Search

Background: Perfusion imaging is used to determine eligibility for mechanical thrombectomy (MT) in patients presenting with large vessel occlusion (LVO) in acute ischemic stroke. The utility of CT perfusion (CTP) imaging in spoke hospitals remains controversial. The aim of the study was to compare the imaging parameters, functional and safety outcomes in patients selected for MT based on CTP availability in regional hospitals. Methods: Consecutive adult patients treated for LVO after initially presenting to regional spokes between Jan 2021 to Dec 2021 were included in the analysis. Our standard Acute Stroke Imaging Protocol included CT/CT angiogram. CT perfusion was added to the protocol in April 2021. Demographics, NIHSS, imaging metrics (ASPECTS, CTA collateral score, reperfusion (mTICI 2b/3) rates), hemorrhage (per ECASS III) and functional outcomes (modified Rankin Scale of 0-2 at 90 days) were compared between patients who did vs. did not undergo CTP at regional. Core infarct volume growth and rate (ml/min) was assessed for patients with CTP data both at the regional and at the Comprehensive Stroke Center (CSC). Results: A total of 29 patients met inclusion during the study period. Of these, 14 patients were transferred to CSC for possible MT without CTP at the regional facility and 15 were transferred with CTP at regional. Age (p=0.44), NIHSS (p=0.08) and onset-to-regional facility arrival time (p=0.54) were similar in both groups. For all patients, repeat imaging at the CSC showed significantly decreased ASPECTS (p=0.002) but stable CTA collateral score (p=0.94). For patients who underwent CTP both at the regional facility and CSC, median core infarct volume grew significantly (0[0, 7] ml vs. 7 [0, 14] ml, p=0.022), with a median growth rate of 0 [0, 0.07] ml/min. No patients were excluded from thrombectomy due to core infarct growth. Rates of successful reperfusion (p=0.83) and hemorrhagic transformation (p=0.49) did not differ significantly between the groups. Rates of mRS 0-2 were similar in both groups (p=0.85). Conclusion: Our study suggests that core infarct grows during interfacility transfers without an impact on selection for thrombectomy and outcomes. A larger study is needed to assess the need for repeating perfusion imaging at CSC.

Background: Mechanical thrombectomy (MT) is effective for reducing morbidity in patients presenting within 24 hours of symptom onset with large vessel occlusion (LVO). Perfusion imaging is helpful in selecting patients in this extended time window. We report our experience of implementing CTP imaging in all our spokes and its impact on quality metrics and outcomes. Methods: Acute Stroke Imaging Protocol was updated to include CTP in addition to CT and CTA for all suspected stroke patients presenting within 24 hours of symptoms in April 2021 processed by RAPID AI (iSchema View inc, Menlo Park, CA, USA). Consecutive patients aged ≥18 years with suspected stroke presenting to regional spokes between Jan 2021 to Dec 2021 were included in the analysis. Demographics, NIHSS, quality metrics at regional facility (Door-to-thrombolysis, door-in-door-out times) and Comprehensive Stroke Center-CSC (door-to-arterial access, reperfusion rates (TICI IIb/III), hemorrhagic transformation) and functional outcomes (modified Rankin Scale, mRS) at 90 days were compared between the groups. Results: CTP studies were performed on 1113 patients between April 2021 and December 2021. RAPID AI processing identified LVO in 203 (18.7%) patients, of which 29 (14%) patients were included in this analysis; 14 patients were transferred to CSC for possible MT without CTP at regional facility and 15 were transferred with CTP at regional. Age (p=0.44), NIHSS (p=0.08) and onset-arrival time (p=0.54) were similar in both groups at the regional facility. Door to thrombolysis (p=0.13) and door-in-door-out (p=0.17) times were similar in both groups at the regional facility. CSC metrics of door-to-arterial access times (p=0.84), reperfusion (TICI IIB/III) (p=0.83) and hemorrhagic transformation (p=0.49) rates were similar between groups. 90 day good functional outcomes (mRS 0-2) were similar in patients undergoing thrombectomy (p=0.86). Conclusion: Our study suggests that AI supported perfusion imaging is feasible in a large stroke network and does not appear to have a significant impact on process times, quality metrics and outcomes. A larger study would need to be conducted to validate the observed results.

