Abstract
Background
Few studies have examined the prehospital presentation, assessment, or treatment of patients diagnosed with coronavirus disease 2019 (COVID-19). The objective of this preliminary report is to describe prehospital encounters for patients with a COVID-19 hospital diagnosis and/or COVID-19 EMS suspicion versus those with neither a hospital diagnosis nor EMS suspicion of the disease.
Methods
This case series evaluated electronic patient care records from EMS agencies participating in a large national bi-directional data exchange. All records for 9-1-1 responses between March 1 and April 19, 2020, resulting in transport to a hospital, with at least one ICD-10 outcome returned via the data exchange were included. Hospital ICD-10 codes used to determine COVID-19 diagnoses included B97.2, B97.21, B97.29, B34.2, and U07.1. COVID-19 EMS suspicion was defined as a documented EMS primary or secondary impression of COVID-19, or indication of COVID-19 suspicion in the prehospital free-text narrative. Comparisons were stratified by COVID-19 hospital diagnosis and COVID-19 EMS suspicion. Descriptive and comparative statistics are presented.
Results
There were 84,540 EMS patient records with linked hospital ICD-10 codes included. Of those, 814 (1%) patients had a COVID-19 hospital diagnosis. Overall, COVID-19 EMS suspicion was documented for 3,204 (4%) patients. A COVID-19 EMS suspicion was documented for 636 (78%) of hospital diagnosed COVID-19 patients. Those with COVID-19 hospital diagnoses were more likely to present with tachycardia, tachypnea, hypoxia, and fever during the EMS encounter. EMS responses for patients diagnosed with COVID-19 were also more likely to originate from a skilled nursing/assisted living facility. EMS PPE (eye protection, mask, or gown) use was more frequently documented on records of patients who had hospital diagnosed COVID-19.
Conclusion
In this large sample of prehospital encounters, EMS COVID-19 suspicion demonstrated sensitivity of 78% and positive predictive value of 20% compared with hospital ICD-10 codes. These data indicate that EMS suspicion alone is insufficient to determine appropriate utilization of PPE.
Introduction
As the coronavirus disease 2019 (COVID-19) pandemic continues to evolve, Emergency Medical Services (EMS) providers are often a first line of healthcare response for patients experiencing complications of this disease. The burden of disease is significant, with over 10% of administered tests being positive in some emergency departments as well as some community settings in the United States (Citation1, Citation2). Patients with COVID-19 may be asymptomatic, experience mild to moderate symptoms like fever, cough, and chills, or experience more severe and even life-threatening symptoms such as shortness of breath and hypoxia (Citation3). Patients experiencing symptoms related to COVID-19 may request EMS for emergent treatment and transport to a hospital. Prehospital identification of patients experiencing signs and symptoms related to COVID-19 is vital to ensure appropriate use of personal protective equipment (PPE) and reduce risk of exposure to EMS providers. Nevertheless, the characteristics of prehospital encounters for patients with COVID-19 have not been thoroughly explored.
The objective of this preliminary report is to describe a case series of prehospital 9-1-1 encounters for patients with a COVID-19 hospital diagnosis and/or COVID-19 EMS suspicion versus those with neither a hospital diagnosis nor EMS suspicion of the disease. This preliminary report is intended to provide additional information for the ongoing evaluation, treatment, and transport by EMS of patients with suspected COVID-19, with an emphasis on informing PPE utilization.
Methods
Study Design and Data Source
All records were obtained using the de-identified research database maintained by ESO, Inc. (Austin, TX), a large provider of prehospital electronic health record software in the US. The prehospital electronic health record software collects event dispatch data, patient demographic characteristics, clinical presentation as well as course, intervention, treatment, and outcome at the transfer of care. In 2019, there were more than 1,350 EMS agencies that opted to share their de-identified data for the purposes of research and benchmarking in the database. A subset of these agencies participated in the health data exchange (HDE), which allowed for integration of hospital outcome data, including diagnoses, with the prehospital patient care record. Electronic patient care records were included in this case series if the 9-1-1 request was received between March 1 and April 19, 2020, the request for service resulted in transport to a hospital, and at least one emergency department or hospital ICD-10 outcome code was returned from the hospital via the data exchange.
The institutional review board at St. David’s HealthCare (Austin, TX) determined that this study was exempt.
