Greater spatial access to care is associated with lower mortality for emergency general surgery : Journal of Trauma and Acute Care Surgery

Secondary Logo

Journal Logo

AAST PODIUM 2022

Greater spatial access to care is associated with lower mortality for emergency general surgery

McCrum, Marta L. MD, MPH; Allen, Chelsea M. PhD; Han, Jiuyin MS; Iantorno, Stephanie E. MD; Presson, Angela P. PhD; Wan, Neng PhD

Author Information
Journal of Trauma and Acute Care Surgery 94(2):p 264-272, February 2023. | DOI: 10.1097/TA.0000000000003837
  • Free
  • SDC
  • Infographic
  • Best Of

The state of emergency general surgery (EGS) care in the United States is increasingly regarded as a growing public health crisis; progressive increases in EGS hospital admissions, worsening workforce shortages, and hospital closures have contributed to demand that is quickly outpacing system capacity and significantly affecting patient access to care.1–4 While EGS diseases are a heterogeneous group of conditions with wide-ranging severity, they are by definition time-sensitive, requiring urgent evaluation and often emergent operative intervention. Physical access to hospitals with EGS services is therefore critical to the diagnosis and treatment of these diseases. This concept—the ease with which residents of a certain area can reach needed health services and facilities—is referred to as “potential spatial access.”

Spatial access to care encompasses both the accessibility of service providers (commonly measured by distance or time to nearest hospital) and availability, which describes the number of service providers and their capacity to meet nearby population demand (often reported as provider-to-population ratios).5 Distance to the nearest hospital is a known barrier to spatial access for surgical care for many rural patients; however, inadequate hospital capacity to serve patients requiring surgical care is pervasive throughout both urban and rural areas.6 The Acute Care Congress in their 2009 report on “The Future of Emergency Surgical Care in the United States” noted that emergency department overcrowding and staff shortages have led to unacceptably long waits for emergent surgery, specifically highlighting the importance of considering hospital capacity in assessments of spatial access.2 More recently, advanced geospatial methods such as gravity models have emerged as a combined metric for both accessibility and availability.7 Regardless of method, considerable population-level spatial access disparities have been noted in the United States, with rural, minority, and uninsured populations disproportionately affected.8–12 Despite this growing body of evidence documenting disparities, it remains unclear how spatial access to emergency surgical care contributes to outcomes for patients with EGS disease.

To address this critical gap in our knowledge of the connection between spatial access and clinical outcomes, we analyzed 12 inpatient state databases and used advanced geospatial modeling to evaluate the association of spatial access to emergency surgical care with in-hospital mortality and major morbidity for eight common EGS conditions. Given the time-sensitive nature of these diseases, we hypothesized that greater spatial access to EGS care would be associated with improved clinical outcomes.

PATIENTS AND METHODS

Data Sources and Patient Selection

We used the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project 2014 State Inpatient Databases from 11 states (Arizona, Colorado, Florida, Iowa, Kentucky, New York, North Carolina, Oregon, Utah, and Washington) and the 2014 California Inpatient Discharge Dataset from the California Office of Statewide Health Planning and Development to identify all adults (18 years or older) with an urgent or emergent admission for a primary diagnosis of one of eight common EGS conditions: appendicitis, cholecystitis, diverticulitis, abdominal wall or intra-abdominal hernia, intestinal obstruction, mesenteric ischemia, peptic ulcer disease, and pancreatitis using the International Classification of Diseases, Ninth Revision (ICD-9), codes (Supplementary Table 1, https://links.lww.com/TA/C785).13,14 These diseases were chosen for both their overall frequency and the presence of American Association for the Surgery of Trauma severity scales for each; states were selected based on availability of necessary data elements in the State Inpatient Databases. We used the dichotomized schema described by Scott et al.14 to define “complex” versus “uncomplicated” disease for each diagnosis, which were developed by mapping ICD-9 codes to American Association for the Surgery of Trauma severity scales. Patients were excluded if they were younger than 18 years at time of EGS admission, were transferred out to another acute care hospital, or were missing home ZIP code (total missing, n = 5,750 [0.49%]). Patients who were admitted as an interfacility transfer were included in the analysis and analyzed at the terminal location of acute care admission.

Patient-level data was linked to the American Hospital Association 2015 Annual Survey to obtain hospital-level data, including geographic location and clinical resources.15 Spatial access was calculated as described hereinafter using our previously constructed Geographic Information Science platform for EGS-capable hospitals in the United States, which uses data from the Census Topologically Integrated Geographic Encoding and Referencing file, and the StreetMap North America network data set from the Environmental Systems Research Institute.10,16,17

Patient and Hospital Characteristics

Patient characteristics examined included age, sex, race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, Asian or Pacific Islander, Native American, or other race), rurality, rate of poverty in home census block group, primary payer (private, Medicaid, Medicare, self pay, or other), and Charlson Comorbidity Index (CCI) derived from ICD-9 diagnosis codes. Hospital characteristics assessed included teaching status, hospital bed size, and presence of advance clinical resources, which we have previously defined as number of ICU beds >25th percentile nationally, presence of computed tomography and ultrasound imaging, advanced gastroenterology serviced (identified by endoscopic retrograde cholangiopancreatography), or were in the top national quartile of inpatient operations (>2,753 procedures annually).10

Spatial Access

We used the spatial access ratio (SPAR), an enhanced two-step floating catchment area (E2SFCA) model, to measure spatial access to EGS-capable hospitals.10,16,18 E2SFCA models are a widely validated type of gravity model, which incorporate both accessibility and availability of health services into a single measure of spatial access for a given population site.6,7,19–21 In brief, this method uses provider-to-population ratios weighted by hospital capacity, potential patient volume, and travel impedance (i.e., travel time). Spatial access ratio is presented as a ratio of the spatial access for the specific census block group relative to the national mean. Greater values of SPAR denote better spatial access, and SPAR values greater than one indicate better-than-national-average spatial access. We used an equal-interval scheme to stratify patients into four categories of access: very low (SPAR, 0), low (0 < SPAR < 0.5), moderate (0.5 ≤ SPAR ≤ 1) and greater than average (SPAR, >1).