Study Objective: Though select inpatient-based performance measures exist for the care of patients with nontraumatic intracranial hemorrhage (ICH), Emergency Departments (EDs) lack measurement instruments designed to support and improve care processes in the hyperacute phase. To address this, we propose a set of measures applying a syndromic (rather than diagnosis-based) approach informed by performance data from a national sample of community EDs participating in the Emergency Quality Network Stroke Initiative (E-QUAL). Methods: We convened a workgroup of experts in acute neurologic emergencies. The group considered the appropriate use case for each proposed measure (internal quality improvement [QI], benchmarking, or accountability) and examined data from E-QUAL-participating EDs to consider the validity and feasibility of proposed measures. The initially conceived set included 14 measure concepts of which 7 were selected for inclusion in the measure set after review of data and further deliberation. To inform the process, we collected data from E-QUAL participating EDs. EDs used a web-based portal to submit data from chart reviews on patients with ICH during 2020-2021. Descriptive statistics characterize performance and identify variation. Results: Proposed measures include 2 for QI, benchmarking, and accountability, 3 for QI and benchmarking and 2 for QI only (Table). Of the E-QUAL participating EDs, 35 had sufficient case volume (≥25) and were included for review of preliminary performance data. Median annual ED volume was 44,000 (IQR 31,337-67,000) and 16 (48%) reported typically transferring patients with ICH. The Table includes available performance data. Conclusion: Application of these measures may identify opportunities for improvement and focus QI resources on evidence-based targets. The proposed measures warrant further development and validation to support broader implementation and advance national healthcare quality goals.

Background: Endovascular therapy (EVT) offers dramatic benefit to selected patients with large vessel occlusion (LVO) ischemic stroke. However, identification of EVT candidates requires advanced imaging and often interfacility transfer. We sought to quantify the yield of such testing as well as identify clinical predictors of EVT candidacy. Methods: This retrospective cohort study identified consecutive Emergency Department (ED) patients with stroke symptoms who underwent CT angiogram and brain perfusion (CTA/P) imaging to assess for EVT candidacy. Demographics, medical history, clinical characteristics, final diagnosis, and outcomes were abstracted. We compared clinical characteristics among those who did and did not undergo EVT. Multivariable logistic regression was used to identify independent clinical predictors of EVT and derive a clinical prediction rule to quantify the probability of EVT. Results: Over a 12-month period, 835 patients underwent CTA/P imaging in the ED. EVT was undertaken for 116 (13.9%) patients; 321 (38.4%) ultimately received a non-stroke diagnosis. Patients who received EVT were older and had higher stroke scores (Table). Patients with an unknown last known well (LKW) time were less likely to receive EVT, however increasing time form LKW to door did not predict EVT (test for trend p=0.976). Multivariable analysis results are presented in the Table. A clinical decision rule based on the regression coefficients demonstrated moderately high discrimination for predicting EVT with an AUC of 0.79 (0.74 to 0.83). Among 102 patients transferred for CTA/P, 24 (24%) had and a score Conclusions: EVT Candidates are common among ED patients screened with CTA/P. Clinical factors can predict the likelihood of EVT candidacy. If validated in other populations, a simple clinical prediction rule may assist in triaging patients in need of urgent transfer to a thrombectomy-capable facility.

