Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Factors Predicting Misidentification of Acute Ischemic Stroke and Large Vessel Occlusion by Paramedics

Factors Predicting Misidentification of Acute Ischemic Stroke and Large Vessel Occlusion by... LWW Original Study advanced life support (ALS) and basic life support (BLS) providers at peak times and completes 85,000 transports per year. The IEMS Factors Predicting Misidentification of Acute Ischemic Stroke protocol directs both BLS and ALS providers to obtain a blood glu- cose level and to perform the CPSS on any patient with suspected and Large Vessel Occlusion by Paramedics stroke. There are no specific mandates for symptoms that trigger the CPSS, but the protocol does suggest that patients with strokes may have “fallen, (be) unable to walk, have new balance problems Nancy K. Glober, MD,* Tyler Fulks, MD,† Michael Supples, MD,* Peter Panagos, MD,‡ and or (have) acute altered level of consciousness.” The protocol further David Kim, MD, PHD§ directs that RACE be recorded for all patients with suspected stroke, though this scale is not currently used by IEMS to direct transport. Emergency Medical Dispatch (EMD) is performed using Association 1–5 Abstract: The emergence of thrombectomy for large vessel occlusions has with LVO. Most strokes are not LVO, however, and patients with of Public-Safety Communications Officials guidecards. increased the importance of accurate prehospital identification and triage of acute non-LVO strokes benefit from rapid transport to the nearest acute ischemic stroke (AIS). Despite available clinical scores, prehospital thrombolysis-capable facility, which is not generally a thrombec- Patient and Public Involvement identification is suboptimal. Our objective was to improve the sensitivity tomy-capable comprehensive stroke center (CSC). Emergency No patient was involved. of prehospital AIS identification by combining dispatch information with medical services (EMS) providers must therefore accurately identify paramedic impression. We performed a retrospective cohort review of and transport patients with strokes to appropriate receiving centers, Measures emergency medical services and hospital records of all patients for whom a which may be the nearest primary stroke center or may be a more We collected demographic data (gender and age), EMS run stroke alert was activated in 1 urban, academic emergency department from distant CSC. descriptors (location of patient pick-up, date of service, response and January 1, 2018, to December 31, 2019. Using admission diagnosis of acute Most EMS systems approach this dilemma either by transport- transport times, dispatch code, prehospital primary impression, level stroke as outcome, we calculated the sensitivity and specificity of dispatch ing all patients with suspected stroke to the nearest thrombolysis- of service, prehospital CPSS and RACE), hospital evaluation and and paramedic impression in identifying AIS and large vessel occlusion. We capable facility, leaving subsequent transfer of LVO patients to the treatment data (physician National Institute of Health Stroke Scale, identified factors that, when included together, would improve the sensitivity emergency physician, or by using prehospital stroke severity scales thrombolysis, thrombectomy), and diagnosis of ischemic (not hemor - of prehospital AIS identification. Two-hundred twenty-six stroke alerts were to identify patients with potential LVO for transport directly to a rhagic) stroke on hospital admission as recorded in stroke neurology activated by emergency department physicians after transport by Indianapolis CSC, often bypassing the nearest thrombolysis-capable facility The . notes. Noncontrast computed tomography (CT) was completed on all emergency medical services. Forty-four percent (99/226) were female, most recent American Heart Association guidelines recommend a patients with suspected stroke. The stroke neurologist determined the median age was 58 years (interquartile range, 50–67 years), and median validated stroke screen and stroke severity score to assess for pos- need for CT angiography or magnetic resonance imaging. Data were National Institutes of Health Stroke Scale was 6 (interquartile range, 2–12). sible LVO and transport to a CSC if: last known well time is less collected in Excel (Microsoft Corporation, Redmond, WA). Paramedics demonstrated superior sensitivity (59% vs. 48%) but inferior than 24 hours, transport time to CSC will not disqualify a patient for Analysis specificity (56% vs. 73%) for detection of stroke as compared with dispatch. thrombolysis, and total transport time to the CSC is under 30 min- We described count frequencies and percentages and calcu- A strategy incorporating dispatch code of stroke, or paramedic impression utes. Systems applying this approach have demonstrated improved lated continuous variable medians and interquartile ranges (IQRs). of altered mental status or weakness in addition to stroke, would be 84% times from scene departure to thrombectomy and improved patient Chi-square compared counts, and independent-sample t test com- sensitive and 27% specific for identification of stroke. To optimize rapid and functional outcomes. However, meta-analyses demonstrate that only pared continuous variables. sensitive stroke detection, prehospital systems should consider inclusion 26%–51% of patients identified by stroke severity scales as poten- 7,10 We calculated the sensitivity and specificity of dispatch and of patients with dispatch code of stroke and provider impression of altered tial LVOs are in fact diagnosed with LVOs. Furthermore, stroke paramedic impressions of stroke in identifying strokes definitively mental status or generalized weakness. severity scores are limited in that they are only applied when a stroke diagnosed in the ED. Paramedic-suspected strokes that were not is suspected by prehospital providers. Thus, if an EMS system has Key Words: emergency medical services, large vessel occlusion, stroke called as a stroke alert by the ED physician were considered false limited sensitivity for stroke detection, stroke and LVO detection will (Crit Pathways in Cardiol 2022;21: 172–175) positives. We identified the dispatch codes and paramedic impres- necessarily be similarly limited. sions associated with prehospital false negatives and calculated the While many studies have focused on scores to optimize pre- test characteristics of alternative strategies for prehospital stroke hospital identification of acute ischemic stroke (AIS) and LVO, few identification incorporating non-stroke dispatch and paramedic have explored the characteristics of stroke and LVO cases missed by arly detection of acute stroke by paramedics in the prehospital set- impressions. Data analysis performed with SAS University Edition EMS. In this study, we identified all stroke alerts in 1 academic urban Eting facilitates appropriate triage and rapid treatment of this time- (SAS Institute, Inc., Cary, NC). emergency department (ED) after transport by Indianapolis EMS critical condition. Large vessel occlusions (LVOs) can be effectively (IEMS). We characterized paramedic sensitivity and specificity for treated with mechanical thrombectomy, which is only available at identifying stroke and LVO with their protocol-directed Cincinnati RESULTS some facilities, and which improves functional outcomes in patients Prehospital Stroke Scale (CPSS) and Rapid Arterial oCclusion Between January 1, 2018, and December 31, 2019, IEMS Evaluation (RACE) scale for identification of LVO. We identified the transported 45,339 patients to Eskenazi Hospital. Of those, 211 had dispatch codes and paramedic impressions associated with missed a paramedic impression of stroke. Median public-safety answering From the *Department of Emergency Medicine, Indiana University, Indianapolis, AISs and suggest a strategy to optimize sensitivity for prehospital point call to dispatch time was 1 minute (IQR, 0–2 minutes), median IN; †Department of Emergency Medicine, Southern Illinois University, stroke detection. Springfield, IL; ‡Department of Emergency Medicine, Washington University scene time was 14 minutes (IQR, 11–18 minutes), and median trans- at St. Louis, St. Louis, MO; and §Department of Emergency Medicine, port time to Eskenazi ED was 11 minutes (IQR, 8–14 minutes). Stanford