Article, Neurology

Rates and factors associated with admission in patients presenting to the ED with TIA in the United States—2006 to 2008

a b s t r a c t

Background: The estimates of patients who present with transient ischemic attacks (TIA) in the emergency departments (EDs) of United States and their disposition and factors that determine hospital admission are not well understood.

Objective: We used a nationally representative database to determine the rate and predictors of admission in TIA patients presenting to EDs.

Methods: We analyzed data from the National Emergency Department Sample (2006-2008) for all patients presenting with a primary diagnosis of TIA in the United States. Samples were weighted to provide national estimates of TIA hospitalizations and identify factors that increase the odds of hospital admission including age, sex, type of insurance, median Household income, and hospital type (urban teaching, urban nonteaching, and nonurban). Multivariate logistic regression analysis was used to identify independent predictors of hospital admission.

Results: There were 812908 ED visits for primary diagnosis of TIA; mean age (+-SD), 70.3 +- 14.9 years; and 57.9% were women from 2006 to 2008. Of these ED visits, 516837 (63.5%) were admitted to the hospital, whereas 296071 (36.5%) were discharged from the ED to home. In the multivariate logistic regression analysis adjusting age, sex, and medical comorbidities, independent factors associated with hospital admissions were median household income $64000 or higher (odds ratio [OR], 1.33; 95% confidence interval [CI], 1.22-1.44; P = .003), Medicare Insurance type (OR, 1.19; 95% CI, 1.14-1.26; P b .0001), and metropolitan

teaching hospital ED (OR, 2.17; 95% CI, 1.90-2.48; P b .0001).

Conclusion: From 2006 to 2008, approximately 64% of all patients presenting with TIAs to the EDs within United States were admitted to the hospital. Factors unrelated to patients’ condition such as median household income, insurance status, and ED affiliated hospital type play an important role in the decision to admit TIA patients to the hospitals.

(C) 2013

? All authors have read and approved submission of the manuscript.

?? The material in the manuscript has not been published and is not being

considered for publication elsewhere in whole or in part in any language except as an abstract.

? Dr Qureshi has received funding from the National Institutes of Health RO1-

NS44976-01A2 (medication provided by ESP Parma), American Heart Association Established Investigator Award 0840053N, National Institutes of Health U01- NS062091-01A2.

?? Contribution of authors: All authors have equally assisted in the synthesis and

discussion of ideas, interpretation of data, and endoVascular procedures. The majority of research, data collection, and writing of the manuscript were performed by these authors. Each person shared equal responsibility for the information written in the manuscript above.

* Corresponding author. Michigan State University, 804 Service Road, A-217, East Lansing, MI 48824. Tel.: +1 612 626 1825; fax: +1 612 625 7950.

E-mail address: [email protected] (S.A. Chaudhry).

Introduction

The annual incidence of Transient ischemic attacks is estimated to range between 200000 and 500000 [1,2] and accounts for approximately 1.1 per 1000 emergency department (ED) visits [3] in the United States. Previous studies have found that patients with an index TIA have a 4% to 5% risk of ischemic stroke within the next 48 hours [1,4-7]. rapid identification and timely management of patients with TIA can reduce the rate of subsequent stroke by 80% [8-11]. Numerous criteria and guidelines [7,12-18] are provided to encourage prompt initiation of secondary prevention and ensure rapid access to diagnosis and acute treatments such as thrombolysis, if subsequent Ischemic events occur. Such intensity of care is considered best provided by hospital admission in most settings. There is enormous variability in the management of TIA patients, particularly in regard to

0735-6757/$ – see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2012.10.003

S.A. Chaudhry et al. / American Journal of Emergency Medicine 31 (2013) 516519 517

hospital admission among physicians [5,19-24], and some studies have suggested that patients may benefit from hospital admission [25].

The purpose of our study is to provide national estimates of patients who present with TIA in EDs and to determine the rate and predictors of hospital admission obtained from the National Emer- gency Department Sample (NEDS).

