Search

Background Disparate racial/ethnic burdens of the Coronavirus Disease 2019 (COVID-19) pandemic may be attributable to higher susceptibility to Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) or to factors such as differences in hospitalization and care provision. Methods In our cross-sectional analysis of lab-confirmed COVID-19 cases from a tertiary, eight-hospital healthcare system across greater Houston, multivariable logistic regression models were fitted to evaluate hospitalization and mortality odds for non-Hispanic Blacks (NHBs) vs. non-Hispanic Whites (NHWs) and Hispanics vs. non-Hispanics. Results Between March 3rd and July 18th, 2020, 70,496 individuals were tested for SARS-CoV-2; 12,084 (17.1%) tested positive, of whom 3536 (29.3%) were hospitalized. Among positive cases, NHBs and Hispanics were significantly younger than NHWs and Hispanics, respectively (mean age NHBs vs. NHWs: 46.0 vs. 51.7 years; p p p p p = 0.001). Both minority groups resided in lower median income (median income [USD]; NHBs vs. NHWs: 63,489 vs. 75,793; p p p p Conclusions Our data did not demonstrate racial and ethnic differences in care provision and hospital outcomes. Higher susceptibility of racial and ethnic minorities to SARS-CoV-2 and subsequent hospitalization may be driven primarily by social determinants.

BACKGROUND The COVID-19 pandemic has exacerbated the challenges of meaningful health care digitization. The need for rapid yet validated decision-making requires robust data infrastructure. Organizations with a focus on learning health care (LHC) systems tend to adapt better to rapidly evolving data needs. Few studies have demonstrated a successful implementation of data digitization principles in an LHC context across health care systems during the COVID-19 pandemic. OBJECTIVE We share our experience and provide a framework for assembling and organizing multidisciplinary resources, structuring and regulating research needs, and developing a single source of truth (SSoT) for COVID-19 research by applying fundamental principles of health care digitization, in the context of LHC systems across a complex health care organization. METHODS Houston Methodist (HM) comprises eight tertiary care hospitals and an expansive primary care network across Greater Houston, Texas. During the early phase of the pandemic, institutional leadership envisioned the need to streamline COVID-19 research and established the retrospective research task force (RRTF). We describe an account of the structure, functioning, and productivity of the RRTF. We further elucidate the technical and structural details of a comprehensive data repository—the HM COVID-19 Surveillance and Outcomes Registry (CURATOR). We particularly highlight how CURATOR conforms to standard health care digitization principles in the LHC context. RESULTS The HM COVID-19 RRTF comprises expertise in epidemiology, health systems, clinical domains, data sciences, information technology, and research regulation. The RRTF initially convened in March 2020 to prioritize and streamline COVID-19 observational research; to date, it has reviewed over 60 protocols and made recommendations to the institutional review board (IRB). The RRTF also established the charter for CURATOR, which in itself was IRB-approved in April 2020. CURATOR is a relational structured query language database that is directly populated with data from electronic health records, via largely automated extract, transform, and load procedures. The CURATOR design enables longitudinal tracking of COVID-19 cases and controls before and after COVID-19 testing. CURATOR has been set up following the SSoT principle and is harmonized across other COVID-19 data sources. CURATOR eliminates data silos by leveraging unique and disparate big data sources for COVID-19 research and provides a platform to capitalize on institutional investment in cloud computing. It currently hosts deeply phenotyped sociodemographic, clinical, and outcomes data of approximately 200,000 individuals tested for COVID-19. It supports more than 30 IRB-approved protocols across several clinical domains and has generated numerous publications from its core and associated data sources. CONCLUSIONS A data-driven decision-making strategy is paramount to the success of health care organizations. Investment in cross-disciplinary expertise, health care technology, and leadership commitment are key ingredients to foster an LHC system. Such systems can mitigate the effects of ongoing and future health care catastrophes by providing timely and validated decision support.

