Cardiac & Resuscitation, Patient Care, Special Topics

Study Shows Race and Sex Disparities in Prehospital Stroke Recognition

Issue 9 and Volume 40.

Subconscious Judgment

Govindarajan P, Friedman BT, Delgadillo JQ, et al. Race and sex disparities in prehospital recognition of acute stroke. Acad Emerg Med. 2015;22(3):264–272.

The U.S. Department of Health and Human Services has been tracking and reporting significant healthcare disparities since 2003. Unfortunately, a decade later, the 2014 National Healthcare Disparities Report describes prevailing and significant racial and socioeconomic disparities in the delivery of care.

Although we’d like to think EMS provides the same high-quality patient care regardless of ethnicity or socioeconomic status, there’s been mounting evidence that we’re just as prone to inequities as the rest of healthcare. Stroke care is one example, but EMS personnel should be hypervigilant to avoid subconscious judgments that may provide substandard care for any patient population.

Methods: In this study, Govindarajan, from the University of California, San Francisco, led a diverse team of researchers comparing hospital discharge diagnosis to field impressions documented in EMS electronic charts from two California counties.

The team used probabilistic linking to retrospectively associate EMS and hospital database records. They analyzed records from 14 hospitals in these two counties from 2005– 2007, and identified 10,719 stroke patients.

Retrospective “data-mining” practices with large databases pose serious threats to the validity of a study. In this case however, researchers did a great job of finding all patients who were discharged with a diagnosis of stroke, and then narrowing those down to patients who were transported by ambulance. They linked 3,787 patient records.

Results: EMS documented stroke as their field impression in only 1,223 patients (sensitivity = 32%). Unfortunately this means that either EMS didn’t document well, the patients hadn’t yet developed identifiable symptoms or, more alarmingly, EMS didn’t recognize the stroke. The good news is that crews were extremely accurate (specificity = 99%) in those they did identify.

The authors report that correct recognition of stroke was lower in Asians and Hispanics (27% and 29%, respectively), and women (30%). This was statistically significant (adjusted odds ratios of 0.77 for Hispanics, 0.66 for Asians and 0.82 for women).

Discussion: Other studies have reported better identification of strokes in some EMS system (such as North Carolina). We also know that women present with nontraditional symptoms, such as fatigue, weakness and disorientation. The authors of this paper should be commended for their efforts to provide targeted education to improve stroke recognition.

Interestingly, the difference in stroke recognition for African American patients in this study wasn’t statistically significant. We should be careful not to dismiss this lack of statistical significance, since it may still be clinically significant. As Govindarajan and his co-authors note, other studies identified EMS disparities for African American stroke patients.

This data set didn’t have data on the ethnicity of the EMS provider or language barriers. The team did, however, factor language into their analysis and that independent variable alone is unlikely to account for the failed stroke identifications.

While this data is relatively old, and care may have already improved, it indicates EMS clinicians should have a much higher index of suspicion for stroke, especially in non-white and female patients.


GLOSSARY

Probabilistic linking of database records involves sophisticated computer algorithms that compare numerous data fields (e.g., birthdate, sex, date of service, etc.), leading to a high probability that two records refer to the same patient. As EMS focuses on evidence-based care, this technique could be key to prehospital studies focused on patient outcome. Researchers can remain blinded to private patient data and can analyze the anonymous records linked to discharge data. The technique has been used successfully in multiple EMS studies and should become more popular as our electronic charting continues to evolve and improve.

BOTTOM LINE

What we know: Early recognition of acute strokes improves outcomes. EMS identification, transport to an appropriate receiving center and proper notification make a significant difference in outcomes. National statistics show that, overall, patients of low socioeconomic status and minorities receive a lesser quality of care than their white and affluent counterparts.

What this study adds: New evidence shows that prehospital EMS crews don’t recognize strokes in female, Asian and Hispanic populations as often as they do with male and non-Hispanic white populations.

Visit www.pcrfpodcast.org for audio commentary.