University, Palo Alto, CA. Patients were transported by ALS in 199 (88%) cases and by BLS Presented at the National Association of Emergency Medical Services Physicians, METHODS in 26 (12%) cases. Among all stroke codes, using admission diagno- virtual abstract presentation January 2021. This study was deemed exempt from review by the Indiana Supplemental digital content is available for this article. Direct URL citations sis of stroke as the gold standard, paramedics demonstrated superior University Institutional Review Board, protocol number 2003587456. appear in the printed text and are provided in the HTML and PDF versions of sensitivity (58% vs. 48%) but inferior specificity (56% vs. 73%) for this article on the journal’s Web site (www.critpathcardio.com). detection of stroke as compared with dispatch ( Table 1). Study Design and Setting Reprints: Nancy K. Glober, MD, Department of Emergency Medicine, Indiana When strokes were not identified in the prehospital setting, University, 1701N Senate Ave, Indianapolis, IN 46202. E-mail: nglober@ We retrospectively reviewed the in-hospital and prehospital iu.edu. the most common dispatch codes were sick person (21), chest pain electronic medical records from January 1, 2018, to December 31, Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. This (12), and syncope (10). When paramedics failed to identify AIS, their 2019, for every patient transported by IEMS on whom a stroke alert is an open-access article distributed under the terms of the Creative Commons most common impressions were altered mental status (14), general- Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), (ie, potential acute stroke) was activated by an ED physician based ized weakness (11), and chest pain (6) ( Table 2). where it is permissible to download and share the work provided it is properly on patient evaluation in 1 academic urban ED (Eskenazi Hospital). cited. The work cannot be changed in any way or used commercially without Using either dispatch or paramedic impression of stroke In Marion County, Indianapolis, EMS care is provided to the permission from the journal. would improve sensitivity to 77.2% (115/149) at a specificity of population of about 900,000 by IEMS as well as paramedics based ISSN: 1003-0117/22/2104-0172 8 46.8% (36/77) among stroke codes. Identifying possible strokes with DOI: 10.1097/HPC.0000000000000307 at fire stations. IEMS operates 31 ambulances staffed with both 172 | www.critpathcardio.com Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 Factors Predicting Misidentification of AIS Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 and LVO by Paramedics advanced life support (ALS) and basic life support (BLS) providers TABLE 1. Sensitivity and Specificity for Identification of at peak times and completes 85,000 transports per year. The IEMS Acute Ischemic Stroke by Dispatch and Paramedics protocol directs both BLS and ALS providers to obtain a blood glu- Test Characteristics Dispatch Paramedics cose level and to perform the CPSS on any patient with suspected stroke. There are no specific mandates for symptoms that trigger Sensitivity 47.7% 58.4% the CPSS, but the protocol does suggest that patients with strokes Specificity 72.7% 54.5% may have “fallen, (be) unable to walk, have new balance problems or (have) acute altered level of consciousness.” The protocol further directs that RACE be recorded for all patients with suspected stroke, TABLE 2. When Strokes Were Not Identified by though this scale is not currently used by IEMS to direct transport. (A) Dispatch or (B) Paramedic, Impressions Were Emergency Medical Dispatch (EMD) is performed using Association Varied, But Most Commonly “Sick Person” and of Public-Safety Communications Officials guidecards. “Altered Mental Status” Patients With Strokes That Patient and Public Involvement (A) Dispatch Code Were Not Identified (%) No patient was involved. Sick person 21 (9.3) Measures Chest pain 12 (5.3) We collected demographic data (gender and age), EMS run Syncope 10 (4.4) descriptors (location of patient pick-up, date of service, response and Diabetic problem 8 (3.5) transport times, dispatch code, prehospital primary impression, level (B) Primary Provider ImpressionPatients With Strokes That Were Not of service, prehospital CPSS and RACE), hospital evaluation and Identified (%) treatment data (physician National Institute of Health Stroke Scale, Altered mental status 14 (6.2) thrombolysis, thrombectomy), and diagnosis of ischemic (not hemor - rhagic) stroke on hospital admission as recorded in stroke neurology Generalized weakness 11 (4.9) notes. Noncontrast computed tomography (CT) was completed on all Chest pain 6 (2.7) patients with suspected stroke. The stroke neurologist determined the Dizziness 4 (1.8) need for CT angiography or magnetic resonance imaging. Data were Diabetic hypoglycemia 3 (1.3) collected in Excel (Microsoft Corporation, Redmond, WA). Headache 3 (1.3) Analysis Seizure 2 (0.9) We described count frequencies and percentages and calcu- lated continuous variable medians and interquartile ranges (IQRs). paramedic impressions of stroke, altered mental status, or generalized Chi-square compared counts, and independent-sample t test com- weakness would achieve 75.2% sensitivity (112/149) at a specificity pared continuous variables. of 31.2% (24/77). A strategy using either dispatch code of stroke, We calculated the sensitivity and specificity of dispatch and or prehospital impression of stroke, altered mental status, or weak- paramedic impressions of stroke in identifying strokes definitively ness would be 83.9% sensitive (125/149) and 27.3% specific (21/77) diagnosed in the ED. Paramedic-suspected strokes that were not for identification of stroke, among patients activated as stroke alerts. called as a stroke alert by the ED physician were considered false When calculated among all patients dispatched as stroke or with an positives. We identified the dispatch codes and paramedic impres- EMS primary provider impression of stroke, altered mental status, or sions associated with prehospital false negatives and calculated the weakness, the specificity would improve to 65.9% (29,812/45,214) test characteristics of alternative strategies for prehospital stroke for identification of stroke. identification incorporating non-stroke dispatch and paramedic RACE was documented for 47% (106/226) of stroke alert impressions. Data analysis performed with SAS University Edition patients; paramedics documented “unable to complete” RACE in (SAS Institute, Inc., Cary, NC). 5% (12/226). In our system, RACE ≥ 5 was 71% sensitive and 57% specific for identification of an LVO confirmed by CT angiography. RESULTS All 11 patients determined to have LVO by CT angiography received Between January 1, 2018, and December 31, 2019, IEMS mechanical thrombectomy. Of the patients who had LVOs, 8 (73%) transported 45,339 patients to Eskenazi Hospital. Of those, 211 had had a primary provider impression of stroke. The other 3 had various a paramedic impression of stroke. Median public-safety answering primary provider impressions and dispatch codes ( Table 3). point call to dispatch time was 1 minute (IQR, 0–2 minutes), median Applying RACE to all patients with dispatch code of stroke scene time was 14 minutes (IQR, 11–18 minutes), and median trans- or paramedic impression of stroke, altered mental status or general- port time to Eskenazi ED was 11 minutes (IQR, 8–14 minutes). ized weakness would have potentially identified 1/11 (9%) additional Patients were transported by ALS in 199 (88%) cases and by BLS LVO in the prehospital setting. in 26 (12%) cases. Among all stroke codes, using admission diagno- Of the 211 patients with paramedic impression of stroke, ED sis of stroke as the gold standard, paramedics demonstrated superior physicians activated 122 (57.8%) stroke alerts upon arrival in the ED. sensitivity (58% vs. 48%) but inferior specificity (56% vs. 73%) for detection of stroke as compared with dispatch ( Table 1). When strokes were not identified in the prehospital setting, TABLE 3. LVO Not Identified As Strokes by the Prehospital the most common dispatch codes were sick person (21), chest pain Provider (12), and syncope (10). When paramedics failed to identify AIS, their most common impressions were altered mental status (14), general Case - Dispatch