Materials and methods

Data selection

We used the NEDS files from 2006 to 2008 for our analysis. NEDS is the largest all-payer ED database that is publicly available in the United States. NEDS is derived from a combination of Healthcare Cost and Utilization Project, State Emergency Department and the State Inpatient Databases and reports upon every ED visit, regardless of whether they result in admission. The NEDS contains information of approximately 28 million ED visits at 980 hospitals within the United States, approximately a 20% stratified sample of hospital-based EDs. The NEDS contain clinical and nonclinical variables associated with hospital stays, including primary and secondary diagnoses/procedures, patients’ admission and discharge status, patient demographic infor- mation (eg, sex, age, expected payment source and Total charges, and nature of visit, eg, acute and chronic conditions and injuries). Information on discharge destination is classified as: treat-and-release, transfer to another hospital or treated in the ED and admitted to the same hospital. Extensive documentation and resources are available to facilitate data analysis at http://www.hcup-us.ahrq.gov/ databases.jsp and http://www.hcup-us.ahrq.gov/db/nation/neds/

discharge weights of the NEDS following Healthcare Cost and Utilization Project recommendations.

We used the ?2 test for categorical data and analysis of variance for continuous data to detect any significant differences in clinical variables among patients who were treated and released or admitted to the same hospital. We also evaluated the difference in hospital location/teaching status, trauma level, geographical region, and primary health insurance between the 2 groups. We performed statistical analyses using SAS software, version 9.1 (SAS Institute, Cary, NC) to convert raw counts into weighted counts to generate national estimates. SAS procedures SURVEYMEANS, SURVEYFREQ, and SURVEYLOGISTIC were used for

descriptive analyses and modeling efforts. We performed univariate analysis followed by multivariate logistic regression to identify differences in study variables and end points between the 2 patient groups. We included comorbid conditions, age, hospital location/ teaching status, primary payer information, and patient’s median income by zip code as potential factors associated with admission in the model.

Results

There were 812908 ED visits for the primary diagnosis of TIAs in United States from 2006 to 2008: mean age (+-SD), 70.3 +- 14.9 years, and 57.9% were women. Of these ED visits, 516837 (63.5%) were

Table 1 Baseline demographics and clinical characteristics of patients with TIA presenting to ED in United States

NEDS2008Introductionv3.pdf.

2.2. Identification of patients with TIA

TIA visit

(not admitted to hospital)

TIA visit (admitted to hospital)

P Value

We used appropriate International Classification of Disease, Ninth Revision, Clinical Modification, codes (435xx) and Medicare severity diagnosis-related group codes (524) to identify patients presenting to the ED with a primary diagnosis of TIA. This approach has been used in various studies for the identification of patients with TIA [3,26]. We further divided the patients into 2 groups based on ED disposition. Patients who were discharged to home were included in 1 group, whereas patients who were admitted to the same hospital were included into second group. We excluded patients who were transferred from the ED to another hospital (b 3% all encounters). International Classification of Disease, Ninth Revision, Clinical Modifica-

tion, diagnostic code fields were screened for specific codes to identify

Nonmetropolitan

Metropolitan nonteaching

76387 (25.8)

137744 (46.5)

74841 (14.4) b.0001

253642 (49.0)

patients with underlying diabetes mellitus (249 and 250), hyperten-

Metropolitan teaching

81940 (27.6)

188354 (36.4)

sion (401-405), atrial fibrillation (427.31), and tobacco dependence

(305.1) as secondary diagnoses.

