C​LINICAL ​P​RACTICE

Hospital and Demographic Influences on the

Disposition of Transient Ischemic Attack
Jeffrey H. Coben, MD, Pamela L. Owens, PhD, Claudia A. Steiner, MD, MPH, Todd J. Crocco, MD

Abstract
Objectives: There is substantial variation in the emergency department (ED) disposition of patients with transient
ischemic attack (TIA), and the factors responsible for this variation have not been determined. In this study, the
authors examined the influence of clinical, sociodemographic, and hospital characteris tics on ED disposition.
Methods: All ED-treated TIA cases from community hospitals in 11 states were identified from the 2002 Healthcare
Cost and Utilization Project (HCUP). Using the aggregate data, descriptive analyses com pared admitted and
discharged cases. Pearson’s chi-square test was used to determine the statistical sig nificance of these
comparisons. Based on the results of the bivariate analyses, logistic regression models of the likelihood of hospital
admission were derived, using a stepwise selection process. Adjusted risk ratios and 95% confidence intervals (CI)
were calculated from the logistic regression models.
Results: A total of 34,843 cases were identified in the 11 states, with 53% of cases admitted to the hospi tal. In
logistic regression models, differences in admission status were found to be strongly associated with clinical
characteristics such as age and comorbidities. After controlling for comorbidities, differ ences in admission status
were also found to be associated to hospital type and with sociodemographic characteristics, including county of
residence and insurance status.
Conclusions: While clinical factors predictably and appropriately impact the ED disposition of patients diagnosed
with TIA, several nonclinical factors are also associated with differences in disposition. Addi tional research is
needed to better understand the basis for these disparities and their potential impact on patient outcomes.
ACADEMIC EMERGENCY MEDICINE 2008; 15:171–176 ª 2008 by the Society for Academic Emergency Medicine
Keywords: transient ischemic attack, emergency care, disease management
Address for correspondence and reprints: Jeffrey H. Coben, MD;
e-mail: jcoben@hsc.wvu.edu.

I​
subsequent stroke and death. Prior research has demon
strated a 90-day stroke risk of between 9.5 and 10.5% fol
schemic stroke is among the leading causes of death lowing TIA.​1,2 ​A recent population-based study found an
overall 6-month ischemic stroke rate of 17%, with more than
and disability in the United States and many other 65% occurring within 30 days of the initial TIA, and a 6% rate
industrialized nations. Transient ischemic attacks (TIAs) have of stroke or recurrent TIA within the first 48 hours following a
been shown to be a strong predictor of TIA.​3
Patients with TIA are most commonly evaluated in hospital
emergency departments (EDs). The annual overall rate of ED
From the Department of Emergency Medicine (JHC, TJC), the visits for TIA is 1.1 per 1,000 U.S. population, corresponding
Department of Community Medicine (JHC), and the Center for Rural to approximately 300,000 annual ED visits.​4 ​Guidelines
Emergency Medicine (JHC), West Virginia University, Morgantown, relating to the ED man agement of TIA have been published
WV; and the Center for Delivery, Organization and Markets, by the American Heart Association,​5 ​the National Stroke
Healthcare Cost and Utilization Project (HCUP), Agency for Association,​6 ​and other organizations.​7 ​While the initial ED
Healthcare Research and Quality (AHRQ) (JHC, PLO, CAS), evalua tion of TIA is generally straightforward and includes a
Rockville, MD. history and physical examination, electrocardiogram, routine
Received August 27, 2007; revision received October 24, 2007; blood work, and diagnostic brain imaging,​8 ​the final ED
accepted October 27, 2007. disposition of these cases is highly variable. In the 10-year
Presented at the Annual Meeting of the Society for Academic period from 1992 through 2001, Edlow and colleagues​4 ​found
Emergency Medicine, Chicago, IL, May 16, 2007. The views herein that just over half (54%) of TIA cases were admitted to the
are the authors. They do not necessarily reflect the views or policies hospital following their ED
of the AHRQ or the U.S. Depart ment of Health and Human Services.

