Abstract Background and aims Telestroke aims to provide rapid expert evaluation and evidence-based treatment to patients with acute ischemic stroke (AIS) in hospitals without on-site stroke specialists. However, intravenous (IV) thrombolysis and interhospital transfers are often delayed in telestroke-managed patients. We evaluated perceived drivers of these delays by surveying hospitals from a statewide acute stroke registry. Methods The Paul Coverdell Michigan Acute Stroke Registry (MASR) has collected stroke data since 2003 to measure and improve care. Currently MASR includes over 50 hospitals, representing 64% of the state’s stroke cases. Our team developed telestroke-specific questions for the 2025 MASR Hospital Inventory Survey, administered annually to hospital stroke program coordinators. We asked spoke sites (where telestroke call is initiated from) to identify the main causes of delay in 1) IV thrombolytic therapy delivery and 2) interhospital transfer when using telestroke. Results Of 53 MASR hospitals, 46 (87%) participated in the Inventory Survey. Of these, 18 were telestroke spoke sites. The top 3 identified sources of delay in thrombolytic therapy at the spoke sites were: determining patient eligibility (50%), obtaining consent (38.9%), and diagnosing stroke (22.2%). The top three sources of delay in interhospital transfer were: securing interhospital transport (44.4%), determining patient eligibility for transfer (27.8%), and identifying an accepting facility (22.2%). Conclusions A state-level stroke registry survey identified key perceived drivers of thrombolytic and transfer delays within telestroke systems. These results highlight modifiable targets to improve telestroke care and support future work incorporating diverse telestroke provider perspectives. Conflict of interest Brian Stamm: nothing to disclose. Ghada Ibrahim: nothing to disclose. Adrienne Nickles: nothing to disclose. Regina Royan: reported receiving a grant from the National Institute of Neurological Disorders and Stroke (K12NS137516) during the conduct of the study. Rodney Hayward: nothing to disclose. Mollie McDermott: nothing to disclose. Phillip Scott: reported receiving grants from NIH during the conduct of the study. Kevin Sheth: reported receiving grants from NIH during the conduct of the study; grants from Hyperfine, Genentech, and the American Heart Association; and personal fees from Astrocyte, Bexorg, and BrainQ outside the submitted work; and an issued patent for Alva. Mathew Reeves: nothing to disclose. Deborah Levine: reported receiving grants from NIH and consulting fees on NIH grants from Tufts University and Northwestern University outside the submitted work. Figure 1 - belongs to Conclusions

Introduction: There has been an ongoing debate regarding the effectiveness of inpatient rehabilitation facility (IRF) and skilled nursing facility (SNF) in promoting functional recovery. Home time is a valid measure of functional recovery in stroke patients that is often used in outcome studies. Considering that SNF patients have twice the length of stay compared to IRF patients, our objective was to explore whether home time is a suitable measure to compare the effectiveness between IRF and SNF in achieving functional recovery. Methods: We probabilistically linked data from Michigan’s Coverdell Stroke Program and Michigan Value Collaborative multipayer claims database for Medicare FFS beneficiaries hospitalized with acute stroke (ICD-10 I61-I63) between 2016-2020. Patients admitted to IRF or SNF after hospital discharge were confirmed using claims data. Home time was calculated over 90-days and 1-year following hospital discharge by subtracting the number of days spent in inpatient setting (i.e., IRF, SNF, and long-term care) from the number of days alive. We calculated the crude and inverse probability of treatment weighted (IPTW) mean difference of home time between IRF and SNF groups. We conducted a sensitivity analysis to examine the effect of time spent in the same rehabilitation setting over 30-days post discharge on home time. Results: From a cohort of 14,316 linked patients, we identified 2,995 (20.9%) and 2,948 (20.6%) patients directly admitted to IRF or SNF following stroke hospitalization, respectively. Compared to SNF patients, IRF patients were younger, more likely to be male, had minor strokes (NIHSS 1-4), and were able to ambulate at discharge. The unadjusted 90-day and 1-year mean home time were 15.6 and 67.6 days higher among IRF patients compared to SNF patients, respectively (Table). After accounting for rehabilitation time during 30-days post discharge, 90-day and 1-year unadjusted mean difference in home time remained higher among IRF patients compared to SNF patients but was reduced to 4.6 and 56.5 days, respectively. Using the amended home time, the adjusted 90-days mean difference was almost zero and not significantly different (0.5 days) but remained significantly different over 1-year (35.7 days). Conclusions: Home time is heavily impacted by rehabilitation length of stay. Future rehabilitation related studies should be cautious when using home time as a measure of functional recovery, especially over short duration of follow-up.