Hospital trauma level Nontrauma center

213941 (72.3)

356938 (69.1) 0.008

Trauma center I-III

82130 (27.7)

159899 (30.9)

Primary health insurance

2.3. Statistical analysis and estimation of national estimates

Private including health

97129 (32.8)

144078 (28.1)

b.0001

maintenance organization/other

The NEDS database provides weights for national estimation. The Medicare 186424 (63.2) 346531 (67.1)

Medicaid 11562 (3.9) 25714 (4.8)

Overall number 296071 (36.5%) 516837 (63.5%)

Age, mean years (95% CI) 69.7 (69.6-69.8) 70.5 (70.5-70.6) b.0001

Age strata

b60 y 75167 (25.4) 121806 (23.5) b.0001

>=60 y 220904 (74.6) 395031 (76.5)

Sex

Men 128694 (43.5) 213249 (41.2) b.0001

Women 167147 (56.5) 303576 (58.8)

Comorbid conditions

Hypertension 146448 (49.4) 386741 (74.8) b.0001

Diabetes mellitus 53862 (18.2) 148318 (28.7) b.0001

Atrial fibrillation 17294 (5.8) 70723 (13.6) b.0001

cigarette smoking 21534 (7.2) 56266 (10.9) b.0001 Hospital location and teaching status

discharge weights are calculated for NEDS data by first stratifying the Hospital geographic region

NEDS hospitals on the same variables that were used for creating the West

57962 (19.5)

80360 (15.4)

b.0001

sample. These variables were geographic region, trauma center Northeast

34637 (11.6)

104065 (20.1)

designation, urban/rural location, teaching status, and ownership. A Midwest

79291 (26.8)

101770 (19.6)

South

124181 (41.9)

230642 (44.6)

weight was then calculated for each stratum by dividing the number Median household income by zip code

of universe discharges in that stratum–obtained from American

$1-$48999

169870 (58.6)

269024 (53.1)

b.0001

Hospital Association (AHA) data–by the number of NEDS discharges

$49000-$63999

67918 (23.4)

121686 (24.1)

in the stratum. Weighted estimates were calculated by uniformly applying the stratum weights to the discharges according to the stratum from which the discharge was drawn. The discharge weights were applied to the unweighted NEDS observations, and the result is an estimate of the number of discharges for the entire sample. All analyses accounted for the complex sampling design and sample

>=$64000 52020 (17.9) 115199 (22.7)

Discharge status

Discharged home NA 319993 (85.7)

Discharge to short-term facility NA 4313 (1.1)

Discharged to other facility NA 47237 (12.6)

In-hospital mortality NA 1402 (0.4)

Unknown NA 58 (0.02)

518 S.A. Chaudhry et al. / American Journal of Emergency Medicine 31 (2013) 516519

Table 2

Results of multivariate analysis evaluating factors associated with hospital admission among patients presenting to ED with TIA

Variable OR 95% CI

Sex

Men Reference

Women 1.06 1.04-1.09

Age strata

b60 y Reference

>=60 y 0.79 0.76-0.84

Comorbid conditions

Hypertension 2.84 2.73-2.96

Diabetes mellitus 1.47 1.43-1.52

Atrial fibrillation 2.52 2.39-2.65

Hospital location and teaching status

Nonmetropolitan Reference

Metropolitan nonteaching 1.75 1.57-1.94

Metropolitan teaching 2.17 1.90-2.48

Hospital trauma level

Nontrauma center Reference

Trauma center I-III 0.98 (0.87-1.11) Primary health insurance

based on audits of several EDs [2,3,27]. In our study, 63% of patients were admitted to the hospital. This admission rate did not increase during the study period but is higher than the 54% admission rate reported previously derived from data acquired from 1991 to 2001 [3]. We found that the type of insurance and the socioeconomic status of patients play an important role in the decision to admit patients to the hospital after TIA. Patients with primary health insurance of Medicare and Medicaid compared with private and other payment methods and with higher median annual income were more likely to be admitted to the hospital. If the patient was presenting to a metropolitan teaching or nonteaching hospital, the odds of admission to the hospital were statistically higher compared with other nonmetropolitan hospitals. This finding is contrary to another study of TIA predictors for hospital admission [3]. Patients with higher underlying comorbid conditions, for example, diabetes mellitus, hypertension, atrial fibrillation, and cigarette smoking were more likely to be admitted to the hospital.