To calculate SPAR, we used measures of supply, demand, and travel impedance. Supply or hospital capacity was measured using number of inpatient hospital beds. Demand was approximated by the population of each census block group and located at the population-weighted centroid. We set the catchment area to 60 minutes driving time, as this represents a common benchmark for access to surgical care.8,22–25 Distance and travel time from population-weighted zip code tabulation area centroid to hospital site was calculated using ArcMap software (Esri, Redlands, CA). Further details on calculation of E2SFCA models and SPAR can be found in Supplemental Digital Content (Supplementary Data 1, Supplementary Methods, https://links.lww.com/TA/C784).

Primary and Secondary Outcomes

Our primary outcome was all-cause in-hospital mortality. Our secondary outcome was a composite measure of major morbidity consisted of respiratory failure, myocardial infarction, stroke/cerebrovascular accident, venous thromboembolism (deep vein thrombosis or pulmonary embolism), acute renal failure, surgical site infections, and urinary tract infection/cystitis identified using previously published ICD-9 codes (Supplementary Table 2, https://links.lww.com/TA/C785).26,27

Statistical Analysis

We assessed differences in patient and hospital characteristics using χ2 and analysis of variance tests for categorical and continuous variables as appropriate. Mixed-effects univariable logistic regressions were used to calculate the unadjusted odds of mortality and morbidity (separately) for SPAR and all covariates of interest. All models included random effects for state and hospital, nested within state. A multivariable logistic regression model was then constructed using the SPAR of the patient's home geographic location (ZIP code) and the following base variables: age, sex, primary payer, CCI, hospital teaching status, and presence of advanced hospital resources (Model 1). A second model (Model 2) was then constructed by adding the presence of complex disease to the variables in Model 1. While complex disease potentially falls on the causal pathway between spatial access to care and our mortality and major morbidity outcomes, it is also known to be strongly associated with the outcome. Variables that were highly related to the calculation of SPAR (e.g., rurality, distance to hospital) or advanced hospital resources (e.g., bed number) were not considered as separate covariates. Multicollinearity was assessed using variance inflation factors: all variance inflation factor values were less than 5, and thus, no covariates were eliminated or combined because of multicollinearity. Model accuracy or the ability of our models to discriminate outcomes was measured using the area under the receiver operating characteristic curve with 95% confidence intervals (CIs). We analyzed all EGS conditions as a group and then performed separate subanalyses for each EGS condition for both mortality and major morbidity outcomes. Furthermore, because patients who require operative intervention and those who undergo interhospital transfer might be particularly sensitive to the effects of spatial access to surgical care, we performed separate subgroup analyses for each of these two groups. Finally, we examined the differential effects of spatial access on populations known to be vulnerable to effects of decreased access to care through interactions between SPAR and each of race/ethnicity, insurance, and poverty. Odds ratios and adjusted odds ratios (aORs), 95% CIs, and p values were reported.

All statistical analyses were performed using R version 4.2.1 (R-studio, Vienna, Austria), with statistical significance set at p < 0.05.28 The study followed the Strengthening of the Reporting of Observational Studies in Epidemiology guidelines and was classified as not human subjects research by policy of our institutional review board as all data were deidentified (Supplemental Digital Content, Supplementary Data 2, https://links.lww.com/TA/C786). All research activities followed regulations within the Healthcare Cost and Utilization Project and Agency for Healthcare Research and Quality data use agreements.

RESULTS

Patient Cohort

Among 877,928 adult patients admitted with appendicitis (n = 67,190 [6.8%]), cholecystitis (n = 44,251 [4.5%]), diverticulitis (n = 104,066 [11%]), abdominal wall or intra-abdominal hernia (n = 274,791 [28%]), intestinal obstruction (n = 141,182 [14%]), mesenteric ischemia (n = 38,710 [3.9%]), peptic ulcer disease (n = 139,706 [14%]), or pancreatitis (n = 176,275 [18%]), nearly one half were male (n = 409,519 [47%]) and majority were White (n = 544,651 [62%]), with private insurance (299,004 [26%]) or Medicare (418,861 [48%]) and a median SPAR of 1.6 (interquartile range, 0.9–2.4). Compared with patients with spatial access at or above the national average (SPAR, ≥1), very-low-access (SPAR, 0) patients had a higher proportion of males (SPAR 0, 52% vs. SPAR ≥1, 47%), White race (66% vs. 59%), and higher rates of Medicaid coverage (22% vs. 20%). Median CCI was similar. Very-low-access patients were less likely to be admitted to teaching hospitals (SPAR 0, 13% vs. SPAR ≥1, 17%) and hospitals with advanced clinical resources (SPAR 0, 59% vs. SPAR ≥1, 67%). (Table 1 for all). Distribution of EGS conditions was roughly similar across all categories of SPAR, with slightly higher proportions of patients with cholecystitis (SPAR 0, 10% vs. SPAR ≥1, 5%) and diverticulitis (SPAR 0, 15% vs. SPAR ≥1, 12%) and lower proportion of intra-abdominal or abdominal wall hernia (SPAR 0, 26% vs. SPAR ≥1, 31%) in the very-low-access group (Table 2). Patients from very low (SPAR, 0) and low (SPAR, 0–0.5) areas had a higher proportion of complex disease compared with patients with average or greater access (SPAR 0, 31%; SPAR 0–0.5, 15%; SPAR ≥1, 12%), a pattern that was consistent for each individual EGS disease (Table 2).

