Association of Race/Ethnicity with Emergency Department Wait Times
http://www.100md.com
《小儿科》
the Division of Emergency Medicine, Children's Hospital Boston, and Department of Pediatrics, Harvard Medical School, Boston, Massachusetts
ABSTRACT
Objective. To determine whether wait times for children treated in emergency departments (EDs) nationally are associated with patient race/ethnicity.
Methods. Data were obtained from the National Hospital Ambulatory Medical Care Survey, which collects information on patient visits to EDs throughout the United States. We examined data for patients 15 years of age who presented to EDs during the 4-year period of 1997–2000. Sample weights were applied to the identified patient records to yield national estimates. For the purposes of this study, race/ethnicity was analyzed for 3 major groups, ie, non-Hispanic white (NHW), non-Hispanic black (NHB), and Hispanic white (HW).
Results. During the 4-year study period, 20633 patient visits were surveyed, representing a national sample of 92.9 million children 15 years of age. The race/ethnicity distribution included 9019 NHW children (59.5%), 3910 NHB children (23.9%), and 2991 HW children (16.6%). The wait time for all groups was 43.6 ± 1.7 minutes (mean ± SEM). There were significant unadjusted intergroup differences in wait times (38.5 ± 1.6 minutes, 48.7 ± 0.5 minutes, and 54.5 ± 0.1 minutes for NHW, NHB, and HW children, respectively). Visit immediacy (triage status), when reported, was categorized as <15 minutes for 2203 children (17.1%), 15 to 60 minutes for 5324 (41.4%), 1 to 2 hours for 3010 (25.1%), and >2 to 24 hours for 1910 (16.4%). There were significant unadjusted differences in triage status according to race, with 14.6% of NHW patients being placed in the >2-hour immediacy range, compared with 18.8% of NHB patients and 20.0% of HW patients. In a linear regression analysis with logarithmically transformed wait time as a dependent variable and with adjustment for potential confounders, including hospital location, geographic region, and payer status, both NHB and HW patients waited longer than NHW patients, although the results were statistically significant only for HW patients.
Conclusions. These nationally representative data suggest that children who come to EDs have wait times that vary according to race/ethnicity. There are several potential explanations for this observation, including discrimination, cultural incompetence, language barriers, and other social factors. These data and similar data from the National Hospital Ambulatory Medical Care Survey are useful in identifying nonclinical influences on the delivery of pediatric emergency care.
Key Words: race/ethnicity emergency department wait time
Abbreviations: NHAMCS, National Hospital Ambulatory Medical Care Survey NHW, non-Hispanic white NHB, non-Hispanic black HW, Hispanic white ED, emergency department CI, confidence interval
Emergency department (ED) overcrowding has become a national crisis that parallels increasing ED volume. The 2001 National Hospital Ambulatory Medical Care Survey (NHAMCS) reported that the annual number of ED visits in the United States increased 20% from 1992 to 2001, to 107.5 million visits per year.1 This includes 22 million visits per year for children 15 years of age (20.7% of total annual ED visits). The causes of overcrowding are multifactorial and include not only increased volume but also an aging population with increased health care needs, increased complexity of the conditions of patients presenting to the ED, many of whom require hospital and/or intensive care unit admission, and the relative lack of inpatient and intensive care unit beds. Other factors include shortages of nurses and other clinical personnel, increased demands for ancillary services, decreased numbers of EDs in the United States, and saturation of the primary care network.2–6 The mean wait time for ED patients has also increased, with a mean of 23.9 minutes for visits classified as emergency (immediacy of <15 minutes), 38.1 minutes for urgent visits (immediacy of 15–60 minutes), 56.5 minutes for semiurgent visits (immediacy of 1–2 hours), and 67.7 minutes for nonurgent visits (immediacy of 2–24 hours).7 Racial and ethnic differences in ED use have also been reported. The NHAMCS demonstrated that ED utilization by blacks of all ages was 76% higher than that by whites; whereas white children had 34.7 ED visits per 100 persons per year, black children had 56.5 visits per 100 persons per year.1 However, the percentages of visits classified as semiurgent or nonurgent were reportedly comparable for black children and white children (28.1% and 26.3%, respectively).
Longer wait times in the ED contribute to patient and family dissatisfaction with the care received. Conversely, patients who perceive that the wait time to see a provider was shorter than expected are more satisfied with their ED visits than are patients whose wait was longer than expected.8 Clinically, prolonged ED wait times may result in protracted pain and suffering and in delays in diagnosis and treatment.2,6 Overcrowding in EDs may also place patients at greater risk for medical errors. In 2004, the Joint Commission on Accreditation of Healthcare Organizations developed a standard regarding ED overcrowding that emphasized that patients in overcrowded EDs are at high risk of experiencing treatment delays or inadequate care.3 Similarly, the American Academy of Pediatrics released a policy statement in 2004 that discussed the ways in which ED overcrowding threatens access to care for America's children.4 The purpose of this study was to determine whether wait times for children treated in EDs throughout the nation are associated with race or ethnicity, with adjustment for social, demographic, and clinical factors.
METHODS
The NHAMCS is a national probability-sample survey of visits to EDs and outpatient departments in the United States. It is conducted by the National Center for Health Statistics, a division of the Centers for Disease Control and Prevention. Findings are based on a national sample of visits to the EDs and outpatient departments of noninstitutional general (medical, surgical, and children's) and short-stay hospitals, exclusive of federal, military, and Veterans Affairs hospitals, located in the 50 States and the District of Columbia. The survey uses a 4-stage probability design involving samples of geographically defined areas, hospitals within those areas, clinics within the hospitals, and patient visits within the clinics. EDs and outpatient clinics are sampled separately. Annual data collection began in 1992. Hospital staff members are instructed by the US Bureau of the Census to complete patient record forms for a systematic random sample of patient visits during a randomly assigned, 4-week reporting period. Hospital staff members complete the patient record forms in most cases, but NHAMCS field representatives complete the forms from medical records if necessary. Data are obtained regarding demographic characteristics of the patients, expected sources of payment, chief complaints, physicians' diagnoses, diagnostic/screening services, procedures, medication therapy, disposition, types of health care professionals consulted, causes of injury (where applicable), and certain characteristics of the hospital, such as type of ownership.
