Predictors Of Outcome In ICH Patients Initially Admitted To The ICU Who Are Transferred To Inpatient Rehab (P5-7.002)

Abstract

<h3>Objective:</h3> Assess factors associated with good functional outcomes after ICH requiring ICU admission. <h3>Background:</h3> Intracerebral hemorrhage (ICH) is associated with high mortality, with patients requiring intensive care unit (ICU) admission having worse outcomes. However, current scores are limited in their ability to predict functional outcomes. <h3>Design/Methods:</h3> We collected data using a prospectively maintained registry of stroke patients admitted to 5 inpatient rehabs (IR)s in our hospital system between 1/2018 to 6/2019. We performed predictive modeling with RIDGE regression analysis to predict the rehab discharge motor functional independence measure (FIM). Variables used to perform predictive modeling included age, ICU length of stay (LOS), concurrent IVH, ICH expansion, decompressive surgery, laterality of ICH, symptom onset to rehab admission, rehab admission motor FIM score, initial NIHSS, and total therapy days. <h3>Results:</h3> We identified 166 patients with ICH admitted to IR, of which a significant number (51%) were initially admitted to the ICU. After excluding patients with incomplete variables, 61 patients were included in the model. The median age was 59[47–71] years, and 44% were female. The median initial NIHSS was 12 [5–18], and the length of stay in the ICU was 6 [3–10] days. 34 (55.7%) of the patient had IVH, and 7 (11.5%) had hematoma expansion. 7 (11.5%) needed decompressive surgery. The median admit motor FIM score was 24 [17–36], and after a median of 14 [14–14] days in IR, the median discharge motor FIM score more than doubled to 58 [46–69]. Our predictive model had an r-squared of 0.65. Decompressive surgery patients had better scores by 4 points than patients without decompressive surgery. As therapy days increase by one day, discharge motor FIM scores increase by approximately 2 points. <h3>Conclusions:</h3> Our model can satisfactorily predict rehab discharge motor FIM. However, additional factors involved in the process will be explored to improve the model. <b>Disclosure:</b> Dr. Ahmad has nothing to disclose. Dr. Tariq has nothing to disclose. Christin Silos has nothing to disclose. Ms. Hernandez has nothing to disclose. Ms. Contreras has nothing to disclose. Dr. Savitz has nothing to disclose.