ª 2008 by the Society for Academic Emergency Medicine ISSN 1069-6563 doi: 10.1111/j.1553-2712.2008.00041.x PII ISSN 1069-6563583
171
172 Coben et al. • DISPOSITION OF TIA
 nursing home or left against medical advice were excluded
from the analysis (n = 1,575 or 4.3% of ED TIA visits among
evaluation. They also described regional variation in
individuals 30 years and older). These cases were omitted
admission rates.
either because they would be dou ble counted or because it
While variation in the ED disposition of TIA cases has been
was not possible to determine whether there was an
documented, the factors responsible for this variation have
admission at another hospital. Additional sensitivity analyses
not been determined. In this study, we examined the
suggested that the omis sion of these records had no effect
associations between disposition status and several clinical
on the results of the study.
and nonclinical factors. Specifi cally, our objective was to
determine the influence of hospital characteristics and
Measures
sociodemographic factors on ED disposition, after controlling
The primary outcome of interest was whether or not the ED
for clinical char acteristics.
visit resulted in admission to the hospital. ED visits were
grouped into two mutually exclusive catego ries: those in
METHODS which patients were treated and released and those in which
patients were admitted to the hospi tal following ED
Study Design
evaluation.
This was a retrospective cross-sectional study of TIA cases
using Healthcare Cost and Utilization Project (HCUP) data. The primary independent variables of interest included
The Agency for Healthcare Research and Quality (AHRQ) patient demographics, clinical and visit char acteristics
including expected payer, and hospital char acteristics.
Institutional Review Board deter mined that this project was
Patient demographics included age (30–44, 45–54, 55–64,
exempt from informed consent.
75–84, and 85+ years vs. 65–74 years), gender, and
residential area. A patient’s county of resi dence, as
Study Setting and Population determined by the centroid of a zip code, is classified into
This study used administrative data obtained from nearly all four geographic areas based on the 2003 Urban Influence
community, non-Federal hospitals in 11 states in 2002, Codes (UICs) established by the Eco nomic Research
including Connecticut, Georgia, Maine, Massa chusetts,
Service:​11 ​large metropolitan (large metro), small
Minnesota, Missouri, Nebraska, South Caro lina, Tennessee,
metropolitan (small metro), micropolitan (or large rural
Utah, and Vermont. The data were obtained from the HCUP,
county), and noncore (or small rural county).
sponsored by the AHRQ. HCUP is based on a voluntary
In addition to patient demographics, selected second ary
partnership of statewide data organizations (including state
diagnoses were examined using the ICD-9-CM codes and the
agencies, hospital associations, and private industry) and the
Clinical Classifications Software (CCS;
Federal gov ernment. This study included both outpatient ED
data and inpatient hospital discharge data for all patients http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs. jsp).​12
seen at the participating hospitals. These data are derived The CCS is a classification system that groups ICD-9-CM
from hospital discharge abstracts collected for billing codes into 260 mutually exclusive, clinically meaningful
purposes and when combined contain informa tion on the categories. The following secondary diag noses were
universe of ED visits.​9 ​Across the 11 states, 93% of all examined: cardiac conditions (atrial fibrilla tion, cardiac
dysrhythmias, chest pain, congestive heart failure, coronary
community hospitals were included in the HCUP in 2002.
atherosclerosis, heart valve disorders, myocardial infarction),
Data elements in these databases included patient
hypertension, anticoagulant use, coagulation disorders, prior
demographics, all-listed diagnoses and procedures, expected
stroke, syncope, organic brain syndrome, diabetic conditions,
payer (including self-pay and uninsured), total charges, and
sickle cell disor der, and significant anemia. In addition, the
disposition. Additional information about HCUP can be
number of these secondary conditions was categorized (none
obtained at http:// www.hcup-us.ahrq.gov/.​10
vs. one diagnosis, two to three diagnoses, four to eight