Measures
COVID-19 Hospital Diagnosis
We used hospital ICD-10 codes to determine diagnoses of COVID-19. We included codes: U07.1 (COVID-19), B.97.2 (Coronavirus as the cause of diseases), B97.21 (SARS-associated coronavirus as the cause of diseases), B97.29 (Other coronaviruses as the cause of diseases), and B34.2 (Coronavirus infection, unspecified).
COVID-19 EMS Suspicion
COVID-19 EMS suspicion was determined using the discrete data fields for EMS primary or secondary impression as well as manual review of the open text EMS narrative. Within the prehospital electronic health record, EMS providers document their clinical impressions of the patient’s problem or condition using the “primary impression” and “secondary impression” data fields. These data fields are populated using a pre-determined drop-down list of clinical impressions. Three impressions related to COVID-19 were incorporated into the primary and secondary impressions lists on March 5, 2020: COVID-19 – Confirmed by testing, COVID-19 – Exposure to confirmed patient, and COVID-19 – Suspected – no known exposure. Additionally, an impression for Severe Acute Respiratory Syndrome (SARS) already existed and was included as this may have been selected in the absence of the COVID-19 specific impressions in the first few days of the study period. For analysis, presence of any of these four impressions as either an EMS primary or secondary impression was grouped into the category of COVID-19 related impressions. In addition to the discrete data fields for the primary and secondary clinical impression, EMS providers also complete a free-text narrative section of the electronic health record. Because it is possible that EMS providers did not choose one of the COVID-19 related impressions but may have indicated suspicion of COVID-19 in the free-text portion of the record, a narrative review was conducted.
For records that did not have a COVID-19 related impression documented, three investigators (SB, RPC, JBM) performed a review of the free-text prehospital narratives. Based on consensus from the reviewers, the following criteria were developed at the start of the narrative review process to determine whether an EMS suspicion of COVID-19 was documented in the narrative: 1) a confirmed positive COVID-19 test result or person under investigation, 2) an increase in PPE after encountering the patient, 3) placement of a surgical mask on the patient following history-taking, 4) listing of other sick persons in the home with a focus on respiratory/hypoxia symptoms, 5) viral alert called to the hospital. Sampling of narratives was stratified by COVID-19 hospital diagnosis status. Narratives for all records with a COVID-19 hospital diagnosis and no documented COVID-19 EMS impression were reviewed. Because reviewing all prehospital narratives for patients with neither a COVID-19 hospital diagnosis nor a documented COVID-19 EMS impression was not feasible, a sample size calculation was performed to determine the number of narrative reviews needed to make estimates with 95% confidence. Assuming a 5% sampling error and a conservative 50/50 split (Citation4), the calculation determined that at least 384 randomly-selected narrative reviews were needed for the group of records without a COVID-19 hospital diagnosis. Using a statistical software package (Stata version 15.1: StataCorp LLC, College Station, TX), we randomly selected 450 narratives from records with no COVID-19 hospital diagnosis and no documented COVID-19 EMS impression. All narratives selected for review were loaded into the same statistical software and randomly assigned to one of the three reviewers. Reviewers were blinded to the COVID-19 hospital diagnosis status of records. A set of 50 narratives (25 with COVID-19 hospital diagnosis and 25 without COVID-19 hospital diagnosis) was assigned to all three reviewers to assess interrater agreement. Reviewers were blinded to which narratives were assigned to all three reviewers. Interrater agreement was calculated as percentage agreement (the number of records for which all three reviewers agreed divided by 50).
For records with documented COVID-19 hospital diagnosis, we defined COVID-19 EMS suspicion as either a documented primary or secondary impression related to COVID-19 or presence of COVID-19 suspicion identified in the narrative review. For patients without a COVID-19 hospital diagnosis, EMS suspicion was defined as either a documented primary or secondary impression related to COVID-19 only. The sample of narrative reviews conducted among this group was used to estimate the potential for misclassification due to narrative only inclusion of COVID-19 EMS suspicion.
For patients with a COVID-19 hospital diagnosis and no documented COVID-19 EMS suspicion, the most common EMS provider primary and secondary impressions were described.