Introduction: Two-thirds of US stroke patients undergo rehabilitation post-discharge with about 20% and 25% receiving care at an inpatient rehabilitation facility (IRF) and skilled nursing facility (SNF), respectively. Total rehabilitation time for stroke patients discharged to SNF are about twice as long as those of IRF, but the effect of hospital readmission (to acute care) on the continuity of rehabilitation care in either setting is not known. Methods: We probabilistically linked data from Michigan’s Coverdell Stroke Program and Michigan Value Collaborative claims database for Medicare FFS beneficiaries following acute stroke (ICD-10 I61-I63) between 2016-2020. Patients admitted to IRF or SNF after hospital discharge were confirmed using claims data. We followed patients for 30 days post-discharge and compared the all-cause readmission rate, initial rehabilitation length of stay, 30-day total rehabilitation length of stay (in the same setting), and number of admissions to the same rehabilitation setting between IRF and SNF patients. Results: From an initial cohort of 14,316 patients, we identified 2,995 (20.9%) and 2,948 (20.6%) directly admitted to IRF or SNF following stroke hospitalization, respectively. Compared to SNF patients, IRF patients were younger, and more likely to be male, have minor strokes (NIHSS 1-4), and be able to ambulate at hospital discharge. Over 30 days of follow up, 12.6% (n=376) of IRF and 19.6% (n=577) of SNF patients were readmitted at least once to an acute hospital setting (Table). Of the patients who experienced readmission, SNF patients were more likely to be readmitted to a SNF rehabilitation setting compared to IRF patients being readmitted to an IRF setting (mean number of SNF admissions = 2.1 vs mean number of IRF admissions = 1.3). The mean length of stay of the initial IRF and SNF care settings were 14.6 (SD=8.0) and 11.5 (SD=8.2) days, respectively. However, the mean cumulative length of stay in the same rehabilitation setting over the 30-day period increased slightly to 15.3 (SD=8.1) days for IRF patients but increased substantially to 26.4 (SD=21.3) days for SNF patients. Conclusions: Readmission to the acute hospital has a disrupting effect on the continuity of rehabilitation care especially for SNF patients who are less likely to complete their initial rehabilitation stay. However, because most SNF patients return to SNF, the cumulative amount of rehabilitation care is close to the theoretical 30 day maximum.

Introduction: Telestroke has the potential to revolutionize acute stroke treatment by improving access to optimal stroke care, including time-sensitive care such as thrombolysis. Yet few studies have compared acute stroke treatment metrics and outcomes in patients treated using telestroke versus standard in-person stroke evaluation. Methods: This was a retrospective cohort study of acute ischemic stroke patients age ≥18 presenting to 53 Paul Coverdell Michigan hospitals between 2022 and 2023 who were potentially eligible for thrombolysis (i.e., presented ≤ 4 hours of last known well, no contraindications to thrombolysis). The primary exposure was telestroke (vs non-telestroke), and primary outcomes were receipt of thrombolysis and door-to-needle (DTN) time. Secondary outcomes included discharge ambulatory status and door-in-door-out (DIDO) time in transferred patients. Multivariable hierarchical models evaluated associations between the telestroke (vs. non-telestroke) activation and outcomes, sequentially adjusted for demographics, medical history, presenting/arrival, and hospital characteristics. Results: Among the 4974 stroke patients potentially eligible for thrombolysis (mean age 69.2 [SD: 14.6], 48.3% female), 1078 (21.7%) were evaluated using telestroke and 3896 (78.3%) without telestroke. Telestroke patients were more commonly at primary stroke centers (71.1% vs 39.0%) and less at comprehensive stroke centers (13.3% vs 53.9%; P<0.001). Thrombolysis was administered to 56.8% of telestroke patients (at the site of telestroke initiation) versus 54.7% of patients without telestroke (P=0.23). Telestroke patients had longer DTN times (55 vs. 47 minutes, P<0.001), longer DIDO times (166 vs. 142 minutes, P<0.001), and a lower likelihood of ambulating independently at discharge (P<0.001). After adjusting for patient demographics, medical history, and presenting/arrival factors, telestroke patients had significantly longer DTN times (8.1 minutes longer, 95% CI 1.9, 14.3). This difference was attenuated after adjustment for hospital characteristics, including stroke center status. Discussion: Acute stroke treatment metrics, including DTN and DIDO times, were significantly worse in telestroke vs. non-telestroke cases from the Paul Coverdell Michigan stroke registry. Differences in DTN were partially explained by hospital-level systems factors, such as stroke center status, which may serve as targets for future studies and quality improvement initiatives.