diagnoses). The primary expected payer was classified as
Study Protocol Medicare, Medicaid, commercial insurance (private), or
The study sample was limited to those ED visits for uninsured (self-pay, no charge, other). In addition, whether
individuals 30 years of age and older who had a princi pal the ED visit occurred on a weekday or a week end was
diagnosis of TIA, as indicated by an International recorded.
Classification of Diseases, Ninth Revision, Clinical Modi Hospital characteristics included the geographic loca tion,
fication (ICD-9-CM) diagnosis code of ‘‘other specified bed size, ownership, and teaching status. The hos pital
transient cerebral ischemia’’ (435.8) or ‘‘unspecified transient location was characterized by the state and the modified
cerebral ischemia’’ (435.9; n = 36,418 ED visits for TIA of 9.9 version of the UICs described above. The cod ing of the
million ED visits for individuals 30 years and older). Since hospital’s UIC was based on the hospital county, as recorded
hospital admission is generally recom mended for patients by the American Hospital Associa tion (AHA) Annual Survey.
manifesting symptoms of vertebro basilar ischemia, those Information on the hospi tal’s bed size (<100, 100–299, and
diagnosed with basilar artery syndrome, vertebral artery 300–499 beds vs. 500+ beds), teaching status (teaching vs.
syndrome, subclavian steal syndrome, and⁄ or vertebrobasilar nonteaching), and ownership (private for profit, private
artery syndrome were excluded. Thus, only cases diagnosed not-for-profit vs. public) were similarly obtained from the AHA
as TIAs involving the carotid circulation were included. ED vis Annual Survey.
its that resulted in a transfer to another hospital or
ACAD EMERG MED •​ ​February 2008, Vol. 15, No. 2 •​ ​www.aemj.org 173
Cary, NC). Descriptive analyses were performed to compare
differences between admission status in patient
Data Analysis
demographics and clinical character istics and hospital
All analyses were conducted using the aggregate data for all
attributes. A Pearson’s chi-square test was used to determine
cases identified, using PC-SAS Version 9.1 (SAS Institute,
 the statistical significance of these comparisons. After criteria for TIA were identified across the 11 states (range
assessing the correlation among all independent variables, across hospitals 0.1%–2.7%). Of these, 18,575 (53.3%) were
multivariate analyses were performed using a modified admitted to the hospital following ED evaluation, and 16,268
Poisson regression approach with a robust standard error (46.7%) were discharged follow ing ED evaluation. Table 1
variance. This method has been shown to be effective in illustrates the distribution of cases across the 11 states and
estimating relative risk for common binary outcomes, such as additional characteristics of the hospitals where patients were
the risk of hospital admission following a TIA.​13–15 ​Patient evaluated. Signifi cant differences were noted in ED
demographics and clinical characteristics were entered in the disposition by state and by hospital. Eight states had nearly
multivariate model first, followed by hospital characteristics. half of ED cases for TIA result in admission (range
Models were built in a step wise fashion, assessing the 47.3%–62.0%), while three states (Maine, Utah, Vermont)
change in parameter esti mates and model fit. Variables that had approxi mately one-third of ED cases for TIA result in
were significant at the p < 0.10 level were retained in the final admis sion (range 31.3%–37.2%). Hospitals located in small
models. Potential interactions between geographic location of rural counties were more likely to discharge patients to home
the patient and the hospital and hospital bed size were than hospitals located in large rural or metropoli tan areas.
tested. Relative risks and 95% confidence intervals (CIs) Similarly, smaller hospitals were more likely to discharge
were calculated from the regression models. patients home following their ED visit than larger hospitals.
ED disposition also varied according to hospital teaching
status and ownership (Table 1).
Patient demographic and clinical characteristics also
Table 1
demonstrated significant associations with ED disposi tion
RESULTS (Table 2). Patients residing in rural counties were more likely
to be discharged following ED evaluation,
A total of 34,843 (0.4%) ED visits meeting our inclusion

