Prehospital time for patients with acute cardiac complaints: A rural health disparity

Abstract

Objective:

Delays in care for patients with acute cardiac complaints are associated with increased morbidity and mortality. The objective of this study was to quantify rural and urban differences in prehospital time intervals for patients with cardiac complaints.

Methods:

The ESO Data Collaborative dataset consisting of records from 1332 EMS agencies was queried for 9–1–1 encounters with acute cardiac problems among adults (age ≥ 18) from 1/1/2013–6/1/2018. Location was classified as rural or urban using the 2010 United States Census. The primary outcome was total prehospital time. Generalized estimating equations evaluated differences in the average times between rural and urban encounters while controlling for age, sex, race, transport mode, loaded mileage, and patient stability.

Results:

Among 428,054 encounters, the median age was 62 (IQR 50–75) years with 50.7% female, 75.3% white, and 10.3% rural. The median total prehospital, response, scene, and transport times were 37.0 (IQR 29.0–48.0), 6.0 (IQR 4.0–9.0), 16.0 (IQR 12.0–21.0), and 13.0 (IQR 8.0–21.0) minutes. Rural patients had an average total prehospital time that was 16.76 min (95%CI 15.15–18.38) longer than urban patients. After adjusting for covariates, average total time was 5.08 (95%CI 4.37–5.78) minutes longer for rural patients. Average response and transport time were 4.36 (95%CI 3.83–4.89) and 0.62 (95%CI 0.33–0.90) minutes longer for rural patients. Scene time was similar in rural and urban patients (0.09 min, 95%CI −0.15–0.33).

Conclusion:

Rural patients with acute cardiac complaints experienced longer prehospital time than urban patients, even after accounting for other key variables, such as loaded mileage.

1. Introduction

1.1. Background

Acute cardiac problems such as acute coronary syndrome (ACS) account for nearly 10% of all emergency medical services (EMS) encounters [1,2]. Furthermore, emergency departments (ED) in the United States (US) care for between 8 and 10 million patients with acute cardiac complaints annually, with almost half of these patients arriving by EMS [3–5].

1.2. Importance

There is a paucity of data examining total prehospital, response, scene, and transport times for patients with acute cardiac complaints in rural and urban settings. Studies evaluating patients with ST-elevation myocardial infarction (STEMI) suggest that almost half of rural EMS STEMI patients do not meet guideline-based time goals [6–8]. It is unclear if a prehospital time disparity exists for patients with acute cardiac complaints. In 2020, the American Heart Association underscored the importance of mitigating cardiovascular rural health disparities. Quantifying differences in prehospital times for patients in rural vs. urban settings is a first step in reducing these disparities.

1.3. Goals of this investigation

We hypothesized that rural patients with acute cardiac complaints would have prolonged prehospital times compared to their urban counterparts. Therefore, the primary objective was to determine if total prehospital, response, scene, and transport time differed between rural and urban patients. A secondary objective was to assess for associations between age, sex, race, transport mode, and loaded mileage with each time interval.

2. Methods

2.1. Study design

A national retrospective cross-sectional study of US EMS encounters from 1332 EMS agencies using the ESO Data Collaborative (Austin, TX) was conducted for 9–1–1 calls with an impression, chief complaint, or supporting documentation of an acute cardiac problem among adult patients (age ≥ 18) from January 1st, 2013 to June 1st, 2018. Keywords used to identify an acute cardiac complaint included “chest pain,” “heart problem,” “chest discomfort,” “heart failure,” “myocardial infarction,” “STEMI,” “NSTEMI,” “chest pressure,” and “arrythmia.” ESO provides prehospital electronic health records (EHR) and generates a deidentified public-use dataset. The Wake Forest University Health Sciences Institutional Review Board deemed this investigation exempt. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines helped direct the research process [9].

2.2. Study setting and population

Both advanced and basic life support agencies were included. Only patients transported from the scene to a hospital by ground ambulance were included. Encounters with patients who presented in or went into cardiac arrest while in the care of EMS were excluded.