Collection of patient-level outcomes data following hospital discharge is challenging for stroke registries. Data linkage to administrative claims data is a potential solution to obtain outcomes data. We aimed to generate data on 30-day, 90-day and 1-year outcome events following hospitalization for stroke using linked data in Michigan. We probabilistically linked clinical data from a 5-year cohort (2016-2020) of all index acute stroke discharges (ICD-10 I61-I63) from 31 hospitals participating in Michigan’s Acute Stroke program (MiSP) to a representative statewide multi-payer claims database. We used the linked data to generate data on 30-day, 90-day, and 1-year event rates including hospital readmissions, stroke recurrence, post-acute care services (i.e., facility-based rehabilitation and home health), and out-patient visits. Mortality data was only available for Medicare fee-for-service beneficiaries. Outcomes were stratified by age, race, stroke type, and stroke severity. Of the 46,330 MiSP stroke discharges, 23,918 (51.6%) were linked to the claims database. Readmission and stroke recurrence rates were 14.1% and 3.3%, respectively, at 30 days, increasing to 42.2% and 8.3% at one year. By 30 days about a quarter of subjects had used facility-based rehab and another quarter had used home health; home health utilization increased to 44.7% by one year. At all time points Black patients had significantly higher readmission rates compared to whites, but higher stroke recurrence rates were only observed at the 1-year mark. At 30 days, utilization of post-acute care services did not differ by race, but utilization rates were significantly higher in Blacks at 90 days and one year. In contrast utilization of outpatient services was significantly higher among White patients at all time points. Linkage between acute stroke registry and claims data provides an important source of surveillance data for stroke outcomes up to 1-year post discharge. This data allows for real-time monitoring of healthcare outcomes and potentially leads to interventions to improve stroke care.

Intro: The Michigan Stroke Program (MiSP) monitors statewide EMS stroke care performance and data quality. Starting in 2018, quarterly performance and outcome benchmark data were shared with EMS agency partners to support prehospital quality improvement (QI) for stroke patients. Objective: To improve EMS data quality and prehospital stroke care performance. Methods: This is a retrospective before and after study of the impact of tailored QI interventions to improve EMS compliance with three prehospital stroke quality metrics selected by MiSP EMS partners: (prehospital stroke scale documentation, last known well documentation, and prehospital notification). Five MiSP EMS partners participated in the intervention which consisted of provision of quarterly performance reports generated from linked data from Michigan’s EMS Information System and the Michigan Stroke Registry. The MiSP team reviewed agency performance with EMS leadership to select the target measures for intervention and then assisted in development of a QI plan to improve agency performance. Measure compliance was compared from three months prior to and six months after data quality solutions were implemented. Results: Through data sharing and communication with EMS agencies and EMS software data vendors, common themes were identified related to poor data quality beginning in January 2020. A sample of agencies and resolutions to address these issues are displayed in the Table. Compliance with the targeted metric improved significantly at each agency at three and six months following intervention. Conclusions: Analysis of EMS data and sharing performance data directly with EMS agency partners identified correctable data quality issues negatively impacting EMS quality measures. Correcting these issues led to dramatic and sustained improvement in EMS quality metric compliance. This approach may serve as a model for improving the accuracy of EMS performance data in state-level EMS registries.

