Telemedicine impact on post-stroke outpatient follow-up in an academic healthcare network during the COVID-19 pandemic.

BACKGROUND: The expansion of telemedicine associated with the COVID-19 pandemic has influenced outpatient medical care. The objective of our study was to determine the impact of telemedicine on post-acute stroke clinic follow-up. METHODS: We retrospectively evaluated the impact of telemedicine in Emory Healthcare, an academic healthcare system of comprehensive and primary stroke centers in Atlanta, Georgia, on post-hospital stroke clinic follow-up. We compared the frequency of 90-day follow-up in a centralized subspecialty stroke clinic among patients hospitalized before the local COVID-19 pandemic (January 1, 2019- February 28, 2020), during (March 1- April 30, 2020) and after telemedicine implementation (May 1- December 31, 2020). A comparison was made across hospitals less than 1 mile, 10 miles, and 25 miles from the stroke clinic. RESULTS: Of 1096 ischemic stroke patients discharged home or to a rehab facility during the study period, 342 (31%) had follow-up in the Emory Stroke Clinic (comprehensive stroke center 46%, primary stroke center 10 miles away 18%, primary stroke center 25 miles away 14%). Overall, 90-day follow-up increased from 19% to 41% after telemedicine implementation (p<0.001) with telemedicine appointments amounting for up to 28% of all follow-up visits. In multivariable analysis, factors associated with teleneurology follow-up (vs no follow-up) included discharge from the comprehensive stroke center, thrombectomy treatment, private insurance, private transport to the hospital, NIHSS 0-5 and history of dyslipidemia. CONCLUSIONS: Despite telemedicine implementation at an academic healthcare network successfully increasing post-stroke discharge follow-up in a centralized subspecialty stroke clinic, the majority of patients did not complete 90-day follow-up during the COVID-19 pandemic.

Accuracy of Prehospital Identification of Stroke in a Large Stroke Belt Municipality.

OBJECTIVE: Strokes are a leading cause of morbidity and mortality in the United States, especially in the "stroke belt" of the southeast. Up to 65% of stroke patients access care by calling 9-1-1. The primary objective of this study is to measure the accuracy of emergency medical dispatchers (EMD) and paramedics, in the prehospital identification of stroke. METHODS: The study was based at Grady Emergency Medical Services, which is Atlanta, Georgia's public emergency medical services (EMS) provider. A retrospective analysis of all medically related 9-1-1 calls to Grady EMS classified as "stroke" between January 1, 2012, and December 31, 2012 was performed. A database was created using deterministic linkage between records from Grady EMS, Grady Hospital Emergency Department (ED), and the Grady Hospital Stroke Registry. Patients excluded were less than 18 years of age, had previous or concurrent head injuries, were transferred from another inpatient facility, and/or had incomplete patient records in any one of the three databases. Descriptive analysis, linear regression, and logistic multivariable regression were performed to discover the accuracy of stroke identification and contributory prehospital factors. RESULTS: A total of 548 patients were included: 475 were transported with EMS impression of stroke and 73 with an impression other than stroke. The median age was 59 years, 87.4% were black, and 52.6% were female. Paramedics adhered to all seven elements of the Grady EMS stroke protocol in 76.4% (n = 363) of suspected stroke cases. Sensitivity and positive predictive value for paramedic stroke identification was 76.2% and 49.3%, respectively, and for EMD, was 48.9% and 24%, respectively. Identification of hemorrhagic strokes had a relatively lower sensitivity. Paramedics were more likely to positively identify strokes when the Cincinnati Prehospital Stroke Scale (CPSS) screen was positive, or when classified by EMD as stroke. Paramedics were less likely to identify stroke in female patients. Paramedics' diagnostic accuracy was similar regardless of their adherence to the EMS stroke care protocol. CONCLUSIONS: EMD and EMS personnel in a large city in the Southeastern United States, with high stroke prevalence, had a relatively high sensitivity in identifying acute stroke patients. Paramedic accuracy was augmented by positive CPSS screening and by EMD recognition of stroke.

The impact of IV alteplase on long-term patient survival: The Georgia Coverdell acute stroke registry's experience.