Objective To quantify in-hospital systolic blood pressure variability among patients with intracerebral hemorrhage, determine the association between high systolic blood pressure variability (HSBPV) and 90-day severe disability or death, and examine the association between pre-hospital factors and HSBPV. Methods Adult, radiologically confirmed, intracerebral hemorrhage patients enrolled in a multi-site cohort were included. Using a semi-automated algorithm, systolic blood pressure values recorded from routine non-invasive systolic blood pressure monitoring in critical and acute care settings were extracted for the duration of hospitalization. Inter and intra-patient systolic blood pressure variability was quantified using generalized estimating equation methods. Modified Poisson and logistic regression models were fit to determine the association between HSBPV and 90-day severe disability or death and between pre-hospital characteristics and HSBPV, respectively. Results A total of 566 patients managed at four certified stroke centers were included. Over 120,000 systolic blood pressure readings were analyzed, and a standard deviation (SD) of 13.0 was parameterized as a cut-off point to categorize HSBPV. Patients with HSBPV had a greater risk of 90-day severe disability or death (relative risk: 1.20, 95% confidence interval: 1.04–1.39), after controlling for age, pre-morbid functional status, and other disease severity measures. Greater likelihood of in-hospital HSBPV was independently observed in elderly, female patients, and in patients with high admission systolic blood pressure. Conclusion Quantification of HSBPV is feasible utilizing routinely collected systolic blood pressure readings, and a singular cut-off parameter for systolic blood pressure variability demonstrated association with 90-day severe disability or death. Elderly, female, and patients with high admission systolic blood pressure may be more likely to demonstrate HSBPV during hospitalization.

Introduction: We characterize 10-year nationwide trends in utilization of palliative care (PC) for ischemic stroke (IS) patients and describe in-hospital outcomes and costs for IS patients receiving PC, particularly those with acute thrombolytic therapy (TT) i.e. IV tPA and endovascular thrombectomy (EVT). Methods: We analyzed the National Inpatient Sample (90% US hospitalizations) from 2006 - 2015 and identified adult (age ≥18) IS patients, with or without TT and PC using validated ICD-9 codes. We used survey design methods to report nationally representative adjusted odds ratios (aOR) of PC utilization across five 2-year time periods. Results: Of 4,249,201 IS encounters, 4% were coded for PC utilization. PC utilization increased over time (aOR 4.80, 95% CI 4.15 - 5.55), regardless of acute treatment type (Figure 1a). Advanced age, female sex, White race, non-Medicare insurance, higher income, disease severity and comorbidity burden, and discharge from teaching hospitals in non-Northeastern regions were independently associated with receiving PC. In the fully adjusted model, treatment with IV tPA and EVT respectively confer a 5% and 10% greater likelihood of receiving PC. Overall, PC utilization is associated with higher mortality (aOR 15.14, 95% CI 14.36 - 15.96). However, 10-year outcome trends demonstrate a significant decline in in-hospital mortality compared to all other dispositions (aOR 0.46, 95% CI 0.38 - 0.56) with an increasing proportion of PC patients receiving long-term care and home health/hospice (aOR 2.42, 95% CI 2.03 - 2.88) (Figure 1b). Despite longer length of stay, PC hospitalizations incurred 16% lower adjusted costs. Conclusions: Though PC utilization has increased in IS patients, considerable disparities exist. Acute TT is associated with higher PC utilization and outcome trends indicate lower in-hospital mortality with higher long-term care and home health/hospice. Protocolized PC for IS patients may optimize outcomes and costs.

Background: In the US, one third of strokes occur among younger adults (< 65 years). While stroke incidence among elderly has declined, trends in younger adults by age, sex and region are unknown. We report recent trends in stroke incidence among commercially-insured young adults aged 20 - 64. Methods: Stroke incidence was estimated using 2011 - 2017 IBM MarketScan® Commercial Database stratified by age (20 - 34, 35 - 44, 45 - 54, 55 - 64), sex, and US Census region and was extrapolated to the US population using census data. Stroke cases were identified based on a primary diagnosis of either ischemic stroke (IS) or intracerebral hemorrhage (ICH) or subarachnoid hemorrhage (SAH) using validated ICD-9 / 10 codes. We report incidence rates (IR) per 100,000 person-years, 95% confidence intervals and proportional change in IR between 2011 and 2017. Results: We identified 115,276 stroke admissions resulting in 172 million person-years. The overall stroke IR increased by 7% from 60.0 to 64.2 per 100,000 person-years between 2011 and 2017. The increase was greater among males as compared to females (9.8% vs 3.3%) and across both sexes the largest increase was observed for age category of 35 - 44 years (Table 1). Age-specific IR increased in all age categories except 20 - 34. Across US regions, the South showed the highest IR increase (10.0%) (Figure 1). IR for IS and ICH increased while a decrease in SAH IR was observed. The IR for IS was higher for males vs females (12.7% vs 5.4%), whereas it was higher for females vs males for ICH (14.5% vs 3.6%). Conclusions: Our analyses indicate an overall increase in stroke incidence for commercially-insured younger adults during a contemporary time period. Males between the ages of 35 - 44 have experienced the greatest increase in stroke incidence. Furthermore, the Southern US continues to experience highest increases in stroke incidence and there are gender differences in IR for stroke subtypes. Targeted strategies for stroke prevention may be warranted.