Michigan's CHRONICLE, the Chronic Disease Registry Linking Electronic Health Record Data, is a near-real-time disease monitoring system designed to harness electronic health record (EHR) data and existing health information exchange (HIE) infrastructure for transformative public health surveillance. Strong evidence indicates that using EHR data in chronic disease monitoring will provide rapid insight over time on health care use, outcomes, and public health interventions. We examined the potential of EHR data for chronic disease surveillance through close collaboration with our statewide HIE network and 2 participating health systems. We describe the development of CHRONICLE, the promising findings from its implementation, the identified challenges, and how those challenges will inform the next steps in testing, refining, and expanding the system. By detailing our approach to developing CHRONICLE and the considerations and early steps required to build an innovative, EHR-based chronic disease registry, we aim to inform public health leaders and professionals on the value of EHR data for chronic disease surveillance. With systematic testing, evaluation, and enhancement, our goal for CHRONICLE, as a fully realized and comprehensive surveillance system, is to model how collaborative health information exchange can support evidence-based strategies, resource allocation, and precision in disease monitoring.

Introduction: Hospital readmissions are often used as an indicator of quality of care. However, identifying patients at risk of readmission after stroke is challenging and predictive models have historically not performed well, in part because they often rely on single data sources. Data linkage might offer a solution. Methods: We probabilistically linked data from the Michigan’s Get With The Guidelines Stroke registry and Michigan Value Collaborative multipayer claims database from Medicare and Blue Cross Blue Shield beneficiaries discharged alive following acute stroke (ICD-10 I61-I63) between 2016-2020. The registry dataset included 64 variables covering demographics, stroke presentation, medical history, procedures, and complications. The hospital dataset included 20 variables from the American Hospital Association’s database. The claims dataset included payer and 79 HCC comorbidity codes. Using combinations of the 3 data sources, we examined the performance of multivariable LASSO logistic regression models to predict all cause readmission at 30-days and 1-year post discharge. We generated hospital-specific testing and training models and reported the mean model discrimination (AUC) of all combinations of the 3 datasets. Results: Of 19,382 linked stroke discharges, 2,724 (14.1%) and 8,169 (42.2%) were readmitted within 30-days and 1-year, respectively. For 30-day readmission, the model based on only registry data produced the best performance (M1, Table). However, for prediction of 1-year readmission, the combination of registry and claims data produced the best performing model (M13, Table). Hospital level characteristics did not have any significant impact on prediction accuracy. Conclusions: Clinical registry data was the best data source for predicting 30-day readmission. However, claims based data were additive when predicting readmission within 1-year, probably because HCC codes add information about total comorbidity burden.

Background: Thirty-day readmission following hospital discharge for stroke is an important quality measure for US hospitals. Current US prediction models for post stroke readmission based on electronic medical records from single healthcare systems or hospitals have modest discrimination (AUC range 0.64 - 0.74). Aim: To develop 30-day all-cause readmission prediction model using a machine learning (ML) based method trained on linked stroke registry and administrative claims data. Methods: Using probabilistic linking, we matched acute stroke (ICD-10 I61-I63) discharges from 31 hospitals participating in the Michigan Acute Stroke registry between 2016-2020 to multipayer administrative claims data provided by the Michigan Value Collaborative for Medicare and Blue Cross Blue Shield of Michigan commercial beneficiaries. Stroke registry data included patient demographics, clinical characteristics, past medical history, and treatments. Claims data was used to identify readmissions within 30 days of discharge. We used multivariable LASSO logistic regression- a simple ML technique to predict 30-day all-cause-readmission and evaluated the prediction accuracy using a hospital-split internal validation scheme to generate hospital-specific and pooled AUC estimates with 95% confidence intervals (Figure 1). Results: Of 19,382 linked stroke discharges, 2,724 (14.1%) were readmitted within 30-days. Readmitted patients were older, more likely to be male, black, and have higher stroke severity (NIHSS >5). Registry hospitals were either primary (64%) or comprehensive (26%) stroke centers. Hospital-specific 30-day readmission ranged between 9.9%-23.1% ( P <.001) with an average of 14.1% (95% CI:13.6%-14.5%) (Figure 2). Hospital specific AUC estimates ranged between 0.60-0.80 with a pooled AUC of 0.68 (95% CI:0.65-0.70) (Figure 2). Conclusions: ML prediction model fitted to linked registry data can be used to predict hospital-specific and statewide readmissions post stroke.