Patient race and ethnicity recorded in the survey are determined by the study personnel when they complete the patient record forms. Unless it is hospital policy to ask patients directly for this information, the personnel are instructed to record race/ethnicity on the basis of their observations, marking the box that is, in their opinion, the most appropriate. Race categories include white, black or African American, Asian, Native Hawaiian or other Pacific Islander, and American Indian or Alaska Native. Ethnicity is categorized as Hispanic or non-Hispanic and is recorded separately from race (1997 standards of the Office of Management and Budget).9 The 2001 visit rates for age and race are based on Census 2000 population data, and the 1997–2000 visit rates were calculated with 1990 census data.
NHAMCS data for 1992–2002 are available for public use. We examined data for patients 15 years of age who presented to EDs between 1997 and 2000, a period during which wait time was a recorded variable. Wait time was determined by the NHAMCS from the time the patient arrived in the ED and the time the patient was examined by a physician, both of which were recorded on the patient record form. For the purposes of our study, race/ethnicity was analyzed according to 3 categories, ie, non-Hispanic white (NHW), non-Hispanic black (NHB), or Hispanic white (HW). Other groups had limited representation and were therefore not included. Visit immediacy (triage status) was assigned by triage personnel at the time of arrival at the ED and was categorized into 4 levels, ie, emergency (<15 minutes), urgent (15–60 minutes), semiurgent (1–2 hours), or nonurgent (2–24 hours). The NHAMCS provides guidelines to assist hospital personnel in deciding how their institution's triage categories correspond to these levels. Payer status for patients was categorized as private insurance, government source of payment (Medicare, Medicaid, or Worker's Compensation), self-pay, other (no charge or other), or unknown.
Data were analyzed with SAS version 10.1 software (SAS Institute, Cary, NC). SUDAAN statistical software (Research Triangle Institute, Research Triangle Park, NC) was used to account for the sampling technique used in NHAMCS, which includes oversampling of certain groups. Sample weights were applied to the patient records identified, to yield national estimates. The wait times were positively skewed and were therefore normalized with a natural logarithm transformation before analysis. The 2 test was used to measure the association between race/ethnicity and triage status, and Student's t test of logarithmically transformed data was used for intergroup comparisons of wait times, because nonparametric testing cannot be performed with SUDAAN software. Linear regression analysis using logarithmically transformed wait time as a dependent variable and adjusting for potential confounders, including hospital location, geographic region, and payer status, was performed to examine adjusted wait times.
RESULTS
For the 4-year study period, 20633 patient visits were surveyed, representing a national sample of 92.9 million children 15 years of age between 1997 and 2000. The 3 race/ethnicity categories of interest included a total of 15920 children. The study sample consisted of 9019 (59.5%) NHW children, 3910 (23.9%) NHB children, and 2991 (16.6%) HW children. The mean ages for the 3 groups were 6.4 years for NHW, 5.6 years for HW, and 4.6 years for NHB. The mean (±SD) wait time for all race/ethnicity groups was 43.6 ± 1.7 minutes (Table 1). There were significant unadjusted differences in the wait times among the groups, with HW children waiting 54.5 ± 0.1 minutes, compared with 48.7 ± 0.5 minutes and 38.5 ± 1.6 minutes for NHB and NHW children, respectively (P < .01 for NHW patients, compared with the other groups).
The wait times for the 3 race/ethnicity groups were examined also on the basis of the location of the hospital. Hospital locations were designated as urban or nonurban and as Northeastern, Southern, Midwestern, or Western. Among children who presented to urban hospitals, NHW children waited a significantly shorter time than did HW children (P < .01) (Table 2). The same difference was found in each of the 4 regions (P = .01 for the Western region and P < .01 for all other regions). In addition, Midwestern NHW children waited significantly shorter times than did NHB children (P < .01), and Southern NHB children had shorter waits than did HW children (P = .03).
The triage status distribution was <15 minutes for 2203 children, 15 to 60 minutes for 5324, >1 to 2 hours for 3010, and >2 to 24 hours for 1910 (Table 2). The average ages were very similar among the triage groups, ranging from 5.5 to 5.8 years. There were significant differences in triage assignments among the race/ethnicity groups, with NHB and HW children being triaged to less urgent levels at a significantly higher rate than NHW patients (P = .025). A triage status of >2 hours was assigned to 14.6% of NHW patients, compared with 18.8% of NHB patients and 20.0% of HW patients (P = .04 and .07 for NHW compared with NHB and HW, respectively).
Patient wait times were examined in terms of patient demographic features (gender and race/ethnicity), hospital characteristics (geographic region, metropolitan status, and hospital ownership), source of payment, and characteristics of the ED visit (triage status and disposition) (Table 3). In the univariate analyses, factors significantly associated with longer wait times were NHB (change in wait time: 23%; 95% confidence interval [CI]: 9–38%) and HW (change in wait time: 45%; 95% CI: 29–64%) race/ethnicity, all nonprivate sources of payment (government source: change in wait time: 10%; 95% CI: 1–21%; self-pay: change in wait time: 12%; 95% CI: 0–27%; other sources: change in wait time: 29%; 95% CI: 11–50%), and all triage groups other than the <15-minute group (P < .01 for overall variable). Factors associated with shorter wait times included hospital admission (change in wait time: –31%; 95% CI: –16 to –43%), hospital location in the Midwest (change in wait time: –32%; 95% CI: –17 to –43%), and nonurban setting (change in wait time: –49%; 95% CI: –40 to –56%).