Introduction: A priori sample size determination is an essential step in designing randomized controlled trials (RCTs). Failure to reach pre-planned sample size introduces risk of both falsely negative and spuriously positive findings. We undertook a systematic review of contemporary acute stroke trials to document the prevalence and reasons for termination of trials prior to completion of enrollment. Methods: We searched MEDLINE for RCTs of acute stroke therapy published between 2013 and 2018 in 9 major journals. Manuscripts describing the final primary results of phase 3 and large phase 2 trials of any therapeutic intervention were eligible for inclusion. Study characteristics, including the presence of a data monitoring committee (DMC) and stopping rules, risk-of-bias assessment, funding sources and conflicts of interest, were abstracted from published manuscripts and trial protocols by two independent reviewers. The prevalence of and reasons for early termination were quantified. Multivariable logistic regression was used to identify study-level predictors of early termination. Results: Of 756 hits, 60 were eligible for inclusion, 21 (35%) of which were terminated early. Among the trials stopped early, 10 (48%) reported stopping for benefit or newly available evidence while 11 (52%) were terminated for futility; 20 (95%) reported a DMC and 17 (81%) reported the use of a pre-specified statistical stopping rule. Factors associated with early termination included study location in North America, larger planned sample size, and industry funding (Table). Study location in North America and larger planned sample size retained statistical significance in a multivariable model. Conclusions: One in three contemporary stroke trials were terminated prior to completion of enrollment. Reasons for termination were evenly split between benefit and futility. Further study is needed to understand the reasons for and impact of early termination on study results.

Background: Emergency Medical Services (EMS) transport facilitates rapid ischemic stroke treatment. We hypothesized that EMS stroke recognition is critical to delivery of optimal stroke care and favorably impacts downstream outcomes. Methods: We identified consecutive EMS-transported ischemic stroke (IS) and transient ischemic attack (TIA) patients at 4 primary stroke center hospitals over a 2-year period. Cases were categorized as EMS recognized or unrecognized based on documented EMS impression. Compliance with EMS performance measures (stroke screen use, onset time documentation, prenotification), hospital-based measures (door-to-CT [DTCT] times, t-PA delivery), and hospital discharge disposition were compared among EMS recognized and unrecognized cases. Logistic regression was used to adjust for confounders such as age, gender, stroke severity, and clustering by agency or hospital. Results: EMS transported 1,056 IS/TIA patients during the study period; 719 (68%) were recognized and 337 (32%) unrecognized. EMS recognized cases were more likely to receive recommended prehospital care, early CT, and t-PA (Table). EMS recognized cases also had higher odds of being discharged home and lower odds of death in adjusted analysis. Conclusion: EMS recognition of IS/TIA was associated with prehospital quality measure compliance, faster ED evaluations, t-PA delivery, and superior outcomes. Quality improvement efforts should focus on improving EMS stroke recognition.

Background: Emergency dispatchers represent the first point of contact for patients activating an acute stroke response. Accurate dispatcher stroke recognition is associated with faster emergency medical services response time; however, stroke is often unrecognized during initial emergency calls. Stroke screening tools such as the Cincinnati Prehospital Stroke Scale have been shown to improve on-scene stroke recognition and thus have been proposed as a means to improve dispatcher accuracy. We conducted a systematic review of the accuracy of emergency dispatcher stroke recognition when employing stroke screening tools. Methods: We conducted a comprehensive search of Medline, EMBASE, CINAHL, and Cochrane databases to identify studies of dispatcher stroke recognition accuracy. Those that specifically reported dispatcher utilization of any validated stroke screening tools in isolation or in the context of a comprehensive screening algorithm such as the Medical Priority Dispatch System (MPDS) were potentially eligible. Studies that reported data sufficient for calculation of dispatcher sensitivity or positive predictive value (PPV) using a hospital-based stroke/transient ischemic attack diagnosis as the reference standard were included. Two independent reviewers determined study eligibility, assessed quality using the QUADAS 2 instrument, and abstracted data. Results: We identified 1,413 potential studies; 54 underwent full text review. Three retrospective and 4 prospective cohort studies enrolling a total of 16,382 patients met the inclusion criteria. Stroke screening tools included MPDS (n = 4), Face Arm Speech Time (n = 2), and a novel screening algorithm developed after analysis of emergency calls for stroke (n = 1). Regardless of the screening tool employed, dispatcher stroke recognition sensitivity was suboptimal (5 studies, range 41-83%) as was the PPV (7 studies, range 42-68%). Primary study limitations included application of variable reference standards and questions regarding exclusion of subjects. No studies directly compared stroke screening algorithms and no studies specifically examined stroke recognition among potential candidates for acute stroke therapies. Conclusion: Even when utilizing a stroke screening tool, the accuracy of stroke recognition by emergency dispatchers was suboptimal. More research is needed to identify the causes of poor dispatcher stroke recognition and should focus on potential candidates for time-dependent stroke treatment.