Introduction: In recent years, Medicare Advantage (MA) enrollment in the US has increased dramatically relative to traditional Medicare (TM). There is evidence to suggest that MA stroke patients are less likely to receive inpatient rehabilitation facility (IRF) based care in favor of home health. To evaluate potential gaps in quality of care, we compared post-acute care use patterns following hospitalization for acute stroke between TM and MA populations. Methods: We probabilistically linked data from Michigan’s Get With The Guidelines-Stroke registry and Michigan Value Collaborative multipayer claims data registry for 16,231 TM and MA beneficiaries discharged alive following acute stroke (ICD-10 I61-I63) between 2016-2020. We used discharge claims to identify the initial discharge destination classified as home, home health, IRF, skilled nursing facility (SNF) or other. Unadjusted differences in proportions and time trends between TM and MA populations were assessed using chi square statistical tests. Results: TM and MA-insured patients constituted 75.1% and 24.9% of the 16,231 stroke discharges, respectively. Compared with TM, MA beneficiaries were older (77.4 vs 75.8, P <0.001), less likely to be female (50.3% vs 55.3%, P <0.001), and more likely to be white (87.0% vs 82.4%, P <0.001). Over the study period, the proportion of MA patients increased from 21.2% in 2016 to 32.4% in 2020 ( P <0.001) (Table). Overall, significant differences in discharge destination were due to more MA patients discharged to home and home health ( P <0.001). Despite significant differences between MA and TM populations, there were no consistent time trends in discharge patterns. However, the exception was in 2020 where increased utilization of home health was accompanied by decreases in IRF and SNF discharge. Conclusions: We did not find clinically meaningful differences in discharge patterns between TM and MA populations. Changes in 2020 were likely attributed to COVID-19.

Background: Despite substantial to minimize ischemic stroke treatment delays, hospital door-to-needle (DTN) time variations persist. We sought to identify sources of variation in practice as well as understand stroke team member perceptions of factors that influence DTN performance in a sample of hospitals that participate in a statewide stroke registry. Methods: We conducted a series of semi-structured interviews with stroke coordinators and emergency department (ED) staff. Participants were invited from hospitals with the fastest or slowest DTN times in 2022. Transcripts of the interviews were coded to identify differences in hospital acute stroke code processes as well as interviewee perceptions of facilitators and barriers to optimal performance. Results: Ten individuals from 6 hospitals participated in interviews (6 stroke coordinators, 3 ED physicians, 1 ED nurse). Variations in stroke code processes included location of initial patient evaluation, stroke team composition (inclusion of pharmacy or in-person neurology provider), imaging strategy for large-vessel occlusion stroke screening, thrombolytic medication utilized, and team members primarily responsible for thrombolytic decision-making, mixing, and delivery. Common facilitators of rapid DTN times included integration of EMS, multidisciplinary engagement with quality improvement teams (especially physician engagement), embedding critical care nurses in stroke teams, and providing rapid, individualized feedback to clinical personnel. Common barriers included identifying stroke in triage, staff turnover, and inadequate stroke coordinator resources. Participants from higher performing hospitals reported having larger stroke teams with in-person neurology support. Participants from lower performing hospitals highlighted staffing shortages and difficulties with video-based teleneurology. Conclusion: Stroke code process variation is common and largely driven by hospital resources. Highly engaged, physician-led multidisciplinary teams and timely, actionable feedback were seen as key elements of achieving DTN goals. Staffing shortages and lack of access to timely neurological consultation are system-level barriers to optimal stroke care.