INTRODUCTION: Intravenous alteplase reduces disability and improves functionality among acute ischemic stroke patients. Two decades after its approval, only a small fraction of patients get the treatment, and demonstrating its impact on mortality may make a strong case for its wider use. This study assessed the impact of thrombolytic treatment by alteplase on 1-year mortality and readmission among acute ischemic stroke patients. METHOD: The 2008-2013 Georgia Coverdell Acute Stroke Registry data were linked with the 2008-2013 hospital discharge and the 2008-2014 death data in Georgia. Multiple imputation was applied; a propensity score measuring the probability of receiving intravenous alteplase was calculated and used for matching. A conditional logistic regression was applied to compare 1-year mortality and readmission among propensity score matched pairs. RESULTS: Overall, 20.3% of 9620 acute ischemic stroke patients died and 22.4% were readmitted in one year. The multivariable regression result showed that patients who did not receive IV alteplase had a 1.49 (95%CI: 1.09-2.04; p-value=0.01) times higher odds of dying at one year than those who were treated with the thrombolytic agent. Among patients discharged home, no statistically significant difference was documented in the odds of being readmitted at least once within 365days post-stroke discharge. DISCUSSION AND CONCLUSION: After accounting for patient differences and missing value, intravenous alteplase is associated with reduction in long-term mortality. The results of this study suggest that patients who are identified as eligible for intravenous alteplase need to be offered the treatment.

Door to Intravenous Tissue Plasminogen Activator Time and Hospital Length of Stay in Acute Ischemic Stroke Patients, Georgia, 2007-2013.

BACKGROUND: Ischemic stroke patients benefit most from intravenous thrombolysis when they receive the treatment as quickly as possible after symptom onset. Hospitals participating in the Georgia Coverdell Acute Stroke Registry reduced the time from patient arrival to administration of intravenous tissue plasminogen activator. This study evaluates the benefit of reducing door-to-treatment (DTT) time as measured by hospital length of stay (LOS). METHODS: Data from 3154 ischemic stroke patients treated with intravenous thrombolysis from 2007 to 2013 were analyzed. The impact of door-to-treatment time on patients' length of hospital stay, discharge disposition, ambulatory status at discharge, and bleeding complications was assessed, controlling for patient-, hospital- and event-related characteristics. RESULTS: Patients who received intravenous thrombolysis within 30 minutes of hospital arrival had a 19% shorter (95% confidence interval [CI]: 2%-32%, P value = .04) hospital LOS than those treated for more than 120 minutes after arrival. Patients treated within 60 minutes of arrival were 27% more likely (odds ratio = 1.28, 95% CI: 1.06-1.56, P = .01) to have a better discharge disposition than patients treated after 60 minutes of arrival while having a similar rate of bleeding complications. CONCLUSIONS: Shortening the door-to-treatment time is associated with a decrease in patient LOS and better patient outcomes. Hospitals should be encouraged to measure, monitor, and reduce DTT time progressively for a better patient outcome.

Presenting Symptoms and Dysphagia Screen Predict Outcome in Mild and Rapidly Improving Acute Ischemic Stroke Patients.

OBJECTIVE: There are limited data on which patients not treated with intravenous (IV) tissue-type plasminogen activator (tPA) due to mild and rapidly improving stroke symptoms (MaRISS) have unfavorable outcomes. MATERIALS AND METHODS: Acute ischemic stroke (AIS) patients not treated with IV tPA due to MaRISS from January 1, 2009 to December 31, 2013 were identified as part of the Georgia Coverdell Acute Stroke Registry. Multivariable regression analysis was used to identify factors associated with a lower likelihood of favorable outcome, defined as discharge to home. RESULTS: There were 1614 AIS patients who did not receive IV tPA due to MaRISS (median National Institutes of Health stroke scale [NIHSS] 1], of which 305 (19%) did not have a favorable outcome. Factors associated with lower likelihood of favorable outcome included Medicare insurance status (odds ratio [OR]: .53, 95% confidence interval [CI]: .34-.84), arrival by emergency medical services (OR: .46, 95% CI: .29-.73), increasing NIHSS score (per unit OR: .89, 95% CI: .84-.93), weakness as the presenting symptom (OR: .50, 95% CI: .30-.84), and a failed dysphagia screen (OR: .43, 95% CI: .23-.80). During the study period, <1% of patients presenting to participating hospitals with MaRISS within the eligible time window received IV tPA. CONCLUSIONS: Baseline characteristics, presenting symptoms, and response to dysphagia screen identify a subgroup of patients who are more likely to have an unfavorable outcome. Whether IV tPA treatment can improve the outcome in this subgroup of patients needs to be evaluated in a randomized placebo-controlled trial.