Our study included TIA patients presenting to the ED between

Private including health maintenance

organization/other

Medicare

1.19

(1.14-1.26)

Medicaid

1.42

(1.33-1.52)

Median household income by zip code

$1-$48999

Reference

$49000-$63999

1.08

(1.02-1.34)

>=$64000

1.33

(1.22-1.44)

Reference

2006 and 2008. Subsequent studies that were published afterward including a study that reported that patients with TIA may benefit from hospital admission with the intent to reduce the risk of subsequent stroke and Associated costs in 2009 [28]. Therefore, the impact of more recent studies on TIA admission patterns may not be reflected in our results. Such reduction is attributed to early Diagnostic investigations and timely initiation of stroke preventive measures such as vascular risk factors control and lifestyle modification [3]. In 1

admitted to the hospital, whereas 296071 (36.5%) were discharged from the ED to home (Table 1). Mean age in years (+-SD) and proportion of women were higher for patients admitted to the hospital compared with those who were discharged home after ED visit: 70.5 +- 14.6 vs 69.7 +- 15.3, P b .0001, and 58.8% vs 56.5%, P b

.0001, respectively. The proportion of patients with hypertension (74.8% vs 49.4%, P b .0001), diabetes mellitus (28.7% vs 18.2%, P b

.0001), atrial fibrillation (13.3% vs 5.8%, P b .0001), and cigarette smoking (10.9% vs 7.2%, P b .0001) were higher in those admitted to the hospital compared with those who were discharged after ED visit. Patients with Medicare and Medicaid insurance were more likely to be admitted to the hospital as compared with those with private medical insurance and other Insurance types (odds ratio [OR], 1.25 [95% CI, 1.20-1.30] and OR, 1.49 [95% C.I. 1.39-1.60],

respectively). Patients presenting to metropolitan teaching hospital EDs were more likely to be admitted to the hospital (36.4% vs 27.6%, P b .0001). Patients with higher socioeconomic status determined by median income by zip code and those presenting to the ED of hospital with trauma facilities were more likely to be admitted (22.7% vs 17.9%; P b .0001; OR, 1.39) and (30.9% vs

27.7%; P b .0001; OR, 1.16), respectively.

In the multivariate analysis, after adjusting for age, gender, medical co-morbidities, and hospital characteristics, the predictors for admission to the hospital were history of diabetes mellitus (OR, 1.47; 95% CI, 1.43-1.52), atrial fibrillation (OR, 2.52; 95% CI, 2.39-2.65),

and hypertension (OR, 2.84; 95% CI, 2.73-2.96) (see Table 2). Primary health insurance (Medicare OR, 1.19; 95% CI, 1.14-1.26) (Medicaid OR, 1.42; 95% CI, 1.33-1.52) and socioeconomic status were independent predictors of hospital admission (annual income >=64000; OR, 1.33; 95% CI, 1.22-1.44). Hospital location and teaching status were also independent predictors of admission to the hospital following TIA (metropolitan nonteaching OR, 1.75; 95% CI, 1.57-1.94) (metropolitan teaching OR, 2.17; 95% CI, 1.90-2.48).

Discussion

We estimated the national annual incidence of TIA to be 270969, an estimate that is consistent with previous estimates

study, the rate of subsequent stroke was significantly higher for patients diagnosed with TIA in the ED with no subsequent hospital admission [25]. Findings supporting better outcomes related to early comprehensive management have been reported in 2 recent studies from the United Kingdom [8] and France [29]. Referral to a rapidly accessible and comprehensive TIA service or clinic, possibly guided by risk stratification, may be an alternative to hospital admission if admission is not feasible [8,29].

We are unable to determine whether the proposed benefit is related to the hospitalization or whether hospitalization is simply a marker of rapid patient assessment and care. There have been no large randomized study to illustrate the risks or benefits of hospitalization for TIA. Our study demonstrates that factors unrelated to patient’s clinical condition such as insurance status, hospital teaching status, and socioeconomic status all play an important role in the decision-making process to admit TIA patients to the hospital. If hospital admission is considered a quality marker in patients presenting to ED with TIA based on future studies, steps may be required to reduce disparities in implementation of such practice.