Bivariate Analysis of State and Hospital Characteristics of ED Visits for TIA by Hospital Admission Status

Treated and

Admitted to the Hospital from the ED (N = 16,268)​

Unadjusted
ED (N = 18,575)
Released from Chi-Square
Hospital Demographics State
n n %* n %* p-Value Risk Ratio

Connecticut 2,571 1,393 54.2 1,178 45.8 1.02 <0.001 Georgia 5,893 2,986 50.7 2,907 49.3 0.94 Maine 1,317 472 35.8 845 64.2 0.66
Massachusetts 3,939 2,385 60.5 1,554 39.5 1.16 Minnesota 2,987 1,853 62.0 1,134 38.0 1.18 Missouri 5,395 3,057 56.7 2,338 43.3
1.08 Nebraska 1,007 476 47.3 531 52.7 0.88 South Carolina 3,476 1,984 57.1 1,492 42.9 1.08 Tennessee 6,730 3,471 51.6 3,259 48.4
0.96 Utah 1,187 371 31.3 816 68.7 0.58 Vermont 341 127 37.2 214 62.8 0.70 Location
Large metro 12,564 7,514 59.8 5,050 40.2 1.20 <0.001 Small metro 13,924 7,069 50.8 6,855 49.2 0.92 Large rural 5,225 2,667 51.0
2,558 49.0 0.95 Small rural 3,130 1,325 42.3 1,805 57.7 0.78 No. of beds
<100 8,240 3,657 44.4 4,583 55.6 0.79 <0.001 100–299 13,974 7,399 52.9 6,575 47.1 0.99 300–499 7,675 4,544 59.2 3,131 40.8
1.15 500+ 4,954 2,975 60.1 1,979 39.9 1.15 Teaching status
Teaching 11,423 6,744 59.0 4,679 41.0 1.17 <0.001 Non teaching 23,420 11,831 50.5 11,589 49.5 0.86 Ownership
Public 5,287 2,715 51.4 2,572 48.6 0.96 <0.001 Private not-for-profit 25,298 13,646 53.9 11,652 46.1 1.04 Private for-profit 4,258
2,214 52.0 2,044 48.0 0.97

ED ¼ emergency department; TIA ¼ transient ischemic attack.

*Row-percent.
174 Coben et al. • DISPOSITION OF TIA

Table 2
Bivariate Analysis of Patient and Clinical Characteristics of ED Visits for TIA by Hospital Admission Status

Treated and
Released
Admitted to the Hospital (n = 16,268)​
Unadjusted
from the ED (n = 18,575)
from ED Chi-Square
Characteristics n n % p-Value *​ ​n %​*​ ​Risk Ratio
Patient demographics Age (years)

30–44 1,401 720 51.4 681 48.6 0.96 <0.001 45–54 3,291 1,718 52.2 1,573 47.8 0.98 55–64 5,261 2,635 50.1 2,626 49.9 0.93
65–74 7,863 4,193 53.3 3,670 46.7 1.00 75–84 11,036 6,055 54.9 4,981 45.1 1.04 85+ 5,991 3,254 54.3 2,737 45.7 1.02 Gender
 Female 20,320 10,936 53.8 9,384 46.2 1.02 0.025 Male 14,521 7,638 52.6 6,883 47.4 0.98 Patient residence
Large metro 13,460 7,713 57.3 5,747 42.7 1.13 <0.001 Small metro 12,016 6,344 52.8 5,672 47.2 0.99 Large rural 5,054 2,468
48.8 2,586 51.2 0.90 Small rural 4,313 2,050 47.5 2,263 52.5 0.88 Clinical and visit characteristics
No. of secondary conditions
None 10,073 2,335 23.2 7,738 76.8 0.35 <0.001 One 10,840 5,883 54.3 4,957 45.7 1.03 Two or three 12,029 8,704 72.4 3,325 27.6
1.67 Four to eight 1,901 1,653 87.0 248 13.0 1.69 Secondary conditions
Use of anticoagulants 1,292 618 47.8 674 52.2 0.89 <0.001 Cardiac conditions 11,341 8,428 74.3 2,913 25.7 1.72 <0.001
Coagulation and hemorrhagic disorder 277 219 79.1 58 20.9 1.49 <0.001 Prior stroke 1,535 1,238 80.7 297 19.3 1.55 <0.001
Syncope 551 374 67.9 177 32.1 1.28 <0.001 Organic brain syndrome 1,114 739 66.3 375 33.7 1.25 <0.001 Diabetic disorders
6,829 4,584 67.1 2,245 32.9 1.34 <0.001 Significant anemia 1,796 1,484 82.6 312 17.4 1.60 <0.001 Expected payer
Medicare 24,485 13,534 55.3 10,951 44.7 1.13 <0.001 Medicaid 1,568 778 49.6 790 50.4 0.93 Commercial (private) 7,374 3,585
48.6 3,789 51.4 0.89 Uninsured 1,292 633 49.0 659 51.0 0.91 Visit day
Weekend 9,491 5,181 54.6 4,310 45.4 1.03 0.003 Weekday 25,352 13,394 52.8 11,958 47.2 0.97

ED ¼ emergency department; TIA ¼ transient ischemic attack.