Administrative data linkage to evaluate a quality improvement program in acute stroke care, Georgia, 2006-2009.

INTRODUCTION: Tracking the vital status of stroke patients through death data is one approach to assessing the impact of quality improvement in stroke care. We assessed the feasibility of linking Georgia hospital discharge data with mortality data to evaluate the effect of participation in the Georgia Coverdell Acute Stroke Registry on survival rates among acute ischemic stroke patients. METHODS: Multistage probabilistic matching, using a fine-grained record integration and linkage software program and combinations of key variables, was used to link Georgia hospital discharge data for 2005 through 2009 with mortality data for 2006 through 2010. Data from patients admitted with principal diagnoses of acute ischemic stroke were analyzed by using the extended Cox proportional hazard model. The survival times of patients cared for by hospitals participating in the stroke registry and of those treated at nonparticipating hospitals were compared. RESULTS: Average age of the 50,579 patients analyzed was 69 years, and 56% of patients were treated in Georgia Coverdell Acute Stroke Registry hospitals. Thirty-day and 365-day mortality after first admission for stroke were 8.1% and 18.5%, respectively. Patients treated at nonparticipating facilities had a hazard ratio for death of 1.14 (95% confidence interval, 1.03-1.26; P = .01) after the first week of admission compared with patients cared for by hospitals participating in the registry. CONCLUSION: Hospital discharge data can be linked with death data to assess the impact of clinical-level or community-level chronic disease control initiatives. Hospitals need to undertake quality improvement activities for a better patient outcome.

Remote assessment of stroke using the iPhone 4.

Therapy with recombinant tissue plasminogen activator is underused in the treatment of ischemic stroke in rural hospitals, due to a lack of local stroke expertise. Telemedicine solutions for stroke are a level I, class A recommendation when a vascular neurologist is absent. However, current solutions require exorbitant startup costs, which are prohibitive for the rural hospitals in which they are needed most. This study demonstrates the efficacy of using the relatively inexpensive iPhone 4 in telestroke management. Twenty patients with stroke were assessed at the bedside using an iPhone 4, and each examination was directed remotely on another iPhone 4. Both the physician performing the bedside exam and the remote physician calculated a National Institutes of Health Stroke Scale (NIHSS) score for each patient. Each physician was blinded to the other's NIHSS score. In the 20 patients assessed, NIHSS scores ranged from 0 to 22. Interrater reliability assessed using the κ statistic demonstrated excellent agreement in 10 items (level of consciousness, month and age, visual fields, right motor arm, left motor arm, right motor leg, left motor leg, sensation, language, and neglect), moderate agreement in 3 items (gaze, facial palsy, and dysarthria), and poor agreement in 1 item (ataxia). Total NIHSS scores obtained remotely and at bedside showed an excellent level of agreement (intraclass correlation coefficient, 0.98). Our findings indicate that the iPhone 4 is an economical mobile solution that can be used to assess stroke patients remotely with high fidelity and can be readily incorporated into a telestroke network.

Factors associated with stroke after COVID-19 vaccination: a statewide analysis.