Encounter location was classified as rural or urban by matching the encounter zip code to the US Census Bureau's 2010 census. The US Census Bureau defines an Urbanized Area (UA) as an area of at least 50,000 people, an Urbanized Cluster (UC) as an area of at least 2500 people but less than 50,000, and a Rural Area as all areas that are not UAs or UCs [10]. Urban was defined as the composite of a UA or a UC. A patient was considered unstable if systolic blood pressure was <90 mmHg, the heart rate was ^>120 or < 40 beats per minute, or pulse oximetry was <90% [11]. Ethnicity was not routinely captured separately from race. Therefore, race was collapsed into a four-level variable (White, Hispanic or Latino, Black, and Other). EMS transport mode was defined as emergent (lights and siren) or non-emergent (no lights or siren). Loaded mileage was defined as the distance traveled from the scene to the hospital.

2.3. Outcomes

The primary outcome was total prehospital time, defined as the time from EMS dispatch to hospital arrival. Secondary outcomes included response time (dispatch to scene arrival), scene time (scene arrival to depart scene), and transport time (depart scene to hospital arrival). These time points were recorded in the ESO EHR and then typically validated by each local prehospital agency.

2.4. Statistical analysis

Descriptive statistics were used to characterize the population. The unit of analysis was the EMS encounter. Age, sex, race, location, stability, and transport mode were treated as categorical variables and reported as frequencies and percentages. Total prehospital time, response time, scene time, transport time, and loaded mileage were treated as continuous variables and reported as medians and interquartile ranges (IQR). A complete case analysis was performed under the assumption that data were missing completely at random (MCAR). Encounters with a missing time interval were excluded. Extreme outliers were excluded and defined as a response time of 0 or > 60 min, scene time of 0 or > 120 min, transport time of 0 or > 180 min, transport distance of 0 miles, rural encounters with a loaded mileage <0.5 miles, calculated transport speed ^>100 miles per hour (MPH), and those with a loaded mileage ^>25 miles with a transport speed <5 MPH. Encounters with missing covariates (sex, race, rurality, or instability) were also excluded.

Generalized estimating equations (using the normal distribution with identity link and assuming an exchangeable correlation structure) evaluated differences in average total prehospital time, response time, scene time, and transport time while controlling for age, sex, race, location, transport mode, loaded mileage, and patient stability while accounting for clustering of patients within EMS agencies. The Quasilikelihood under the Independence model Criterion (QIC) was used to select the covariance structure and Zeng's marginal R² statistic was calculated to evaluate model fit [12,13]. Parameter estimates with 95% confidence intervals (95%CI) were determined.

3. Results

3.1. Characteristics of study subjects

There were 774,174 acute cardiac prehospital encounters. Of these, 514,642 met the inclusion criteria and 428,054 were included in the analysis after exclusions (Fig. 1). The population had a median age of 62 years (IQR 50–75) years and was 50.7% female, 75.3% white race, and 10.3% rural. The median total prehospital, response, scene, and transport times were 37.0 (IQR 29.0–48.0), 6.0 (IQR 4.0–9.0), 16.0 (IQR 12.0–21.0), and 13.0 (IQR 8.0–21.0) minutes. Table 1 provides demographic information and Table 2 provides time interval findings. Fig. 2 graphically presents the time intervals and the rural and urban time interval medians.

Fig. 1.

Case-selection flow diagram.

Table 1.

Demographics of the overall study population. Reported as median (IQR) or N(%)

Variable	Rural N = 44,176 (10.3%)	Urban N = 383,878 (89.7%)	Total N = 428,054
Age (continuous)	62 (50–75)	62 (50–75)	62 (50–75)
Age (categorical)
18–44	7039 (15.9%)	64,962 (16.9%)	72,001 (16.8%)
45–64	17,200 (38.9%)	146,285 (38.1%)	163,485 (38.2%)
65+	19,937 (45.1%)	172,631 (45.0%)	192,568 (45.0%)
Sex
Female	21,807 (49.4%)	195,092 (50.8%)	216,899 (50.7%)
Race/Ethnicity
White	39,094 (88.5%)	283,243 (73.8%)	322,337 (75.3%)
Black or African American	4048 (9.2%)	75,536 (19.7%)	79,584 (18.6%)
Hispanic or Latino	857 (1.9%)	20,444 (5.3%)	21,301 (5.0%)
Other	177 (0.4%)	4655 (1.2%)	4832 (1.1%)
Transport Mode
Lights/Siren	12,217 (27.7%)	77,749 (20.3%)	89,966 (21.0%)
Unstable
Yes	14,390 (32.6%)	129,474 (33.7%)	143,864 (33.6%)
Loaded Mileage	20.8 (14.0–29.7)	5.5 (2.9–10.5)	6.2 (3.0–12.8)

IQR – interquartile range.