In a multivariate linear regression model, adjustments were made for effects of patient gender, race/ethnicity, hospital location and ownership, payment type, triage status, and disposition of the visit. The race/ethnicity of the patient remained a significant determinant of wait time (P = .01 for overall variable) (Table 4). HW children waited 18% longer than did NHW children (95% CI: 6-30%), whereas NHB patients waited 6% longer than did NHW patients, although this difference was not statistically significant. Additional factors predicting longer wait times were a payment source classified as other (change in wait time: 24%; 95% CI: 6-44%) and any triage assignment less acute than the most urgent group (P < .01 for overall variable). Shorter wait times were associated with patients requiring admission (change in wait time: –20%; 95% CI: –6 to –32%), Midwestern location (change in wait time: –21%; 95% CI: –9 to –32%), nonurban setting (change in wait time: –41%; 95% CI: –35 to –47%), and government ownership of the hospital (change in wait time: –22%; 95% CI: –9 to 34%).
DISCUSSION
The race/ethnicity of children treated in EDs in the United States appears to be significantly associated with wait time. Although the mean wait time for children in all groups was 43.6 minutes, both HW and NHB patients waited longer, with wait times of 54.5 and 48.7 minutes, respectively, although the results were statistically significant only for HW patients. These findings may be linked to patient-, provider-, or system-related variables. Patient-related variables potentially include language spoken, socioeconomic status, type of insurance, geographic location, level of literacy, and cultural values. Provider-related variables include bias, prejudice, and stereotyping, which might play a role in triage decisions. System-related variables include availability of primary care services, lack of available interpreter services, and ED volume.
The finding that HW and NHB patients wait longer than NHW patients may reflect the tendency for minority patients to use the ED for less urgent problems, perhaps because of real or perceived barriers to primary care. Several studies have found that lesser continuity of care is associated with greater risk of ED utilization10 and that children who are members of underrepresented minority groups, are poor, or are uninsured are at greater risk than other children of experiencing obstacles to health care.11 The fact that higher percentages of HW and NHB children fell into the lowest-acuity group supports this theory. However, other studies have suggested that, after adjustment for other factors, including socioeconomic status, regular source of primary care, and insurance status, there are no significant differences in ED use among blacks, whites, and Hispanics.12 In 2002, 38.8% of US children 0 to 18 of age who were Medicaid recipients were NHW, compared with 25.2% who were NHB and 28.3% who were HW. Similarly, 40.5% of uninsured children in the United States are NHW, 18.0% are NHB, and 35.2% are HW.13
Another explanation for wait time disparities could be that patients from underrepresented minority groups are more likely to be treated in EDs that have longer wait times for all patients, such as those located in densely populated, urban areas. Several studies have found that hospitals in more densely populated areas have higher rates of overcrowding than do hospitals serving smaller populations.6,14 We found that, in general, hospitals in urban areas had longer wait times than did those in nonurban areas. Even with adjustment for hospital location, however, race/ethnicity remained an important predictor of wait time in the ED.
Although triage status was found to be an important independent factor in wait time, minority children had longer wait times in the ED even when triage status was controlled. This could result from the fact that triage classification is not standard among institutions. Studies have shown that great variability in triage practice occurs among nurses, physicians, and software programs, even in the same ED.15,16 This may indicate that triage decisions are influenced by factors other than standardized objective assessments of the patient's medical condition. Subjective factors could result in patients being given either greater or lesser immediacy than other patients with the same presentation. Assigning a patient an inappropriately low triage level could lead to a delay in care for that patient. Conversely, assigning an inappropriately high triage level to a less severely ill patient might lead to delays in care for other, possibly sicker patients. Our data suggest that, if there is a bias in the triage process, then patients from underrepresented minorities could be given lower acuity and consequently wait longer than nonminority patients with the same symptoms. When providers are forced to make triage decisions quickly and without complete information, providers' own (sometimes unconscious) beliefs likely enter into triage assignments. Additional investigation of the effect of race/ethnicity on triage decision-making is needed. In addition, a recent study showed that pediatric emergency medicine and general emergency medicine attending physicians may assign different triage levels to the same patients, with general emergency medicine physicians assigning higher levels to febrile children than do their pediatric emergency medicine counterparts.17 Because we could not determine from the NHAMCS data whether children were treated in general or pediatric EDs, it is unclear what effect this factor would have on the results of our study; few children's hospitals participate in this survey.
Our study found that, among children who presented to the EDs of urban hospitals in all regions of the United States, both NHB and NHW children were treated by physicians more quickly than were HW patients. This finding may reflect language barriers between the patients and ED providers. One study of language barriers and resource utilization in pediatric EDs found that non–English-speaking patients had a mean total length of stay in the ED that was 28 minutes longer than that of English-speaking patients.18 The absolute wait times were not reported in that study. The same study found that the overall charge for diagnostic tests was higher when there was a language barrier between the patient and the ED physician. In our investigation, the proportions of higher-acuity cases were similar in the 2 groups, which suggests that the level of acuity did not account for the differences between the 2 groups. Wait times for non—English-speaking patients may be longer because the provider must wait for a translator before evaluating the patient. It is also possible that non—English-speaking patients or parents may be less persistent in insisting that the child be evaluated sooner because they cannot communicate with the triage nurse or may be less likely to complain to ED staff members about the wait because of cultural values. The primary language spoken was not recorded on the NHAMCS data sheets; therefore, we were unable to analyze the effect of a provider-patient language barrier on patient wait times.