Code Primary Provider Impression ized weakness (11), and chest pain (6) ( Table 2). Patient 1 Stroke Altered mental status Using either dispatch or paramedic impression of stroke Patient 2 Breathing problem Acute respiratory distress would improve sensitivity to 77.2% (115/149) at a specificity of Patient 3 Psych problem Behavioral/psych episode 46.8% (36/77) among stroke codes. Identifying possible strokes with © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. www.critpathcardio.com | 173 Glober et al Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 The other 89 (42.2%) were determined by the ED physician to not primary impression of stroke. We noted differences in demograph- identification of the subset of stroke patients experiencing LVO in have sustained a stroke. There were an additional 104 ED stroke aler ics. P t atients transported to Eskenazi Hospital versus any other the prehospital setting is essential to directing appropriate transport activations after transport by IEMS that were not identified as strokhospital w es ere significantly younger (median [IQR] age = 59.0 destination decisions. We described the ability of EMS dispatch and by the IEMS provider (Fig. 1). Of the 122 patients activated as strok[50.0–67.0] vs. 68.0 [58.0–70.0]; e P < 0.001), more often male paramedics to identify strokes in the prehospital setting in 1 large alerts in the ED, 99 (44%) were female, median age was 58 years (116 [56.0%] vs. 595 [43.5%]; P < 0.001), and less often White urban EMS system. The rate of missed stroke by both EMD and (IQR, 50–67 years), median National Institute of Health Stroke Scale (87 [42.0%] vs. 860 [62.8%]; P < 0.001). However, we did not prehospital professionals is higher than the ED Our data suggest . was 6 (IQR, 2–12). find significant difference in CPSS positivity (182 [87.9%] vs. combinations of dispatch and paramedic impressions that could One-hundred forty-nine (65.9%) of ED stroke alerts were 1157 [84.5%]; P = 0.212) (Supplemental Table 1, http://links.lww. improve the prehospital detection of AIS and marginally improve admitted with a diagnosis of AIS, and 77 (34.1%) were deter - com/HPC/A244). prehospital detection of LVO. mined not to be strokes after imaging and evaluation by a stroke Numerous prehospital stroke scales aim to identify patients DISCUSSION neurologist ( Table 4). Of the 149 patients admitted for stroke, 44 experiencing LVO, and in our system, both RACE and the CPSS (30%) patients received thrombolysis, and 11 (7%) underwent Rapid and accurate prehospital identification of AIS is are used. The RACE scale demonstrates similar predictiv-e per thrombectomy for LVO. All patients found to have an LVO under - critically important, given the time-sensitive nature of available formance to the CPSS, Los Angeles Motor Screen, and Vision, 7 10,11 14 went mechanical thrombectomy.Although our efforts were basedtreatments, and is associated with improved outcomes. The Aphasia, and Neglect instruments, but may have inferior per - on a convenience sample of patients taken to 1 ED, we compareda vailability of CSCs that offer endovascular therapy for patients formance to other prehospital stroke scales. The sensitivity of characteristics of potential stroke patients taken to Eskenazi with with L VO is limited as compared with more common primary stroke paramedics in this study to identify AIS was similar to previously 12 16–20 those taken to other hospitals. During the study period, 1572 centers that can pro vide medical thrombolysis. Given that endo- reported sensitivities for large metropolitan EMS systems, 1,7 patients were transported to any hospital by IEMS with paramedic vascular therap y for LVO is superior to medical therapy alone, and IEMS utilization of the RACE score demonstrated similar 21–23 sensitivity and specificity as previously reported. Failure to recognize a patient having a stroke likely represents a significant 45,339 paents transported barrier to applying a prehospital stroke scale. Our findings sug- to Eskenazi Hospital by IEMS gest that in addition to primary provider impression of stroke, all EMS responses with a dispatch of stroke or with primary provider impression of altered mental status or weakness should be con- sidered as possible AISs. This underscores the importance of both 211 paramedic primary 104 acvated as stroke paramedic and EMD evaluation to identify AIS in the prehospital impression of stroke alert by emergency setting. physician, but not 89 determined not to have a While it is not practical to transport every patient with a dis- idenfied as stroke by stroke by EM physician patch or prehospital impression of altered mental status and gener - paramedic primary 122 acve as stroke alized weakness to a CSC, these are the most common patients in impression alert by EM physician whom AIS is missed, and special attention should be paid to these patients, for whom a more detailed prehospital stroke assessment should be performed, and for whom early hospital notification might 37 EMS suspected strokes be considered. Some over-triage of patients with potential AIS to determined not to have a 226 IEMS paents idenfied as higher levels of care may be appropriate given the time-sensitive stroke by neurology possible stroke by EM physician 24,25 nature of stroke treatment. This study has limitations. The specific features associated FIGURE 1. Flow chart of patients with incorrect prehospital impressions cannot be ascertained from identified for the study. EM indi- 126 EMS false posive 87 EMS true posive 61 EMS false negave 41 EMS true negave the available data. Of the subset of patients with a false-negative cates Emergency medicine. EMS primary impression, it is unclear whether a more detailed stroke assessment would have improved detection of LVO. Including TABLE 4. ED Stroke Alert Patient Demographic and Prehospital Factors Stratified by Neurologist Confirmed Acute Ischemic only patients transported to an academic center with a CSC may bias Stroke Status the study population toward sicker patients. While there were - differ ences in demographics of patients with paramedic-suspected stroke Confirmed Stroke No Stroke by transport to the study hospital versus other area hospitals, there Patient Factors Median IQR Median IQR P was no difference in positivity of CPSS. Expanding the use of a pre- hospital stroke scale to the most common false-negative dispatch Age 59.5 51.3–67.0 56.5 44.8–65.3 0.083 codes and provider impressions should be investigated with a future NIHSS 5.0 2.0–10.0 5.5 2.0–14.0 0.358 prospective study and in other EMS systems to ascertain generaliz- Sex n (%) 95% CI n (%) 95% CI 0.574 ability of our findings. Male 62 (41.9) 33.9–49.8 36 (46.2) 35.1–57.2 Female 86 (58.1) 50.2–66.1 42 (53.8) 42.8–64.9 CONCLUSIONS CPSS 0.181 Conducting detailed prehospital stroke assessments on Positive 95 (64.5) 56.5–71.9 41 (52.6) 41.5–63.6 patients with dispatch codes of stroke or with prehospital provider Negative 15 (10.1) 5.3–15.0 8 (10.3) 3.5–17.0 impression of altered mental status or weakness in addition to pre- Not performed 38 (25.7) 18.6–32.7 29 (37.2) 26.5–47.9 hospital provider impression of stroke may improve the prehospital Paramedic impression stroke 0.103 detection of AIS. Further prospective research is needed to evaluate Yes 87 (58.8) 50.9–66.7 37 (47.4) 36.4–58.5 this assessment model. No 61 (41.2) 33.3–49.1 41 (52.6) 41.5–63.6 Dispatch code was stroke <0.01 DISCLOSURES Yes 71 (48.0) 39.9–56.0 21 (26.9) 17.1–36.8 This research was funded, in part, by Agency for Healthcare Research and Quality K12 HS026390. This study was deemed exempt No 77 (52.0) 44.0–60.1 57 (73.1) 63.2–82.9 from review by the Indiana University Institutional Review Board, CI indicates confidence interval; NIHSS, National Institute of Health Stroke Scale. protocol number 2003587456. Nothing to declare. 174 | www.critpathcardio.com © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. Factors Predicting Misidentification of AIS Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 and LVO by Paramedics identification of the subset of stroke patients experiencing LVO in REFERENCES the prehospital setting is essential to directing appropriate transport 1. Albers GW, Lansberg MG, Kemp S, et al. A multicenter randomized con- destination decisions. We described the ability of EMS dispatch and trolled trial of endovascular therapy following imaging evaluation for isch- emic stroke (DEFUSE 3). Int J Stroke. 