Patterns of epinephrine prescribing at discharge and length of stay for adult anaphylaxis in U.S. emergency departments

Abstract

Background:

Anaphylaxis is a life-threatening allergic reaction that requires prompt epinephrine and post-treatment management. Guidelines recommend prescribing epinephrine at discharge from the emergency department (ED), yet results of studies show lower-than-recommended prescribing rates and variable ED lengths of stay (LOS). Large-scale analyses of adult anaphylaxis care in U.S. EDs remain limited.

Objective:

The objective was to estimate the proportion of adults discharged from U.S. EDs with an epinephrine prescription, characterize ED LOS, and identify predictors of these outcomes.

Methods:

We analyzed 2013–2015 and 2018–2022 National Hospital Ambulatory Medical Care Survey–ED visits for adults ages ≥ 18 years diagnosed with anaphylaxis who were discharged home. Outcomes were (1) epinephrine prescription at discharge and (2) ED LOS < 4 hours. Patients who arrived by emergency medical services (EMS) were excluded for LOS analysis. Multivariable log-binomial regression estimated risk ratios (RR) between outcomes and clinical predictors.

Results:

Among 609 ED visits, 14% resulted in an epinephrine prescription at discharge (95% confidence interval [CI], 10%–18%). The median (interquartile range) LOS was 138 minutes (80–199 minutes), with 83% discharged within 4 hours (95% CI, 78%–87%). Older adults (≥42 years) and Hispanic patients were less likely to receive an epinephrine prescription (RR 0.47 [95% CI, 0.25–0.90]; RR 0.34 [95% CI, 0.12–0.95], respectively). Rural visits were more likely to have LOS < 4 hours (RR 1.12 [95% CI, 1.02–1.23]).

Conclusion:

Epinephrine prescribing at discharge was infrequent across the full cohort, whereas ED LOS was typically brief among the patients who did not arrive by EMS. Older adults and Hispanic patients were less likely to receive epinephrine, and rural visits had shorter LOS. These findings highlight potential disparities in discharge prescribing and ED management of adult anaphylaxis in U.S. EDs.

Anaphylaxis is a severe, potentially life-threatening systemic allergic reaction, commonly triggered by foods, medications, or insect stings, that generally involves multiple organ systems.¹ Prompt recognition and treatment in the emergency department (ED) are critical. The current standard of care for anaphylaxis management includes the following: (1) prompt administration of epinephrine, (2) adequate observation period, (3) a prescription for epinephrine at discharge, and (4) follow-up with an allergist.¹ The optimal duration of post-treatment observation has not been clearly established, and observation practices are often individualized, with extended monitoring for patients at higher risk of recurrent or severe anaphylaxis.²

Although guidelines outline recommended practices for anaphylaxis management, adherence in U.S. EDs varies.^1,3 Previous U.S. studies, largely single- or multicenter cohorts, have reported low rates of epinephrine prescription at discharge, variable observation periods, and inconsistent referral to allergy specialists after discharge.^4,5 Of note, previous multicenter studies of ED visits for food- and insect sting–related anaphylaxis have demonstrated underuse of epinephrine and inconsistent discharge practices.^6,7 A more recent analysis that examined concordance with food-induced anaphylaxis guidelines over time suggest modest improvements in acute management, although gaps in epinephrine prescribing at discharge persist.^8,9 Although these studies provide important insight, most have relied on single- or multicenter cohorts, which limits the ability to assess patterns across a broad spectrum of adult ED settings. Analyses by using probability-based, multistage sampling offer a complementary perspective on epinephrine prescribing and ED length of stay (LOS) among adults treated for anaphylaxis.

To address this gap, we used data from the National Hospital Ambulatory Medical Care Survey (NHAMCS), conducted by the National Center for Health Statistics, which is a national probability sample survey of ED visits.¹⁰⁻¹² We used this data base to generate national estimates of the proportion of adult patients discharged with an epinephrine prescription and the proportion of visits with an LOS < 4 hours. The 4-hour threshold reflects commonly reported ED visit durations for patients with anaphylaxis in the United States and is used in our study as a descriptive measure of visit duration rather than a fixed guideline recommendation.^13,14 We also evaluated patient- and visit-level factors associated with epinephrine prescribing at discharge and those associated with an ED LOS < 4 hours. For examining these outcomes by patient demographics and ED characteristics, this study uses a probability-sampled dataset to provide insights and identify potential sociodemographic disparities in anaphylaxis care.