There are several issues related to Data interpretation in our study. The NEDS is a large-sized data set, which has been used previously [30-32] to provide national estimates of US hospital- based EDs. We might be underestimating the rates of TIA occurrence in the United States as not all patients with TIA present to the ED, as studies illustrate that some TIA patients present to primary care physicians and not necessarily to EDs, which are equipped to deliver rapid stroke assessment [24,33]. Furthermore, the database does not include conventional severity scales including ABCD score or National Institutes of Health Stroke Scale score, diagnostic study results, time spent in the ED, and confounding factors influencing decision making such as the level of neurologic expertise of the ED staff and perceived general health status of the patient. We have attempted to adjust for comorbidities to extent permissible by the design of the database in our multivariate model to minimize confounding errors. Accuracy of medical record coding of TIA as a diagnosis is another potential confounder, although these codes have been used in several studies in the past to obtain national estimates [3,26].

S.A. Chaudhry et al. / American Journal of Emergency Medicine 31 (2013) 516519 519

Conclusion

Our study found that factors unrelated to the patient’s medical condition, such as insurance, hospital type, and socioeconomic status have an important impact on the decision-making process to admit TIA patients presenting to the ED. These results support further investigation of variations in ED management of TIA, with particular attention to the influence of socioeconomic status in the decision- making process.

Acknowledgments

This study was performed independently of any financial support. None of the authors have any conflict of interest to disclose and there are no financial conflicts to disclose.