Background Emergency medical services (EMS) compliance with recommended prehospital care for patients with acute stroke is inconsistent; however, sources of variability in compliance are not well understood. The current analysis utilizes a linkage between a statewide stroke registry and EMS information system data to explore patient and EMS agency‐level contributions to variability in prehospital care. Methods and Results This is a retrospective analysis of a cohort of confirmed stroke cases transported by EMS to hospitals participating in a statewide stroke registry. Using EMS information system data, the authors quantified EMS compliance with 6 performance measures derived from national guidelines for prehospital stroke care: prehospital stroke scale performance, glucose check, stroke recognition, on‐scene time ≤15 minutes, time last known well documentation, and hospital prenotification. Multilevel multivariable logistic regression analysis was then used to examine associations between patient‐level demographic and clinical characteristics and EMS compliance while accounting for and quantifying the variation attributable to agency of transport and recipient hospital. Over an 18‐month period, EMS and stroke registry records were linked for 5707 EMS‐transported stroke cases. Compliance ranged from 24% of cases for last known well documentation to 82% for documentation of a glucose check. The other measures were documented in approximately half of cases. Older age, higher National Institutes of Health Stroke Scale, and earlier presentation were associated with more compliant prehospital care. EMS agencies accounted for more than half of the variation in EMS prehospital stroke scale documentation and last known well documentation and 27% of variation in glucose check but <10% of stroke recognition and prenotification variability. Conclusions EMS stroke care remains highly variable across different performance measures and EMS agencies. EMS agency and electronic medical record type are important sources of variability in compliance with key prehospital performance metrics for stroke.

BACKGROUND: Emergency medical services (EMS) is an important link in the stroke chain of recovery. Various prehospital quality metrics have been proposed for prehospital stroke care, but their individual impact is uncertain. We sought to measure associations between EMS quality metrics and downstream stroke care. METHODS: This is a retrospective analysis of a cohort of EMS-transported stroke patients assembled through a linkage between Michigan’s EMS and stroke registries. We used multivariable regression to quantify the independent associations between EMS quality metric compliance (dispatch within 90 seconds of 911 call, prehospital stroke screen documentation [Prehospital stroke scale], glucose check, last known well time, maintenance of scene times ≤15 minutes, hospital prenotification, and intravenous line placement) and shorter door-to-CT times (door-to-CT ≤25), accounting for EMS recognition, age, sex, race, stroke subtype, severity, and duration of symptoms. We then developed a simple EMS quality score based on metrics associated with early CT and examined its associations with hospital stroke evaluation times, treatment, and patient outcomes. RESULTS: Five thousand seven hundred seven EMS-transported stroke cases were linked to prehospital records from January 2018 through June 2019. In multivariable analysis, prehospital stroke scale documentation (adjusted odds ratio, 1.4 [1.2–1.6]), glucose check (1.3 [1.1–1.6]), on-scene time ≤15 minutes (1.6 [1.4–1.9]), hospital prenotification ([2.0 [1.4–2.9]), and intravenous line placement (1.8 [1.5–2.1]) were independently associated with a door-to-CT ≤25 minutes. A 5-point quality score (1 point for each element) was therefore developed. In multivariable analysis, a 1-point higher EMS quality score was associated with a shorter time from EMS contact to CT (−9.2 [−10.6 to −7.8] minutes; P <0.001) and thrombolysis (−4.3 [−6.4 to −2.2] minutes; P <0.001), and higher odds of discharge to home (adjusted odds ratio, 1.1 [1.0–1.2]; P =0.002). CONCLUSIONS: Five EMS actions recommended by national guidelines were associated with rapid CT imaging. A simple quality score derived from these measures was also associated with faster stroke evaluation, greater odds of reperfusion treatment, and discharge to home.

This cross-sectional study compares trends in out-of-hospital cardiac arrests and fatalities in the Detroit area during the COVID-19 pandemic with year-earlier events for the same period.