2017;12:896–905. paramedics to identify strokes in the prehospital setting in 1 large 2. Berkhemer OA, Fransen PSS, Beumer D, et al; MR CLEAN Investigators. A urban EMS system. The rate of missed stroke by both EMD and 13 randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J prehospital professionals is higher than the ED Our data suggest . Med. 2015;372:11–20. combinations of dispatch and paramedic impressions that could 3. Saver JL, Goyal M, Bonafe A, et al; SWIFT PRIME Investigators. Stent- improve the prehospital detection of AIS and marginally improve retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke. N Engl prehospital detection of LVO. J Med. 2015;372:2285–2295. Numerous prehospital stroke scales aim to identify patients 4. Goyal M, Demchuk AM, Menon BK, et al; ESCAPE Trial Investigators. Randomized assessment of rapid endovascular treatment of ischemic stroke. experiencing LVO, and in our system, both RACE and the CPSS N Engl J Med. 2015;372:1019–1030. are used. The RACE scale demonstrates similar predictiv-e per 5. Jovin TG, Chamorro A, Cobo E, et al; REVASCAT Trial Investigators. formance to the CPSS, Los Angeles Motor Screen, and Vision, Thrombectomy within 8 hours after symptom onset in ischemic stroke. N Engl Aphasia, and Neglect instruments, but may have inferior per - J Med. 2015;372:2296–2306. formance to other prehospital stroke scales. The sensitivity of 6. Adeoye O, Albright KC, Carr BG, et al. Geographic access to acute stroke care paramedics in this study to identify AIS was similar to previously in the United States. Stroke. 2014;45:3109–3024. 16–20 reported sensitivities for large metropolitan EMS systems, 7. Lima FO, Mont’Alverne FJA, Bandeira D, et al. Pre-hospital assessment of large vessel occlusion strokes: implications for modeling and planning stroke and IEMS utilization of the RACE score demonstrated similar 21–23 systems of care. Front Neurol. 2019;10:955. sensitivity and specificity as previously reported. Failure to 8. American Heart Association; Mission: Lifeline Stroke. Emergency Medical recognize a patient having a stroke likely represents a significant Services Acute Stroke Triage and Routing. Available at: https://www.heart. barrier to applying a prehospital stroke scale. Our findings sug- org/idc/groups/ahaecc-public/@wcm/@gwtg/documents/downloadable/ gest that in addition to primary provider impression of stroke, all ucm_492341.pdf. Accessed 2021. EMS responses with a dispatch of stroke or with primary provider 9. Jayaraman MV, Hemendinger ML, Baird GL, et al. Field triage for endovas- cular stroke therapy: a population-based comparison. J NeuroIntervent Surg. impression of altered mental status or weakness should be con- 2020;12:233–239. sidered as possible AISs. This underscores the importance of both 10. Lin CB, Peterson ED, Smith EE, et al. Emergency medical service hospital paramedic and EMD evaluation to identify AIS in the prehospital prenotification is associated with improved evaluation and treatment of acute setting. ischemic stroke. Circ Cardiovasc Qual Outcomes. 2012;5:514–522. While it is not practical to transport every patient with a dis- 11. Patel MD, Rose KM, O’Brien EC, et al. Prehospital notification by emergency patch or prehospital impression of altered mental status and gener - medical services reduces delays in stroke evaluation: findings from the North Carolina stroke care collaborative. Stroke. 2011;42:2263–2268. alized weakness to a CSC, these are the most common patients in 12. Schieb LJ, Casper ML, George MG. Mapping primary and comprehensive whom AIS is missed, and special attention should be paid to these stroke centers by Certification Organization. Circ Cardiovasc Qual Outcomes. patients, for whom a more detailed prehospital stroke assessment 2015;8(6 Suppl 3):S193–S194. should be performed, and for whom early hospital notification might 13. Madsen TE, Khoury J, Cadena R, et al. Potentially missed diagnosis of isch- be considered. Some over-triage of patients with potential AIS to emic stroke in the Emergency Department in the Greater Cincinnati/Northern higher levels of care may be appropriate given the time-sensitive Kentucky Stroke Study. Acad Emerg Med. 2016;23:1128–1135. 24,25 nature of stroke treatment. 14. Crowe RP, Myers JB, Fernandez AR, et al. The Cincinnati Prehospital Stroke Scale compared to stroke severity tools for large vessel occlusion stroke pre- This study has limitations. The specific features associated diction. Prehosp Emerg Care. 2020;1:9. with incorrect prehospital impressions cannot be ascertained from 15. Zhelev Z, Walker G, Henschke N, et al. Prehospital stroke scales as screening the available data. Of the subset of patients with a false-negative tools for early identification of stroke and transient ischemic attack. Cochrane EMS primary impression, it is unclear whether a more detailed Database Syst Rev. 2019;4:CD0114–C27-CD. stroke assessment would have improved detection of LVO. Including 16. Brandler ES, Sharma M, McCullough F, et al. Prehospital stroke identifica- only patients transported to an academic center with a CSC may bias tion: factors associated with diagnostic accuracy. J Stroke Cerebrovasc Dis. 2015;24:2161–2166. the study population toward sicker patients. While there were - differ 17. Oostema JA, Konen J, Chassee T, et al. Clinical predictors of accurate prehos- ences in demographics of patients with paramedic-suspected stroke pital stroke recognition. Stroke. 2015;46:1513–1517. by transport to the study hospital versus other area hospitals, there 18. Mould-Millman N-K, Meese H, Alattas I, et al. Accuracy of prehospital iden- was no difference in positivity of CPSS. Expanding the use of a pre- tification of stroke in a large stroke belt municipality Prehosp . Emerg Care. hospital stroke scale to the most common false-negative dispatch 2018;22:734–742. codes and provider impressions should be investigated with a future 19. Gropen TI, Gokaldas R, Poleshuck R, et al. Factors related to the sensitiv- prospective study and in other EMS systems to ascertain generaliz- ity of emergency medical service impression of stroke. Prehosp Emerg Care. 2014;18:387–392. ability of our findings. 20. Ramanujam P, Guluma KZ, Castillo EM, et al. Accuracy of stroke recognition by emergency medical dispatchers and paramedics—San Diego experience. CONCLUSIONS Prehosp Emerg Care. 2008;12:307–313. Conducting detailed prehospital stroke assessments on 21. Dickson RL, Crowe RP, Patrick C, et al. Performance of the RACE score for patients with dispatch codes of stroke or with prehospital provider the prehospital identification of large vessel occlusion stroke in a suburban/ rural EMS service. Prehosp Emerg Care. 2019;23:612–618. impression of altered mental status or weakness in addition to pre- 22. Carrera D, Gorchs M, Querol M, et al. Revalidation of the RACE scale after its hospital provider impression of stroke may improve the prehospital regional implementation in Catalonia: a triage tool for large vessel occlusion. detection of AIS. Further prospective research is needed to evaluate J Neurointerv Surg. 2019;11:751–756. this assessment model. 23. Vidale S, Agostoni E. Prehospital stroke scales and large vessel occlusion: a systematic review. Acta Neurol Scand. 2018;138:24–31. DISCLOSURES 24. Sun CH, Nogueira RG, Glenn BA, et al. “Picture to puncture”: a novel time metric to enhance outcomes in patients transferred for endovascular reperfu- This research was funded, in part, by Agency for Healthcare sion in acute ischemic stroke. Circulation. 2013;127:1139–1148. Research and Quality K12 HS026390. This study was deemed exempt 25. Prabhakaran S, Ward E, John S, et al. Transfer delay is a major factor lim- from review by the Indiana University Institutional Review Board, iting the use of intra-arterial treatment in acute ischemic stroke. Stroke. protocol number 2003587456. Nothing to declare. 2011;42:1626–1630. © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. www.critpathcardio.com | 175 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Critical Pathways in Cardiology Wolters Kluwer Health

Factors Predicting Misidentification of Acute Ischemic Stroke and Large Vessel Occlusion by Paramedics

Download PDF
4 pages

Loading next page...