Background: Emergency medical services (EMS) stroke recognition is critical for appropriate triage and rapid activation of stroke systems. Posterior strokes represent up to 25% of acute stroke admissions. Prehospital stroke screening tools such as the Cincinnati Prehospital Stroke Scale (CPSS) rely on stroke signs that may not be present in patients with posterior stroke, resulting in delays in diagnosis and treatment. We hypothesized that addition of the finger to nose test (FNT) to the standard CPSS would improve EMS recognition of posterior stroke. Methods: Over a 21-month period consecutive ischemic stroke cases transported by 3 EMS agencies in a single county in southwestern Michigan were identified and classified as anterior, posterior, or indeterminate based upon the final hospital discharge diagnoses. Following a 12-month baseline data collection period, all paramedics in the county completed a 30-minute on-line training module targeting recognition of stroke symptoms. Additionally, paramedics in 1 of the 3 EMS agencies also received in-person training in the performance of the finger to nose test (FNT group). We calculated the change in EMS posterior stroke recognition between the 12-month preintervention period and the 9-month postintervention period in the agency that received FNT training, and compared this to the change in the 2 control agencies that did not received FNT education intervention. Results: Over 21 months, 798 ischemic stroke cases were transported by EMS; 114 (14%) were posterior circulation strokes. Before training, 9/26 (38%) of posterior strokes were recognized by paramedics in the intervention agency and 10/36 (28%) in the 2 control agencies. Following training, the FTN group recognized 12/16 (75%, p=0.02 compared to 38% baseline) posterior strokes while the control group recognized 13/23 (46%, p=0.2 compared to 28% baseline). The net increase in posterior stroke recognition was therefore greater in the CPSS-A group than in the control agencies (37% vs. 18%, p=0.03). Conclusion: Posterior strokes were more likely to be recognized by paramedics when the finger to nose test was added to the CPSS. These encouraging initial results demonstrate the potential value of incorporating this simple test into EMS training.

Background: EMS stroke recognition and hospital prenotification facilitate rapid stroke treatment. However, EMS stroke recognition is suboptimal. To promote stroke recognition, we deployed a 30-minute online EMS educational module in a single southwestern Michigan county. Methods: Following 12 months of baseline data collection, the online module was completed by all paramedics in the study county. An interrupted time series design was used to measure the impact of the educational intervention in the subsequent 9 months. Prehospital and hospital data were collected for all EMS suspected stroke cases and hospital confirmed strokes that were missed by EMS. We compared the sensitivity and overcall (false positive) rate of an EMS stroke impression during the pre- and post-intervention periods. We also compared stroke scale documentation, prehospital notification, and door-to-CT (DTCT) times. Results: Over a 21-month period, 1656 cases were included: 744 EMS identified strokes, 408 EMS missed strokes, and 504 EMS overcalled cases (not strokes). Following the intervention, EMS sensitivity for stroke recognition increased from 65.3% to 71.5% (p=0.048) without change in the proportion of overcalls (39.4% to 40.8%, p=0.65). However, improvement in sensitivity was limited to the first post-intervention quarter (65.3% to 74.0%, p=0.027), and returned to baseline thereafter (figure). Among 1109 confirmed stroke cases, EMS hospital prenotification improved following the intervention (57.1% to 81.6%, p Conclusions: A brief, web-based educational intervention was associated with higher EMS stroke recognition without an increased rate of stroke overcalls. Although improvements were transient, they coincided with improved DTCT times.

