Intro: While thrombolysis in stroke mimics (SM) is considered safe, recent data highlights the excess cost associated with treatment of these patients. Several studies have identified common demographic features of SM, however less is known about whether language barriers between patient and physician influence SM treatment rates. We sought to evaluate the role of physician-patient language discordance on the rate of SM treatment at a single center serving a large Spanish-speaking population.
Hypothesis: Stroke mimic treatment rates are higher when there are language barriers between physician and patient due to greater diagnostic uncertainty
Methods: We reviewed the electronic medical record (EMR) for all patients who received tPA in the ED from 7/2011 to 7/2015. Patient’s primary language was obtained from the EMR; language fluency of treating neurologists was obtained via questionnaire; final diagnosis (SM, imaging negative, imaging confirmed) was the attending physician’s impression at discharge. We compared baseline characteristics and SM rates between encounters where the treating neurologist and patient spoke the same language (concordant group) versus those where they did not (discordant group). Means were compared via t test, medians via Mann Whitney U test and dichotomized variables via chi square test.
Results: During this period 311 patients received tPA. English was the primary language for 158 (51%), Spanish for 144 (46%), and other languages for 9 (3%); 183 (59%) encounters were classified as concordant and 128 (41%) as discordant. Final diagnosis was SM for 37 (12%); among those with a final diagnosis of stroke, 65 (24%) were imaging negative. There were no significant differences in mean age (67 vs. 70, p=0.1), male sex (38% vs. 32%, p=0.3), and median NIHSS (7 vs. 7, p=0.4) between concordant and discordant groups. We found higher rates of SM in the concordant group (16% vs. 6%, p=0.01). When imaging negative strokes were included with SM, these differences were no longer significant (33% vs. 32%, p=0.9).
Conclusion: At our institution, language discordancy does not contribute to higher rates of SM treatment. Careful observation of how language discordant pairs communicate is needed to understand the role of interpreters in these findings