A retrospective study comparing ED analgesic use for adult patients with isolated long-bone fractures revealed that HW patients were twice as likely as NHW patients to receive no pain medications. Only 58% of the HW patients in the study spoke English, as opposed to 97% of the NHW patients. With controlling for all covariates, Hispanic ethnicity was the strongest predictor of withheld analgesics. Controlling for primary language spoken or insurance status did not eliminate the differential use of analgesics.19 In a similar study of adult patients, only 57% of NHB patients received analgesia for isolated long-bone fractures, compared with 74% of NHW patients with similar fractures.20 A follow-up study of adult patients did not reveal any differences in physician pain assessments for HW patients versus NHW patients.21
The NHAMCS provides a wealth of data on ED visits in the United States but has several limitations. One is that race and ethnicity are sometimes recorded on the basis of the perceptions of hospital personnel, rather than patient self-reports. Several studies of Department of Veterans Affairs data showed only fair agreement between administrative data and self-reported race/ethnicity.22,23 Patients whose race was other than white were most likely to be misclassified. If patients who identify themselves as HW or NHB are classified as NHW by hospital personnel, then the true differences in wait times between groups might be smaller than our data revealed.
Another limitation is that the wait times derived from the patient record forms vary according to how each hospital records its data. For example, patients in some EDs are registered in the treatment area, sometimes even after being examined by a provider; patients in other EDs register even before they undergo triage. This might cause wait times to vary among facilities, which might affect our results.
Another limitation is that the NHAMCS does not record any measure of socioeconomic status except for type of insurance. Therefore, although we controlled for payer status, we were not able to control fully for socioeconomic status. The NHAMCS does not record whether patients were referred to the ED by their primary care physicians. Patients who are referred to the ED by their primary care physicians might be evaluated with higher acuity status than patients who self-refer to the ED; if more HW and NHB patients have more difficulty in obtaining access to primary care, then this might also result in longer ED wait times.
CONCLUSIONS
These data suggest that children who are treated in EDs in the United States have wait times that vary according to race/ethnicity. There are several potential explanations for this observation, including language barriers, availability of primary care, differences in triage practices, payer status, and hospital characteristics. Additional study is urgently needed, to determine root causes and to identify solutions addressing this disparity in the delivery of pediatric emergency care.
FOOTNOTES
Accepted Oct 19, 2004.
No conflict of interest declared.
REFERENCES
McCaig LF, Burt CW. National Hospital Ambulatory Medical Care Survey: 2001 emergency department summary. Adv Data. 2003; (335)1-25
Derlet RW, Richards JR. Overcrowding in the nation's emergency departments: complex causes and disturbing effects. Ann Emerg Med. 2000;35 :63 –68
American Academy of Pediatrics, Committee on Pediatric Emergency Medicine. Overcrowding crisis in our nation's emergency departments: is our safety net unraveling Pediatrics. 2004;114 :878 –888
Richardson LD, Hwang U. Access to care: a review of the emergency department literature. Acad Emerg Med. 2001;8 :1030 –1036
Derlet RW, Richards JR, Kravitz RL. Frequent overcrowding in U.S. emergency departments. Acad Emerg Med. 2001;8 :151 –155
McCaig LF, Ly N. National Hospital Ambulatory Medical Care Survey: 2000 emergency department summary. Adv Data. 2002;(326) :1 –32
Thompson DA, Yarnold PR, Williams DR, Adams SL. Effects of actual waiting time, perceived waiting time, information delivery, and expressive quality on patient satisfaction in the emergency department. Ann Emerg Med. 1996;28 :657 –665
Christakis DA, Mell L, Koepsell TD, Zimmerman FJ, Connell FA. Association of lower continuity of care with greater risk of emergency department use and hospitalization in children. Pediatrics. 2001;103 :524 –529
Newacheck PW, Hughes DC, Stoddard JJ. Children's access to primary care: differences by race, income, and insurance status. Pediatrics. 1996;97 :26 –32
Baker DW, Stevens CD, Brook RH. Determinants of emergency department use: are race and ethnicity important Ann Emerg Med. 1996;28 :677 –682
Brillman JC, Doezema D, Tandberg D, et al. Triage: limitations in predicting need for emergent care and hospital admission. Ann Emerg Med. 1996;27 :493 –500
Wuerz R, Fernandes CMB, Alarcon J. Inconsistency of emergency department triage. Ann Emerg Med. 1998;32 :431 –435
Maldonado T, Avner JR. Triage of the pediatric patient in the emergency department: are we all in agreement Pediatrics. 2004;114 :356 –360
Hampers LC, Cha S, Gutglass DJ, Binns HJ, Krug SE. Language barriers and resource utilization in a pediatric emergency department. Pediatrics. 1999;103 :1253 –1256
Todd KH, Samaroo N, Hoffman JR. Ethnicity as a risk factor for inadequate emergency department analgesia. JAMA. 1993;269 :1537 –1539
Todd KH, Deaton C, D'Adamo AP, Goe L. Ethnicity and analgesic practice. Ann Emerg Med. 2000;35 :11 –16
Todd KH, Lee T, Hoffman JR. The effect of ethnicity on physician estimates of pain severity in patients with isolated extremity trauma. JAMA. 1994;271 :925 –928
Kressin NR, Chang B, Hendricks A, Kazis LE. Agreement between administrative data and patients' self-reports of race/ethnicity. Am J Public Health. 2003;93 :1734 –1739
Boehmer U, Kressin NR, Berlowitz DR, et al. Self-reported vs. administrative race/ethnicity data and study results. Am J Public Health. 2002;92 :1471 –1473(Catherine A. James, MD, F)
ABSTRACT
Objective. To determine whether wait times for children treated in emergency departments (EDs) nationally are associated with patient race/ethnicity.