Introduction: Emergency Medical Services (EMS) are an important component of stroke systems of care. EMS transport has been associated with faster emergency department (ED) stroke evaluations and higher t-PA delivery rates, especially among EMS recognized cases with hospital prenotification. However, delivery of high quality care including hospital prenotification requires that paramedics received regular education and training to improve their ability to identify and treat stroke cases. Methods: We identified opportunities for improvement in EMS care of suspect stroke patients through a review of local EMS stroke-related performance data and focus group discussions with paramedics. We then developed a 20-minute online educational module that reviewed basic stroke pathophysiology, discussed acute stroke therapies, and stressed the importance of performing stroke screening examinations and hospital prenotification for suspect stroke cases. The training was provided to all paramedics in a single agency. Scores on 10 item multiple-choice assessment were compared before and after receipt of the online training module using Fisher Exact tests and Wilcoxon Rank Sum Tests. Results: A total of 40 paramedics underwent the training. Overall quiz scores improved from a median of 9 (interquartile range [IQR] 7 to 9) to 10 (IQR 9 to 10, p=0.0002). Prior to training, only 43% of paramedics correctly identified 4.5 hours as the window for t-PA treatment, with 43% choosing 6 hours. Correct t-PA window identification improved significantly following education (43% to 87.5%, p Conclusions: Overall, paramedics demonstrated good baseline knowledge regarding stroke, however important misconceptions regarding the time frame for t-PA eligibility were common. A brief, on-line educational intervention improved stroke knowledge. Further work is needed to determine the impact of stroke education on real world paramedic performance.

Background: Acute ischemic stroke patients that are recognized by Emergency medical services (EMS) receive expedited care are more likely to receive t-PA. In order to study the continuum of acute stroke care and ensure accurate case ascertainment, prehospital and inpatient medical records must be linked. We tested the feasibility of probabilistic matching to link prehospital EMS data and in-hospital records using data from existing de-identified registries, and compared results to manual linkage. Methods: A registry containing all EMS-transported suspect or confirmed stroke cases in Kent County, Michigan was used to identify subjects discharged from either of two primary stroke centers with a diagnosis of stroke or TIA. The registry contains EMS data obtained through direct query of EMS electronic medical record systems which is linked manually to hospital data using patient identifiers (name, date of birth). For the probabilistic match, de-identified data for EMS suspected stroke transports from the EMS information system (EMSIS) was provided by the county medical control authority. Probabilistic linkage between hospital stroke discharges and EMSIS data was done using LinkPlus software. Age, sex, transport agency, and date of service were used as matching variables. Probabilistic match rates were compared to manual registry match rates among EMS suspect stroke cases using a chi square test. Results: Over a 6-month period in 2015, 172 hospital-confirmed stroke and TIA patients arrived by EMS at the target hospitals. Over the same period, 214 EMS suspect stroke cases were identified from the registry and 199 EMS suspect stroke cases were identified in EMSIS. Probabilistic linkage correctly matched 64.5% (111/172) of confirmed strokes to prehospital records in EMSIS, which was not significantly different from the manual registry match rate of 73.8% (127/172, p=0.24). There were 16 cases in the manual match that were missed by the probabilistic match: 5 (2.9%) due to age mismatch and 11 (6.4%) were missing from the EMSIS data. The remaining non-matched cases represented EMS missed stroke cases. Conclusion: Probabilistic matching of de-identified pre-hospital and hospital data is a feasible approach for the evaluation of prehospital stroke care.

