Search

Background: Clinical trials of patent foramen ovale (PFO) closure for stroke prevention have produced controversial results. We investigated differences in presenting the net benefits of PFO closure or description of the PFO in a randomized vignette-based internet survey. Methods: Adults aged 18 to 60 inclusive were recruited via Amazon’s Mechanical Turk. Participants read a vignette asking them to imagine that they had a minor stroke. The vignettes included three randomized factors in a 2 x 2 x 3 factorial design: a) benefit of PFO closure (no proven benefit versus ~3% absolute risk reduction, analogous to difference in individual trials versus subsequent meta-analysis); b) normalization intervention (PFO described as very common and benign for most individuals versus no such description); and c) labeling of the PFO (?PFO? versus ?hole in the heart? versus ?difference in heart structure?). Participants were then asked their choice of treatment (PFO closure or aspirin). Secondary outcomes included self-perceived stroke risk and level of anxiety about the PFO diagnosis. Regression analysis was used to assess predictors of treatment choice (logistic) and secondary outcomes (linear). Results: 2157 participants completed the survey (54.6% female, 76.2% non-Hispanic white, median age 32 with IQR 27 – 41). Participants shown data suggesting 3% absolute risk reduction for PFO closure were more likely to choose PFO closure (30.9% vs. 7.5%, p < 0.001). The normalization intervention decreased anxiety about the PFO. The labeling intervention increased self-perceived stroke risk and had differential effects on patient anxiety (see Table for details). Conclusion: Differences in how information is presented can influence choice of treatment, perceived stroke risk, and level of anxiety surrounding PFO-associated stroke. Future work should address the impact of information presentation on potential variability in use of PFO closure and patient psychological outcomes.

Introduction: Stroke Core Measures are intended to ensure that patients receive high quality, guideline-concordant acute stroke care. Compliance data is collected and reported to the public and to regulatory agencies for benchmarking and for reimbursement. Hypothesis: Core Measures noncompliance within one academic institution will often reflect ambiguous or incomplete documentation instead of true failure to provide medically appropriate care. Methods: We retrospectively reviewed Core Measures data for all ischemic and hemorrhagic stroke patients discharged from the University of Michigan between January 2013 and May 2014. Core Measures data was collected and reported per routine practice and contemporaneously reviewed by the institutional team. For this study, Core Measures failures were cross-referenced with the full medical chart and classified as “true failures” when care was not compliant with the Core Measure standard or “documentation failures” when chart review revealed poor documentation of otherwise appropriate care--for example, where a normal neurological examination was not explicitly linked with a decision to defer assessment for rehabilitation. Determinations of the basis for noncompliance on chart review were made by two different reviewers, with 100% agreement. Results: A total of 40 failures in 872 patients were identified and reviewed. Core Measures failures were documentation failures in 20 patients. Additional details are provided in Table 1. Conclusion: The high number of documentation-based failures in our experience illustrates potential problems in the use of administratively-defined measures as a marker of the quality of clinical care.