METHODS

Study Design

We conducted a repeated cross-sectional study by using NHAMCS-ED data from 2013 to 2015 and 2018 to 2022.¹⁰ NHAMCS-ED is a probability-based survey of U.S. ED visits conducted by the U.S. Centers for Disease Control and Prevention. The survey uses a multistage sampling design that selects nationally representative hospitals, EDs within hospitals, and patient visits within EDs, with each visit assigned a probability weight.¹¹ When survey weights and design variables are applied, analyses generate national estimates of ED visit characteristics rather than estimates limited to sampled hospitals. Although the unweighted number of anaphylaxis visits is modest, the sampling design of NHAMCS-ED allows each sampled encounter to represent a known number of ED visits nationally, enabling estimation of population-level proportions and associations.

NHAMCS-ED collects standardized data on patient demographics, visit characteristics, vital signs, diagnoses, medications, procedures, and disposition across a broad range of ED settings, including academic, community, urban, and rural hospitals.¹² This design differs from multicenter observational studies, which typically rely on convenience samples of selected hospitals and are not intended to generate population-level estimates. Because NHAMCS-ED data are publicly available and de-identified, this study was exempt from institutional review board approval.¹⁰ Our study did not involve direct contact with human subjects and was considered not human subjects research; therefore, informed consent was not required.

Study Population

We included ED visits among adult patients ages ≥ 18 years with a diagnosis of anaphylaxis, identified by using relevant International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) and ICD-10-CM codes, and who were discharged home.¹⁵ Relevant codes included the following diagnoses: “Anaphylactic reaction due to food,” “Other adverse food reactions, not elsewhere classified,” “Anaphylactic shock, unspecified,” “Angioneurotic edema,” “Other and unspecified allergy.” The corresponding ICD-9-CM (2013–2015) were 9950, 9953, 9951, 9956; and ICD-10-CM (2018–2022) were T78.0, T78.1, T78.2, T78.3, and T78.4.

ED LOS was defined a priori as the time from ED arrival to discharge. Although this measure is an imperfect proxy for clinical observation time, it represents an objective, routinely recorded metric in NHAMCS-ED that enables standardized comparisons across visits and studies.^16,17 Visits with a known LOS were included, and those with missing LOS data were excluded because LOS was a primary exposure in our analyses. Primary analyses were restricted to patients who did not arrive via emergency medical services (EMS), given that NHAMCS-ED does not capture prehospital care or observation time, which limits comparability between EMS and non-EMS arrivals. To assess robustness, we performed a sensitivity analysis that included EMS arrivals. In a separate analysis, we used an alternative LOS definition, time from first provider contact (physician, physician assistant, or advanced practice nurse) to discharge to approximate clinical observation time and account for variability in waiting time.

Epinephrine prescription at discharge was coded by using the Cerner Multum Lexicon Plus and were classified by using the Multum Lexicon Therapeutic Classification Scheme (Cerner Multum Denver, CO). We excluded data from years 2016–2017 because information on LOS was not available. We also excluded visits with missing information on oxygen saturation or systolic blood pressure because these vital signs were included as covariates in the multivariable regression models to account for clinical severity.

For analyses that involved ED LOS, visits with EMS arrival were excluded in the primary analyses. Prehospital care, including medication administration and timing of treatment before ED arrival, is not captured in NHAMCS-ED and may influence ED observation duration. Primary LOS analyses were restricted to visits with patients not arriving via EMS to reduce residual confounding from unmeasured prehospital interventions. To ensure robustness, we also performed a supplemental LOS analysis with the overall cohort, including those who arrived by EMS.

Study Variables

The primary outcomes of this study were (1) epinephrine prescription at discharge, and (2) ED LOS of <4 hours, in which LOS was used as an indirect proxy for observation time. LOS was dichotomized at 4 hours to distinguish shorter versus longer ED visits, which reflects a pragmatic operational threshold commonly used in U.S. ED research and throughput metrics, rather than a clinically defined observation period for recurrent reactions.^13,14

Covariates included sociodemographic variables and clinical variables. Sociodemographic variables included age, sex, race and/or ethnicity, insurance type, and study period (2013–2015, 2018–2019, 2020–2022). The geographic regions of ED visits were also determined, which were either metropolitan (located within a U.S. metropolitan statistical area) or rural (located outside a metropolitan statistical area). Clinical variables included oxygen saturation and systolic blood pressure, which were incorporated as covariates to account for baseline patient severity at presentation. In sensitivity analyses, we also characterized the ED triage acuity level for each visit, which was reported on a 5-level scale, categorized as 1–2 (immediate/emergent), 3 (urgent), 4 (semi-urgent), or 5 (nonurgent).