TABLE 1 - Patient Demographics
SPAR = 0 (n = 21,358) 0 < SPAR < 0.5 (N = 82,974) 0.5 ≤ SPAR < 1.0 (N = 194,649) 1.0 ≤ SPAR (N = 578,947)
Age, median (IQR), y 63 (49–74) 62 (49–75) 62 (48–76) 62 (48–76)
Sex
 Male 11,007 (52%) 39,218 (47%) 89,883 (46%) 269,411 (47%)
Race/ethnicity
 White 14,067 (66%) 60,899 (73%) 130,400 (67%) 339, 285 (59%)
 Asian or Pacific Islander 469 (2%) 1,117 (1%) 8, 449 (4%) 28,281 (5%)
 Black 2,066 (10%) 3,959 (5%) 12,796 (7%) 68,002 (12%)
 Hispanic 2,662 (12%) 13,250 (16%) 33,100 (17%) 115,268 (20%)
 Native American 534 (3%) 1,114 (1%) 1,046 (1%) 23,89 (0%)
 Not Reported 533 (2%) 1,268 (2%) 3,507 (2%) 6,149 (1%)
 Other 1,027 (5%) 1,367 (2%) 5,351 (3%) 19,573 (3%)
Expected primary payor
 Private 4,852 (23%) 22,098 (27%) 55,474 (29%) 146,580 (25%)
 Medicaid 4,787 (22%) 14,388 (17%) 35,737 (18%) 113,341 (20%)
 Medicare 9,814 (46%) 40,825 (49%) 91,866 (47%) 276,356 (48%)
 No charge 160 (1%) 283 (0%) 428 (0%) 3,520 (1%)
 Other 548 (3%) 1863 (2%) 4,142 (2%) 12,003 (2%)
 Self pay 1,191 (6%) 3,483 (4%) 6,966 (4%) 26,978 (5%)
CCI, median (IQR) 1 (0–2) 1 (0–3) 1 (0–3) 1 (0–3)
Transfer in from acute care hospital 1,842 (9%) 5,772 (7%) 7,389 (4%) 14,857 (3%)
Major operative intervention 6,454 (30%) 21,433 (26%) 46,405 (24%) 140,860 (24%)
Poverty of census block group, median (IQR) 0.2 (0.1–0.3) 0.1 (0.1–0.2) 0.1 (0.1–0.2) 0.1 (0.1–0.2)
Rurality
 Urban 7,732 (36%) 47,719 (58%) 165,156 (85%) 530,352 (92%)
 Rural 13,626 (64%) 35,255 (42%) 29,489 (15%) 48,531 (8%)
Teaching hospital 2,826 (13%) 6,464 (8%) 13,299 (7%) 99,694 (17%)
Hospital bed size
 Small 2,389 (11%) 14,384 (17%) 19,801 (10%) 27,344 (5%)
 Medium 10,830 (51%) 49,224 (59%) 126,017 (65%) 315,196 (54%)
 Large 7,505 (35%) 15,330 (18%) 38,654 (20%) 215,623 (37%)
Advanced resources 12,683 (59%) 38,842 (47%) 105,584 (54%) 388,966 (67%)
*Missing values: sex, 277; expected primary payer, 314; hospital bed size unknown, 46,838.
IQR, interquartile range.

TABLE 2 - Emergency General Surgery Diagnosis of Study Cohort
SPAR = 0 (n = 21,358) 0 < SPAR < 0.5 (n = 82,974) 0.5 ≤ SPAR < 1.0 (n = 194,649) 1.0 ≤ SPAR (n = 578,947)
Appendicitis
 All 2,128 (10%) 5,948 (7%) 14,182 (7%) 44,932 (8%)
 Uncomplicated 1,305 (6%) 4,375 (5%) 12,965 (7%) 36,597 (6%)
 Complex 823 (4%) 1,573 (2%) 1,217 (1%) 8,335 (1%)
Cholecystitis
 All 2,106 (10%) 4,893 (6%) 7,374 (4%) 29,878 (5%)
 Uncomplicated 310 (1%) 1,555 (2%) 4,235 (2%) 10,630 (2%)
 Complex 1,796 (8%) 3,338 (4%) 3,139 (2%) 19,248 (3%)
Diverticulitis
 All 3,152 (15%) 10,587 (13%) 21,506 (11%) 68,821 (12%)
 Uncomplicated 1,658 (8%) 7,734 (9%) 19,187 (10%) 54,189 (9%)
 Complex 1,494 (7%) 2,853 (3%) 2,319 (1%) 14,632 (3%)
Abdominal wall or intra-abdominal hernia
 All 5,475 (26%) 25,477 (31%) 62,450 (32%) 181,389 (31%)
 Uncomplicated 4,556 (21%) 23,577 (28%) 60,759 (31%) 171,357 (30%)
 Complex 919 (4%) 1,900 (2%) 1,691 (1%) 10,032 (2%)
Intestinal obstruction without hernia
 All 3,390 (16%) 13,894 (17%) 32,807 (17%) 91,091 (16%)
 Uncomplicated 2,512 (12%) 12,283 (15%) 31,437 (16%) 82,707 (14%)
 Complex 878 (4%) 1,611 (2%) 1,370 (1%) 8,384 (1%)
Mesenteric ischemia
 All 1,300 (6%) 4,147 (5%) 8,151 (4%) 25,112 (4%)
 Uncomplicated 595 (3%) 2,856 (3%) 7,102 (4%) 18,697 (3%)
 Complex 705 (3%) 1,291 (2%) 1,049 (1%) 6,415 (1%)
Peptic ulcer disease
 All 2,890 (14%) 12,022 (14%) 30,923 (16%) 93,871 (16%)
 Uncomplicated 2,394 (11%) 11,126 (13%) 29,984 (15%) 88,280 (15%)
 Complex 496 (2%) 896 (1%) 939 (0%) 5,591 (1%)
Pancreatitis
 All 4,106 (19%) 16,971 (20%) 40,428 (21%) 114,770 (20%)
 Uncomplicated 3,107 (15%) 14,943 (18%) 38,674 (20%) 104,118 (18%)
 Complex 999 (5%) 2,028 (2%) 1,754 (1%) 10,652 (2%)
Total
 Uncomplicated 14,814 (69%) 70,374 (85%) 183,682 (94%) 510,703 (88%)
 Complex 6,544 (31%) 12,600 (15%) 10,967 (6%) 68,244 (12%)