Background Understanding and improving EMS stroke care requires linking data from both the prehospital and hospital settings. In the US, such data is collected in separate de-identified registries that cannot be directly linked due to lack of a common, unique patient identifier. In the absence of unique patient identifiers two common approaches to linking databases are deterministic matching, which uses combinations of non-unique matching variables to define matches, and probabilistic matching, which generates estimates of match probability based on the degree of similarity between records. This analysis seeks to compare these two approaches for matching EMS and stroke registry data. Methods Stroke cases transported by EMS to Michigan hospitals participating in the Michigan Coverdell Acute Stroke Registry were linked to records from Michigan's EMS Information System (MI-EMSIS) between January 2018 and June 2019. Destination hospital, date-of-service, patient age, date-of-birth, and sex were used to perform deterministic and probabilistic linkages. Match rates and representativeness of the matched samples were compared between the two matching strategies. Multivariable logistic regression was used to identify characteristics associated with successful matching. Results During the 18-month study period there were 8,828 EMS transported confirmed stroke cases in the registry and 620,907 EMS transports to 38 Coverdell registry-participating hospitals. The probabilistic match linked 5985 (67.7%) strokes to EMS records; the deterministic match linked 4012 (45.5%). Within each strategy the characteristics of matched and unmatched cases were similar, with the exception that deterministically matched cases were less likely to be older than 89 (adjusted odds ratio [aOR]=0.3), white (aOR=0.8), and more likely to have subarachnoid hemorrhage (aOR=1.4) than unmatched cases. Conclusion Probabilistic matching resulted in higher match rates and a more representative sample of EMS transported strokes, suggesting it may be superior in assessing EMS stroke care compared to a deterministic approach.

Background: Hypertension is a leading risk factor for stroke, and its management is key for stroke prevention. Michigan’s Coverdell Acute Stroke Registry collects stroke data including patient history of hypertension and discharge medications. Our aim was to identify independent predictors of having antihypertensive medication prescribed at discharge (AMPD) and to assess hospital-level performance in order to focus future quality improvement efforts. Methods: Thirty-one hospitals contributed data to MOSAIC in 2016-2017. Patients with a diagnosis of acute stroke or TIA were included; subjects receiving comfort measures only (CMO), died prior to discharge, or were discharged to hospice were excluded. Patients were considered eligible for AMPD if they had a history of hypertension. Independent factors associated with no AMPD were identified using multivariable logistic regression. Variation among participating hospitals in the treatment of hypertension at discharge was assessed by calculating the percent of hypertension patients with no AMPD. Results: Out of 19,645 patients with stroke or TIA, 15,253 (77.6%) had a history of hypertension; 12,707 (83.3%) of these cases did not receive CMO, hospice, were not discharged to an acute care facility and were discharged alive. Of these eligible cases, 934 (7.3%) had no AMPD. No AMPD was significantly higher among subjects who were younger, had hemorrhagic stroke or TIA. No AMPD was significantly lower among blacks (Table). Rates of no AMPD varied widely among hospitals; five hospitals had rates below an achievable benchmark of 15% no AMPD (Range: 2.5 - 70.0%). Conclusions: Prescription of antihypertensive medication to patients with a history of hypertension at discharge is high but use occurs less often in younger patients, and those with hemorrhagic stroke or TIA. Hospital-based data identifies substantial variability and identified individual hospitals for follow-up education and quality improvement efforts.