Introduction: Utilization of EMS is associated with faster emergency department evaluation and treatment of patients with stroke, especially among EMS-recognized patients. However, most research focuses on patients with ischemic stroke. We sought to identify factors associated with prehospital recognition of hemorrhagic stroke and the impact of recognition on prehospital and in-hospital care. Methods: A cohort of hospital-confirmed hemorrhagic stroke cases transported by EMS to 2 primary stroke centers in Kent County, Michigan over a 12-month time period was assembled. Data regarding prehospital care (Cincinnati stroke screen [CPSS] documentation, GCS, clinical signs and symptoms, transportation times, and paramedic impression) were linked to in-hospital data on door-to-CT (DTCT) times, mortality, and discharge disposition. We examined the relationships between clinical factors and stroke recognition by paramedics as well as between recognition and in-hospital outcomes. Results: Over 12 months, 73 confirmed hemorrhagic stroke patients arrived by EMS. Forty-seven (64.4%) were correctly identified by EMS as stroke; 26 (35.6%) were missed. EMS recognition was associated with greater likelihood CPSS documentation, intracerebral hemorrhage, dispatch impression of stroke, absence of seizure, and higher systolic blood pressure (Table). Multiple logistic regression confirmed a strong independent relationship between CPSS documentation and stroke recognition (OR 40.3 [5.0 to 323.5]). EMS recognized cases had shorter on-scene times (17 vs. 21 minutes, p=0.004), total transport times (33 vs. 43 minutes, p=0.003), and DTCT times (30 vs. 48 minutes, p=0.004). Recognition was not associated with mortality or discharge disposition. Conclusion: CPSS documentation is strongly associated with hemorrhagic stroke recognition by EMS providers. EMS recognition is associated with more efficient transportation and faster DTCT times upon hospital arrival.

Background A number of emergency medical services (EMSs) performance measures for stroke have been proposed to promote early stroke recognition and rapid transportation to definitive care. This study examined performance measure compliance among EMS-transported stroke patients and the relationship between compliance and in-hospital stroke response. Methods Eight quality indicators were derived from American Stroke Association guidelines. A prospective cohort of consecutive, EMS-transported patients discharged from 2 large Midwestern stroke centers with a diagnosis of acute ischemic stroke was identified. Data were abstracted from hospital and EMS records. Compliance with 8 prehospital quality indicators was calculated. Univariate and multivariable logistic regression analysis were performed to measure the association between prehospital compliance and a binary outcome of door-to-computed tomography (CT) time less than or equal to 25 minutes. Results Over the 12 month study period, 186 EMS-transported ischemic stroke patients were identified. Compliance was highest for prehospital documentation of a glucose level (86.0%) and stroke screen (78.5%) and lowest for on-scene time less than or equal to 15 minutes (46.8%), hospital prenotification (56.5%), and transportation at highest priority (55.4%). After adjustment for age, time from symptom onset, and stroke severity, transportation at highest priority (odds ratio [OR], 13.45) and hospital prenotification (OR, 3.75) were both associated with significantly faster door-to-CT time. No prehospital quality metric was associated with tissue-plasminogen activator delivery. Conclusions EMS transportation at highest priority and hospital prenotification were associated with faster in-hospital stroke response and represent logical targets for EMS quality improvement efforts.