*Row-percent.
Similarly, the presence of
cardiac conditions, hypertension, and anemia demon strated
an increased likelihood of admission following ED evaluation
whereas patients from metropolitan areas were more likely to for a TIA, while patients with prior use of anticoagulants, prior
be admitted. Patients with Medicare were more likely to be stroke, and diabetes were less likely to be admitted. After
admitted to the hospital, whereas patients with other payment controlling for coexisting conditions, differences in admission
sources were more likely to be dis charged home from the status, although smaller in magnitude compared to clinical
ED. Disposition status varied according to the number of conditions, continued to be associated with patient
comorbidities and second ary conditions, with the proportion sociodemo graphic and hospital characteristics. Rural
of cases admitted to the hospital increasing with the number residence of the patient and small numbers of beds at the
of listed sec ondary conditions. Preexisting use of treating facility were associated with a lower likelihood of
anticoagulants was associated with a lower likelihood of admission, while Medicare coverage increased the like lihood
hospital admission following ED evaluation. of admission. The youngest patients in the sam ple were
Results of the multivariate analyses are summarized in most likely to be admitted to the hospital, and patients who
Table 3. Clinical characteristics were strongly asso ciated were evaluated in the ED on the weekend were more likely to
with ED disposition. Hospital admission was nearly three be admitted than those seen during weekdays.
times as likely among patients with four to eight coexisting
illnesses compared to patients with no coexisting illnesses.
ACAD EMERG MED •​ ​February 2008, Vol. 15, No. 2 •​ ​www.aemj.org 175
deem it safe to discharge these patients, given some
indication of ongoing outpatient support. However, after
Table 3 controlling for these conditions, we found a num ber of
Modified Poisson Regression Model of Patient, Clinical, and
nonclinical factors associated with ED disposi tion. Patients
Hospital Characteristics Associated with Hospital Admission
from an ED Visit for TIA residing in more rural communities and those presenting to
smaller hospitals were particularly more likely to be
discharged home from the ED. While
Adjusted
Relative
Characteristics* Risk 95% CI recommendations regarding ED
Patient characteristics current guidelines do not make firm disposition, the high risk of recurrent TIA
Age of the patient, years (reference: 65–74 years) 30–44 1.30
1.23, 1.38 45–54 1.17 1.13, 1.22 55–64 1.05 1.01, 1.09 ED ¼ emergency department; TIA ¼ transient ischemic attack; CI
75–84 0.99 0.96, 1.01 85+ 0.97 0.94, 0.99 Female 1.03 1.01, ¼ confidence interval.
1.04 *Model also adjusted for state.
Patient residence (reference: large metro county) Small metro
county 0.94 0.91, 0.97 Large rural county 0.88 0.85, 0.92 Small
rural county 0.91 0.88, 0.95 Clinical and visit characteristics
No. of secondary diagnoses (reference: none)
One 2.11 2.02, 2.21 Two or three 2.59 2.46, 2.73 Four to DISCUSSION
eight 2.93 2.75, 3.12 Secondary conditions
Cardiac conditions 1.17 1.14, 1.20 Hypertension 1.12 1.09, In this large cross-sectional study of TIA cases, we found
1.15 Use of anticoagulants 0.69 0.65, 0.73 Prior stroke 0.86 significant variation in patient disposition, with approximately
0.84, 0.89 Diabetes 0.94 0.92, 0.96 Anemia 1.19 1.16, 1.22 53% of cases admitted to the hospital and 47% discharged
Expected payer (reference: private insurance) home following their ED visit. Clinical factors, including the
Medicare 1.07 1.04, 1.11 Medicaid 0.93 0.89, 0.98 Uninsured type and total number of coexisting conditions were strongly