Methods. Data were obtained from the National Hospital Ambulatory Medical Care Survey, which collects information on patient visits to EDs throughout the United States. We examined data for patients 15 years of age who presented to EDs during the 4-year period of 1997–2000. Sample weights were applied to the identified patient records to yield national estimates. For the purposes of this study, race/ethnicity was analyzed for 3 major groups, ie, non-Hispanic white (NHW), non-Hispanic black (NHB), and Hispanic white (HW).
Results. During the 4-year study period, 20633 patient visits were surveyed, representing a national sample of 92.9 million children 15 years of age. The race/ethnicity distribution included 9019 NHW children (59.5%), 3910 NHB children (23.9%), and 2991 HW children (16.6%). The wait time for all groups was 43.6 ± 1.7 minutes (mean ± SEM). There were significant unadjusted intergroup differences in wait times (38.5 ± 1.6 minutes, 48.7 ± 0.5 minutes, and 54.5 ± 0.1 minutes for NHW, NHB, and HW children, respectively). Visit immediacy (triage status), when reported, was categorized as <15 minutes for 2203 children (17.1%), 15 to 60 minutes for 5324 (41.4%), 1 to 2 hours for 3010 (25.1%), and >2 to 24 hours for 1910 (16.4%). There were significant unadjusted differences in triage status according to race, with 14.6% of NHW patients being placed in the >2-hour immediacy range, compared with 18.8% of NHB patients and 20.0% of HW patients. In a linear regression analysis with logarithmically transformed wait time as a dependent variable and with adjustment for potential confounders, including hospital location, geographic region, and payer status, both NHB and HW patients waited longer than NHW patients, although the results were statistically significant only for HW patients.
Conclusions. These nationally representative data suggest that children who come to EDs have wait times that vary according to race/ethnicity. There are several potential explanations for this observation, including discrimination, cultural incompetence, language barriers, and other social factors. These data and similar data from the National Hospital Ambulatory Medical Care Survey are useful in identifying nonclinical influences on the delivery of pediatric emergency care.
Key Words: race/ethnicity emergency department wait time
Abbreviations: NHAMCS, National Hospital Ambulatory Medical Care Survey NHW, non-Hispanic white NHB, non-Hispanic black HW, Hispanic white ED, emergency department CI, confidence interval
Emergency department (ED) overcrowding has become a national crisis that parallels increasing ED volume. The 2001 National Hospital Ambulatory Medical Care Survey (NHAMCS) reported that the annual number of ED visits in the United States increased 20% from 1992 to 2001, to 107.5 million visits per year.1 This includes 22 million visits per year for children 15 years of age (20.7% of total annual ED visits). The causes of overcrowding are multifactorial and include not only increased volume but also an aging population with increased health care needs, increased complexity of the conditions of patients presenting to the ED, many of whom require hospital and/or intensive care unit admission, and the relative lack of inpatient and intensive care unit beds. Other factors include shortages of nurses and other clinical personnel, increased demands for ancillary services, decreased numbers of EDs in the United States, and saturation of the primary care network.2–6 The mean wait time for ED patients has also increased, with a mean of 23.9 minutes for visits classified as emergency (immediacy of <15 minutes), 38.1 minutes for urgent visits (immediacy of 15–60 minutes), 56.5 minutes for semiurgent visits (immediacy of 1–2 hours), and 67.7 minutes for nonurgent visits (immediacy of 2–24 hours).7 Racial and ethnic differences in ED use have also been reported. The NHAMCS demonstrated that ED utilization by blacks of all ages was 76% higher than that by whites; whereas white children had 34.7 ED visits per 100 persons per year, black children had 56.5 visits per 100 persons per year.1 However, the percentages of visits classified as semiurgent or nonurgent were reportedly comparable for black children and white children (28.1% and 26.3%, respectively).
Longer wait times in the ED contribute to patient and family dissatisfaction with the care received. Conversely, patients who perceive that the wait time to see a provider was shorter than expected are more satisfied with their ED visits than are patients whose wait was longer than expected.8 Clinically, prolonged ED wait times may result in protracted pain and suffering and in delays in diagnosis and treatment.2,6 Overcrowding in EDs may also place patients at greater risk for medical errors. In 2004, the Joint Commission on Accreditation of Healthcare Organizations developed a standard regarding ED overcrowding that emphasized that patients in overcrowded EDs are at high risk of experiencing treatment delays or inadequate care.3 Similarly, the American Academy of Pediatrics released a policy statement in 2004 that discussed the ways in which ED overcrowding threatens access to care for America's children.4 The purpose of this study was to determine whether wait times for children treated in EDs throughout the nation are associated with race or ethnicity, with adjustment for social, demographic, and clinical factors.
METHODS
The NHAMCS is a national probability-sample survey of visits to EDs and outpatient departments in the United States. It is conducted by the National Center for Health Statistics, a division of the Centers for Disease Control and Prevention. Findings are based on a national sample of visits to the EDs and outpatient departments of noninstitutional general (medical, surgical, and children's) and short-stay hospitals, exclusive of federal, military, and Veterans Affairs hospitals, located in the 50 States and the District of Columbia. The survey uses a 4-stage probability design involving samples of geographically defined areas, hospitals within those areas, clinics within the hospitals, and patient visits within the clinics. EDs and outpatient clinics are sampled separately. Annual data collection began in 1992. Hospital staff members are instructed by the US Bureau of the Census to complete patient record forms for a systematic random sample of patient visits during a randomly assigned, 4-week reporting period. Hospital staff members complete the patient record forms in most cases, but NHAMCS field representatives complete the forms from medical records if necessary. Data are obtained regarding demographic characteristics of the patients, expected sources of payment, chief complaints, physicians' diagnoses, diagnostic/screening services, procedures, medication therapy, disposition, types of health care professionals consulted, causes of injury (where applicable), and certain characteristics of the hospital, such as type of ownership.