Impact of Spatial Access on Mortality and Major Morbidity

For the study population, the overall incidence of in-hospital mortality was 2.49%, and composite major morbidity was 27.1%. Mortality was greater in the very-low-access group compared with the high-access group (SPAR 0, 4.4%; SPAR ≥1, 2.5%). Major morbidity did not differ substantially across groups; however, very-low-access patients had a slight increase in incidence of pulmonary failure (SPAR 0, 7.5%; SPAR ≥1, 5.6%) and acute renal failure (SPAR 0, 16%; SPAR ≥1, 14%) (Table 3). In univariable analysis, increasing SPAR 1 point (e.g., from 0 to the national average) was associated with 4% decrease in odds of mortality (odds ratio, 0.96; 95% CI, 0.95–0.97; p < 0.001) (Supplementary Table 3, https://links.lww.com/TA/C785); it was not associated with composite major morbidity. After controlling for competing risk factors in multivariable analysis, the association of spatial access with reduced in-hospital mortality remained significant (Model 1: aOR, 0.95; 95% CI, 0.94–0.97; p < 0.001). Addition of complex disease to the model attenuated this relationship slightly but not entirely (Model 2: aOR, 0.97; 95% CI, 0.95–0.98; p < 0.001) (Table 4). Within individual EGS conditions, SPAR was significantly associated with in-hospital mortality for those with cholecystitis (aOR, 0.93; 95% CI, 0.88–0.99; p = 0.015), diverticulitis (aOR, 0.93; 95% CI, 0.88–0.99; p = 0.02), abdominal wall or intra-abdominal hernia (aOR, 0.96; 95% CI, 0.93–0.98; p = 0.001), and peptic ulcer disease (aOR, 0.96; 95% CI, 0.93–0.99; p = 0.0018), but not appendicitis, intestinal obstruction, mesenteric ischemia, or pancreatitis (Fig. 1). There was no association of spatial access with major morbidity in univariable (Supplementary Table 4, https://links.lww.com/TA/C785) or multivariable models (Table 5).

TABLE 3 - Primary and Secondary Outcomes
SPAR = 0 (n = 21,358) 0 < SPAR < 0.5 (n = 82,974) 0.5 ≤ SPAR < 1.0 (n = 194,649) 1.0 ≤ SPAR (n = 578,947)
In-hospital mortality 940 (4%) 2,510 (3%) 3,693 (2%) 14,542 (3%)
Composite major morbidity 5,667 (27%) 22,864 (28%) 50,722 (26%) 156,052 (27%)
 Pulmonary failure 1,592 (7%) 5,275 (6%) 9,648 (5%) 32,276 (6%)
 Pneumonia 673 (3%) 2,420 (3%) 5,085 (3%) 16,197 (3%)
 Myocardial infarction 422 (2%) 1,649 (2%) 3,205 (2%) 10,115 (2%)
 Cerebrovascular accident 137 (1%) 660 (1%) 1,492 (1%) 4,635 (1%)
 Venothromboembolic disease (DVT/PE) 256 (1%) 1,182 (1%) 2,419 (1%) 7,163 (1%)
 Acute renal failure 3,412 (16%) 12,076 (15%) 25,888 (13%) 83,306 (14%)
 Surgical site infection 109 (1%) 475 (1%) 823 (0%) 2,533 (0%)
 Urinary tract infection/cystitis 1,737 (8%) 7,887 (10%) 18,317 (9%) 57,276 (10%)
DVT, deep vein thrombosis; PE, pulmonary embolism.