More granular measures of reaction severity (e.g., the number of epinephrine doses administered or anaphylaxis history) were not available and, therefore. could not be included in multivariable analyses. We analyzed study periods to assess potential temporal changes in anaphylaxis care, including those occurring before, during, and after the COVID-19 pandemic. Age was analyzed as a binary variable (≥42 years) based on the approximate cohort median. We recognize that clinically meaningful differences may exist in older adults (≥65 years), but the limited sample size precluded more granular age stratification.

Statistical Analyses

We presented means for symmetrically distributed continuous variables and medians for skewed variables. As such, age and oxygen saturation were summarized as medians with interquartile ranges (IQR), and systolic blood pressure was summarized as a mean ± standard deviation (SD). Categorical variables were summarized as unweighted counts with weighted percentages. As the main analyses, associations between exposures and outcomes were assessed by using multivariable log-binomial regressions, adjusting for sociodemographic and clinical variables. We report risk ratios (RR) with associated 95% confidence intervals (CI) and p-values.

We conducted sensitivity analyses to support cohort validity and assess robustness of modeling assumptions. To support the validity of the anaphylaxis case definition, we examined ED administration of medications commonly used in acute allergic reactions, including epinephrine, antihistamines, and systemic glucocorticoids.¹⁸ These analyses were conducted for cohort validation purposes only and were not intended to serve as diagnostic criteria for anaphylaxis because these medications are not specific to anaphylactic reactions. However, use of at least one of these therapies epinephrine, antihistamines, and/or systemic glucocorticoids would be expected in the management of patients experiencing a suspected anaphylactic reaction. We compared medication administration rates between visits coded with and without anaphylaxis diagnoses. Then, we conducted sensitivity analyses restricted to anaphylaxis-coded visits in which at least one of these medications was administered. To evaluate the potential influence of prehospital transport on ED LOS, we used additional models that incorporated EMS arrival status as a covariate. For visits with missing EMS status, a missingness indicator was included. We also examined whether the severity of presentation, assessed by using the ED triage level, influenced predictors of epinephrine prescription at discharge or ED LOS.

All statistical analyses were performed by using R software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Demographics

A total of 609 ED visits for anaphylaxis were identified in the NHAMCS-ED (Table 1). The median (IQR) age among visits was 41 years (28–55 years), and female patients accounted for 68% of visits. Most visits were among non-Hispanic White patients (57%) and occurred in metropolitan settings (86%), with a similar distribution between private insurance (39%) and public insurance (Medicare/Medicaid, 39%). The mean ± SD systolic blood pressure SBP at presentation was 139 ± 24 mm Hg, and the median (IQR) oxygen saturation was 98% (97%–100%). Among all the visits, including nonresponse as a separate category, triage level 3 was the most frequent response category (level 3, 36%). Patient demographics and visit characteristics were similar across study periods.

Table 1.

Demographic and clinical characteristics of all visits with patients diagnosed with anaphylaxis*

IQR = Interquartile range; CHIP = Children’s Health Insurance Program; LOS = length of stay; BP = blood pressure; SD = standard deviation; ESA = emergency service area.

*Data are weighted estimates from 2013–2015 and 2018–2022 of the National Hospital Ambulatory Medical Care Survey–Emergency Department; percentages in parentheses may not sum to 100% due to rounding. Adapted from Refs. 10 and 12.

#Visits with missing data for triage level.

§Defined as either “No triage” reported for this visit but ESA does conduct nursing triage or the visit occurred in the ESA that does not conduct nursing triage.

In the ED LOS analyses, visits that involved patients transported by EMS were excluded, which yielded a cohort of 536 visits (Table 2). The demographic characteristics of this subgroup were similar to the overall cohort. The median (IQR) age was 40 years (28–56 years) and 70% of visits involved female patients.

Table 2.

Demographic and clinical characteristics of visits with patients not transported by EMS^*

EMS = Emergency medical services; IQR = interquartile range; CHIP = Children’s Health Insurance Program; LOS = length of stay; BP = blood pressure; SD = standard deviation.; ESA = emergency service area.