Table 2.

Time interval descriptive statistics

Time (min)	Rural Median (IQR)	Urban Median (IQR)	Total Median (IQR)
Total Prehospital Time	55.0 (44.0–68.0)	35.0 (28.0–45.0)	37.0 (29.0–48.0)
Response Time	10.0 (5.0–16.0)	5.0 (4.0–8.0)	6.0 (4.0–9.0)
Scene Time	16.0 (12.0–21.0)	16.0 (12.0–21.0)	16.0 (12.0–21.0)
Transport Time	27.0 (19.0–36.0)	12.0 (8.0–18.0)	13.0 (8.0–21.0)

IQR – interquartile range.

Fig. 2.

The prehospital time intervals of interest presented with median rural and urban times.

3.2. Main results

In univariable analysis, rural patients had an average total prehospital time that was 16.76 min (95%CI 15.15–18.38) longer than urban patients. Rural patients had longer response times (5.42 min, 95%CI 4.86–5.97), scene times (0.49 min, 95%CI 0.26–0.72), and transport times (10.92 min, 95%CI 9.69–12.14). The median rural loaded mileage distance was 20.8 miles (IQR 14.0–29.7) while the median urban distance was 5.5 miles (IQR 2.9–10.5).

After adjusting for covariables, including loaded mileage, rural patients still had longer average total prehospital time (5.08 min, 95%CI4.37–5.78). Response time (4.36 min, 95%CI 3.83–4.89) and transport time (0.62 min, 95%CI 0.33–0.90) remained longer for rural patients while scene time was no longer associated with location (0.09 min, 95%CI −0.15–0.33).

In the adjusted model, every 10-mile increase in loaded mileage was associated with longer total prehospital time (11.46 min, 95%CI 11.05–11.88), response time (1.02 min, 95%CI 0.85–1.19), scene time (0.38 min, 95%CI 0.29–0.47), and transport time (10.06 min, 95%CI 9.82–10.30). Unstable patients had longer total prehospital time (1.50 min, 95%CI 1.35–1.65) and scene time (1.45 min, 95%CI 1.32–1.58). Using lights and siren during transport was associated with less total prehospital time (−3.51 min, 95%CI −3.81 to −3.21), response time (−0.22 min, 95%CI −0.29 to −0.15), scene time (−0.27 min, 95%CI −0.45 to −0.08), and transport time (−3.02 min, 95%CI −3.20 to −2.85 min). Older patients, females, and white patients had longer total prehospital times. Table 3 presents the multivariable models.

Table 3.

Multivariable model for each prehospital time interval. Results reported as the change in minutes with the associated 95%CI.

Variable	Total Prehospital Time (marginal R2 = 0.599)	Response Time (marginal R2 = 0.059)	Scene Time (marginal R2 = 0.048)	Transport Time (marginal R2 = 0.826)
Age (categorical)
45–64 vs. 18–44	0.89 (0.79–0.99)	−0.09 (−0.14– −0.05)	0.90 (0.82–0.97)	0.09 (0.04–0.14)
65+ vs. 18–44	2.91 (2.76–3.06)	0.01 (−0.06–0.07)	2.67 (2.56–2.78)	0.23 (0.17–0.30)
Sex
Female vs. Male	1.01 (0.88–1.14)	−0.06 (−0.10– −0.01)	0.88 (0.79–0.98)	0.18 (0.15–0.21)
Race
Black vs. White	−0.18 (−0.45–0.09)	−0.34 (−0.48– −0.20)	0.17 (0.05–0.29)	−0.01 (−0.17–0.16)
Hispanic/Latino vs. White	−0.16 (−0.35–0.03)	−0.13 (−0.24– −0.03)	0.05 (−0.08–0.18)	−0.08 (−0.18–0.01)
Other vs. White	−0.09 (−0.47–0.29)	−0.003 (−0.20–0.20)	0.03 (−0.17–0.23)	−0.11 (−0.27–0.05)
Location
Rural vs. Urban	5.08 (4.37–5.78)	4.36 (3.83–4.89)	0.09 (−0.15–0.33)	0.62 (0.33–0.90)
Transport Mode
Lights/Siren vs. No Lights/Siren	−3.51 (−3.81– −3.21)	−0.22 (−0.29– −0.15)	−0.27 (−0.45– −0.08)	−3.02 (−3.20– −2.85)
Unstable
Yes vs. No	1.50 (1.35–1.65)	0.02 (−0.01–0.06)	1.45 (1.32–1.58)	0.03 (−0.004–0.07)
Loaded Mileage (10-mile increase)	11.46 (11.05–11.88)	1.02 (0.85–1.19)	0.38 (0.29–0.47)	10.06 (9.82–10.30)