Patient race and ethnicity recorded in the survey are determined by the study personnel when they complete the patient record forms. Unless it is hospital policy to ask patients directly for this information, the personnel are instructed to record race/ethnicity on the basis of their observations, marking the box that is, in their opinion, the most appropriate. Race categories include white, black or African American, Asian, Native Hawaiian or other Pacific Islander, and American Indian or Alaska Native. Ethnicity is categorized as Hispanic or non-Hispanic and is recorded separately from race (1997 standards of the Office of Management and Budget).9 The 2001 visit rates for age and race are based on Census 2000 population data, and the 1997–2000 visit rates were calculated with 1990 census data.
NHAMCS data for 1992–2002 are available for public use. We examined data for patients 15 years of age who presented to EDs between 1997 and 2000, a period during which wait time was a recorded variable. Wait time was determined by the NHAMCS from the time the patient arrived in the ED and the time the patient was examined by a physician, both of which were recorded on the patient record form. For the purposes of our study, race/ethnicity was analyzed according to 3 categories, ie, non-Hispanic white (NHW), non-Hispanic black (NHB), or Hispanic white (HW). Other groups had limited representation and were therefore not included. Visit immediacy (triage status) was assigned by triage personnel at the time of arrival at the ED and was categorized into 4 levels, ie, emergency (<15 minutes), urgent (15–60 minutes), semiurgent (1–2 hours), or nonurgent (2–24 hours). The NHAMCS provides guidelines to assist hospital personnel in deciding how their institution's triage categories correspond to these levels. Payer status for patients was categorized as private insurance, government source of payment (Medicare, Medicaid, or Worker's Compensation), self-pay, other (no charge or other), or unknown.
Data were analyzed with SAS version 10.1 software (SAS Institute, Cary, NC). SUDAAN statistical software (Research Triangle Institute, Research Triangle Park, NC) was used to account for the sampling technique used in NHAMCS, which includes oversampling of certain groups. Sample weights were applied to the patient records identified, to yield national estimates. The wait times were positively skewed and were therefore normalized with a natural logarithm transformation before analysis. The 2 test was used to measure the association between race/ethnicity and triage status, and Student's t test of logarithmically transformed data was used for intergroup comparisons of wait times, because nonparametric testing cannot be performed with SUDAAN software. Linear regression analysis using logarithmically transformed wait time as a dependent variable and adjusting for potential confounders, including hospital location, geographic region, and payer status, was performed to examine adjusted wait times.
RESULTS
For the 4-year study period, 20633 patient visits were surveyed, representing a national sample of 92.9 million children 15 years of age between 1997 and 2000. The 3 race/ethnicity categories of interest included a total of 15920 children. The study sample consisted of 9019 (59.5%) NHW children, 3910 (23.9%) NHB children, and 2991 (16.6%) HW children. The mean ages for the 3 groups were 6.4 years for NHW, 5.6 years for HW, and 4.6 years for NHB. The mean (±SD) wait time for all race/ethnicity groups was 43.6 ± 1.7 minutes (Table 1). There were significant unadjusted differences in the wait times among the groups, with HW children waiting 54.5 ± 0.1 minutes, compared with 48.7 ± 0.5 minutes and 38.5 ± 1.6 minutes for NHB and NHW children, respectively (P < .01 for NHW patients, compared with the other groups).
The wait times for the 3 race/ethnicity groups were examined also on the basis of the location of the hospital. Hospital locations were designated as urban or nonurban and as Northeastern, Southern, Midwestern, or Western. Among children who presented to urban hospitals, NHW children waited a significantly shorter time than did HW children (P < .01) (Table 2). The same difference was found in each of the 4 regions (P = .01 for the Western region and P < .01 for all other regions). In addition, Midwestern NHW children waited significantly shorter times than did NHB children (P < .01), and Southern NHB children had shorter waits than did HW children (P = .03).
The triage status distribution was <15 minutes for 2203 children, 15 to 60 minutes for 5324, >1 to 2 hours for 3010, and >2 to 24 hours for 1910 (Table 2). The average ages were very similar among the triage groups, ranging from 5.5 to 5.8 years. There were significant differences in triage assignments among the race/ethnicity groups, with NHB and HW children being triaged to less urgent levels at a significantly higher rate than NHW patients (P = .025). A triage status of >2 hours was assigned to 14.6% of NHW patients, compared with 18.8% of NHB patients and 20.0% of HW patients (P = .04 and .07 for NHW compared with NHB and HW, respectively).
Patient wait times were examined in terms of patient demographic features (gender and race/ethnicity), hospital characteristics (geographic region, metropolitan status, and hospital ownership), source of payment, and characteristics of the ED visit (triage status and disposition) (Table 3). In the univariate analyses, factors significantly associated with longer wait times were NHB (change in wait time: 23%; 95% confidence interval [CI]: 9–38%) and HW (change in wait time: 45%; 95% CI: 29–64%) race/ethnicity, all nonprivate sources of payment (government source: change in wait time: 10%; 95% CI: 1–21%; self-pay: change in wait time: 12%; 95% CI: 0–27%; other sources: change in wait time: 29%; 95% CI: 11–50%), and all triage groups other than the <15-minute group (P < .01 for overall variable). Factors associated with shorter wait times included hospital admission (change in wait time: –31%; 95% CI: –16 to –43%), hospital location in the Midwest (change in wait time: –32%; 95% CI: –17 to –43%), and nonurban setting (change in wait time: –49%; 95% CI: –40 to –56%).