^*The characteristics of the patients not transported by EMS, used for analyses of emergency department observation duration; data are weighted estimates from 2013–2015 and 2018–2022 of National Hospital Ambulatory Medical Care Survey–Emergency Department; percentages may not sum to 100% due to rounding. Adapted from Refs. 10 and 12.

^#Visits with missing data for triage level.

^§Defined as either “No triage” reported for this visit but ESA does conduct nursing triage or the visit occurred in an ESA that does not conduct nursing triage.

Epinephrine Prescribing at Discharge

Among all the ED visits for anaphylaxis, 14% of the patients had an epinephrine prescription at discharge (95% CI, 10%–18%) (Table 1). Patients ages ≥ 42 years were less likely to receive an epinephrine prescription at discharge compared with younger patients (RR 0.47 [95% CI, 0.25–0.90]; p = 0.02) (Table 3). Hispanic patients were less likely to be prescribed epinephrine at discharge compared with non-Hispanic White patients (RR 0.34 [95% CI, 0.12–0.95]; p = 0.04).

Table 3.

Predictors of epinephrine prescription at discharge

CHIP = Children’s Health Insurance Program; BP = blood pressure.

Data in bold represent statistically significant values (P-value < 0.05).

ED LOS

Among visits included in LOS analyses (n = 536), the median (IQR) ED LOS was 138 minutes (80–199 minutes) (Table 2). Most patients (83%) were in the ED for <4 hours (95% CI, 78%–87%).

In multivariable analyses, excluding EMS arrivals, visits to rural EDs were more likely to have LOS of <4 hours compared with visits to metropolitan EDs (RR 1.12 [95% CI, 1.02–1.23]; p = 0.02) (Table 4). Visits among the patients categorized as “other” race and/or ethnicity were also more likely to have an ED LOS of <4 hours (RR 1.20 [95% CI, 1.09–1.33]; p < 0.001). In addition, higher initial systolic blood pressure, modeled in 10 mm Hg increments, was associated with a greater likelihood of LOS of <4 hours (RR 1.03 per 10 mm Hg [95% CI, 1.01–1.04]; p < 0.001).

Table 4.

Predictors of emergency department length of stay of <4 hours^*

CHIP = Children’s Health Insurance Program; BP = blood pressure.

^*Excluding patients who arrived by emergency medical services.

Data in bold represent statistically significant values (P-value < 0.05).

Sensitivity Analyses

To support the validation of the anaphylaxis case definition, we compared ED medication administration between visits coded with and without anaphylaxis. Among visits coded as anaphylaxis, 52% involved administration of epinephrine, antihistamines, and/or glucocorticoids during their ED visit (95% CI, 47%–57%), compared with 6% of patients without anaphylaxis (95% CI, 6%–7%; p < 0.001). We, in addition, conducted sensitivity analyses restricted to anaphylaxis-coded visits in which at least one of these medications was administered. These analyses were performed to assess the internal consistency of our results. In terms of specific treatments, 1.2% received epinephrine alone, 10.6% received epinephrine in combination with other medications, 67.3% received other medications alone, and 20.9% received no pharmacologic treatment.

For epinephrine prescription at discharge, inclusion of the triage level in multivariable models did not significantly change RR estimates for other covariates (Supplemental Table ST1). In the analyses, there was a nonsignificant trend toward lower epinephrine prescribing among visits with lower triage acuity. When the triage level was modeled as a continuous variable, the association was not statistically significant (RR 0.84 [95% CI, 0.69–1.02] per 1-level increase); missing and no-triage responses were included as separate indicator categories. For LOS, adjustment for EMS arrival, including a missingness indicator for visits with unknown EMS status (n = 13), resulted in minimal changes to the RRs (Supplemental Table ST2). When adjusting for the triage level, visits with a triage level of 4 (semi-urgent) and triage level of 5 (nonurgent) were more likely to have an LOS of <4 hours compared with higher acuity visits (RR 1.25 [95% CI, 1.07–1.46], p = 0.01], and RR 1.32 [95% CI, 1.07–1.62], p = 0.01], respectively) (Supplemental Table ST2). Most patients had a time of first provider contact to discharge of <4 hours (80%) (Supplemental Table ST3).