Background: Higher incidence of mild cognitive impairment and dementia (MCID) has been reported in patients with acute neurological injuries (ANI) who are exposed to in-hospital delirium. However, race and sex differences for delirium associated MCID have not been described. Methods: We analyzed State Inpatient and Emergency Department Databases for NY, FL and CA (2006 - 2014) for four categories of adult (≥18) ANI patients (ischemic stroke, intracerebral hemorrhage, sub-arachnoid hemorrhage and traumatic brain injury). Delirium and MCID were tagged using validated diagnostic codes. Patients diagnosed with MCID prior to or within 90-day after ANI event were excluded. We performed time-to-event analyses to explore race and sex differences in development of MCID among delirious ANI patients. Interaction of sex and race with age was also examined. Hazard Ratios (HR) and 95% CIs are reported. Results: We identified 911,380 eligible ANI patients of whom 5.4% were documented to have delirium. After excluding patients without follow-up, we analyzed a total of 34,990 ANI patients with initial in-hospital delirium, among whom 11.5% subsequently developed MCID. In the fully adjusted multivariable model, males had a higher risk of developing MCID (HR, 95% CI: 1.09, 1.01 - 1.16). African Americans, Hispanics and patients in other racial groups also had a higher independent risk of developing MCID, as compared to Caucasians. (Figure 1a). The interaction between age and sex was statically significant (p = 0.04), indicating higher risk of incident MCID among males with increasing age as compared to females. This difference was pronounced beyond 65 years. (Figure 1b). Conclusions: Certain demographic groups hospitalized for ANI may be differentially susceptible to the risk of long-term cognitive decline associated with in-hospital delirium. Targeted and personalized strategies for the early identification and management of delirium during hospitalization may be warranted.

Background: Prior observational studies suggest lower in-hospital mortality (IHM) among intracerebral hemorrhage (ICH) patients treated at higher levels of care (LOC). However, comparisons across non-certified (NC), primary certified (PSC), and comprehensive certified centers (CSC) have not been reported; particularly accounting for changing certification status. Also, unmeasured confounding in observational designs can bias estimates. We performed an instrumental variable (IV) analysis of population level TX state data for association between IHM and LOC. Methods: We analyzed TX in-patient discharges for years 2010 - 2015. Adult, non-traumatic and non-transferred ICH discharges were tagged using ICD-9 codes. Patients’ location was estimated using population weighted centroid for their zip code. Hospitals’ location was geo-coded and certification status (by quarter) was determined from publicly available data. The differential distance of the patient to their closest certified hospital vs closest NC hospital was used as a ‘randomization’ variable (IV). We tested for IV strength and its association with risk of mortality. The IV model was a 2-stage least squares regression (2-SLSR). We report exponentiated IV estimator (IVE) and 95% CI. Results: Based on our criteria and non-missing data, 18,178 discharges were included, mean (SD) age 67.7 (15.5), 50.2% female, 60.0% White and 27.7% Hispanic. Overall, 46.1% discharges were from certified hospitals and IHM was 24.6%. The 2-SLSR demonstrated a significantly lower IHM for ICH discharges from certified hospitals as compared to NC hospitals (IVE, CI: 0.80, 0.74 - 0.97). However, this effect was not observed for comparison between PSC and CSC hospitals (IVE, CI: 0.98, 0.82 - 1.17). The F-static for IV strength was 118 and 249 (recommended > 10). The IV was neither associated with outcome (IHM) nor with measured confounders (severity of illness, age). Conclusion: In a state with large urban - rural divide, IHM reduction was associated with care of ICH patients at certified hospitals. However, similar benefits were not observed across two levels of certification. Factors leading to reduced IHM at certified hospitals need to be studied further and incorporated in health systems redesign for ICH patients.