In a multivariate linear regression model, adjustments were made for effects of patient gender, race/ethnicity, hospital location and ownership, payment type, triage status, and disposition of the visit. The race/ethnicity of the patient remained a significant determinant of wait time (P = .01 for overall variable) (Table 4). HW children waited 18% longer than did NHW children (95% CI: 6-30%), whereas NHB patients waited 6% longer than did NHW patients, although this difference was not statistically significant. Additional factors predicting longer wait times were a payment source classified as other (change in wait time: 24%; 95% CI: 6-44%) and any triage assignment less acute than the most urgent group (P < .01 for overall variable). Shorter wait times were associated with patients requiring admission (change in wait time: –20%; 95% CI: –6 to –32%), Midwestern location (change in wait time: –21%; 95% CI: –9 to –32%), nonurban setting (change in wait time: –41%; 95% CI: –35 to –47%), and government ownership of the hospital (change in wait time: –22%; 95% CI: –9 to 34%).
DISCUSSION
The race/ethnicity of children treated in EDs in the United States appears to be significantly associated with wait time. Although the mean wait time for children in all groups was 43.6 minutes, both HW and NHB patients waited longer, with wait times of 54.5 and 48.7 minutes, respectively, although the results were statistically significant only for HW patients. These findings may be linked to patient-, provider-, or system-related variables. Patient-related variables potentially include language spoken, socioeconomic status, type of insurance, geographic location, level of literacy, and cultural values. Provider-related variables include bias, prejudice, and stereotyping, which might play a role in triage decisions. System-related variables include availability of primary care services, lack of available interpreter services, and ED volume.
The finding that HW and NHB patients wait longer than NHW patients may reflect the tendency for minority patients to use the ED for less urgent problems, perhaps because of real or perceived barriers to primary care. Several studies have found that lesser continuity of care is associated with greater risk of ED utilization10 and that children who are members of underrepresented minority groups, are poor, or are uninsured are at greater risk than other children of experiencing obstacles to health care.11 The fact that higher percentages of HW and NHB children fell into the lowest-acuity group supports this theory. However, other studies have suggested that, after adjustment for other factors, including socioeconomic status, regular source of primary care, and insurance status, there are no significant differences in ED use among blacks, whites, and Hispanics.12 In 2002, 38.8% of US children 0 to 18 of age who were Medicaid recipients were NHW, compared with 25.2% who were NHB and 28.3% who were HW. Similarly, 40.5% of uninsured children in the United States are NHW, 18.0% are NHB, and 35.2% are HW.13
Another explanation for wait time disparities could be that patients from underrepresented minority groups are more likely to be treated in EDs that have longer wait times for all patients, such as those located in densely populated, urban areas. Several studies have found that hospitals in more densely populated areas have higher rates of overcrowding than do hospitals serving smaller populations.6,14 We found that, in general, hospitals in urban areas had longer wait times than did those in nonurban areas. Even with adjustment for hospital location, however, race/ethnicity remained an important predictor of wait time in the ED.
Although triage status was found to be an important independent factor in wait time, minority children had longer wait times in the ED even when triage status was controlled. This could result from the fact that triage classification is not standard among institutions. Studies have shown that great variability in triage practice occurs among nurses, physicians, and software programs, even in the same ED.15,16 This may indicate that triage decisions are influenced by factors other than standardized objective assessments of the patient's medical condition. Subjective factors could result in patients being given either greater or lesser immediacy than other patients with the same presentation. Assigning a patient an inappropriately low triage level could lead to a delay in care for that patient. Conversely, assigning an inappropriately high triage level to a less severely ill patient might lead to delays in care for other, possibly sicker patients. Our data suggest that, if there is a bias in the triage process, then patients from underrepresented minorities could be given lower acuity and consequently wait longer than nonminority patients with the same symptoms. When providers are forced to make triage decisions quickly and without complete information, providers' own (sometimes unconscious) beliefs likely enter into triage assignments. Additional investigation of the effect of race/ethnicity on triage decision-making is needed. In addition, a recent study showed that pediatric emergency medicine and general emergency medicine attending physicians may assign different triage levels to the same patients, with general emergency medicine physicians assigning higher levels to febrile children than do their pediatric emergency medicine counterparts.17 Because we could not determine from the NHAMCS data whether children were treated in general or pediatric EDs, it is unclear what effect this factor would have on the results of our study; few children's hospitals participate in this survey.
Our study found that, among children who presented to the EDs of urban hospitals in all regions of the United States, both NHB and NHW children were treated by physicians more quickly than were HW patients. This finding may reflect language barriers between the patients and ED providers. One study of language barriers and resource utilization in pediatric EDs found that non–English-speaking patients had a mean total length of stay in the ED that was 28 minutes longer than that of English-speaking patients.18 The absolute wait times were not reported in that study. The same study found that the overall charge for diagnostic tests was higher when there was a language barrier between the patient and the ED physician. In our investigation, the proportions of higher-acuity cases were similar in the 2 groups, which suggests that the level of acuity did not account for the differences between the 2 groups. Wait times for non—English-speaking patients may be longer because the provider must wait for a translator before evaluating the patient. It is also possible that non—English-speaking patients or parents may be less persistent in insisting that the child be evaluated sooner because they cannot communicate with the triage nurse or may be less likely to complain to ED staff members about the wait because of cultural values. The primary language spoken was not recorded on the NHAMCS data sheets; therefore, we were unable to analyze the effect of a provider-patient language barrier on patient wait times.