 
/lp/wolters-kluwer-health/factors-predicting-misidentification-of-acute-ischemic-stroke-and-Qyv4dFNwZz

References (56)

Publisher
Wolters Kluwer Health
Copyright
Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
ISSN
1535-282X
eISSN
1535-2811
DOI
10.1097/hpc.0000000000000307
Publisher site
See Article on Publisher Site

Abstract

LWW Original Study advanced life support (ALS) and basic life support (BLS) providers at peak times and completes 85,000 transports per year. The IEMS Factors Predicting Misidentification of Acute Ischemic Stroke protocol directs both BLS and ALS providers to obtain a blood glu- cose level and to perform the CPSS on any patient with suspected and Large Vessel Occlusion by Paramedics stroke. There are no specific mandates for symptoms that trigger the CPSS, but the protocol does suggest that patients with strokes may have “fallen, (be) unable to walk, have new balance problems Nancy K. Glober, MD,* Tyler Fulks, MD,† Michael Supples, MD,* Peter Panagos, MD,‡ and or (have) acute altered level of consciousness.” The protocol further David Kim, MD, PHD§ directs that RACE be recorded for all patients with suspected stroke, though this scale is not currently used by IEMS to direct transport. Emergency Medical Dispatch (EMD) is performed using Association 1–5 Abstract: The emergence of thrombectomy for large vessel occlusions has with LVO. Most strokes are not LVO, however, and patients with of Public-Safety Communications Officials guidecards. increased the importance of accurate prehospital identification and triage of acute non-LVO strokes benefit from rapid transport to the nearest acute ischemic stroke (AIS). Despite available clinical scores, prehospital thrombolysis-capable facility, which is not generally a thrombec- Patient and Public Involvement identification is suboptimal. Our objective was to improve the sensitivity tomy-capable comprehensive stroke center (CSC). Emergency No patient was involved. of prehospital AIS identification by combining dispatch information with medical services (EMS) providers must therefore accurately identify paramedic impression. We performed a retrospective cohort review of and transport patients with strokes to appropriate receiving centers, Measures emergency medical services and hospital records of all patients for whom a which may be the nearest primary stroke center or may be a more We collected demographic data (gender and age), EMS run stroke alert was activated in 1 urban, academic emergency department from distant CSC. descriptors (location of patient pick-up, date of service, response and January 1, 2018, to December 31, 2019. Using admission diagnosis of acute Most EMS systems approach this dilemma either by transport- transport times, dispatch code, prehospital primary impression, level stroke as outcome, we calculated the sensitivity and specificity of dispatch ing all patients with suspected stroke to the nearest thrombolysis- of service, prehospital CPSS and RACE), hospital evaluation and and paramedic impression in identifying AIS and large vessel occlusion. We capable facility, leaving subsequent transfer of LVO patients to the treatment data (physician National Institute of Health Stroke Scale, identified factors that, when included together, would improve the sensitivity emergency physician, or by using prehospital stroke severity scales thrombolysis, thrombectomy), and diagnosis of ischemic (not hemor - of prehospital AIS identification. Two-hundred twenty-six stroke alerts were to identify patients with potential LVO for transport directly to a rhagic) stroke on hospital admission as recorded in stroke neurology activated by emergency department physicians after transport by Indianapolis CSC, often bypassing the nearest thrombolysis-capable facility The . notes. Noncontrast computed tomography (CT) was completed on all emergency medical services. Forty-four percent (99/226) were female, most recent American Heart Association guidelines recommend a patients with suspected stroke. The stroke neurologist determined the median age was 58 years (interquartile range, 50–67 years), and median validated stroke screen and stroke severity score to assess for pos- need for CT angiography or magnetic resonance imaging. Data were National Institutes of Health Stroke Scale was 6 (interquartile range, 2–12). sible LVO and transport to a CSC if: last known well time is less collected in Excel (Microsoft Corporation, Redmond, WA). Paramedics demonstrated superior sensitivity (59% vs. 48%) but inferior than 24 hours, transport time to CSC will not disqualify a patient for Analysis specificity (56% vs. 73%) for detection of stroke as compared with dispatch. thrombolysis, and total transport time to the CSC is under 30 min- We described count frequencies and percentages and calcu- A strategy incorporating dispatch code of stroke, or paramedic impression utes. Systems applying this approach have demonstrated improved lated continuous variable medians and interquartile ranges (IQRs). of altered mental status or weakness in addition to stroke, would be 84% times from scene departure to thrombectomy and improved patient Chi-square compared counts, and independent-sample t test com- sensitive and 27% specific for identification of stroke. To optimize rapid and functional outcomes. However, meta-analyses demonstrate that only pared continuous variables. sensitive stroke detection, prehospital systems should consider inclusion 26%–51% of patients identified by stroke severity scales as poten- 7,10 We calculated the sensitivity and specificity of dispatch and of patients with dispatch code of stroke and provider impression of altered tial LVOs are in fact diagnosed with LVOs. Furthermore, stroke paramedic impressions of stroke in identifying strokes definitively mental status or generalized weakness. severity scores are limited in that they are only applied when a stroke diagnosed in the ED. Paramedic-suspected strokes that were not is suspected by prehospital providers. Thus, if an EMS system has Key Words: emergency medical services, large vessel occlusion, stroke called as a stroke alert by the ED physician were considered false limited sensitivity for stroke detection, stroke and LVO detection will (Crit Pathways in Cardiol 2022;21: 172–175) positives. We identified the dispatch codes and paramedic impres- necessarily be similarly limited. sions associated with prehospital false negatives and calculated the While many studies have focused on scores to optimize pre- test characteristics of alternative strategies for prehospital stroke hospital identification of acute ischemic stroke (AIS) and LVO, few identification incorporating non-stroke dispatch and paramedic have explored the characteristics of stroke and LVO cases missed by arly detection of acute stroke by paramedics in the prehospital set- impressions. Data analysis performed with SAS University Edition EMS. In this study, we identified all stroke alerts in 1 academic urban Eting facilitates appropriate triage and rapid treatment of this time- (SAS Institute, Inc., Cary, NC). emergency department (ED) after transport by Indianapolis EMS critical condition. Large vessel occlusions (LVOs) can be effectively (IEMS). We characterized paramedic sensitivity and specificity for treated with mechanical thrombectomy, which is only available at identifying stroke and LVO with their protocol-directed Cincinnati RESULTS some facilities, and which improves functional outcomes in patients Prehospital Stroke Scale (CPSS) and Rapid Arterial oCclusion Between January 1, 2018, and December 31, 2019, IEMS Evaluation (RACE) scale for identification of LVO. We identified the transported 45,339 patients to Eskenazi Hospital. Of those, 211 had dispatch codes and paramedic impressions associated with missed a paramedic impression of stroke. Median public-safety answering From the *Department of Emergency Medicine, Indiana University, Indianapolis, AISs and suggest a strategy to optimize sensitivity for prehospital point call to dispatch time was 1 minute (IQR, 0–2 minutes), median IN; †Department of Emergency Medicine, Southern Illinois University, stroke detection. Springfield, IL; ‡Department of Emergency Medicine, Washington University scene time was 14 minutes (IQR, 11–18 minutes), and median trans- at St. Louis, St. Louis, MO; and §Department