Background: The Nationwide Inpatient Sample (NIS) represents 90% of US hospitalizations. Beginning fourth quarter (Q4) 2015, NIS transitioned to ICD 10. ICD 10 codes for Intracerebral Hemorrhage (ICH) are hemorrhage location specific, which may predict outcomes. We examine the utilization of ICD 10 codes for ICH patients across various demographic, hospital, and disease severity factors in Q4 2015 and 2016 NIS. We report factors associated with use of non-specific codes (NSC) and assess potential coding validation based on expectations of in-hospital mortality (IHM). Methods: We used ICD 10 codes (i610-16, i618-19) to identify non-traumatic ICH discharges. We categorized hemorrhage locations as (1) Hemispheric (cortical or subcortical) [i610-12], (2) Brain Stem or Cerebellar [i613-14], (3) Intraventricular Hemorrhage (IVH) [i615], (4) Multiple Localized [i616], and (5) Other and Not Specified [i618-19]. We considered ‘Other and Not Specified’ category as NSC and fit survey design logistic regression models for factors associated with NSC. Results: We identified 79, 290 ICH discharges of which 38.9% were NSC. Proportion of discharges by lCH location and corresponding IHM is shown in Figure 1. Highest IHM was observed for IVH. In the fully adjusted model advanced age and white race (vs. African American) were independently associated with higher NSC use. ICH discharges from urban teaching hospitals were less likely to be NSC as compared to rural hospitals (OR, 95% CI: 0.7, 0.6 - 0.8). Patients discharged from large and medium size hospitals were 28% and 17% more likely to be provided a specific code as compared to smaller hospitals. Hospitals in Northeast US were more likely to use NSC. Conclusion: ICD 10 codes provide an opportunity for risk adjustment in administrative data for ICH, we demonstrate that up to 40% of discharges had NSC which is more prevalent in smaller/rural/Northeast hospitals. Our work sets a foundation for examining bias caused by NSC in future studies.

Background: Healthcare costs related to delirium in the elderly and in the post-operative patient populations exceed $150 billion a year. Stroke patients are at a higher risk of developing delirium, which can significantly increase their healthcare costs. Incremental hospitalization and post-discharge costs associated with delirium have not been described in stroke patients. Methods: We analyzed Nationwide Readmission Data from 2010 - 2015 and identified adult (≥18) ischemic and hemorrhagic stroke patients. In-hospital delirium was identified using a validated algorithm. Patients who died during index hospitalization were excluded. Costs for index hospitalization and readmissions were determined using cost to charge ratios, and inflation adjusted to 2015 USD. We used generalized linear models to compare costs between delirious and non-delirious patients adjusting for demographics, comorbidities, illness severity, treatment intensity, in-hospital complications, and surgical procedures. Results: A total of 2,853,175 stroke discharges were included among whom 7.1% were documented to havein-hospital delirium on index admission. Delirious patients had a longer mean length of stay (12.01 vs 7.38 days), and higher 30-day readmission rates (16.7% vs 12.2%). At index admission, delirious patients had 9.9% higher adjusted costs of hospitalization as compared to non-delirious patients. For all admissions (including readmissions), delirious stroke patients had 12.3% higher adjusted costs of hospitalization, which translates into an additional $10,544.2 (2015) per hospitalization. (Figure) Conclusion: Our analyses highlight the potential cost savings in stroke care associated with prevention, recognition, early mitigation and effective management of delirium. Even though delirium prevalence may be underestimated in administrative data, these results indicate potential annual nationwide cost savings of more than $350 million for stroke care.

Background Intracerebral hemorrhage is a devastating disease with no specific treatment modalities. A significant proportion of patients with intracerebral hemorrhage are transferred to large stroke treatment centers, such as Comprehensive Stroke Centers, because of perceived need for higher level of care. However, evidence of improvement in patient-centered outcomes for these patients treated at larger stroke treatment centers as compared to community hospitals is lacking. Methods / design “Efficient Resource Utilization for Patients with Intracerebral Hemorrhage (EnRICH)” is a prospective, multisite, state-wide, cohort study designed to assess the impact of level of care on long-term patient-centered outcomes for patients with primary / non-traumatic intracerebral hemorrhage. The study is funded by the Texas state legislature via the Lone Star Stroke Research Consortium. It is being implemented via major hub hospitals in large metropolitan cities across the state of Texas. Each hub has an extensive network of “spoke” hospitals, which are connected to the hub via traditional clinical and administrative arrangements, or by telemedicine technologies. This infrastructure provides a unique opportunity to track outcomes for intracerebral hemorrhage patients managed across a health system at various levels of care. Eligible patients are enrolled during hospitalization and are followed for functional, quality of life, cognitive, resource utilization, and dependency outcomes at 30 and 90 days post discharge. As a secondary aim, an economic analysis of the incremental cost-effectiveness of treating intracerebral hemorrhage patients at higher levels of care will be conducted. Discussion Findings from EnRICH will provide much needed evidence of the effectiveness and efficiency of regionalized care for intracerebral hemorrhage patients. Such evidence is required to inform policy and streamline clinical decision-making.