A retrospective study comparing ED analgesic use for adult patients with isolated long-bone fractures revealed that HW patients were twice as likely as NHW patients to receive no pain medications. Only 58% of the HW patients in the study spoke English, as opposed to 97% of the NHW patients. With controlling for all covariates, Hispanic ethnicity was the strongest predictor of withheld analgesics. Controlling for primary language spoken or insurance status did not eliminate the differential use of analgesics.19 In a similar study of adult patients, only 57% of NHB patients received analgesia for isolated long-bone fractures, compared with 74% of NHW patients with similar fractures.20 A follow-up study of adult patients did not reveal any differences in physician pain assessments for HW patients versus NHW patients.21
The NHAMCS provides a wealth of data on ED visits in the United States but has several limitations. One is that race and ethnicity are sometimes recorded on the basis of the perceptions of hospital personnel, rather than patient self-reports. Several studies of Department of Veterans Affairs data showed only fair agreement between administrative data and self-reported race/ethnicity.22,23 Patients whose race was other than white were most likely to be misclassified. If patients who identify themselves as HW or NHB are classified as NHW by hospital personnel, then the true differences in wait times between groups might be smaller than our data revealed.
Another limitation is that the wait times derived from the patient record forms vary according to how each hospital records its data. For example, patients in some EDs are registered in the treatment area, sometimes even after being examined by a provider; patients in other EDs register even before they undergo triage. This might cause wait times to vary among facilities, which might affect our results.
Another limitation is that the NHAMCS does not record any measure of socioeconomic status except for type of insurance. Therefore, although we controlled for payer status, we were not able to control fully for socioeconomic status. The NHAMCS does not record whether patients were referred to the ED by their primary care physicians. Patients who are referred to the ED by their primary care physicians might be evaluated with higher acuity status than patients who self-refer to the ED; if more HW and NHB patients have more difficulty in obtaining access to primary care, then this might also result in longer ED wait times.
CONCLUSIONS
These data suggest that children who are treated in EDs in the United States have wait times that vary according to race/ethnicity. There are several potential explanations for this observation, including language barriers, availability of primary care, differences in triage practices, payer status, and hospital characteristics. Additional study is urgently needed, to determine root causes and to identify solutions addressing this disparity in the delivery of pediatric emergency care.
FOOTNOTES
Accepted Oct 19, 2004.
No conflict of interest declared.
REFERENCES
McCaig LF, Burt CW. National Hospital Ambulatory Medical Care Survey: 2001 emergency department summary. Adv Data. 2003; (335)1-25
Derlet RW, Richards JR. Overcrowding in the nation's emergency departments: complex causes and disturbing effects. Ann Emerg Med. 2000;35 :63 –68
American Academy of Pediatrics, Committee on Pediatric Emergency Medicine. Overcrowding crisis in our nation's emergency departments: is our safety net unraveling Pediatrics. 2004;114 :878 –888
Richardson LD, Hwang U. Access to care: a review of the emergency department literature. Acad Emerg Med. 2001;8 :1030 –1036
Derlet RW, Richards JR, Kravitz RL. Frequent overcrowding in U.S. emergency departments. Acad Emerg Med. 2001;8 :151 –155
McCaig LF, Ly N. National Hospital Ambulatory Medical Care Survey: 2000 emergency department summary. Adv Data. 2002;(326) :1 –32
Thompson DA, Yarnold PR, Williams DR, Adams SL. Effects of actual waiting time, perceived waiting time, information delivery, and expressive quality on patient satisfaction in the emergency department. Ann Emerg Med. 1996;28 :657 –665
Christakis DA, Mell L, Koepsell TD, Zimmerman FJ, Connell FA. Association of lower continuity of care with greater risk of emergency department use and hospitalization in children. Pediatrics. 2001;103 :524 –529
Newacheck PW, Hughes DC, Stoddard JJ. Children's access to primary care: differences by race, income, and insurance status. Pediatrics. 1996;97 :26 –32
Baker DW, Stevens CD, Brook RH. Determinants of emergency department use: are race and ethnicity important Ann Emerg Med. 1996;28 :677 –682
Brillman JC, Doezema D, Tandberg D, et al. Triage: limitations in predicting need for emergent care and hospital admission. Ann Emerg Med. 1996;27 :493 –500
Wuerz R, Fernandes CMB, Alarcon J. Inconsistency of emergency department triage. Ann Emerg Med. 1998;32 :431 –435
Maldonado T, Avner JR. Triage of the pediatric patient in the emergency department: are we all in agreement Pediatrics. 2004;114 :356 –360
Hampers LC, Cha S, Gutglass DJ, Binns HJ, Krug SE. Language barriers and resource utilization in a pediatric emergency department. Pediatrics. 1999;103 :1253 –1256
Todd KH, Samaroo N, Hoffman JR. Ethnicity as a risk factor for inadequate emergency department analgesia. JAMA. 1993;269 :1537 –1539
Todd KH, Deaton C, D'Adamo AP, Goe L. Ethnicity and analgesic practice. Ann Emerg Med. 2000;35 :11 –16
Todd KH, Lee T, Hoffman JR. The effect of ethnicity on physician estimates of pain severity in patients with isolated extremity trauma. JAMA. 1994;271 :925 –928
Kressin NR, Chang B, Hendricks A, Kazis LE. Agreement between administrative data and patients' self-reports of race/ethnicity. Am J Public Health. 2003;93 :1734 –1739
Boehmer U, Kressin NR, Berlowitz DR, et al. Self-reported vs. administrative race/ethnicity data and study results. Am J Public Health. 2002;92 :1471 –1473(Catherine A. James, MD, F)