Introduction: Emergency dispatchers are usually the first contact point for stroke patients accessing the healthcare system. Dispatcher stroke recognition is associated with better paramedic stroke recognition and faster emergency medical services (EMS) response time, but stroke is often unrecognized during the initial call. We conducted a systematic review to quantify the accuracy of emergency dispatcher stroke recognition when employing stroke screening tools. Methods: We conducted a comprehensive search Medline, EMBASE, CINAHL, and Cochrane databases to identify studies of dispatcher stroke recognition accuracy using a stroke screening tools such as the Medical Priority Dispatch System (MDPS) or Face Arm Speech Time (FAST) mnemonic. Studies were included if they reported dispatcher sensitivity or positive predictive value (PVP) using hospital stroke/TIA diagnosis as the gold standard. Study quality was assessed using QUADAS 2. Two independent reviewers determined study eligibility, assessed quality, and abstracted data. Results: We identified 1450 abstracts; seven cohort studies met inclusion criteria. Stroke screening tools included MPDS (n=4), FAST (n=2), and a novel screening algorithm (n=1). The sensitivity of dispatcher stroke recognition (5 studies, range 41-83%) and PPV (7 studies, range 42% -68%) were variable. Specificity was high (3 studies, 95-99%). No screening tool demonstrated superiority. Included studies had moderate risk of bias and dramatically variable stroke prevalence (range 3-68%). Conclusion: Published reports of screening tools utilized by emergency dispatchers demonstrate suboptimal sensitivity and PPV. More research is needed to identify the causes of poor dispatcher stroke recognition and to improve stroke screening tools.

Introduction: Ischemic stroke patients who arrive by emergency medical service (EMS) receive faster emergency department evaluations and improved rates of thrombolytic treatment. However, EMS stroke recognition and compliance with prehospital stroke quality measures are inconsistent. We hypothesized that EMS stroke care is influenced by a complex interaction of knowledge, beliefs, and system-level variables that influence behavior. Methods: Focus groups of paramedics from a single urban/suburban county were assembled to discuss their experiences identifying and transporting stroke patients. Focus groups were conducted using a semi-structured interview format and audio recorded. Transcripts of focus groups were qualitatively analyzed to identify themes, subthemes, and patterns of paramedic responses. The Clinical Practice Guidelines Framework provided the initial coding scheme, which was modified during the coding process by three coders using grounded theory methods, who came to consensus on which codes to apply. Results: Three focus groups (n=13) were conducted to reach theme saturation. Overall, paramedics reported high confidence in clinical gestalt for assessing stroke patients and a strong desire to “do the right thing,” but were unfamiliar with published guidelines. Paramedics identified variability in the clinical presentations of stroke, inadequate or inconsistent hospital guidance, and lack of feedback regarding care as principle barriers to ideal prehospital stroke care. Participants reported conflicting hospital guidance regarding the appropriate time frame for a high priority transport and hospital prenotification. Feedback regarding final diagnosis was viewed as critical for developing improved clinical acumen. Direct to CT protocols were cited as an effective way to integrate EMS into hospital stroke response. Conclusion: In this qualitative analysis, paramedics expressed a desire for clear, hospital-directed guidance and consistent feedback regarding outcomes for suspected stroke patients.

Background: Prehospital activation of in-hospital stroke response depends on accurate Emergency Medical Services (EMS) provider stroke recognition. While screening tools such as the Cincinnati Prehospital Stroke Scale (CPSS) have been validated, their real world impact is less clear. We determined the accuracy of prehospital stroke recognition and its relationship to CPSS documentation in a cohort of EMS transported suspect and confirmed stroke patients. Methods: We identified a prospective cohort of all suspected or confirmed ischemic stroke or transient ischemic attack (TIA) patients transported by EMS to a large primary stroke center. Data on prehospital and hospital care were abstracted from medical records. The sensitivity and positive predictive value (PPV) of EMS stroke recognition were calculated using the final hospital discharge diagnosis of ischemic stroke or TIA as the gold standard. We compared the sensitivity and PPV of EMS recognition among cases with and without a documented CPSS using chi square tests. We performed multivariable logistic regression analysis to determine the independent relationship between CPSS use and the odds of EMS stroke recognition among confirmed stroke and TIA cases. Results: Over a 12 month period, 434 EMS transported patients met inclusion criteria. The median age was 78 and 60% were female. EMS transported 371 (85.5%) patients as suspect strokes and documented a CPSS in 343 (79.0%). A total of 246 cases (56.7%) were confirmed ischemic strokes (n=186) or TIA (n=60). The sensitivity of EMS stroke recognition was 74.4% (95% CI 68.9 to 79.8) and PPV was 49.3% (95% CI 44.2 to 54.4). Sensitivity was higher among cases with CPSS documentation than without (85.9% vs. 25.5% p Conclusion: EMS providers recognized three quarters of all stroke or TIA admissions, while EMS suspicion of stroke was correct only half of the time. Documentation of a CPSS was associated with substantially higher EMS stroke recognition.