Introduction: Evidence of association between delirium during initial hospitalization of ischemic stroke (IS) patients, and development of long-term mild cognitive impairment or dementia (MCID) is lacking. We examined the prospective incident risk of MCID among IS patients who experienced in-hospital delirium in a large multi-state population-based cohort. Methods: We utilized NY, CA, and FL (2005 - 2014) State Inpatient and Emergency Department Databases, and established a cohort by selecting MCID free patients with a primary diagnosis of IS (ICD-9: 433.x1, 434.x1, 436) during a 2yr period. Delirium during the initial IS event was tagged using a validated algorithm with high specificity for the confusion assessment method. The cohort was followed for MCID diagnoses. Patients that died within 90 days of initial IS event were excluded. We conducted time-to-event analyses and report cumulative incidence, 95% Confidence Interval (CI) and hazard ratios (HR) for risk of MCID among IS patients with and without delirium. Results: A total of 133,815 IS patients (age: 70.5, female: 51.7%, white: 65.5%) were followed for up to 8yrs, resulting in 357,500 person-years. Patients who experienced delirium were significantly older and had a higher comorbidity burden. A significantly smaller proportion of delirium patients underwent IV or IA thrombolysis. Cumulative incidence (95% CI) for development of MCID was significantly higher among delirium patients (9.01, 8.32 - 9.74) as compared to non-delirium patients (3.99, 3.93 - 4.06). Likewise, the HR for MCID among delirium patients was statistically significant (1.86, 1.71 - 2.02) in the fully adjusted Cox proportional model (Figure). Conclusion: In-hospital delirium was independently associated with a higher risk of developing MCID in this large population-based cohort. Further investigation is warranted to elucidate the mechanistic basis for the role of delirium in long-term cognitive decline among IS patients.

Introduction: Recent data suggest that early high systolic blood pressure variability (HSPBV) is associated with poor long term outcomes in intracerebral hemorrhage (ICH) patients. We employed a Machine Learning (ML) based approach to predict HSBPV during hospitalization. Methods: Adult radiologically confirmed ICH patients were enrolled in a multisite cohort. A semi-automated algorithm extracted systolic blood pressure (SBP) data from electronic medical records (EMR) and linkage to database with demographic, clinical, and outcomes information was created. Pre-hospital and early admission variables were used to develop predictive algorithms. Patients who expired during hospitalization were excluded. Generalized estimating equations quantified inter and intra-patient SBPV. An 80/20 data split was used to train and test predictive models for HSBPV. Two supervised ML models were trained using a repeated cross-validation bagged classification and regression tree, the first with a complete set of records and the second utilizing imputation (Figure). Results: A total of 455 patients were included with mean(SD) age 63.6 (14.98), females 36.7%, African American 34.5%, and median ICH score 2.0. The average per-patient observation time was 9.7 days resulting in 152,691 SBP readings. A 25 variable imputed model was parameterized (accuracy: 72.2%, sensitivity: 76.5%, specificity: 66.7%, and kappa: 0.43) (Figure). The top influential variables were age, hemorrhage volume, glucose, LDL, and platelets. Methodological and output contrasts with traditional likelihood-based modeling approaches will be presented. Conclusions: We demonstrate that ML-based pipelines can be developed for predicting HSBPV by utilizing SBP values generated from routine monitoring of ICH patients. Future development will incorporate streaming EMR SBP data points during patient stay, and evaluating high-dimensionality reduction methods such as principal component analysis.