4. Limitations

This study has limitations. It was a retrospective study using deidentified data among a heterogeneous population, making it possible that some patients had multiple encounters. Because this dataset was not linked with outcomes, it is unclear how many patients were diagnosed with ACS, required admission, or experienced other morbidity or mortality that could have been affected by time-to-delivery of care. Non-ESO EHR systems exist and were not included in this study, which limits the study's generalizability and may have led to selection bias. The US Census Bureau's dichotomous rural vs. urban scheme may oversimplify the rural-urban continuum as it does not account for socioeconomic factors, geography, or access to healthcare [14,15]. Regarding response time, scene time, transport time, loaded mileage, and transport MPH, practical yet arbitrary cut-offs were selected to exclude likely erroneous data entries. A complete case analysis was performed under the assumption that missing data are MCAR. If this assumption is incorrect, there could be some bias in the estimates. The analysis was also done under the assumption of normality. Even though the times evaluated were skewed, the normality assumption should be reasonable given the large sample size. Lastly, owing to the size of the dataset, there is a risk of detecting statistically significant but not clinically relevant differences.

5. Discussion

This national study of nearly 500,000 acute cardiac EMS encounters demonstrates that even after controlling for loaded mileage and other potential confounders, total prehospital time is longer for rural patients than urban patients. This suggests that opportunities exist to improve rural prehospital times. Addressing this rural disparity may reduce cardiovascular morbidity and mortality.

The importance of timely prehospital care and transport is well described. The American College of Cardiology recommends that patients with “suspected ACS and high-risk features” be referred immediately to the ED. [16] The Mission: Lifeline program identifies timely prehospital care as key for caring for patients with STEMI [8,17,18]. For these patients, every 30-min delay in reperfusion therapy is associated with a 7.5% increase in mortality [19]. Furthermore, EMS scene times less than 15 min are associated with a ten times greater odds of achieving the 90-min first medical contact to percutaneous coronary intervention benchmark for patients with STEMI [20,21].

A key finding is that rural patients were delayed in reaching hospitalbased care due to prolonged response times and transport times. Rural patients waited almost twice as long (11.4 vs. 6.5 min) for EMS to arrive compared to urban patients. A potential solution to this response time disparity includes optimizing rural EMS resources such as staff, ambulances, quick response vehicles, and base locations. Rural patients also had longer transport times. Distance to hospitals is naturally longer in the rural setting, especially in the era of many rural hospitals closing [22,23]. However, this disparity persisted despite accounting for loaded mileage, suggesting that the problem may stem from potentially modifiable rural EMS factors. Rural agencies may lack quality improvement and assurance programs. Alternatively, perhaps the quality and type of roadway available in rural areas contribute to this disparity. Although rural and urban scene time were similar, average times in both groups failed to meet Studnek's 15-min scene time metric, which has been associated with meeting guideline-based care metrics [21].

This study reveals other care disparities. Older patients had longer scene and transport times. Older patients may require moving assistance and more interventions, thereby accounting for longer scene time. Females received less timely care than males. This sex disparity is pervasive in the cardiovascular literature [24–27]. Previous prehospital literature indicates that use of lights and siren during transport does save time, especially during urban transports [28–31]. However, their use has not been shown to improve patient outcomes [28,32,33]. Unstable patients had longer prehospital time, particularly due to longer scene times. These longer scene times are likely due to performing onscene stabilization and procedures.

6. Conclusion

This national study demonstrates that rural patients with acute cardiac complaints receive less timely care than their urban counterparts. Even after adjusting for distance traveled and other confounders, rural patients still had longer prehospital times. These results suggest that further studies of EMS agency workflows could identify opportunities to improve rural prehospital times. Females and unstable patients also had longer prehospital time. These delays in care may contribute to increased morbidity and mortality.