of Emergency Medicine, port time to Eskenazi ED was 11 minutes (IQR, 8–14 minutes). Stanford University, Palo Alto, CA. Patients were transported by ALS in 199 (88%) cases and by BLS Presented at the National Association of Emergency Medical Services Physicians, METHODS in 26 (12%) cases. Among all stroke codes, using admission diagno- virtual abstract presentation January 2021. This study was deemed exempt from review by the Indiana Supplemental digital content is available for this article. Direct URL citations sis of stroke as the gold standard, paramedics demonstrated superior University Institutional Review Board, protocol number 2003587456. appear in the printed text and are provided in the HTML and PDF versions of sensitivity (58% vs. 48%) but inferior specificity (56% vs. 73%) for this article on the journal’s Web site (www.critpathcardio.com). detection of stroke as compared with dispatch ( Table 1). Study Design and Setting Reprints: Nancy K. Glober, MD, Department of Emergency Medicine, Indiana When strokes were not identified in the prehospital setting, University, 1701N Senate Ave, Indianapolis, IN 46202. E-mail: nglober@ We retrospectively reviewed the in-hospital and prehospital iu.edu. the most common dispatch codes were sick person (21), chest pain electronic medical records from January 1, 2018, to December 31, Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. This (12), and syncope (10). When paramedics failed to identify AIS, their 2019, for every patient transported by IEMS on whom a stroke alert is an open-access article distributed under the terms of the Creative Commons most common impressions were altered mental status (14), general- Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), (ie, potential acute stroke) was activated by an ED physician based ized weakness (11), and chest pain (6) ( Table 2). where it is permissible to download and share the work provided it is properly on patient evaluation in 1 academic urban ED (Eskenazi Hospital). cited. The work cannot be changed in any way or used commercially without Using either dispatch or paramedic impression of stroke In Marion County, Indianapolis, EMS care is provided to the permission from the journal. would improve sensitivity to 77.2% (115/149) at a specificity of population of about 900,000 by IEMS as well as paramedics based ISSN: 1003-0117/22/2104-0172 8 46.8% (36/77) among stroke codes. Identifying possible strokes with DOI: 10.1097/HPC.0000000000000307 at fire stations. IEMS operates 31 ambulances staffed with both 172 | www.critpathcardio.com Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 Factors Predicting Misidentification of AIS Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 and LVO by Paramedics advanced life support (ALS) and basic life support (BLS) providers TABLE 1. Sensitivity and Specificity for Identification of at peak times and completes 85,000 transports per year. The IEMS Acute Ischemic Stroke by Dispatch and Paramedics protocol directs both BLS and ALS providers to obtain a blood glu- Test Characteristics Dispatch Paramedics cose level and to perform the CPSS on any patient with suspected stroke. There are no specific mandates for symptoms that trigger Sensitivity 47.7% 58.4% the CPSS, but the protocol does suggest that patients with strokes Specificity 72.7% 54.5% may have “fallen, (be) unable to walk, have new balance problems or (have) acute altered level of consciousness.” The protocol further directs that RACE be recorded for all patients with suspected stroke, TABLE 2. When Strokes Were Not Identified by though this scale is not currently used by IEMS to direct transport. (A) Dispatch or (B) Paramedic, Impressions Were Emergency Medical Dispatch (EMD) is performed using Association Varied, But Most Commonly “Sick Person” and of Public-Safety Communications Officials guidecards. “Altered Mental Status” Patients With Strokes That Patient and Public Involvement (A) Dispatch Code Were Not Identified (%) No patient was involved. Sick person 21 (9.3) Measures Chest pain 12 (5.3) We collected demographic data (gender and age), EMS run Syncope 10 (4.4) descriptors (location of patient pick-up, date of service, response and Diabetic problem 8 (3.5) transport times, dispatch code, prehospital primary impression, level (B) Primary Provider ImpressionPatients With Strokes That Were Not of service, prehospital CPSS and RACE), hospital evaluation and Identified (%) treatment data (physician National Institute of Health Stroke Scale, Altered mental status 14 (6.2) thrombolysis, thrombectomy), and diagnosis of ischemic (not hemor - rhagic) stroke on hospital admission as recorded in stroke neurology Generalized weakness 11 (4.9) notes. Noncontrast computed tomography (CT) was completed on all Chest pain 6 (2.7) patients with suspected stroke. The stroke neurologist determined the Dizziness 4 (1.8) need for CT angiography or magnetic resonance imaging. Data were Diabetic hypoglycemia 3 (1.3) collected in Excel (Microsoft Corporation, Redmond, WA). Headache 3 (1.3) Analysis Seizure 2 (0.9) We described count frequencies and percentages and calcu- lated continuous variable medians and interquartile ranges (IQRs). paramedic impressions of stroke, altered mental status, or generalized Chi-square compared counts, and independent-sample t test com- weakness would achieve 75.2% sensitivity (112/149) at a specificity pared continuous variables. of 31.2% (24/77). A strategy using either dispatch code of stroke, We calculated the sensitivity and specificity of dispatch and or prehospital impression of stroke, altered mental status, or weak- paramedic impressions of stroke in identifying strokes definitively ness would be 83.9% sensitive (125/149) and 27.3% specific (21/77) diagnosed in the ED. Paramedic-suspected strokes that were not for identification of stroke, among patients activated as stroke alerts. called as a stroke alert by the ED physician were considered false When calculated among all patients dispatched as stroke or with an positives. We identified the dispatch codes and paramedic impres- EMS primary provider impression of stroke, altered mental status, or sions associated with prehospital false negatives and calculated the weakness, the specificity would improve to 65.9% (29,812/45,214) test characteristics of alternative strategies for prehospital stroke for identification of stroke. identification incorporating non-stroke dispatch and paramedic RACE was documented for 47% (106/226) of stroke alert impressions. Data analysis performed with SAS University Edition patients; paramedics documented “unable to complete” RACE in (SAS Institute, Inc., Cary, NC). 5% (12/226). In our system, RACE ≥ 5 was 71% sensitive and 57% specific for identification of an LVO confirmed by CT angiography. RESULTS All 11 patients determined to have LVO by CT angiography received Between January 1, 2018, and December 31, 2019, IEMS mechanical thrombectomy. Of the patients who had LVOs, 8 (73%) transported 45,339 patients to Eskenazi Hospital. Of those, 211 had had a primary provider impression of stroke. The other 3 had various a paramedic impression of stroke. Median public-safety answering primary provider impressions and dispatch codes ( Table 3). point call to dispatch time was 1 minute (IQR, 0–2 minutes), median Applying RACE to all patients with dispatch code of stroke scene time was 14 minutes (IQR, 11–18 minutes), and median trans- or paramedic impression of stroke, altered mental status or general- port time to Eskenazi ED was 11 minutes (IQR, 8–14 minutes). ized weakness would have potentially identified 1/11 (9%) additional Patients were transported by ALS in 199 (88%) cases and by BLS LVO in the prehospital setting. in 26 (12%) cases. Among all stroke codes, using admission diagno- Of the 211 patients with paramedic impression of stroke, ED sis of stroke as the gold standard, paramedics demonstrated superior physicians activated 122 (57.8%) stroke alerts upon arrival in the ED. sensitivity (58% vs. 48%) but inferior specificity (56% vs. 73%) for detection of stroke as compared with dispatch ( Table 1). When strokes were not identified in the prehospital setting, TABLE 3. LVO Not Identified As Strokes by the Prehospital the most common dispatch codes were sick person (21), chest pain Provider (12), and syncope (10). When paramedics failed to