Abstract 312: Evaluating Novel Methods of Post Cardiac Arrest Outcome Assessment

Abstract

Introduction: Recovery of cardiac arrest patients is dynamic over the first year. Loss to follow up is common and validated tools to assess long-term survival and function useful for observational research are lacking. We tested novel methods of long-term follow-up after resuscitation from cardiac arrest. Hypothesis: We hypothesized that post-acute encounters in our electronic health record (EHR) and internet-based searches would each allow determination of long-term vital status and function. Methods: We performed a retrospective study including consecutive patients treated after cardiac arrest at a single academic medical center from 2010 to 2018 who survived to hospital discharge. We used a structured data collection tool and three methods to assess post-discharge vital status and function (modified Rankin Scale score (mRS)): 1) a search of our health system’s EHR for post-discharge encounters; 2) an web-based search of local newspapers, ancestry sites, and other sources; and 3) a search of public social media postings attributed to the patient on Facebook, Twitter, and Instagram. Results: Overall, 900 patients met inclusion criteria. Of these, 544 (60%) were male and 643 (71%) arrested out-of-hospital. A minority, 257 patients, (29%) had good function (mRS 0-2) at hospital discharge. Overall, 862 patients (96%) had EHR data after their index hospitalization. Of these, 307 (36%) had a date of death confirmed in the EHR a median of 196 days [interquartile range (IQR) 59-703] days post-arrest. The remaining 555 patients were alive at their last known follow-up after a median of 830 [IQR 246 - 1661] days. Of these, 357 (64%) had documentation of function consistent with an mRS of 0-2. By contrast, we found 602 (67%) patients using web-based sources: 333 (37%) in local newspapers, 239 (27%) in ancestry sites, 319 (35%) on Facebook, 65 (7%) on Twitter, and 55 (6%) on Instagram. Web-based sources confirmed 25 additional deaths not identified in the EHR. Conclusions: This work supports use of novel methods of outcome assessment of patients’ vital and functional status after cardiac arrest. Further work will test these methods against reference standards.