TABLE 4 - Multivariable Model for In-hospital Mortality
Model 1 aOR (95% CI) p Model 2 aOR (95% CI) p
SPAR 0.95 (0.94–0.97) <0.001 0.97 (0.95–0.98) <0.001
Age 1.04 (1.04–1.04) <0.001 1.04 (1.04–1.04) <0.001
Sex
 Female Reference Reference
 Male 1.10 (1.07–1.13) <0.001 1.05 (1.02–1.09) <0.001
Race/ethnicity
 White Reference Reference
 Asian or Pacific Islander 0.93 (0.87–1.00) 0.037 0.81 (0.76–0.87) <0.001
 Black 0.92 (0.87–0.97) 0.001 0.98 (0.92–1.03) 0.41
 Hispanic 0.85 (0.81–0.89) <0.001 0.83 (0.79–0.88) <0.001
 Native American 1.02 (0.83–1.26) 0.84 0.99 (0.80–1.24) 0.95
 Not reported 1.81 (1.60–2.05) <0.001 1.72 (1.51–1.96) <0.001
 Other 0.95 (0.87–1.04) 0.31 0.91 (0.83–0.99) 0.039
Expected primary payor
 Private Reference Reference
 Medicaid 1.28 (1.21–1.36) <0.001 1.41 (1.33–1.50) <0.001
 Medicare 1.04 (1.00–1.09) 0.073 1.10 (1.05–1.15) <0.001
 No charge 0.76 (0.50–1.15) 0.19 0.85 (0.56–1.29) 0.44
 Other 1.32 (1.18–1.48) <0.001 1.42 (1.26–1.60) <0.001
 Self pay 1.19 (1.06–1.33) 0.002 1.33 (1.19–1.49) <0.001
CCI 1.28 (1.27–1.28) <0.001 1.28 (1.28–1.29) <0.001
Census block group poverty 1.41 (1.16–1.71) <0.001 1.53 (1.28–1.82) <0.001
Teaching hospital
 No Reference Reference
 Yes 1.28 (1.18–1.40) <0.001 1.09 (1.00–1.19) 0.056
Advanced resource hospital
 No Reference Reference
 Yes 1.16 (1.09–1.22) <0.001 1.07 (1.01–1.13) 0.028
Diagnosis category
 Uncomplicated Reference
 Complex 9.41 (9.13–9.69) <0.001

F1
Figure 1:
Association of spatial access with in-hospital mortality by EGS condition. *Denotes significance with p < 0.05.
TABLE 5 - Multivariable Models for Major Morbidity
Model 1 OR (95% CI) p Model 2 OR (95% CI) p
SPAR 0.99 (0.99–1.00) <0.001 1.00 (0.99–1.00) 0.14
Age 1.03 (1.03–1.03) <0.001 1.03 (1.03–1.03) <0.001
Sex
 Female Reference Reference
 Male 0.92 (0.91–0.93) <0.001 0.90 (0.89–0.91) <0.001
Race/ethnicity
 White Reference Reference
 Asian or Pacific Islander 0.83 (0.81–0.86) <0.001 0.80 (0.78–0.82) <0.001
 Black 1.09 (1.07–1.11) <0.001 1.11 (1.09–1.13) <0.001
 Hispanic 0.86 (0.84–0.87) <0.001 0.85 (0.84–0.87) <0.001
 Native American 0.94 (0.87–1.01) 0.091 0.94 (0.87–1.01) 0.11
 Not reported 1.05 (0.99–1.12) 0.092 1.03 (0.97–1.09) 0.34
 Other 0.95 (0.92–0.98) 0.002 0.94 (0.91–0.97) <0.001
Expected primary payer
 Private Reference Reference
 Medicaid 1.35 (1.33–1.38) <0.001 1.39 (1.36–1.42) <0.001
 Medicare 1.31 (1.29–1.33) <0.001 1.34 (1.32–1.36) <0.001
 No charge 1.14 (1.04–1.25) 0.004 1.16 (1.06–1.27) 0.001
 Other 1.20 (1.16–1.25) <0.001 1.23 (1.18–1.28) <0.001
 Self pay 1.11 (1.07–1.15) <0.001 1.14 (1.10–1.17) <0.001
CCI 1.23 (1.22–1.23) <0.001 1.23 (1.22–1.23) <0.001
Census block group poverty 1.44 (1.34–1.56) <0.001 1.47 (1.37–1.59) <0.001
Teaching hospital
 No Reference Reference
 Yes 1.05 (0.98–1.12) 0.17 0.99 (0.93–1.06) 0.77
Advanced resource hospital
 No Reference Reference
 Yes 1.19 (1.15–1.24) <0.001 1.17 (1.12–1.21) <0.001
Diagnosis category
 Uncomplicated Reference
 Complex 2.60 (2.56–2.64)) <0.001
OR, odds ratio.

Subgroup Analysis

Separate analyses were conducted for (a) patients who underwent operating room intervention and (b) patients who underwent interhospital transfer from another acute care hospital (Supplementary Table 5, https://links.lww.com/TA/C785). The operative cohort was consisted of 215,152 total admissions. Effect of SPAR was similar to that of the total cohort: in-hospital mortality aOR of 0.96 (95% CI, 0.94–0.99; p < 0.01) and major morbidity aOR of 0.98 (95% CI, 0.97–0.99; p < 0.01). The interhospital transfer cohort included 29,860 admissions. There was no significant effect of SPAR on either in-hospital mortality or major morbidity.

Interaction of Spatial Access With Patient and Disease Characteristics

We conducted additional analyses to evaluate possible interaction between spatial access and patient and disease characteristics hypothesized to be differentially affected by access to care and did not observe any clear patterns of effect modification between SPAR and each of race/ethnicity, payor, or complex disease.

DISCUSSION

Disparities in spatial access to surgical services have been widely documented in the United States, yet the relationship between spatial access and clinical outcomes has not yet been characterized. In this study of 877,928 EGS patients in 12 states, we found that greater spatial access to hospitals with emergency surgical capability was associated with a small but significant reduction in mortality, even after adjusting for other relevant aspatial factors including insurance status and neighborhood poverty. Increasing spatial access from the lowest access category to that of the current national mean was associated with a 5% decrease in odds of mortality. The higher mortality in low-spatial access areas can be at least partially explained by the markedly increased proportion of complex EGS disease compared with areas with average or above spatial access.