0.99 0.93, 1.04 ED visit on the weekend 1.03 1.01, 1.05 Hospital associated with ED disposition. Clear clinical indicators, such
characteristics
as complicat ing cardiac conditions or current use of
Hospital location (reference: large metro county)
Small metro county 0.97 0.93, 1.00 Large rural county 1.07 anticoagulants, are associated with the likelihood of
1.03, 1.11 Small rural county 1.01 0.96, 1.07 No. of beds admission to the hospital or discharge from the ED as would
(reference: 500+) be clinically appropriate. The lower likelihood of hospital
<100 0.86 0.83, 0.89 100–299 0.94 0.92, 0.97 300–499 1.04 admission for patients with current use of anticoagulants may
1.01, 1.07 Teaching hospital 1.02 1.00, 1.04 be an indirect indicator of patients already under the care of
 an outpatient physician and perhaps receiving pro phylaxis colleagues​4 ​found no change in this rate over the 10-year
against a TIA or stroke. ED personnel may period 1992 through 2001. In clini cal practice, emergency
and ⁄ or stroke in these patients has led some to conclude physicians routinely consult with primary care physicians and
that the emergency physician must ensure that patients who neurologists on patient disposition following the initial ED
present to the ED with TIA are not discharged from the evaluation. Therefore, the variation in ED disposition and lack
hospital prior to establishing if they require a revascularization of change in disposition patterns over time appears to reflect
an overall system of care with ingrained patterns of clinical
procedure.​8 ​Others have suggested that hospital admission is
management variation, including those who prefer to evaluate
recommended if appropriate imaging studies are not
these cases on an outpatient basis. While there are no
‘‘immediately’’ available.​16 ​Given these concerns, and completed large-scale trials of emer gent therapies for TIA,
assuming that smaller facili ties and patients in rural the high clinical instability of these cases and potential
communities are less likely to have timely access to opportunity to initiate emer gent therapy for recurrent
advanced outpatient diagnostic studies, the disposition symptoms have led to con vincing arguments that emergent
pattern we have found appears paradoxical. Furthermore, evaluation, treatment, and (inpatient) monitoring are
those residing in rural com munities may have longer
warranted.​17
transport times to obtain emergent care if a recurrent TIA or
Several other nonclinical factors were found to be
stroke should occur in the outpatient setting. However, it is
associated with ED disposition. Women were more likely to
difficult to ascer tain if these patients are discharged from the
ED and sent to neighboring facilities where these more be admitted to the hospital than men, and patients evaluated
advanced diagnostic and therapeutic interventions are on weekends were more likely admit ted than those seen
available. Additionally, we are unable to determine if patient during weekdays. Patients with Medicare were more likely to
preference, primary care physician access and availability, be admitted than those with either private health insurance or
and family support may have influenced ED disposition in with no health insurance. Although we cannot determine
small rural communities. Finally, the out comes of care for causality or a verified explanation, these associations are of
patients admitted or discharged from the ED with a TIA interest. Although smaller in magnitude, these associations
cannot be assessed in this study. may have implications for utilization of hospital services on
The 53% hospital admission rate reported in our study is weekends and for some public payers.
similar to that reported previously.​4 ​Interest ingly, Edlow and
176 Coben et al. • DISPOSITION OF TIA