References

  1. Kleindorfer D, Panagos P, Pancioli A, et al. Incidence and short-term prognosis of transient ischemic attack in a population-based study. Stroke 2005;36:720-3.
  2. Johnston SC, Fayad PB, Gorelick PB, et al. Prevalence and knowledge of transient ischemic attack among US adults. Neurology 2003;60:1429-34.
  3. Edlow JA, Kim S, Pelletier AJ, Camargo Jr CA. National study on emergency department visits for transient ischemic attack, 1992-2001. Acad Emerg Med 2006;13:666-72.
  4. Lovett JK, Dennis MS, Sandercock PA, Bamford J, Warlow CP, Rothwell PM. Very early risk of stroke after a first transient ischemic attack. Stroke 2003;34:e138-40.
  5. Gladstone DJ, Kapral MK, Fang J, Laupacis A, Tu JV. Management and outcomes of transient ischemic attacks in Ontario. CMAJ 2004;170:1099-104.
  6. Eliasziw M, Kennedy J, Hill MD, Buchan AM, Barnett HJ. Early risk of stroke after a transient ischemic attack in patients with Internal carotid artery disease. CMAJ 2004;170:1105-9.
  7. Johnston SC, Gress DR, Browner WS, Sidney S. Short-term prognosis after emergency department diagnosis of TIA. JAMA 2000;284:2901-6.
  8. Rothwell PM, Giles MF, Chandratheva A, et al. Effect of Urgent treatment of transient ischaemic attack and minor stroke on early recurrent stroke (EXPRESS study): a prospective population-based sequential comparison. Lancet 2007;370: 1432-42.
  9. Luengo-Fernandez R, Gray AM, Rothwell PM. Effect of urgent treatment for transient ischaemic attack and minor stroke on disability and hospital costs (EXPRESS study): a prospective population-based sequential comparison. Lancet Neurol 2009;8:235-43.
  10. Shah KH, Kleckner K, Edlow JA. Short-term prognosis of stroke among patients diagnosed in the emergency department with a transient ischemic attack. Ann Emerg Med 2008;51:316-23.
  11. Hill MD, Yiannakoulias N, Jeerakathil T, Tu JV, Svenson LW, Schopflocher DP. The high risk of stroke immediately after transient ischemic attack: a population- based study. Neurology 2004;62:2015-20.
  12. Rothwell PM, Giles MF, Flossmann E, et al. A simple score (ABCD) to identify individuals at high early risk of stroke after transient ischaemic attack. Lancet 2005;366:29-36.
  13. Johnston SC, Rothwell PM, Nguyen-Huynh MN, et al. Validation and refinement of scores to predict very early stroke risk after transient ischaemic attack. Lancet 2007;369:283-92.
  14. Coutts SB, Sylaja PN, Choi YB, et al. The ASPIRE approach for TIA risk stratification. Can J Neurol Sci 2011;38:78-81.
  15. Carpenter CR, Keim SM, Crossley J, Perry JJ. Post-transient ischemic attack early stroke stratification: the ABCD(2) prognostic aid. J Emerg Med 2009;36:194-8 [discussion 8-200].
  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. Albers GW, Hart RG, Lutsep HL, Newell DW, Sacco RL. AHA Scientific Statement. Supplement to the guidelines for the management of transient ischemic attacks: a statement from the Ad Hoc Committee on Guidelines for the Management of Transient Ischemic Attacks, Stroke Council, American Heart Association. Stroke 1999;30:2502-11.
  18. Culebras A, Kase CS, Masdeu JC, et al. Practice guidelines for the use of imaging in transient ischemic attacks and acute stroke. A report of the Stroke Council, American Heart Association. Stroke 1997;28:1480-97.
  19. Perry JJ, Sharma M, Sivilotti ML, et al. Prospective validation of the ABCD2 score for patients in the emergency department with transient ischemic attack. CMAJ 2011;183:1137-45.
  20. Shah KH, Metz HA, Edlow JA. Clinical prediction rules to stratify Short-term risk of stroke among patients diagnosed in the emergency department with a transient ischemic attack. Ann Emerg Med 2009;53:662-73.
  21. Chang E, Holroyd BR, Kochanski P, Kelly KD, Shuaib A, Rowe BH. Adherence to practice guidelines for transient ischemic attacks in an emergency department. Can J Neurol Sci 2002;29:358-63.
  22. Johnston SC, Smith WS. Practice variability in management of transient ischemic attacks. Eur Neurol 1999;42:105-8.
  23. Volpato S, Maraldi C, Ble A, et al. Prescription of antithrombotic therapy in older patients hospitalized for transient ischemic attack and ischemic stroke: the GIFA study. Stroke 2004;35:913-7.
  24. Goldstein LB, Bian J, Samsa GP, Bonito AJ, Lux LJ, Matchar DB. New transient ischemic attack and stroke: outpatient management by primary care physicians. Arch Intern Med 2000;160:2941-6.
  25. Kehdi EE, Cordato DJ, Thomas PR, et al. Outcomes of patients with transient ischaemic attack after hospital admission or discharge from the emergency department. Med J Aust 2008;189:9-12.
  26. National Vital Statistics Report. (Accessed at http://www.cdc.gov/nchs/data/nvsr/ nvsr49/nvsr49_02.pdf.).
  27. Whisnant JP, Matsumoto N, Elveback LR. Transient cerebral ischemic attacks in a community. Rochester, Minnesota, 1955 through 1969. Mayo Clin Proc 1973;48: 194-8.
  28. Gommans J, Barber PA, Fink J. Preventing strokes: the assessment and management of people with transient ischaemic attack. N Z Med J 2009;122: 3556.
  29. Lavallee PC, Meseguer E, Abboud H, et al. A transient ischaemic attack clinic with round-the-clock access (SOS-TIA): feasibility and effects. Lancet Neurol 2007;6: 953-60.
  30. Wang D, Zhao W, Wheeler K, Yang G, Xiang H. Unintentional fall injuries among US children: a study based on the National Emergency Department Sample. Int J Inj Contr Saf Promot 2012.
  31. Allareddy V, Nalliah RP, Rampa S, Kim MK. Firearm related injuries amongst children: estimates from the nationwide emergency department sample. Injury 2011.
  32. Platts-Mills TF, Hunold KM, Esserman DA, Sloane PD, McLean SA. Motor vehicle collision-related emergency department visits by older adults in the United States. Acad Emerg Med 2012;19:821-7.
  33. Nguyen-Huynh MN, Fayad P, Gorelick PB, Johnston SC. Knowledge and management of transient ischemic attacks among US primary care physicians. Neurology 2003;61:1455-6.