Background: Each year roughly 800,000 Americans experience a stroke. With shortened hospital length of stay, many patients are discharged home with expectations to self-manage follow-up appointments, comply with medications, and modify life style factors. Coordination of transition of care (TOC) post-discharge provides an organized approach to ease treatment and compliance. Michigan’s Ongoing Stroke Registry to Accelerate Improvement of Care (MOSAIC) collects voluntary TOC data from participating hospitals and encourages additional hospitals to participate. Objective: To identify barriers MOSAIC hospitals encounter deterring participation in TOC data collection, provide recommendations for barrier elimination, and improve facilitation of care for discharged patients. Design: Qualitative observational study with use of follow-up survey. Participants: 35 focus group call attendees participating with MOSAIC. Methods: Structured focus group with pre-determined free-response questions were conducted via phone call with participating hospitals. Upon completion, hospitals participating in the call received a follow-up survey to provide additional information not covered in the focus group discussion. Use of thematic analysis to identify commonality among hospitals experiencing barriers was conducted from compiled focus group notes from two recorders and follow-up survey data. Results: Based on a thematic analysis of focus group notes and follow-up survey responses, staffing resources was the most prevalent theme hindering MOSAIC hospitals from participating in TOC collection. Time issues, other required data reporting, data visualization, and form usability were also highlighted as deterring factors. Further, participants indicated that while not currently participating in collection, TOC data could be useful in decision making processes and quality improvement within their respective programs.

Background: Michigan’s Ongoing Stroke Registry to Accelerate Improvement of Care (MOSAIC) supports expansion of statewide stroke systems of care. The Michigan EMS Information System (Mi-EMSIS) is a repository for EMS data developed to standardize collection, storage, sharing and analysis of data. To provide feedback to EMS providers concerning stroke patient identification, care, and outcomes, development of statewide linkage between hospital and EMS registry data is needed. Objectives: Develop data linkage between MOSAIC and MiEMSIS stroke data and assess completeness and accuracy of match results. Method: All patients transported by EMS in 2017 with an impression of stroke (i.e. suspected stroke) were linked to patients who were discharged with confirmed stroke or TIA from 33 participating MOSAIC hospitals with successful linkage defined as suspected stroke transport having confirmed stroke upon discharge. Records were deterministically matched using four variables: receiving hospital code, hospital admission date, date of birth, and sex. Additionally, performance on EMS stroke quality measures and hospital outcomes were calculated. Results: 197 EMS agencies transported 4,580 suspected stroke patients to MOSAIC hospitals. Of the 11,352 confirmed stroke cases discharged from MOSAIC hospitals, 4,924 (43.4%) were documented to have arrived via EMS. However, only 1,287 (26.1%) cases were successfully linked to the MiEMSIS data. Of the 4,580 suspected stroke transports 1,287 (28.1%) were confirmed as acute stroke cases at discharge. Conclusions: A deterministic linkage between Coverdell stroke cases and EMS suspected strokes from a state EMS registry resulted in low match rates, which may imply low rates of EMS stroke recognition or high rates of match failure. Alternative linkage methods such as probabilistic matching and inclusion of all EMS transports should be explored.

Background: There is considerable patient and hospital level variation in the selection of rehabilitation settings at the time of hospital discharge and contributing factors in decision making for post-acute care are poorly defined. Objective: Describe variability in frequency of discharge destination (acute, sub-acute, other facility type) of stroke patients to rehabilitation facilities across hospitals participating in Michigan’s Ongoing Stroke Registry to Accelerate Improvement of Care (MOSAIC), a statewide Coverdell stroke registry. Sample: Patients discharged from hospitals participating in MOSAIC’s hospital inventory between January 1- December 31, 2017 with clinical diagnosis of Ischemic or Hemorrhagic Stroke (N=3,129). Methods: Multinomial logistic regression analysis conducted to quantify associations between patient and hospital level factors related to discharge destination. Discharge destination was utilized as the outcome variable with all other factors (age, gender, race, insurance, functional status- admission/discharge, stroke type, hospital size, and stroke volume) as indicator variables. Results: From the model, 4 variables indicated significance with discharge destination. Patients suffering a hemorrhagic stroke were less likely to be discharge to acute or sub-acute facilities. Patients who were female, unable to ambulate at discharge, or of increased age (60+ years) were more likely to be discharged to a sub-acute facility. As such, hospitals with small bed sizes were more likely to discharge patients to sub-acute facilities than any other destination. Conclusions: Patient and hospital-based factors may predict discharge destination, with hospital bed size, stroke type, and discharge ambulatory status showing the strongest associations. Knowing major indicators related to decision-making processes, hospitals can reduce variability in decision making and improve quality of life outcomes for stroke patients.