In sensitivity analyses, including visits with patients arriving via EMS, the associations between patient/visit characteristics and short LOS (<4 hours) were similar to those observed in the primary analysis (Supplemental Table ST4). A subgroup analysis of visits with patients arriving via EMS was not performed due to the small sample size (n= 73). In a separate analysis that used an alternative definition of LOS, the time from first provider contact to discharge, predictors of short LOS were overall consistent with the primary analysis (Supplemental Table ST5). In sensitivity analyses that used mean arterial pressure instead of SBP, associations with epinephrine prescription at discharge and ED LOS were similar to the primary analyses (Supplemental Tables ST6 and ST7).

DISCUSSION

In this NHAMCS-based analysis of adult ED visits for anaphylaxis, epinephrine prescribing at discharge was infrequent, with only 14% of visits that resulted in a prescription. Among the patients discharged from the ED, most visits (83%) lasted < 4 hours. Older adults and Hispanic patients were less likely to receive epinephrine at discharge. Rural visits and visits with patients who identified as “other” race and/or ethnicity were more likely to have shorter ED stays. Higher initial systolic blood pressure was associated with shorter LOS, which likely reflected less-severe presentations. Lower triage acuity (levels 4–5) was associated with a shorter LOS, whereas higher acuity presentations were associated with longer ED stays.

These findings are broadly consistent with previous studies, highlighting gaps in post-ED anaphylaxis management.^6,7 Previously, Clark et al.⁶ reported that only 16% of patients with food allergy–related anaphylaxis were prescribed self-injectable epinephrine at discharge. In a subsequent multicenter study on insect-sting allergies, the same group found that only 27% of patients who experienced systemic reactions received epinephrine at discharge.⁷ By contrast, Motosue et al.⁴ observed a greater proportion of patients (47%) receiving an epinephrine autoinjector at discharge. Results of our study suggest that epinephrine prescription at discharge remains less frequent than recommended by current guidelines.¹ In our study, sensitivity analyses restricted to visits in which patients received epinephrine, antihistamines, and/or corticosteroids during the ED visit yielded similarly low discharge prescribing rates, which suggests that misclassification of anaphylaxis or inclusion of clinically mild cases alone is unlikely to fully explain these findings.

In addition, most visits in this cohort were associated with an ED LOS < 4 hours. As expected, a shorter LOS was associated with lower triage acuity. ED LOS, defined as time from arrival to discharge, is an imperfect proxy for clinical observation time because it may be influenced by factors such as waiting room time and illness severity. Additional unmeasured factors may influence ED LOS and anaphylaxis treatment, including patient socioeconomic status, ED capacity and staffing, triage practices, patient access to epinephrine outside the ED, needle aversion, and diagnostic uncertainty. As such, our findings should be interpreted in this context. However, LOS is an objective, consistently recorded metric in large ED datasets, and, in combination with sensitivity analyses that use provider-contact times and inclusion of EMS arrivals, provides a reasonable approach to evaluate patterns of ED management.^16,17 To address this limitation, we performed sensitivity analyses by using an alternative definition of LOS based on the time from first provider contact to discharge and separate sensitivity analyses, including visits with patients arriving via EMS. Findings were consistent across these analyses, which support the robustness of the observed associations. Although LOS is an indirect proxy for post-treatment observation time, this pattern is consistent with prior work demonstrating relatively brief ED observation periods after anaphylaxis.⁵ Current practice parameters emphasize individualized observation based on clinical response and risk factors rather than a fixed minimum duration, and short ED stays may be appropriate for selected patients who achieve rapid clinical stability.¹⁹ However, given the risk of recurrent or biphasic reactions, careful post-treatment monitoring remains an important component of anaphylaxis management. The high proportion of brief ED stays observed in this probability-sampled cohort suggests variability in monitoring practices and highlights the need for further study to better inform post-treatment observation practices.

In addition, we observed potential disparities in both discharge prescribing and ED LOS. Older adults and Hispanic patients were less likely to receive an epinephrine prescription at discharge, consistent with previous studies that describe age- and race and/or ethnicity–associated differences in anaphylaxis recognition and management.^20–23 Campbell et al.²⁰ found that older adults who presented with anaphylaxis were less likely to be prescribed epinephrine compared with younger patients, possibly reflecting diagnostic uncertainty or provider hesitation due to cardiovascular comorbidities. Clinicians may also be more cautious in treating older adults with complex comorbidities and polypharmacy, given concerns about cardiovascular risk and challenges with medication adherence, which have been documented as factors that influence prescribing behavior in older adults.^21,22 In addition, Bourgoyne et al.²³ found race-based disparities in anaphylaxis care because Black patients were less likely than White patients to receive an epinephrine prescription at discharge. Hispanic patients may face similar disparities, which could result from a combination of factors, including language barriers, limited access to follow-up care, differences in health literacy, and implicit biases in clinical decision-making.²⁴ Visits to rural EDs were more likely to be shorter, which may reflect differences in staffing, resource availability, or ED throughput pressures in non-metropolitan settings. Patients categorized as “other” race and/or ethnicity were also more likely to have an ED LOS of <4 hours. Of note, associations that involve the “other” race and/or ethnicity category should be interpreted cautiously because this designation is heterogeneous and may reflect variation in data collection or reporting rather than a biologically meaningful grouping.