Dispatch Complaints
At the time a 9-1-1 call is received, the emergency dispatcher selects a dispatch complaint from a discrete drop-down list to be reported to the responding EMS unit. For patients with a COVID-19 hospital diagnosis, the five most commonly reported dispatch complaints were tabulated and stratified by documented EMS COVID-19 suspicion.
Initial Prehospital Vital Signs
Initial prehospital vital signs, defined as the first recorded instance of each measurement in the EMS record, were evaluated. We created clinically meaningful categories to facilitate interpretation and actionability of our findings. We utilized generally accepted thresholds for tachycardia, hypotension, and tachypnea: pulse rate equal to or greater than 100 (beats/minute), systolic blood pressure (SBP) less than or equal 90 millimeters of Mercury (mmHg), and respiratory rate was greater than 20 breaths/minute. SpO2 was categorized based upon thresholds for severity of hypoxia: <85%, 85–90%, 91–95%, and ≥96% (Citation5). Temperature was categorized using the CDC thresholds for below normal, normal, and fever: <98, 98–100.4, and >100.4 degrees Fahrenheit (F) (Citation6).
Personal Protective Equipment (PPE)
The prehospital electronic health record allows for documentation of each article of PPE used by each EMS provider listed on the record. Articles of PPE assessed included use of any eye protection or face shield, any mask (surgical, N95, or powered air-purifying respirator [PAPR]), or gown. PPE was considered to have been used during the encounter if any of the above-mentioned PPE articles were documented for at least one EMS provider listed on the prehospital care record.
Analysis
Encounter and patient characteristics were stratified by records with COVID-19 hospital diagnoses and those without COVID-19 hospital diagnoses. Descriptive statistics were calculated. Statistical comparisons were not conducted between the COVID-19 hospital diagnosed/not diagnosed groups because the large sample size is likely to result in statistically significant differences, when clinical relevance may or may not be present. Similarly, strata for records with COVID-19 EMS suspicion, records with COVID-19 hospital diagnosis, and those who had neither a COVID-19 EMS suspicion nor a COVID-19 hospital diagnosis are presented descriptively.
After narrowing to records with a COVID-19 hospital diagnosis, statistically significant differences in patient and encounter characteristics by COVID-19 EMS suspicion were assessed using Chi-square tests for categorical variables and Wilcoxon rank sum tests for non-normally distributed continuous variables.
To evaluate the predictive performance characteristics of COVID-19 EMS suspicion for the identification of COVID-19 hospital diagnosis, Sensitivity (True Positive/[True Positive + False Negative]), Specificity (True Negative/[True Negative + False Positive]), Positive Predictive Value (True Negative/[True Negative + False Positive]), and Negative Predictive Value (True Negative/[True Negative + False Negative]) were estimated. All analyses were performed using STATA IC version 15.1 (StataCorp LLC; College Station, TX, USA).
Results
There were 796,929 9-1-1 EMS responses from March 1, 2020 to April 19, 2020 included in the research database. Of those, 84,540 (10.6%) prehospital patient care records from 274 EMS agencies had linked hospital ICD-10 codes and were included in this case series. displays the patient and encounter characteristics for records evaluated in this study. The median pulse rate was 90 beats/minute (IQR: 78–107). The median SBP was 140 mmHg (IQR: 122–159). The median respiratory rate was 18 breaths/minute (IQR: 16–20). The median SpO2 was 97% (IQR: 95–99). The median temperature was 98.2 F (IQR: 97.7–98.8).
TABLE 1. Patient and encounter characteristics.
COVID-19 Hospital Diagnosis
There were 814 (1.0%) patients with a COVID-19 hospital diagnosis (). The median age among patients diagnosed with COVID-19 was 67 years (IQR: 54–79 years) compared to 58 years (IQR: 38–73 years) among those not diagnosed with COVID-19. There was a disproportionately higher representation of patients whose race/ethnicity was documented as Black/African American or Hispanic/Latino among those diagnosed with COVID-19. Patients with a COVID-19 hospital diagnosis had a higher median pulse rate compared to those not diagnosed with COVID-19 (98 beats/minute vs. 90 beats/minute, respectively). Median respiratory rate was higher among COVID-19 diagnosed patients (20 breaths/minute vs. 18 breaths/minute, respectively). The median SpO2 was lower for patients diagnosed with COVID-19 compared to patients that did not have a COVID-19 hospital diagnosis (93% vs. 97%, respectively). Median body temperature was higher among patients diagnosed with COVID-19 (99.2 F vs 98.2 F, respectively).