Introduction: The diagnosis of TIA is often challenging in the emergency department (ED) setting. ICD-9 discharge code 435 is often used to identify TIA cases; however, its accuracy is unclear. Our objective was to determine the sensitivity of ICD-9 code 435 to identify confirmed TIA cases treated in 4 community-based EDs. Methods: Over a 12 month period suspect TIA cases were prospectively identified using hot-pursuit mechanisms in 4 EDs. A time-based definition of TIA was used (i.e., acute onset focal unilateral symptoms with duration Results: Of 366 subjects with suspected TIA, 336 (92%) had a confirmed final diagnosis of TIA. Mean age was 66 years, 54% female, 31% non-white and 83% were hospitalized. The relative frequency of primary ICD-9 discharge codes at each of the 4 EDs are shown in the Table. The overall sensitivity of the 435 code was 64.3% (95% CI 58.9-69.4). Alternative codes used included stroke-related (8.0%), other CVD-related (2.7%), miscellaneous (8.3%) and non-specific symptoms codes (16.7%). The frequency of 435 coding varied non-significantly between sites (56% to 72%; p= 0.13); however, sites varied substantially in their use of non-specific symptoms codes (4.8% to 26.7%; p=0.002). Conclusions: In this series of confirmed TIA cases, more than 1/3 rd would not have been identified if the 435 code was used. These findings suggest that the 435 code should not be relied upon to identify TIA cases. Future studies should determine the causes of the variation in coding between sites, and also measure both sensitivity and specificity of the 435 code.

Background: Evaluation of prehospital stroke care quality has been limited by the inability to link EMS and hospital data. The Prehospital Acute Stroke Surveillance System (PASSS) is designed to assess prehospital stroke care in Kent Country, Michigan through linkage of EMS and hospital data for patients transported with suspected or confirmed stroke. Methodology: EMS transported patients were eligible for inclusion if stroke was suspected by EMS dispatchers or providers prehospital or confirmed after hospital admission. Electronic records containing data elements derived from the National EMS Information System (NEMSIS) were obtained from all three EMS agencies in the county and manually linked to in-hospital data from 2 stroke center hospitals. Outcome measures included the accuracy of prehospital stroke recognition by EMS dispatchers and providers and the proportion of runs with documentation of symptom onset time, scene time ≤ 10 minutes, pre-hospital stroke screen, blood glucose, and hospital pre-notification. Results: In the first three months, 167 EMS transported patients were entered into PASSS; 58 (35%) were confirmed as stroke cases after admission. Emergency dispatchers missed 36% (n=21) of confirmed strokes and EMS providers missed 22% (n=13). The false positive rate for stroke identification was 69% (82/119) among dispatchers and 54% (53/98) among EMS providers. Among 98 cases identified by EMS as suspected stroke, documentation rates were 82% for symptom onset time, 86% for prehospital stroke screen, 87% for blood glucose, 10% for scene time ≤10 minutes, and 61% for hospital prenotification. Among 26 patients with a positive prehospital stroke screen and onset of symptoms < 120 minutes, the prenotification rate increased to 77%. Two patients received IV tPA. Conclusions: The PASSS illustrates that substantial opportunities for improvement in EMS prehospital stroke care exist, including stroke recognition and compliance with quality measures. The high rate of false-positive prehospital stroke identification has important implications for the development of regional stroke systems.