Background: The objective of our study was to evaluate vaccine type, COVID-19 infection, and their association with stroke soon after COVID-19 vaccination. Methods: In a retrospective cohort study, we estimated the 21-day post-vaccination incidence of stroke among the recipients of the first dose of a COVID-19 vaccine. We linked the Georgia Immunization Registry with the Georgia Coverdell Acute Stroke Registry and the Georgia State Electronic Notifiable Disease Surveillance System data to assess the relative risk of stroke by the vaccine type. Results: Approximately 5 million adult Georgians received at least one COVID-19 vaccine between 1 December 2020 and 28 February 2022: 54% received BNT162b2, 41% received mRNA-1273, and 5% received Ad26.COV2.S. Those with concurrent COVID-19 infection within 21 days post-vaccination had an increased risk of ischemic (OR = 8.00, 95% CI: 4.18, 15.31) and hemorrhagic stroke (OR = 5.23, 95% CI: 1.11, 24.64) with no evidence for interaction between the vaccine type and concurrent COVID-19 infection. The 21-day post-vaccination incidence of ischemic stroke was 8.14, 11.14, and 10.48 per 100,000 for BNT162b2, mRNA-1273, and Ad26.COV2.S recipients, respectively. After adjusting for age, race, gender, and COVID-19 infection status, there was a 57% higher risk (OR = 1.57, 95% CI: 1.02, 2.42) for ischemic stroke within 21 days of vaccination associated with the Ad26.COV2.S vaccine compared to BNT162b2; there was no difference in stroke risk between mRNA-1273 and BNT162b2. Conclusion: Concurrent COVID-19 infection had the strongest association with early ischemic and hemorrhagic stroke after the first dose of COVID-19 vaccination. Although not all determinants of stroke, particularly comorbidities, were considered in this analysis, the Ad26.COV2.S vaccine was associated with a higher risk of early post-vaccination ischemic stroke than BNT162b2.

Impact of Intravenous Alteplase Door-to-Needle Times on 2-Year Mortality in Patients With Acute Ischemic Stroke.

Objective: We sought to determine whether administration of Intravenous Thrombolysis (IVT) to patients with Acute Ischemic Stroke (AIS) within 60 min from hospital arrival is associated with lower 2-year mortality. Methods: This retrospective study was conducted among patients receiving IVT in hospitals participating in the Georgia Coverdell Acute Stroke Registry (GCASR) from January 1, 2008 through June 30, 2018. Two-year mortality data was obtained by linking the 2008-2018 Georgia Discharge Data System data and the 2008-2020 Georgia death records. We analyzed the study population in two groups based on the time from hospital arrival to initiation of IVT expressed as Door to Needle time (DTN) in a dichotomized (DTN ≤ 60 vs. > 60 min) fashion. Results: The median age of patients was 68 years, 49.4% were females, and the median NIHSS was 9. DTN ≤60 min was associated with lower 30-day [odds ratio (OR), 0.62; 95% CI, 0.52-0.73; P < 0.0001], 1-year (OR, 0.71; 95% CI, 0.61-0.83; P < 0.0001) and 2-year (OR, 0.76; 95% CI, 0.65-0.88; P = 0.001) mortality as well as lower rates of sICH at 36 h (OR, 0.57; 95% CI, 0.43-0.75; P = 0.0001), higher rates of ambulation at discharge (OR, 1.38; 95% CI, 1.25-1.53; P < 0.0001) and discharge to home (OR, 1.36; 95% CI, 1.23-1.52; P < 0.0001). Conclusion: Faster DTN in patients with AIS was associated with lower 2-year mortality across all age, gender and race subgroups. These findings reinforce the need for intensifying quality improvement measures to reduce DTN in AIS patients.

Impact of Participation in a Telestroke Network on Clinical Outcomes.

BACKGROUND: A telestroke program, known as the Remote Evaluation for Acute Ischemic Stroke program, has been implemented in Georgia since 2003. This study examined whether a hospital's participation in a telestroke network was associated with improvement in clinical outcomes and quality indicators. METHODS AND RESULTS: An observational study was conducted using data from the Georgia Coverdell Acute Stroke Registry between September 2005 and September 2016 for patients aged ≥18 years with ischemic stroke. We use a difference-in-differences approach to compare the following clinical outcomes and quality indicators among those admitted at hospitals within and outside of the Remote Evaluation for Acute Ischemic Stroke network: tPA (tissue-type plasminogen activator) use, complications related to tPA use, door-to-needle time, ambulation at discharge, discharge status, and destination. Logistic regression models and a propensity score weighting approach were performed to adjust for patients' age, sex, race, insurance coverage, arrival mode, ambulatory status before the current stroke, stroke severity, medical history, admission time, and hospital bed size. A total of 25 494 patients with ischemic stroke admitted at 15 nonteaching hospitals located outside of the Atlanta metropolitan area were included in the analysis. After propensity score weighting, hospitals participated in a telestroke network was not associated with a significant increase in the rate of tPA use, while it was significantly associated with a modest decline in the rate of complications related to tPA (-5.9%; 95% CI, -9.2% to -2.6%). Telestroke participation showed no significant difference in other clinical outcomes and quality measures except for a marginally significant decrease in in-hospital mortality (-1.1%; 95% CI, -2.2% to -0.1%). CONCLUSIONS: Although a slight decrease in tPA complication was observed among hospitals participating in the telestroke network, overall the impact of telestroke participation on a hospital's stroke care quality was not statistically significant based on our observational study.