COVID-19 EMS Suspicion Identification
Of the 814 patients with COVID-19 hospital diagnoses, 404 (49.6%) had a documented EMS primary or secondary impression related to COVID-19. Of the 410 patients with a COVID-19 hospital diagnosis and no documented COVID-19 EMS impression, 232 (56.6%) records were found to have suspicion of COVID-19 present in the free-text narrative. Thus, 78.1% (636/814) patients with a COVID-19 hospital diagnosis were found to have COVID-19 EMS suspicion documented.
Among the 83,726 patients that did not have COVID-19 hospital diagnosis recorded, 2,568 (3.1%) had a documented EMS primary or secondary COVID-19 impression. In the random sample of 450 records selected for patients that did not have a COVID-19 hospital diagnosis and did not have a documented COVID-19 impression, 28 (6.2%) were found to have suspicion of COVID-19 in the narrative.
Interrater agreement for narrative reviews was 95% (47/50).
COVID-19 EMS Suspicion and COVID-19 Hospital Diagnosis
COVID-19 EMS suspicion demonstrated a sensitivity of 78.1% for those with a hospital diagnosis of COVID-19. The estimation of specificity revealed that among patients who did not have a hospital COVID-19 diagnosis, the probability that EMS did not document COVID-19 suspicion was 96.9%. Positive predictive value estimation showed that among EMS suspected COVID-19 patients, the probability of a COVID-19 hospital diagnosis was 19.9%. Finally, among those patients for whom EMS did not document suspicion of COVID-19, the probability of not having a hospital diagnosis of COVID-19 was 99.8% ().
TABLE 2. Cross tabulation of patients by COVID-19 EMS suspicion and COVID-19 hospital diagnosis.
To estimate the impact of COVID-19 EMS suspicion misclassification for records without COVID-19 hospital diagnosis on estimates of specificity and negative predictive value, a sensitivity analysis was conducted. Based on the results of the narrative review, 6.1% of records with neither a COVID-19 hospital diagnosis nor COVID-19 EMS impression were re-classified as having COVID-19 EMS suspicion. Specificity decreased from 96.9% to 91.0%. The negative predictive value decreased from 99.8% to 93.7% (Appendix A, supplemental material).
Among patients with a COVID-19 hospital diagnosis, gender, race/ethnicity, incident location, pulse rate, SpO2, and temperature were similar for records with and without documented COVID-19 EMS suspicion (). There was a statistically significant difference found in patient age, SBP, and respiratory rate (p < 0.05)
TABLE 3. Patient and encounter characteristics for patients with COVID-19 hospital diagnosis by COVID-19 EMS suspicion
Dispatches for “Breathing Problem” and “Sick Person” were the two most common complaints representing 76% of dispatches for patients with COVID-19 EMS suspicion and 60% of patients without documented COVID-19 EMS suspicion (). For patients with COVID-19 hospital diagnoses who did not have a COVID-19 EMS suspicion documented, the third most common dispatch complaint was “Chest Pain (Non-Traumatic)” (6%) followed by “Unconscious/Fainting” (5%). The top 5 most common dispatch complaints for those who did not have a COVID-19 hospital diagnosis are provided as an appendix (Appendix B, supplemental material).
TABLE 4. Top 5 dispatch complaints for EMS records with COVID-19 hospital diagnosis.
Acute respiratory distress (dyspnea)/shortness of breath (18%) and generalized weakness (14%) were the most commonly documented primary impressions among hospital-diagnosed patients when EMS COVID-19 suspicion was not documented (Appendix C, supplemental material).
displays the patient and encounter characteristics for all records with COVID-19 EMS suspicion, all records with COVID-19 hospital diagnosis, and those without COVID-19 hospital diagnosis or COVID-19 EMS suspicion documented. Records with COVID-19 hospital diagnosis or EMS suspicion documented appear to have lower initial prehospital SpO2, higher body temperature, more often originate from skilled nursing/long term care, and less often originate from a public place when compared to those with neither EMS suspicion nor hospital diagnosis of COVID-19.
