection and data extraction, comparator selection and matching, clinical conclusion classification, independent review, meta-analysis characterization, and statistical analysis. • Supplemental Table I, providing the clinical question, EHR-derived study citation, EHR data source, EHR directional conclusion, comparator study citation, comparator evidence type, comparator directional conclusion, and concordance classification for each of 66 matched clinical questions. • Supplemental Table II, characterizing 15 unique EHR study–meta-analysis pairings, including the EHR data source, comparator meta-analysis, number and design of studies contributing to the relevant evidence synthesis, and whether the matched EHR-derived study was included in the comparator meta-analysis. The study evaluated 34 eligible dermatology EHR-derived studies reporting 144 clinical outcomes, which were matched with 27 comparator studies and consolidated into 66 unique matched clinical questions. These materials are provided to enhance methodological transparency, facilitate independent evaluation of the matching and classification framework, and support reproducibility of the published analysis. No patient-level or identifiable health information is included.
Files
Steps to reproduce
Steps to reproduce 1. Identify dermatology EHR-derived observational studies through a structured literature search. The original PubMed search was conducted July 2, 2026 using broad combinations of dermatology-, disease-, and EHR-related terms and was supplemented by targeted web and journal searches. The contemporaneous search approach, rather than a reconstructed verbatim search string, is reported in the Supplemental Methods. 2. Apply the predefined eligibility criteria described in the Supplemental Methods. Include observational analyses using routinely collected clinical data from institutional, multicenter, or federated EHR sources that address clinically relevant dermatologic questions and provide sufficient information to determine an overall clinical conclusion. 3. Extract the dermatologic disease or population, intervention or exposure, outcome, study design, effect estimate, statistical uncertainty, overall clinical conclusion, and EHR data source. 4. Match each eligible EHR-derived study to comparator evidence addressing the same underlying clinical question according to population or disease, intervention or exposure, clinically relevant outcome, and clinical question. 5. When multiple comparator studies are available, preferentially select the highest available evidence in the following order: meta-analysis, randomized clinical trial, prospective cohort study, retrospective cohort study, and cross-sectional study. 6. Consolidate closely related outcomes representing the same underlying clinical construct into a single matched clinical question when appropriate. 7. Classify EHR-derived and comparator evidence as supporting benefit, harm, or equivocal findings using reported effect estimates, confidence intervals, statistical significance, and the original authors’ overall interpretation. 8. Classify a matched question as concordant when the EHR-derived and comparator evidence reach the same overall clinical conclusion. Define major discordance as direct benefit-versus-harm disagreement. 9. Independently review the final matched clinical questions and concordance classifications. Calculate overall concordance as the proportion of matched clinical questions with agreement, report an exact binomial 95% confidence interval, assess interrater agreement using Cohen κ, and summarize concordance according to comparator evidence type. 10. For EHR study–meta-analysis pairings, review the component evidence within each meta-analysis to determine the number and design of contributing studies and whether the corresponding EHR-derived study was included.
Institutions
- University of FloridaFlorida, Gainesville