Cardiac & Resuscitation, Columns, Patient Care, Special Topics

Study Shows Race and Sex Disparities in Prehospital Stroke Recognition

Issue 9 and Volume 40.

Subconscious Judgment

Govindarajan P, Friedman BT, Delgadillo JQ, et al. Race and sex disparities in prehospital recognition of acute stroke. Acad Emerg Med. 2015;22(3):264–272.

The U.S. Department of Health and Human Services has been tracking and reporting significant healthcare disparities since 2003. Unfortunately, a decade later, the 2014 National Healthcare Disparities Report describes prevailing and significant racial and socioeconomic disparities in the delivery of care.

Although we’d like to think EMS provides the same high-quality patient care regardless of ethnicity or socioeconomic status, there’s been mounting evidence that we’re just as prone to inequities as the rest of healthcare. Stroke care is one example, but EMS personnel should be hypervigilant to avoid subconscious judgments that may provide substandard care for any patient population.

Methods: In this study, Govindarajan, from the University of California, San Francisco, led a diverse team of researchers comparing hospital discharge diagnosis to field impressions documented in EMS electronic charts from two California counties.

The team used probabilistic linking to retrospectively associate EMS and hospital database records. They analyzed records from 14 hospitals in these two counties from 2005– 2007, and identified 10,719 stroke patients.

Retrospective “data-mining” practices with large databases pose serious threats to the validity of a study. In this case however, researchers did a great job of finding all patients who were discharged with a diagnosis of stroke, and then narrowing those down to patients who were transported by ambulance. They linked 3,787 patient records.

Results: EMS documented stroke as their field impression in only 1,223 patients (sensitivity = 32%). Unfortunately this means that either EMS didn’t document well, the patients hadn’t yet developed identifiable symptoms or, more alarmingly, EMS didn’t recognize the stroke. The good news is that crews were extremely accurate (specificity = 99%) in those they did identify.

The authors report that correct recognition of stroke was lower in Asians and Hispanics (27% and 29%, respectively), and women (30%). This was statistically significant (adjusted odds ratios of 0.77 for Hispanics, 0.66 for Asians and 0.82 for women).

Discussion: Other studies have reported better identification of strokes in some EMS system (such as North Carolina). We also know that women present with nontraditional symptoms, such as fatigue, weakness and disorientation. The authors of this paper should be commended for their efforts to provide targeted education to improve stroke recognition.

Interestingly, the difference in stroke recognition for African American patients in this study wasn’t statistically significant. We should be careful not to dismiss this lack of statistical significance, since it may still be clinically significant. As Govindarajan and his co-authors note, other studies identified EMS disparities for African American stroke patients.

This data set didn’t have data on the ethnicity of the EMS provider or language barriers. The team did, however, factor language into their analysis and that independent variable alone is unlikely to account for the failed stroke identifications.

While this data is relatively old, and care may have already improved, it indicates EMS clinicians should have a much higher index of suspicion for stroke, especially in non-white and female patients.

GLOSSARY

Probabilistic linking of database records involves sophisticated computer algorithms that compare numerous data fields (e.g., birthdate, sex, date of service, etc.), leading to a high probability that two records refer to the same patient. As EMS focuses on evidence-based care, this technique could be key to prehospital studies focused on patient outcome. Researchers can remain blinded to private patient data and can analyze the anonymous records linked to discharge data. The technique has been used successfully in multiple EMS studies and should become more popular as our electronic charting continues to evolve and improve.

BOTTOM LINE

What we know: Early recognition of acute strokes improves outcomes. EMS identification, transport to an appropriate receiving center and proper notification make a significant difference in outcomes. National statistics show that, overall, patients of low socioeconomic status and minorities receive a lesser quality of care than their white and affluent counterparts.

What this study adds: New evidence shows that prehospital EMS crews don’t recognize strokes in female, Asian and Hispanic populations as often as they do with male and non-Hispanic white populations.

Visit www.pcrfpodcast.org for audio commentary.