TABLE 5. Patient and encounter characteristics for records with COVID-19 EMS suspicion, COVID-19 hospital diagnosis, and neither a COVID-19 hospital diagnosis nor COVID-19 EMS suspicion
Use of Personal Protective Equipment (PPE)
The documented use of PPE by EMS differed across records of patients who were diagnosed with COVID-19 compared to those who were not diagnosed with COVID-19 (). The rate of prehospital documentation related to use of any eye protection, mask or gown was substantially higher for records with patients that had a COVID-19 hospital diagnosis. Additionally, the documented use of PPE during EMS responses for patients who were diagnosed with COVID-19 at the hospital increased throughout the study period ().
Figure 1. EMS personal protective use among patients with COVID-19 hospital diagnosis.
Among patients with a COVID-19 hospital diagnosis, documented use of PPE varied by COVID-19 EMS suspicion (). Documented use of eye protection or any type of mask was lower when a COVID-19 EMS suspicion was not documented. Documented use of a gown was more than 2 times lower when COVID-19 EMS suspicion was not documented.
Use of eye protection or a face shield, any mask, and gowns were documented in higher percentages among those with EMS COVID-19 suspicion and COVID-19 diagnosed patients when compared to those with neither a COVID-19 hospital diagnosis nor a documented COVID-19 EMS suspicion (). Documented use of PPE when evaluating the EMS provider’s primary or secondary impression alone mirrored the COVID-19 EMS suspicion results (Appendices D and E).
Discussion
This preliminary report described a large case series of EMS patient care records with linked hospital diagnosis data. Characteristics among records with documented COVID-19 EMS suspicion, COVID-19 hospital diagnosis, or both were similar to previously-published characteristics of patients with COVID-19 (7). Specifically, for patients diagnosed with COVID-19, we noted that the most common dispatch complaints, impressions documented by EMS, and age were consistent with reported patient presentations stemming from analyses of hospital data (Citation8, Citation9). Importantly, a notable proportion of patients who were diagnosed in the hospital with COVID-19 presented in the prehospital setting with moderate and severe hypoxia.
Additionally, as has been reported elsewhere, there was disproportionately larger representation of patients from minority racial/ethnic groups, particularly regarding African American and Hispanic patients, when comparing those diagnosed with COVID-19 to those who were not diagnosed with COVID-19 in this evaluation (Citation7, Citation10). These findings regarding difference across racial/ethnic groups should be interpreted with caution and with consideration for the context of socioeconomic factors (Citation11). The proportion of missing race/ethnicity data in this case series further supports a cautious interpretation of this finding.
EMS Assessment and Suspicion
The prehospital index of suspicion of patients at risk of COVID-19 is informed by dispatch information as well as data gathered during the patient encounter. For patients with COVID-19 hospital diagnoses, whether suspected by EMS or not, the most common dispatch complaints were well-aligned with classic COVID-19 presentation. However, dispatch complaints for the remaining patients diagnosed with COVID-19 were less predicted, especially falls, and stroke/CVA. EMS impressions consistent with stroke presentation were also documented often among COVID-19 hospital diagnosed patients for whom EMS did not suspect COVID-19. These prehospital findings suggest that stroke-like symptoms are of interest, particularly in light of emerging reports of COVID-19 patients who manifested stroke signs and symptoms (Citation12). Further research is needed to determine whether patients exhibiting signs and symptoms of stroke should be presumed positive for COVID-19 (with appropriate use of PPE) during EMS evaluation and transport.
It is encouraging to note the rapidity with which EMS providers incorporated documentation of the new EMS impressions related to COVID-19 into their practice. Furthermore, those with EMS suspicion were clinically and demographically more similar to those with confirmed COVID-19 at the hospital than they were to those without hospital confirmation. Notably, COVID-19 EMS suspicion was recorded over 3 times more frequently than hospital confirmed diagnoses, representing approximately 4% of EMS dispatches in our series. During this time frame, the CDC noted a third “peak” in influenza like illness (ILI) as part of their traditional influenza seasonal surveillance, with the proportion of patient encounters for ILI ranging from 3% to 6% (Citation13). Federal officials presumed this third peak to represent COVID-19 disease and have incorporated monitoring of this metric into re-opening analysis (Citation14). The totality of these observations offers reason for further consideration of these patients for public health follow-up testing and tracing (Citation8, Citation13).