Introduction: While there is a documented increased risk of thrombotic events; little is known regarding the risk of intracerebral hemorrhage (ICH) during pregnancy / postpartum period (PP). We explore the risk of ICH in PP utilizing a cohort crossover design in which patients serve as their own control. Methods: We utilized State Inpatient Databases for NY, FL (2005 - 2014) and CA (2005 - 2011), and identified patients with labor/delivery diagnoses using validated ICD-9 codes. The cohort crossover design included a 64wk case period (40wk pregnancy/labor/delivery, 24wk postpartum), followed by a 52wk interim period, and a 64wk crossover period. We excluded patients who were ≤ 12yr, missing linkage information, admitted for false labor, or had a prior ICH. Subsequent pregnancies were not included. We compared the rate of ICH during case and crossover periods and report absolute risk difference and rate ratio with 95% confidence interval (CI) estimated using conditional Poisson regression. Results: Over 3.3 million deliveries were identified, with 238 patients experiencing ICH during the case period. Those who died during the case / interim period or were in PP during the crossover period were removed from analyses. The overall rate of ICH during case vs. crossover period was 8.1 vs 2.5 per 100,000 deliveries / patients, resulting in an absolute risk difference and rate ratio (CI) of 5.5 (4.3 - 6.8) and 3.2 (2.4 - 4.2) respectively. During the case period, the majority of ICH occurred during the 3 rd trimester and the first 12wks postpartum (Figure). Patients experiencing ICH during PP were more likely to be Black or Asian (compared to White), and had a history of hypertension, diabetes, coagulopathy, thrombocytopenia, or substance abuse (Table). Conclusion: Pregnancy confers a significantly higher risk of ICH that peaks during the 3 rd trimester and continues into early postpartum. Further investigation is warranted to characterize individuals at a greater risk of ICH during PP.

Introduction: Little is known regarding the association between acute in-hospital exposures such as delirium and long-term cognitive outcomes in intracerebral hemorrhage (ICH) patients. We sought to explore the risk of mild cognitive impairment or dementia (MCID) among ICH patients who experienced in-hospital delirium. Methods: We utilized State Inpatient and Emergency Department Databases for NY (2006 - 2014), CA (2005 - 2011), and FL (2005 - 2014), and established a cohort by selecting MCID free patients with a primary diagnosis of ICH (ICD-9: 431) during a 2yr period. Patients with concurrent diagnoses of head trauma, arteriovenous malformation or missing linkage information were excluded. Delirium during the initial ICH event was tagged using a validated algorithm with high specificity for the confusion assessment method. The cohort was followed for MCID diagnoses. Patients that died within 90 days of ICH were excluded. We conducted time-to-event analyses and report cumulative incidence, 95% Confidence Interval (CI) and hazard ratios (HR) for risk of MCID among ICH patients with and without delirium. Results: A total of 18,083 ICH patients (age: 67.2, female: 47.6%, white: 59.1%) were included and were followed for up to 8yrs, resulting in 42,100 person-years. Patients who experienced delirium had a significantly higher comorbidity index and intensity of in-hospital treatment. Cumulative incidence (95% CI) for development of MCID was significantly higher among delirium patients (6.07, 5.05 - 7.30) as compared to non-delirium patients (4.07, 3.87 - 4.27). Likewise, the HR for MCID among delirium patients was statistically significant (1.42, 1.17 - 1.73) in the fully adjusted Cox proportional model (Figure). Conclusion: ICH patients with in-hospital delirium are at a significantly higher risk of developing MCID. Further investigation is warranted to understand the mechanism for cognitive decline among ICH patients who experience in-hospital delirium.

Introduction: High in-hospital SBP variability (HSBPV) is an emerging marker for poor outcomes among Intracerebral Hemorrhage (ICH) patients. We aimed to determine the risk of severe disability or death (SDD) at day-90 among ICH patients with HSPBV and explore pre-hospital factors associated with HSPBV. Methods: Adult, radiologically confirmed primary ICH patients were prospectively enrolled and followed-up until day-90. All routinely collected SBP values were recorded for the inpatient stay. Inter and intra-patient SBPV was quantified using generalized estimating equations. Modified Rankin Scale (mRS) Score of 4 - 6 was defined as SDD. Poisson and logistic regression models were fit to determine the risk of day-90 SDD, and the association of pre-hospital characteristics with HSBPV. Results: A total of 566 patients [mean age: 63.5, females 36.6% (207 of 566)] were included. Total in-hospital follow-up period was 4,908 days [median (IQR) per patient = 8.7 (3-11)]. Over 120,500 SBP readings were analyzed. Inter and intra-patient mean SBP standard deviation (SD) was 11.1 and 13.2, respectively. A SD of 13.0 was parameterized as a cut-off for HSBPV. HSBPV patients had a 17% higher adjusted risk of day-90 SDD (Relative Risk, 95% CI: 1.17, 1.02-1.35) (Table). Older age and female sex were independently associated with HSBPV after controlling for hemorrhage volume, pre-morbid mRS, and Glasgow Coma Scale (Figure). Conclusion: Quantification of HSBPV is feasible utilizing routinely collected SBP readings. HSBPV is associated with poor outcomes. Elderly and female patients may be more likely to demonstrate HSBPV during hospitalization.