To our knowledge, this is the first analysis of the relationship between spatial access and outcomes for EGS conditions using a comprehensive metric that captures both accessibility and availability of surgical care. Prior studies of spatial access to surgical care have largely focused on traumatic injury and evaluated only the accessibility component through use of time or distance to hospital.29–32 For trauma, this may be appropriate, both because the development of regional trauma systems have included triage algorithms that direct patients to hospitals with appropriate resources and capacity for their level of injury and because time to intervention for hemorrhagic shock plays an outsized role in averting early mortality in trauma. However, for other emergent disease states like EGS, for which no organized regional systems exist, the influence of hospital capacity relative to population demand becomes more important, supporting the use of comprehensive spatial metrics that take into account both distance to care and hospital capacity. While data on the effect of spatial access on EGS outcomes are sparse, a single-center analysis of EGS patients in Maryland by Diaz et al.33 found distance to their tertiary center to be associated with in-hospital mortality. Our analysis extends these findings using a comprehensive metric of spatial access and a large, multistate patient population; the results suggest that the effect of low-spatial access on mortality is not limited to trauma, and spatial access is relevant to the design of optimal health systems of nontrauma emergency surgical systems as well.

We were surprised by the markedly increased rate of complex EGS disease in patients with low spatial access to surgical care, with a nearly 2.5-fold greater incidence in the very-low access group compared with those with average or greater access. Complex disease on presentation generally represents greater disease progression and may be an indicator of delays in arriving at appropriate care. Studies of oncologic disease using gravity models of spatial access have shown a similar pattern, with poor access associated with increased rates of late-stage breast and colorectal cancer diagnosis in the United States.34–36 In these cases, poor spatial access to primary care physicians as well as aspatial components such as socioeconomic status and minority race/ethnicity were identified as factors underlying this relationship. Given the EGS conditions in our analysis that were most sensitive to the effect of spatial access on outcomes (i.e., cholecystitis, diverticulitis, hernias, and peptic ulcer disease), it is possible that similar forces are at play, with poor spatial access to health care in general contributing to delayed diagnoses and more severe disease at time of presentation to surgical care.

Access to care is a complex topic, which includes not only spatial access but also aspatial components including acceptability (e.g., cultural appropriateness), affordability (e.g., insurance coverage), and accommodation (e.g., service organization, hours of operation). Interactions between spatial and aspatial components of access must also be considered in the development and treatment of EGS conditions, with the individual effects difficult to separate. Minoritized communities, those with high rates of uninsured residents, and neighborhoods with high social vulnerability have been shown to have poor spatial access to both primary care and specialized hospital services including surgical and intensive care.8,12,37 High neighborhood social vulnerability has also been linked to both low primary care utilization and increased risk of presentation with emergent versus elective general surgery disease, greater severity of emergency surgical disease at time of presentation, and worse perioperative outcomes.37–42 Both lack of primary care availability (spatial), and affordability, language and cultural barriers (aspatial) might lead to delays in seeking care for a condition that could initially be treated electively or early in their course but instead progress to more severe or complex emergent surgical conditions. Further research investigating mechanisms underlying delays in diagnosis, development of complex EGS disease, and barriers to realizing care will be critical to reducing burden of mortality and morbidity associated with EGS disease. Together, research suggests that social and structural elements are both likely at play across the spectrum of primary care to emergency surgical care.

Regionalization of care through the development of organized and coordinated systems has been adopted as a solution for other time-sensitive conditions with large disparities in spatial access to care, such as trauma, stroke, and acute myocardial infarctions.43 While our analysis showed that better spatial access to surgical care was associated with decreased mortality for common EGS conditions, the effect size was fairly small, which brings to question whether regionalized systems should be considered for EGS as well. The purpose of regionalization is to match patients with the appropriate clinical resources through policy that directs (a) the distribution of physicians, (b) the distribution of equipment and facilities, and (c) the control of patient movement within the system.44,45 While our data suggest that improving spatial access to surgical facilities and resources alone might result in a small improvement patient outcomes, it is possible that developing systems to improve the movement of patients through the health care system is a more important lever. This extends from primary care through arrival at urgent or emergent care, potential interhospital transfer, and, finally, through definitive hospital care. The overall absence of organized regional systems for EGS care may be one reason why mortality for patients with complex EGS disease in our analysis was similar regardless of potential spatial access to care. Once patients present with complex disease, effective systems are needed to consistently and rapidly direct them to high-quality hospitals with appropriate resources in a timely manner, and these mechanisms are currently lacking across much of the United States. Regionalization offers the potential of addressing disparities in spatial access for many populations (e.g., minoritized or uninsured communities) through identifying locations for additional care, and it would additionally offer the benefits of triage and transfer protocols, benchmarked data registries, and quality improvement programs.43 Further work on the potential benefits of regionalization for EGS should consider the influence of these components on clinical outcomes.

Several limitations should be considered when interpreting the results of this study. As with all studies using administrative data, the results are subject to residual confounding because of unmeasured factors such as unmeasured comorbidities, sociodemographic factors, and severity of disease. We did attempt to control for severe disease by including the presence of “complex disease” in our model; however, this remains a proxy measure and does not fully account for the broad range of physiologic derangement seen in even complex EGS patients. Second, patient severity of disease is further subject to misclassification, as our identification of complex disease was based on ICD-9 codes. Third, SPAR is a measure of potential spatial access and does not account for the myriad of real-world impendences that may occur in realizing access to care, including interrupted or lack of transportation and patient choice of hospital. The construction of the SPAR metric, weighting supply-demand ratios by impedance measures further makes it difficult to separate the effect of distance to care from hospital capacity from any specific region. Fourth, we could not assess spatial access to other forms of healthcare, including primary care availability, which may influence time to diagnosis and arrival at surgical care. Finally, our patient cohort was derived from 12 inpatient state databases, and will not capture patients who are discharged directly home from the postoperative recovery room or who have short duration stays coded as “observation” status, and may not be generalizable to the entire United States given differences in policy and geographic differences in other contexts.