There were several assessment findings that appeared to be more common in EMS patients who had a COVID-19 hospital diagnosis compared to those who did not have a COVID-19 hospital diagnosis. Specifically, SpO2 < 85%, was reported more often in COVID-19 hospital diagnosed patients. This finding identified during prehospital patient assessment should raise an EMS provider’s index of suspicion when considering a patient’s COVID-19 status and the need for appropriate PPE. Temperature >100.4 F was also more common in patients with COVID-19 diagnosis, but the utility of this finding may be limited since 53% of patients had no body temperature recorded by EMS. Obtaining and documenting a temperature should be prioritized during this pandemic. COVID-19 status and the need for PPE should also be considered when transporting 9-1-1 patients from skilled nursing, assisted living, or residential facilities.
While documented EMS suspicion of COVID-19 was associated with nearly 80% sensitivity for hospital diagnosed disease, these data demonstrate that EMS suspicion alone is not sufficient basis for making decisions regarding policy and practice for provider utilization of PPE. Many patients with COVID-19 hospital diagnoses did not have COVID-19 EMS suspicion documented. For patients diagnosed with COVID-19 who did not have a documented EMS suspicion, the most common primary impressions were consistent with typical COVID-19 signs and symptoms (Citation15). This represents a potential opportunity for further education to increase EMS identification of patients with a high probability of COVID-19 in order to improve provider/patient safety and provide enhanced pre-arrival information to the hospital. Additionally, emphasis on atypical dispatch complaints observed in this study, such as chest pain, unconscious/fainting, and stroke may lead to increased sensitivity on the part of EMS provider suspicion (Citation12, Citation16, Citation17). Almost 100% of patient records for those who did not have a COVID-19 hospital diagnosis did not have documented EMS COVID-19 suspicion, which is not surprising, given the overall low burden of disease in this population.
Utilization of Personal Protective Equipment (PPE)
Utilization of PPE likely reflects the broad diversity of PPE policies (which may be related to local prevalence of disease) and availability across EMS organizations, with eye protection or face shields and N95 respirators most commonly used. This study found inconsistent practices that may impact provider and patient safety. While documented PPE use was significantly higher when EMS documented COVID-19 suspicion, about one of every five records with documented EMS provider suspicion of COVID-19 did not document EMS provider use of any mask. This may represent a failure to take adequate precautions (either by choice or because of unavailability of PPE), or a lack of appropriate documentation of PPE and thus loss of a valuable data source for employee quarantine decisions and inventory management.
Interestingly, mask use (any type) was documented on about three of every four patient records with a COVID-19 hospital diagnosis, but no documented EMS suspicion of COVID-19, perhaps indicating that a gestalt or other undocumented cause led to increased PPE utilization. Mask use was documented substantially less often among records where EMS COVID-19 suspicion was not documented and the patient did not have a COVID-19 hospital diagnosis, although mask use was still reported in over a third of these records. It should be noted that documentation of PPE has increased overall throughout the pandemic. More study is needed to understand patterns of actual PPE use by EMS providers during the assessment and management of patients who have COVID-19, including patterns of nonuse.
Limitations
This case series is limited to a convenience sample of EMS agencies who voluntarily contributed de-identified data for research, and therefore results may not be generalizable. Given that the prevalence of disease was 1%, our study does not address areas with very high disease burden such as Wuhan, Northern Italy, and New York City, where the primary challenges relate to healthcare system capacity rather than only the identification of patients with COVID-19. This study is not intended to address such high demand situations. Rather, this study describes the situation where COVID-19 disease is present, but not at critically high levels (Citation18–20).