identify AIS, their most common impressions were altered mental status (14), general Case - Dispatch Code Primary Provider Impression ized weakness (11), and chest pain (6) ( Table 2). Patient 1 Stroke Altered mental status Using either dispatch or paramedic impression of stroke Patient 2 Breathing problem Acute respiratory distress would improve sensitivity to 77.2% (115/149) at a specificity of Patient 3 Psych problem Behavioral/psych episode 46.8% (36/77) among stroke codes. Identifying possible strokes with © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. www.critpathcardio.com | 173 Glober et al Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 The other 89 (42.2%) were determined by the ED physician to not primary impression of stroke. We noted differences in demograph- identification of the subset of stroke patients experiencing LVO in have sustained a stroke. There were an additional 104 ED stroke aler ics. P t atients transported to Eskenazi Hospital versus any other the prehospital setting is essential to directing appropriate transport activations after transport by IEMS that were not identified as strokhospital w es ere significantly younger (median [IQR] age = 59.0 destination decisions. We described the ability of EMS dispatch and by the IEMS provider (Fig. 1). Of the 122 patients activated as strok[50.0–67.0] vs. 68.0 [58.0–70.0]; e P < 0.001), more often male paramedics to identify strokes in the prehospital setting in 1 large alerts in the ED, 99 (44%) were female, median age was 58 years (116 [56.0%] vs. 595 [43.5%]; P < 0.001), and less often White urban EMS system. The rate of missed stroke by both EMD and (IQR, 50–67 years), median National Institute of Health Stroke Scale (87 [42.0%] vs. 860 [62.8%]; P < 0.001). However, we did not prehospital professionals is higher than the ED Our data suggest . was 6 (IQR, 2–12). find significant difference in CPSS positivity (182 [87.9%] vs. combinations of dispatch and paramedic impressions that could One-hundred forty-nine (65.9%) of ED stroke alerts were 1157 [84.5%]; P = 0.212) (Supplemental Table 1, http://links.lww. improve the prehospital detection of AIS and marginally improve admitted with a diagnosis of AIS, and 77 (34.1%) were deter - com/HPC/A244). prehospital detection of LVO. mined not to be strokes after imaging and evaluation by a stroke Numerous prehospital stroke scales aim to identify patients DISCUSSION neurologist ( Table 4). Of the 149 patients admitted for stroke, 44 experiencing LVO, and in our system, both RACE and the CPSS (30%) patients received thrombolysis, and 11 (7%) underwent Rapid and accurate prehospital identification of AIS is are used. The RACE scale demonstrates similar predictiv-e per thrombectomy for LVO. All patients found to have an LVO under - critically important, given the time-sensitive nature of available formance to the CPSS, Los Angeles Motor Screen, and Vision, 7 10,11 14 went mechanical thrombectomy.Although our efforts were basedtreatments, and is associated with improved outcomes. The Aphasia, and Neglect instruments, but may have inferior per - on a convenience sample of patients taken to 1 ED, we compareda vailability of CSCs that offer endovascular therapy for patients formance to other prehospital stroke scales. The sensitivity of characteristics of potential stroke patients taken to Eskenazi with with L VO is limited as compared with more common primary stroke paramedics in this study to identify AIS was similar to previously 12 16–20 those taken to other hospitals. During the study period, 1572 centers that can pro vide medical thrombolysis. Given that endo- reported sensitivities for large metropolitan EMS systems, 1,7 patients were transported to any hospital by IEMS with paramedic vascular therap y for LVO is superior to medical therapy alone, and IEMS utilization of the RACE score demonstrated similar 21–23 sensitivity and specificity as previously reported. Failure to recognize a patient having a stroke likely represents a significant 45,339 paents transported barrier to applying a prehospital stroke scale. Our findings sug- to Eskenazi Hospital by IEMS gest that in addition to primary provider impression of stroke, all EMS responses with a dispatch of stroke or with primary provider impression of altered mental status or weakness should be con- sidered as possible AISs. This underscores the importance of both 211 paramedic primary 104 acvated as stroke paramedic and EMD evaluation to identify AIS in the prehospital impression of stroke alert by emergency setting. physician, but not 89 determined not to have a While it is not practical to transport every patient with a dis- idenfied as stroke by stroke by EM physician patch or prehospital impression of altered mental status and gener - paramedic primary 122 acve as stroke alized weakness to a CSC, these are the most common patients in impression alert by EM physician whom AIS is missed, and special attention should be paid to these patients, for whom a more detailed prehospital stroke assessment should be performed, and for whom early hospital notification might 37 EMS suspected strokes be considered. Some over-triage of patients with potential AIS to determined not to have a 226 IEMS paents idenfied as higher levels of care may be appropriate given the time-sensitive stroke by neurology possible stroke by EM physician 24,25 nature of stroke treatment. This study has limitations. The specific features associated FIGURE 1. Flow chart of patients with incorrect prehospital impressions cannot be ascertained from identified for the study. EM indi- 126 EMS false posive 87 EMS true posive 61 EMS false negave 41 EMS true negave the available data. Of the subset of patients with a false-negative cates Emergency medicine. EMS primary impression, it is unclear whether a more detailed stroke assessment would have improved detection of LVO. Including TABLE 4. ED Stroke Alert Patient Demographic and Prehospital Factors Stratified by Neurologist Confirmed Acute Ischemic only patients transported to an academic center with a CSC may bias Stroke Status the study population toward sicker patients. While there were - differ ences in demographics of patients with paramedic-suspected stroke Confirmed Stroke No Stroke by transport to the study hospital versus other area hospitals, there Patient Factors Median IQR Median IQR P was no difference in positivity of CPSS. Expanding the use of a pre- hospital stroke scale to the most common false-negative dispatch Age 59.5 51.3–67.0 56.5 44.8–65.3 0.083 codes and provider impressions should be investigated with a future NIHSS 5.0 2.0–10.0 5.5 2.0–14.0 0.358 prospective study and in other EMS systems to ascertain generaliz- Sex n (%) 95% CI n (%) 95% CI 0.574 ability of our findings. Male 62 (41.9) 33.9–49.8 36 (46.2) 35.1–57.2 Female 86 (58.1) 50.2–66.1 42 (53.8) 42.8–64.9 CONCLUSIONS CPSS 0.181 Conducting detailed prehospital stroke assessments on Positive 95 (64.5) 56.5–71.9 41 (52.6) 41.5–63.6 patients with dispatch codes of stroke or with prehospital provider Negative 15 (10.1) 5.3–15.0 8 (10.3) 3.5–17.0 impression of altered mental status or weakness in addition to pre- Not performed 38 (25.7) 18.6–32.7 29 (37.2) 26.5–47.9 hospital provider impression of stroke may improve the prehospital Paramedic impression stroke 0.103 detection of AIS. Further prospective research is needed to evaluate Yes 87 (58.8) 50.9–66.7 37 (47.4) 36.4–58.5 this assessment model. No 61 (41.2) 33.3–49.1 41 (52.6) 41.5–63.6 Dispatch code was stroke <0.01 DISCLOSURES Yes 71 (48.0) 39.9–56.0 21 (26.9) 17.1–36.8 This research was funded, in part, by Agency for Healthcare Research and Quality K12 HS026390. This study was deemed exempt No 77 (52.0) 44.0–60.1 57 (73.1) 63.2–82.9 from review by the Indiana University Institutional Review Board, CI indicates confidence interval; NIHSS, National Institute of Health Stroke Scale. protocol number 2003587456. Nothing to declare. 174 | www.critpathcardio.com © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. Factors Predicting Misidentification of AIS Critical Pathways in Cardiology • Volume 21, Number 4, December 2022 and LVO by Paramedics identification of the subset of stroke patients experiencing LVO in REFERENCES the prehospital setting is essential to directing appropriate transport 1. Albers GW, Lansberg MG, Kemp S, et al. A multicenter randomized con- destination decisions. We described the ability of EMS dispatch and trolled trial of endovascular therapy following imaging evaluation for isch- emic stroke (DEFUSE 3). Int J Stroke. 