CONCLUSION

In a large cohort of adult patients admitted for one of eight common EGS conditions, greater spatial access to care was associated with a small but significant decrease in in-hospital mortality, even after accounting for other forces influencing access to care including insurance status and neighborhood poverty. Low spatial access was associated with markedly increased proportion of complex EGS disease compared with patients with spatial access at or above the national mean. These data emphasize the importance of health system design, including geographic placement and relationship of hospital capacity to population need, to EGS outcomes. Further research should seek to investigate how spatial and aspatial factors influence severity of EGS disease at diagnosis, delays in reaching surgical care, and implications for development of effective and equitable systems for emergency surgical care.

AUTHORSHIP

All authors contributed to the study design, data interpretation, and critical revisions of the manuscript. M.L.M. conducted the literature review and manuscript writing. C.M.A., J.H., and N.W. performed data analysis with contribution by MLM.

ACKNOWLEDGMENT

M.L.M. received support from the National Institute on Minority Health and Health Disparities of the National Institutes of Health (R21MD012657). This investigation was supported by the University of Utah Study Design and Biostatistics Center, with funding in part from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through grant UL1TR002538 (formerly 5UL1TR001067-05, 8UL1TR000105, and UL1RR025764).

DISCLOSURE

The authors declare no conflicts of interest.