Introduction: Systolic blood pressure (SBP) regulation is the cornerstone of intracerebral hemorrhage (ICH) management, and SBP variability (SBPV) is associated with poor outcomes. We aimed to determine SBPV patterns and associated factors in a prospective cohort of ICH patients. Methods: Primary ICH patients are consented, assessed in-hospital and followed up at 30 and 90 days. All SBP values and BP management details for the hospital stay were collected from electronic medical records. Day and night intervals were defined as 0601 - 2159hrs and 2200 - 0600hrs, respectively. Mean and standard deviation (SD) were calculated for all intervals in each patient. Overall, between and within patient SBPV for day and night was characterized by generalized estimating equation (GEE)-based methods. We identified patients with a high SBPV (HSBPV) and built logistic regression models to determine associated factors. Results: Thus far, 158 patients have been enrolled, with detailed SBP data was available for 126. Total in-hospital follow-up period is 1,446 days [Median (IQR): 8 (4.25 - 15)]. We analyzed 34,740 SBP readings, yielding 3,010 day/night intervals. The in-hospital mean (SD) SBP was 138.2 (15.6) mmHg. GEE-based estimates for mean night SBP were significantly lower compared to day (137.2 v 139.1 mmHg, p < 0.01). Mean SBP SD was 9.1 mmHg between patients while 13.4 mmHg within patients’ individual readings. HSBPV was defined as SD > 13 mmHg. Age and high admission SBP were independently associated with HSBPV after controlling for sex, race, admission NIHSS, ICH score, hemorrhage volume and Nicardipine infusion use (Figure). Enrollments are continuing and updated data with outcomes will be presented. Conclusion: Characterization of patients with HSBPV can help in triage and management decisions based on risk-stratification. Elderly patients may be at a higher risk of SBPV, warranting exploration of possible SBPV contribution to poor outcomes in the elderly.

Introduction: Readmission (RA) after stroke is an established quality of care metric and is tied to reimbursements. Administrative databases lack stroke-specific severity indicators and information on post-discharge mortality. We report the cumulative incidence of 30-day RA for patients with primary intracerebral hemorrhage (ICH) from a statewide prospective cohort. Methods: Eligible ICH patients are consented to participate in the cohort to assess the impact of level of care on patient-centered outcomes across Texas. Patients undergo inpatient evaluation, followed by 30- and 90-day assessments for functional, cognitive, quality of life, dependency, and resource utilization outcomes. We defined 30-day RA as any RA that was assessed 30 days post-acute care hospitalization for the index ICH event. We used survival analyses to provide hazard ratios (HR) and 95% confidence interval (CI), while modeling post-discharge mortality as a competing risk. Results: Thus far, 158 patients have been enrolled with RA information available for 104. The overall RA rate is 5.1 / 1000 person-days (CI: 3.1-8.3). Among the various factors evaluated (Table 1), 30-day RA is significantly higher for patients with a higher ICH score (HR: 5.67, CI: 1.45-22.19), whereas post-discharge rehabilitation (as compared to discharge to home) appears to reduce the risk of RA, even after accounting for institution-free period of observation (HR: 0.11, CI: 0.01-0.86). Among patients with a high ICH score, those discharged home had significantly higher risk of RA as compared to those who received rehabilitation (Figure 1). Enrollment is continuing; updated analyses will be presented. Conclusion: Stroke-specific disease severity factors are important to identify to develop effective preventive strategies against RA, and need to be controlled for when comparing RA metrics across patient populations. Influence of post-stroke rehab on curtailing RA for ICH patients needs to be explored further.

Introduction: Substance abuse (SA) is a risk factor for stroke. Opioid abuse (OA) can be a precursor for SA, and national prevalence of OA particularly in young stroke patients (YSP) has not been described. We report prevalence and factors associated with OA in YSP. We also explore in-hospital outcomes in YSP with OA. Methods: We identified ischemic stroke, intracerebral hemorrhage, and sub-arachnoid hemorrhage patients and patients abusing Opioids, Cocaine, Alcohol, Cannabis, Amphetamines and Hallucinogens using ICD-9 codes, from the National Inpatient Sample for the years 2006 - 2011. Patients aged 18-45 were categorized as YSP. Logistic regression methods are used. Odds ratios (OR) and 95% Confidence Intervals (CI) are reported. Results: A total of 163 469 YSP were identified. YSP had a significantly higher proportion of SA as compared to non-YSP for all substances (Table 1). Among YSP, 2 320 had been coded for OA. Among OA YSP, 53% patients also abused at least one additional substance. Males were more likely to be opioid abusers (OA) (OR: 1.23, CI: 1.30-1.47) and more OA were in the lower household income group (Table 2). OA was also associated with higher disease severity and mortality risk, likelihood of being on ventilator support, and longer length of stay. Higher proportion of OA was treated at urban and teaching hospitals, and in the Northeast as compared to other national regions. OA had a greater likelihood of in-hospital death (OR: 1.70, CI: 1.24-2.33) and had poorer discharge disposition (Table 1). Conclusions: OA in YSP is associated with greater disease severity, poor in-hospital outcomes including death, and higher utilization of resources. Although OA has not been independently identified as a risk factor for stroke, OA is frequently accompanied by or is a precursor to SA. Routine screening and counselling are required in YSP to prevent recurrence and poor outcomes associated with OA and SA.