Quality of Care and Its Impact on One-Year Mortality: The Georgia Coverdell Acute Stroke Registry.

Although performance measures help monitor the process change in quality improvement, their utility in measuring long-term outcomes is uncertain. This study assessed the 1-year mortality of acute ischemic stroke patients treated by hospitals participating in the Georgia Coverdell Acute Stroke Registry. Using 10 nationally approved performance measures, quality of care was defined both as an all-or-none measure (defect-free care) and as a composite index. A generalized estimating equation was applied to assess the effect of quality of care on 1-year mortality. Defect-free care did not serve the purpose; however, the composite measure showed that patients who received the lowest and intermediate quality care, respectively, had a 3.94 (95% confidence interval: 3.27, 4.75; P < .0001) and a 1.38 (95% confidence interval: 1.12, 1.62; P = .002) times higher odds of dying in 1 year compared to those who got the best-quality stroke care. Therefore, hospitals should be encouraged to implement quality improvement activities for better long-term patient outcome.

Multiple Imputation for General Missing Data Patterns in the Presence of High-dimensional Data.

Multiple imputation (MI) has been widely used for handling missing data in biomedical research. In the presence of high-dimensional data, regularized regression has been used as a natural strategy for building imputation models, but limited research has been conducted for handling general missing data patterns where multiple variables have missing values. Using the idea of multiple imputation by chained equations (MICE), we investigate two approaches of using regularized regression to impute missing values of high-dimensional data that can handle general missing data patterns. We compare our MICE methods with several existing imputation methods in simulation studies. Our simulation results demonstrate the superiority of the proposed MICE approach based on an indirect use of regularized regression in terms of bias. We further illustrate the proposed methods using two data examples.

Hospitalists and intensivists in the medical ICU: a prospective observational study comparing mortality and length of stay between two staffing models.

BACKGROUND: A shortage of critical care specialists or intensivists, coupled with expanding United States critical care needs, mandates identification of alternate qualified physicians for intensive care unit (ICU) staffing. OBJECTIVE: To compare mortality and length of stay (LOS) of medical ICU patients cared for by a hospitalist or an intensivist-led team. DESIGN: Prospective observational study. SETTING: Urban academic community hospital affiliated with a major regional academic university. PATIENTS: Consecutive medical patients admitted to a hospitalist ICU team (n = 828) with selective intensivist consultation or an intensivist-led ICU teaching team (n = 528). MEASUREMENTS: Endpoints were ICU and in-hospital mortality and LOS, adjusted for patient differences with logistic and linear regression models and propensity scores. RESULTS: The odds ratio adjusted for disease severity for in-hospital mortality was 0.8 (95% confidence interval [CI]: 0.49, 1.18; P = 0.23) and ICU mortality was 0.8 (95% CI: 0.51, 1.32; P = 0.41), referent to the hospitalist team. The adjusted LOS was similar between teams (hospital LOS difference 0.9 days, P = 0.98; ICU LOS difference 0.3 days, P = 0.32). Mechanically ventilated patients with intermediate illness severity had lower hospital LOS (10.6 vs 17.8 days, P < 0.001) and ICU LOS (7.2 vs 10.6 days, P = 0.02), and a trend towards decreased in-hospital mortality (15.6% vs 27.5%, P = 0.10) in the intensivist-led group. CONCLUSIONS: The adjusted mortality and LOS demonstrated no statistically significant difference between hospitalist and intensivist-led ICU models. Mechanically ventilated patients with intermediate illness severity showed improved LOS and a trend towards improved mortality when cared for by an intensivist-led ICU teaching team.