REFERENCES

1. Gale SC, Shafi S, Dombrovskiy VY, Arumugam D, Crystal JS. The public health burden of emergency general surgery in the United States: a 10-year analysis of the Nationwide Inpatient Sample—2001 to 2010. J Trauma Acute Care Surg. 2014;77(2):202–208.
2. Representatives of the Participating Organizations in Congress. Acute care congress on the future of emergency surgical care in the United States. J Trauma. 2009;67(1):1–7.
3. Hoyt D. Looking forward — March 2017. Bull Am Coll Surg. 2017;102(3):8–9.
4. Institute of Medicine. Hospital-Based Emergency Care: At the Breaking Point. Washington, DC: The National Academies Press; 2007.
5. Penchansky R, Thomas JW. The concept of access: definition and relationship to consumer satisfaction. Med Care. 1981;19(2):127–140.
6. Drake C, Nagy D, Nguyen T, Kraemer KL, Mair C, Wallace D, et al. A comparison of methods for measuring spatial access to health care. Health Serv Res. 2021;56(5):777–787.
7. Luo W, Qi Y. An enhanced two-step floating catchment area (E2SFCA) method for measuring spatial accessibility to primary care physicians. Health Place. 2009;15(4):1100–1107.
8. Diaz A, Schoenbrunner A, Pawlik TM. Trends in the geospatial distribution of inpatient adult surgical services across the United States. Ann Surg. 2021;273(1):121–127.
9. Khubchandani JA, Shen C, Ayturk D, Kiefe CI, Santry HP. Disparities in access to emergency general surgery care in the United States. Surgery. 2018;163(2):243–250.
10. McCrum ML, Wan N, Lizotte SL, Han J, Varghese T, Nirula R. Use of the spatial access ratio to measure geospatial access to emergency general surgery services in California. J Trauma Acute Care Surg. 2021;90(5):853–860.
11. Hsia R, Shen Y-C. Possible geographical barriers to trauma center access for vulnerable patients in the United States: an analysis of urban and rural communities. Arch Surg. 2011;146(1):46–52.
12. Eberth JM, Hung P, Benavidez GA, Probst JC, Zahnd WE, McNatt M-K, et al. The problem of the color line: spatial access to hospital services for minoritized racial and ethnic groups. Health Aff. 2022;41(2):237–246.
13. Agency for Healthcare Research and Quality (AHRQ). HCUP-US SID Overview. Available at: https://www.hcup-us.ahrq.gov/sidoverview.jsp. Accessed July 31, 2022.
14. Scott JW, Staudenmayer K, Sangji N, Fan Z, Hemmila M, Utter G. Evaluating the association between American Association for the Surgery of Trauma emergency general surgery anatomic severity grades and clinical outcomes using national claims data. J Trauma Acute Care Surg. 2021;90(2):296–304.
15. Agency for Healthcare Research and Quality (AHRQ). American Hospital Association Linkage Files. Available at: https://www.hcup-us.ahrq.gov/db/state/ahalinkage/aha_linkage.jsp. Accessed July 31, 2022.
16. Wan N, McCrum M, Han J, Lizotte S, Su D, Wen M, et al. Measuring spatial access to emergency general surgery services: does the method matter?Health Services and Outcomes Research Methodology. 2022;22:79–95.
17. McCrum ML, Wan N, Han J, Lizotte SL, Horns JJ. Disparities in spatial access to emergency surgical services in the US. JAMA Health Forum. 2022;3(10):e223633.
18. Wan N, Zou B, Sternberg T. A three-step floating catchment area method for analyzing spatial access to health services. Int J Geogr Inf Sci. 2012;26(6):1073–1089.
19. Guagliardo MF. Spatial accessibility of primary care: concepts, methods and challenges. Int J Health Geogr. 2004;3(1):3.
20. McGrail MR. Spatial accessibility of primary health care utilising the two step floating catchment area method: an assessment of recent improvements. Int J Health Geogr. 2012;11:50.
21. Naylor KB, Tootoo J, Yakusheva O, Shipman SA, Bynum JPW, Davis MA. Geographic variation in spatial accessibility of U.S. Healthcare providers. PLoS One. 2019;14(4):e0215016.
22. Branas CC, MacKenzie EJ, Williams JC, Schwab CW, Teter HM, Flanigan MC, et al. Access to trauma centers in the United States. JAMA. 2005;293(21):2626–2633.
23. Wei R, Clay Mann N, Dai M, Hsia RY. Injury-based geographic access to trauma centers. Acad Emerg Med. 2019;26(2):192–204.
24. Carr BG, Bowman AJ, Wolff CS, Mullen MT, Holena DN, Branas CC, Wiebe DJ. Disparities in access to trauma care in the United States: a population-based analysis. Injury. 2017;48(2):332–338.
25. Gomez D, Haas B, Doumouras AG, Zagorski B, Ray J, Rubenfeld G, McLellan BA, Boyes DM, Nathens AB. A population-based analysis of the discrepancy between potential and realized access to trauma center care. Ann Surg. 2013;257(1):160–165.
26. Ingraham AM, Xiong W, Hemmila MR, Shafi S, Goble S, Neal ML, Nathens AB. The attributable mortality and length of stay of trauma-related complications: a matched cohort study. Ann Surg. 2010;252(2):358–362.
27. Shah AA, Haider AH, Zogg CK, Schwartz DA, Haut ER, Zafar SN, Schneider EB, Velopulos CG, Shafi S, Zafar H, et al. National estimates of predictors of outcomes for emergency general surgery. J Trauma Acute Care Surg. 2015;78(3):482–490.
28. R Core Team. R: A language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computer; 2022. Available at: https://www.R-project.org/. Accessed July 29, 2022.
29. Tung EL, Hampton DA, Kolak M, Rogers SO, Yang JP, Peek ME. Race/ethnicity and geographic access to urban trauma care. JAMA Netw Open. 2019;2(3):e190138.
30. Gauss T, Ageron FX, Devaud ML, Debaty G, Travers S, Garrigue D, et al. Association of prehospital time to in-hospital trauma mortality in a physician-staffed emergency medicine system. JAMA Surg. 2019;154(12):1117–1124.
31. Jarman MP, Curriero FC, Haut ER, Pollack Porter K, Castillo RC. Associations of distance to trauma care, community income, and neighborhood median age with rates of injury mortality. JAMA Surg. 2018;153(6):535–543.
32. Crandall M, Sharp D, Unger E, Straus D, Brasel K, Hsia R, et al. Trauma deserts: distance from a trauma center, transport times, and mortality from gunshot wounds in Chicago. Am J Public Health. 2013;103(6):1103–1109.
33. Diaz JJ Jr., Norris P, Gunter O, Collier B, Riordan W, Morris JA Jr. Triaging to a regional acute care surgery center: distance is critical. J Trauma. 2011;70(1):116–119.
34. Wan N, Zhan FB, Zou B, Wilson JG. Spatial access to health care services and disparities in colorectal cancer stage at diagnosis in Texas. Prof Geogr. 2013;65(3):527–541.
35. Freeman VL, Naylor KB, Boylan EE, Booth BJ, Pugach O, Barrett RE, et al. Spatial access to primary care providers and colorectal cancer-specific survival in Cook County, Illinois. Cancer Med. 2020;9(9):3211–3223.
36. Wang F, McLafferty S, Escamilla V, Luo L. Late-stage breast cancer diagnosis and health care access in Illinois. Prof Geogr. 2008;60(1):54–69.
37. Al Rifai M, Jain V, Khan SU, Bk A, Mahar JH, Krittanawong C, et al. State-level social vulnerability index and healthcare access: the behavioral risk factor surveillance system survey. Am J Prev Med. 2022;63(3):403–409.
38. Diaz A, Barmash E, Azap R, Paredes AZ, Hyer JM, Pawlik TM. Association of county-level social vulnerability with elective versus non-elective colorectal surgery. J Gastrointest Surg. 2021;25(3):786–794.
39. Zhang Y, Kunnath N, Dimick JB, Scott JW, Diaz A, Ibrahim AM. Social vulnerability and outcomes for access-sensitive surgical conditions among Medicare beneficiaries. Health Aff (Millwood). 2022;41(5):671–679.
40. Diaz A, Hyer JM, Barmash E, Azap R, Paredes AZ, Pawlik TM. County-level social vulnerability is associated with worse surgical outcomes especially among minority patients. Ann Surg. 2021;274(6):881–891.
41. Carmichael H, Moore A, Steward L, Velopulos CG. Using the social vulnerability index to examine local disparities in emergent and elective cholecystectomy. J Surg Res. 2019;243:160–164.
42. Cain BT, Horns JJ, Huang LC, McCrum ML. Socioeconomic disadvantage is associated with greater mortality after high-risk emergency general surgery. J Trauma Acute Care Surg. 2022;92(4):691–700.
43. Walton NT, Mohr NM. Concept review of regionalized systems of acute care: is regionalization the next frontier in sepsis care?J Am Coll Emerg Physicians Open. 2022;3(1):e12631.
44. United States Health Resources Administration, Ginzberg E, ed. Regionalization & Health Policy. U.S. Department of Health, Education, and Welfare, Public Health Service, Health Resources Administration; 1977.
45. Institute of Medicine (IOM). Regionalizing Emergency Care: Workshop Summary. Washington, DC: National Academies Press; 2010.
Keywords:

Emergency general surgery; spatial access; geospatial methods

Supplemental Digital Content

Copyright © 2022 American Association for the Surgery of Trauma.