The use of COVID-19 ICD-10 codes rather than confirmed COVID-19 test results is a further limitation of this study. This analysis assumed that all patients were screened and/or tested for COVID-19 by the hospital. Previous test results were not necessarily captured in this evaluation. While the de-identified nature of this dataset precludes determining the location of individual hospitals, during the time of our data collection there was documented, widespread lack of available testing for COVID-19; thus, it is likely that limits in testing availability prevented the inclusion of COVID-19 diagnosis in some patient records, although the magnitude of this shortage is impossible to quantify (Citation21). The study is also limited by the timeliness and availability of COVID-19 testing results. Patients transported by EMS toward the end of the study period may not have had test results included in their linked records at the time analysis was completed. Collectively, these limitations may have contributed to underreporting of positive cases in our series. Further, because the ICD-10 code specific for COVID-19 was made available during the study period (Citation22), there may have been inconsistent hospital reporting of COVID-19 diagnosis. Utilization of the dispatch complaint related to pandemic/outbreak versus symptom-specific dispatch complaints may have varied across EMS agencies. Similarly, the EMS provider impressions for COVID-19 were implemented into the prehospital software in early March (3/5/2020). While all EMS agencies had the option to utilize these impressions, awareness and adoption likely varied. It is possible that documentation requirements, protocols, or habits, impacted the ability to identify COVID-19 EMS suspicion. To estimate how often EMS suspicion of COVID-19 was included in the free-text narrative section of the prehospital electronic health record when a COVID-19 EMS impression was not documented, a manual narrative review was conducted. The review revealed a higher rate of narrative-only inclusion of COVID-19 EMS suspicion among those with COVID-19 hospital diagnoses compared to those without COVID-19 hospital diagnoses. Narratives of all records with a COVID-19 hospital diagnosis and no EMS COVID-19 impression were reviewed, which allowed for re-classification of records in this group for EMS COVID-19 suspicion. Conversely, given that a sample of narratives was reviewed for records with neither a hospital COVID-19 diagnosis nor a COVID-19 EMS impression, re-classification of each record for COVID-19 EMS suspicion based on narrative review was not feasible. To estimate the impact of this misclassification on estimates of specificity and negative predictive value, a sensitivity analysis was conducted. Specificity estimates decreased by 5.8% and negative predictive value decreased by 6.1%. We are unable to estimate the impact of misclassification on demographic and encounter characteristics of individual records.
The large number of records that did not have a patient temperature recorded in this case series further highlights limitations of data entry. We could not evaluate the impact of supplemental oxygen vs room air on the initial SpO2, because over 81% of cases evaluated in this study were missing this information. Moreover, it is unknown how often PPE was used, but not documented. Additionally, the timing of utilization of PPE is not a recorded data element, so it is difficult to determine the utilization of PPE because of “protocol” vs. “provider suspicion.” These points are particularly salient for the first few days of the study period after which an overall increase in the report of PPE use was observed. Finally, asymptomatic cases may also have contributed to the discrepancy COVID-19 EMS suspicion and COVID-19 hospital diagnosis.
Conclusion
In this large case series of EMS encounters with linked hospital outcome information, a documented EMS suspicion for COVID-19 disease demonstrated a 20% positive predictive value and a 78% sensitivity when compared to documented hospital ICD-10 diagnosis. These data indicate that EMS suspicion alone is insufficient to determine appropriate utilization of PPE.
Supplemental Material
Download MS Word (55.2 KB)Supplemental Material
Download MS Word (54.9 KB)Supplemental Material
Download MS Word (49.9 KB)Supplemental Material
Download MS Word (50.5 KB)Supplemental Material
Download MS Word (55.3 KB)References
- North Carolina Department of Health & Human Services. NCDHHS COVID-19: testing. 2020 [accessed 2020 Jun 12]. https://covid19.ncdhhs.gov/dashboard/testing.
- Tolia VM, Chan CT, Castillo EM. Preliminary results of initial testing for coronavirus (COVID-19) in the emergency department. West J Emerg Med. 2020;21(3):503–6. doi:10.5811/westjem.2020.3.47348.
- Chen L, Liu HG, Liu W, Liu J, Liu K, Shang J, Deng Y, Wei S. Analysis of clinical features of 29 patients with 2019 novel coronavirus pneumonia. Chinese Journal of Tuberculosis and Respiratory Disease, Zhonghua Jie He He Hu Xi Za Zhi. 2020;43:203–8. doi:10.3760/cma.j.issn.1001-0939.2020.03.013
- Levy PS, Lemeshow S, ed. Sampling of populations: methods and applications. Germany: Wiley; 2013.
- Jubran A. Pulse oximetry. Crit Care. 2015;19(1). doi:10.1186/s13054-015-0984-8.