Introduction: Intracerebral hemorrhage (ICH) patients are routinely transferred to comprehensive stroke centers (CSCs) for neurosurgical and neurocritical care. We compared transferred (TP) and directly presenting (DP) ICH patients at our CSC, and explore the factors associated with non-utilization of CSC Services (NCS). Methods: We identified primary ICH patients, admitted between 01/01/2016 and 03/31/2017, from our Stroke Registry. We used logistic regression to compare demographics, disease severity, and outcomes between TP and DP, and report odds ratios (OR) and 95% confidence intervals (CI). We categorized patients who did not stay in the neurocritical care unit and did not undergo neurosurgical procedures (including extra-ventricular drain) as NCS patients. We used receiver operative curve (ROC) analyses to determine the discriminatory potential of routinely used severity scales in identifying NCS patients and report area under the curve (AUC). Results: We included 958 patients in our analyses. TP had significantly lower disease severity and shorter length of stay. Overall, 33.7% of patients were NCS, and NCS patients were more likely to be TP as compared to DP [OR (CI): 1.60 (1.18-2.16)]. NCS patients also had a significantly lower median National Institutes of Health Stroke Scale (NIHSS) and ICH scores, and higher median Glasgow Coma Scale (GCS) score on presentation (Table 1). All three scales had a fair-good individual discrimination for classifying NCS patients (AUC for GCS, NIHSS, ICH Score: 0.71, 0.77, and 0.80 respectively). After dichotomizing GCS at 10 and categorizing NIHSS at 0-5 / 6-15 / 16+, the combined AUC for all three scales was 0.84 (Figure 1). Conclusion: A third of ICH patients presenting at CSC do not utilize neurosurgical / neurocritical care. Identification and triage of these patients may help optimize ICH care. Disease severity scales may be helpful in classification of these patients. Further validation studies are warranted.

Introduction: Value-based care and patient-perceived outcomes are increasingly important. Standard Gamble (SG) derived utilities directly measure patients’ preferences for health states and form the basis of health economic analyses. We describe distribution of and factors associated with SG utilities (SGU) in a cohort of intracerebral hemorrhage (ICH) patients, and explore changes in SGU over 90 days post-discharge. Methods: Our study is a multisite cohort aiming to evaluate the comparative effectiveness of ICH patient management at various levels of care across Texas. Consented patients undergo assessments including SG in-hospital, and 30 and 90 days post-discharge. The SG assesses patients’ risk-taking behavior toward achieving a perfect health status, and outputs utility on a scale of 0 - 1 (Figure 1). Median and interquartile range (IQR) are reported for inpatient and day-90 SGU. Quantile regression was used to evaluate factors associated with SGU. Difference in median (DIM) and 95% confidence interval (CI) for the difference are reported. Results: 158 patients have been enrolled. Inpatient and day-90 SG was obtained from 132 and 54 patients respectively. Median inpatient SGU are significantly lower for older patients, white patients (compared to black patients), and those with higher ICH scores (Table 1). Median day-90 SGU was higher than inpatient SGU (DIM: 0.27; 95% CI: 0.08-0.46). Age >65 and higher ICH score were independently associated with lower SGU (Age: DIM -0.30; 95% CI -0.49, -0.11) (ICH Score: DIM -0.59; 95% CI -0.97, -0.21). Patients with mRS scores 0-3 at day 90 had higher SGU values compared to those with mRS scores 4-5 (DIM: 0.25, 95% CI: 0.09 - 0.41). Enrollment continues; updated analyses to be presented. Conclusion: Direct assessment of preferences for a morbid condition like ICH provides unique insight into patient values. Assessment of SGU may be routinely conducted in ICH patient studies to generate evidence for comparative effectiveness.