Charting Diagnostic Safety: Exploring Patient-Provider Discordance in Medical Record Documentation

Background :The 21st Century Cures Act enables patients to access their medical records, thus providing a unique opportunity to engage patients in their diagnostic journey. Objective :To explore the concordance between patients' self-reported diagnostic concerns and clinician-interpreted information in their electronic health records. Design :We conducted a mixed-methods analysis of a cohort of 467 patients who completed a structured data collection instrument (the Safer Dx Patient) to identify diagnostic concerns while reviewing their clinician's notes. We conducted a qualitative content analysis of open-ended responses on both the tools and the case summaries. Two clinical chart reviewers, blinded to patient-reported diagnostic concerns, independently conducted chart reviews using a different structured instrument (the Revised Safer Dx Instrument) to identify diagnostic concerns and generate case summaries. The primary outcome variable was chart review-identified diagnostic concerns. Multivariate logistic regression tested whether the primary outcome was concordant with patient-reported diagnostic concerns. Setting :Geisinger, a large integrated healthcare organization in rural and semi-urban Pennsylvania. Participants :Cohort of adult patients actively using patient portals and identified as "at-risk" for diagnostic concerns using an electronic trigger algorithm based on unexpected visit patterns in a primary care setting. Results :In 467 cohort patients, chart review identified 31 (6.4%) diagnostic concerns, of which only 11 (21.5%) overlapped with 51 patient-reported diagnostic concerns. Content analysis revealed several areas of discordant understanding of the diagnostic process between clinicians and patients. Multivariate logistic regression analysis showed that clinician-identified diagnostic concerns were associated with patients who self-reported "I feel I was incorrectly diagnosed during my visit" (odds ratio 1.65, 95% CI 1.17-2.3, p < 0.05). Conclusion :Patients and clinicians appear to have certain differences in their mental models of what is considered a diagnostic concern. Efforts to integrate patient perspectives and experiences with the diagnostic process can lead to better measurement of diagnostic safety.