Our results have important clinical implications. Current expert guidance recommends that patients treated for anaphylaxis be discharged with an epinephrine autoinjector and receive appropriate education with regard to recognition and management of recurrent reactions.^1,19 Of note, the Joint Task Force on Practice Parameters guidelines states that early epinephrine administration is recommended, even when the diagnosis of anaphylaxis is uncertain.¹ The low rates of epinephrine prescribing observed in this study underscore gaps in discharge management and opportunities to enhance the quality of adult anaphylaxis care in EDs. Results of studies have shown that, even when epinephrine is prescribed at discharge, several barriers may limit its use, including challenges in acquisition of the epinephrine autoinjector, recognition of scenarios for use, and proper administration.^25,26 Previous quality improvement initiatives that incorporate standardized anaphylaxis order sets, electronic decision support, and streamlined access to epinephrine autoinjectors demonstrated improvements in both acute treatment and discharge planning.^27,28 Broader implementation of such strategies may help promote more consistent adherence to recommended management.

This study has several notable strengths. By leveraging NHAMCS-ED, a probability-based national survey of U.S. EDs, this analysis provides contemporary, population-level estimates of adult anaphylaxis practices across diverse ED settings. Validation analyses demonstrated substantially higher use of epinephrine, antihistamines, and/or corticosteroids among visits coded as anaphylaxis compared with non-anaphylaxis visits, which supports the clinical relevance of the identified cohort. The sensitivity analysis, when adjusting for EMS arrival and triage level, demonstrated that associations with epinephrine prescribing and ED LOS were largely unchanged, which suggests that robustness of these findings persist after accounting for initial severity or prehospital care. These results reinforce the robustness of our findings with regard to persistent gaps in post-ED management of adult anaphylaxis.

Some limitations merit consideration. The unweighted number of anaphylaxis visits was modest, which may limit precision and the ability to detect smaller associations or perform detailed subgroup analyses. However, the use of survey weights enables nationally representative estimates, and findings should be interpreted within this context. In addition, ED LOS, defined as the time from arrival to discharge, is an indirect measure of posttreatment observation. It includes preprovider time, which may vary due to clinical factors (e.g., illness severity, prehospital interventions) and operational factors (e.g., staffing, ED crowding). Despite this, LOS is a consistently recorded metric in large ED datasets, and our sensitivity analyses by using provider-contact times and inclusion of EMS arrivals support the robustness of our findings.^16,17 Case identification relied on ICD-9-CM and ICD-10-CM, which may misclassify or undercapture atypical or less-severe presentations. NHAMCS-ED also lacks granular clinical data to confirm diagnostic criteria, recurrent reactions, or previous epinephrine use. Although previous ownership of an epinephrine autoinjector may partly explain lower discharge prescribing rates, clinicians are still generally encouraged to prescribe epinephrine because devices may be expired, lost, or improperly stored, and the ED encounter provides an opportunity to reinforce patient education. The dataset also does not capture patient refusal, discharge counseling, or reasons for nonprescribing, nor detailed clinical variables such as the number of epinephrine doses or anaphylaxis history. Further, prehospital treatment could not be evaluated and represents an important area for future study. Finally, reduced hospital response rates during 2020–2022, particularly in rural EDs, may have introduced nonresponse bias.

CONCLUSION

In this analysis of adult anaphylaxis in U.S. EDs, epinephrine was infrequently prescribed at discharge and post-treatment observation periods were often brief. Older adults and Hispanic patients were less likely to receive an epinephrine prescription at discharge, and visits to rural EDs were more likely to have shorter LOS. These findings identify potential targets for system-level efforts to improve adherence to recommended treatment practices and promote more consistent care.