- Centers for Disease Control. Quarantine and isolation definitions: signs, symptoms & conditions. 2020 [accessed 2020 May 14]. https://www.cdc.gov/quarantine/maritime/definitions-signs-symptoms-conditions-ill-travelers.html.
- Centers for Disease Control. Coronavirus disease 2019 (COVID-19): racial and ethnic minority groups. 2020 [accessed 202 May 1]. https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/racial-ethnic-minorities.html.
- CDC COVID-19 Response Team. Severe outcomes among patients with coronavirus disease 2019 (COVID-19) — United States, February 12–March 16, 2020. MMWR Morb Mortal Wkly Rep. 2020;12(69):343–6.
- Myers LC, Parodi SM, Escobar GJ, Liu VX. Characteristics of hospitalized adults with COVID-19 in an Integrated Health Care System in California. JAMA. 2020;323(21):2195–8. doi:10.1001/jama.2020.7202.
- Wadhera RK, Wadhera P, Gaba P, Figueroa JF, Joynt Maddox KE, Yeh RW, Shen C. Variation in COVID-19 hospitalizations and deaths across New York City Boroughs. JAMA. 2020;323(21):2192. doi:10.1001/jama.2020.7197.
- Chowkwanyun M, Reed AL. Racial health disparities and COVID-19 – caution and context. N Engl J Med. 2020 [accessed 2020 May 18]. doi:10.1056/NEJMp2012910.
- Oxley TJ, Mocco J, Majidi S, Kellner CP, Shoirah H, Singh IP, De Leacy RA, Shigematsu T, Ladner TR, Yaeger KA, et al. Large-vessel stroke as a presenting feature of Covid-19 in the young. N Engl J Med. 2020;382(20):e60. doi:10.1056/NEJMc2009787.
- Centers for Disease Control. Percentage of visits for influenza-like illness reported by ILINET 2019-2020 season. 2020 [accessed 2020 May 15]. https://www.cdc.gov/flu/weekly/weeklyarchives2019-2020/data/senAllregt18.html.
- The Whitehouse. Guidelines opening up America again. 2020 [accessed 2020 May 15]. https://www.whitehouse.gov/openingamerica/.
- Centers for Disease Control. Symptoms of coronavirus. 2020 [accessed 2020 Mar 27]. https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html.
- Bangalore S, Sharma A, Slotwiner A, Yatskar L, Harari R, Shah B, Ibrahim H, Friedman GH, Thompson C, Alviar CL, et al. ST-segment elevation in patients with Covid-19 - a case series. N Engl J Med. 2020;382(25):2478–80. doi:10.1056/NEJMc2009020.
- Helms J, Kremer S, Merdji H, Clere-Jehl R, Schenck M, Kummerlen C, Collange O, Boulay C, Fafi-Kremer S, Ohana M, et al. Neurologic features in severe SARS-CoV-2 infection. N Engl J Med. 2020;382(23):2268–70. doi:10.1056/NEJMc2008597.
- Guan W-j, Ni Z-y, Hu Y, Liang W-h, Ou C-q, He J-x, Liu L, Shan H, Lei C-l, Hui DSC, et al. Clinical characteristics of coronavirus disease 2019 in China. N Engl J Med. 2020;382(18):1708–20. doi:10.1056/NEJMoa2002032.
- Richardson S, Hirsch JS, Narasimhan M, Crawford JM, McGinn T, Davidson KW, The Northwell C-RC, Barnaby DP, Becker LB, Chelico JD, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA. 2020;323(20):2052. doi:10.1001/jama.2020.6775.
- Rosenbaum L. Facing Covid-19 in Italy - ethics, logistics, and therapeutics on the epidemic's front line. N Engl J Med. 2020;382(20):1873–5. doi:10.1056/NEJMp2005492.
- Grimm C. Hospital experiences responding to the COVID-19 pandemic: results of a national pulse survey March 23–27, 2020: US Department of Health and Human Services Office of Inspector General. 2020 [accessed 2020 May 15] https://oig.hhs.gov/oei/reports/oei-06-20-00300.pdf.)
- Centers for Disease Control. New ICD-10-CM code for the 2019 novel coronavirus (COVID-19). 2020 Apr 1 [accessed 2020 Mar 15]. https://www.cdc.gov/nchs/data/icd/Announcement-New-ICD-code-for-coronavirus-3-18-2020.pdf.