The authors acknowledge the statewide data organizations that

LIMITATIONS
participated in the Healthcare Cost and Utilization Project (HCUP) State
Inpatient Databases and State Emergency Department Data bases in
Our cross-sectional design allows us to demonstrate
2002: Connecticut Integrated Health Information (Chime, Inc.); Georgia
associations, but not causality. Our results are based on data GHA: An Association of Hospitals & Health Systems; Maine Health Data
from 11 states and may not be generalizable to other Organization; Massachusetts Division of Health Care Finance and Policy;
locations. Examining the admission patterns of hospitals and Minnesota Hospital Association; Missouri Hospital Industry Data Institute;
Nebraska Hospital Association; South Carolina State Budget & Control
states in the aggregate does not allow us to highlight small
Board; Tennessee Hospital Association; Utah Department of Health (Utah
area practice variation that is evident in the results. However, Hospital Inpatient Discharge Data File [2002]; Utah Health Data
the 11 states in our study are from all four regions of the Committee ⁄ Office of Health Care Statistics and Utah Emergency
United States and vary in population and rurality. In an effort Department Encounter Data [2002]; Bureau of Emergency Medical
Services ⁄ Office of Health Care Statistics; Utah Department of Health,
to clearly delineate cases where the hospital visit was due to
Salt Lake City, UT [2003]); and Vermont Association of Hospitals and
TIA, we lim ited our cases to those with TIA as the principal Health Systems.
diag nosis. Differential documentation in ED records and
inpatient charts may have caused us to overestimate the References
association between hospital admission and the number of
secondary conditions, although the direction of association 1. Johnston SC, Gress DR, Browner WS, Sidney S.
was clinically consistent. While the use of anticoagulants may Short-term prognosis after emergency department
be an indirect indicator that the patient has a primary care diagnosis of TIA. JAMA. 2000; 284:2901–6.
physician, our data do not include any variables that directly 2. Hill MD, Yiannakoulias N. Jeerakathil T, Tu JV, Svenson
indicate this. Finally, the cross-sectional design does not LW, Schopflocher DP. The high risk of stroke immediately
allow us to compare outcomes among those with differing after transient ischemic attack: a population-based study.
disposition. Neurology. 2004; 62:2015– 20.
3. Kleindorfer D, Panagos P, Pancioli A, et al. Inci dence
CONCLUSIONS and short-term prognosis of transient ische
mic attack in a population-based study. Stroke. 2005;
We report substantial variation in the ED disposition of TIA 36:720–4.
cases. Beyond clinical issues, several patient soci 4. Edlow JA, Kim S, Pelletier AJ, Camargo CA. National
odemographic factors and hospital characteristics are study on emergency department visits for transient
associated with disposition status. Our findings can help ischemic attack, 1992-2001. Acad Emerg Med. 2006;
guide the subsequent research that is necessary to better 13:666–72.
understand the basis for these disparities and to further 5. Wolf PA, Clagett GP, Easton JD, et al. Preventing
investigate the decision-making process. Most importantly, ischemic stroke in patients with prior stroke and transient
further research is needed to determine if differential ischemic attack: a statement for healthcare professionals
disposition from the ED demonstrates a dif ferential impact on from the Stroke Council of the Ameri can Heart
subsequent patient outcomes. Association. Stroke. 1999; 30:1991–4.
 6. Brott T, Clark W, Fagan S. Stroke: The First Hours:
Guidelines for Acute Treatment. Centennial, CO: National
Stroke Association, 2000.
7. Schwamm LH, Pancioli A, Acker JE III, et al. Rec
ommendations for the establishment of stroke sys tems
of care: recommendations from the American Stroke
Association’s Task Force on the Development of Stroke
Systems. Stroke. 2005; 36:690–703.
8. Ovbiagele B. The emergency department: first line of
defense in preventing secondary stroke. Acad Emerg
Med. 2006; 13:215–22.
9. Moles E, Andrews RM. Emergency Department Data
Evaluation. May 2005. HCUP Methods Series Report
#2005–2. ONLINE. June 3, 2005. U.S. Agency for
Healthcare Research and Quality. Available at:
http://www.hcup-us.ahrq.gov/reports/2005_02.pdf.
Accessed Oct 30, 2007.
10. HCUP. Healthcare Cost and Utilization Project. June
2007. Agency for Healthcare Research and Quality,
Rockville, MD. Available at: http://www.hcup-us.
ahrq.gov. Accessed Jun 26, 2007.
11. U.S. Department of Agriculture Economic Research
Service. Briefing Room: Measuring rurality–Urban
Influence Codes. Available at: http://www.ers.usda.
gov/Briefing/Rurality/UrbanInf/. Accessed Jun 26, 2007.
12. HCUP Clinical Classifications Software (CCS) for
ICD-9-CM Fact Sheet. Agency for Healthcare Research
and Quality, Rockville, MD. Available at:
http://www.hcup-us.ahrq.gov/toolssoftware/ccs/
ccsfactsheet.jsp. Accessed Jun 26, 2007.
13. Zou G. A modified poisson regression approach to
prospective studies with binary data. Am J Epidem iol.
2004; 159:702–6.
14. Greenland S. Model-based estimation of relative risks
and other epidemiologic measures in studies of common
outcomes and in case-control studies. Am J Epidemiol.
2004; 160:301–6.
15. Linquist K. How to estimate relative risk in SAS using
PROC GENMOD for common outcomes in cohort
studies. UCLA Academic Technology Ser vices SAS
FAQ. Available at: http://www.ats.ucla.
edu/STAT/SAS/faq/relative_risk.htm. Accessed Jun 27,
2007.
16. Albers GW, Hart RG, Lutsep HL, Newell DW, Sacco RL.
Addendum to the supplement to the guidelines for the
management of transient ischemic attacks. Stroke. 2000;
31:1001.
17. Johnston SC. Transient ischemic attacks are emer
gencies [abstract]. Stroke. 2005; 36:724.