2017;12:896–905. paramedics to identify strokes in the prehospital setting in 1 large 2. Berkhemer OA, Fransen PSS, Beumer D, et al; MR CLEAN Investigators. A urban EMS system. The rate of missed stroke by both EMD and 13 randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J prehospital professionals is higher than the ED Our data suggest . Med. 2015;372:11–20. combinations of dispatch and paramedic impressions that could 3. Saver JL, Goyal M, Bonafe A, et al; SWIFT PRIME Investigators. Stent- improve the prehospital detection of AIS and marginally improve retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke. N Engl prehospital detection of LVO. J Med. 2015;372:2285–2295. Numerous prehospital stroke scales aim to identify patients 4. Goyal M, Demchuk AM, Menon BK, et al; ESCAPE Trial Investigators. Randomized assessment of rapid endovascular treatment of ischemic stroke. experiencing LVO, and in our system, both RACE and the CPSS N Engl J Med. 2015;372:1019–1030. are used. The RACE scale demonstrates similar predictiv-e per 5. Jovin TG, Chamorro A, Cobo E, et al; REVASCAT Trial Investigators. formance to the CPSS, Los Angeles Motor Screen, and Vision, Thrombectomy within 8 hours after symptom onset in ischemic stroke. N Engl Aphasia, and Neglect instruments, but may have inferior per - J Med. 2015;372:2296–2306. formance to other prehospital stroke scales. The sensitivity of 6. Adeoye O, Albright KC, Carr BG, et al. Geographic access to acute stroke care paramedics in this study to identify AIS was similar to previously in the United States. Stroke. 2014;45:3109–3024. 16–20 reported sensitivities for large metropolitan EMS systems, 7. Lima FO, Mont’Alverne FJA, Bandeira D, et al. Pre-hospital assessment of large vessel occlusion strokes: implications for modeling and planning stroke and IEMS utilization of the RACE score demonstrated similar 21–23 systems of care. Front Neurol. 2019;10:955. sensitivity and specificity as previously reported. Failure to 8. American Heart Association; Mission: Lifeline Stroke. Emergency Medical recognize a patient having a stroke likely represents a significant Services Acute Stroke Triage and Routing. Available at: https://www.heart. barrier to applying a prehospital stroke scale. Our findings sug- org/idc/groups/ahaecc-public/@wcm/@gwtg/documents/downloadable/ gest that in addition to primary provider impression of stroke, all ucm_492341.pdf. Accessed 2021. EMS responses with a dispatch of stroke or with primary provider 9. Jayaraman MV, Hemendinger ML, Baird GL, et al. Field triage for endovas- cular stroke therapy: a population-based comparison. J NeuroIntervent Surg. impression of altered mental status or weakness should be con- 2020;12:233–239. sidered as possible AISs. This underscores the importance of both 10. Lin CB, Peterson ED, Smith EE, et al. Emergency medical service hospital paramedic and EMD evaluation to identify AIS in the prehospital prenotification is associated with improved evaluation and treatment of acute setting. ischemic stroke. Circ Cardiovasc Qual Outcomes. 2012;5:514–522. While it is not practical to transport every patient with a dis- 11. Patel MD, Rose KM, O’Brien EC, et al. Prehospital notification by emergency patch or prehospital impression of altered mental status and gener - medical services reduces delays in stroke evaluation: findings from the North Carolina stroke care collaborative. Stroke. 2011;42:2263–2268. alized weakness to a CSC, these are the most common patients in 12. Schieb LJ, Casper ML, George MG. Mapping primary and comprehensive whom AIS is missed, and special attention should be paid to these stroke centers by Certification Organization. Circ Cardiovasc Qual Outcomes. patients, for whom a more detailed prehospital stroke assessment 2015;8(6 Suppl 3):S193–S194. should be performed, and for whom early hospital notification might 13. Madsen TE, Khoury J, Cadena R, et al. Potentially missed diagnosis of isch- be considered. Some over-triage of patients with potential AIS to emic stroke in the Emergency Department in the Greater Cincinnati/Northern higher levels of care may be appropriate given the time-sensitive Kentucky Stroke Study. Acad Emerg Med. 2016;23:1128–1135. 24,25 nature of stroke treatment. 14. Crowe RP, Myers JB, Fernandez AR, et al. The Cincinnati Prehospital Stroke Scale compared to stroke severity tools for large vessel occlusion stroke pre- This study has limitations. The specific features associated diction. Prehosp Emerg Care. 2020;1:9. with incorrect prehospital impressions cannot be ascertained from 15. Zhelev Z, Walker G, Henschke N, et al. Prehospital stroke scales as screening the available data. Of the subset of patients with a false-negative tools for early identification of stroke and transient ischemic attack. Cochrane EMS primary impression, it is unclear whether a more detailed Database Syst Rev. 2019;4:CD0114–C27-CD. stroke assessment would have improved detection of LVO. Including 16. Brandler ES, Sharma M, McCullough F, et al. Prehospital stroke identifica- only patients transported to an academic center with a CSC may bias tion: factors associated with diagnostic accuracy. J Stroke Cerebrovasc Dis. 2015;24:2161–2166. the study population toward sicker patients. While there were - differ 17. Oostema JA, Konen J, Chassee T, et al. Clinical predictors of accurate prehos- ences in demographics of patients with paramedic-suspected stroke pital stroke recognition. Stroke. 2015;46:1513–1517. by transport to the study hospital versus other area hospitals, there 18. Mould-Millman N-K, Meese H, Alattas I, et al. Accuracy of prehospital iden- was no difference in positivity of CPSS. Expanding the use of a pre- tification of stroke in a large stroke belt municipality Prehosp . Emerg Care. hospital stroke scale to the most common false-negative dispatch 2018;22:734–742. codes and provider impressions should be investigated with a future 19. Gropen TI, Gokaldas R, Poleshuck R, et al. Factors related to the sensitiv- prospective study and in other EMS systems to ascertain generaliz- ity of emergency medical service impression of stroke. Prehosp Emerg Care. 2014;18:387–392. ability of our findings. 20. Ramanujam P, Guluma KZ, Castillo EM, et al. Accuracy of stroke recognition by emergency medical dispatchers and paramedics—San Diego experience. CONCLUSIONS Prehosp Emerg Care. 2008;12:307–313. Conducting detailed prehospital stroke assessments on 21. Dickson RL, Crowe RP, Patrick C, et al. Performance of the RACE score for patients with dispatch codes of stroke or with prehospital provider the prehospital identification of large vessel occlusion stroke in a suburban/ rural EMS service. Prehosp Emerg Care. 2019;23:612–618. impression of altered mental status or weakness in addition to pre- 22. Carrera D, Gorchs M, Querol M, et al. Revalidation of the RACE scale after its hospital provider impression of stroke may improve the prehospital regional implementation in Catalonia: a triage tool for large vessel occlusion. detection of AIS. Further prospective research is needed to evaluate J Neurointerv Surg. 2019;11:751–756. this assessment model. 23. Vidale S, Agostoni E. Prehospital stroke scales and large vessel occlusion: a systematic review. Acta Neurol Scand. 2018;138:24–31. DISCLOSURES 24. Sun CH, Nogueira RG, Glenn BA, et al. “Picture to puncture”: a novel time metric to enhance outcomes in patients transferred for endovascular reperfu- This research was funded, in part, by Agency for Healthcare sion in acute ischemic stroke. Circulation. 2013;127:1139–1148. Research and Quality K12 HS026390. This study was deemed exempt 25. Prabhakaran S, Ward E, John S, et al. Transfer delay is a major factor lim- from review by the Indiana University Institutional Review Board, iting the use of intra-arterial treatment in acute ischemic stroke. Stroke. protocol number 2003587456. Nothing to declare. 2011;42:1626–1630. © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. www.critpathcardio.com | 175

Journal

Critical Pathways in Cardiology – Wolters Kluwer Health

Published: Dec 22, 2022

There are no references for this article.