predictors of early hospital readmission for asthma among inner-city children

4
Journal of Asthma, 43:37–40, 2006 Copyright C 2006 Taylor & Francis Group, LLC ISSN: 0277-0903 print / 1532-4303 online DOI: 10.1080/02770900500446997 ORIGINAL ARTICLE Predictors of Early Hospital Readmission for Asthma Among Inner-City Children MARINA REZNIK, M.D., M.S., 1 SUSAN M. HAILPERN, B.S.N., M.P.H., M.S., 2 AND PHILIP O. OZUAH, M.D., PH.D. 1,1 Children’s Hospital at Montefiore, Department of Pediatrics, Albert Einstein College of Medicine, Bronx, New York, USA 2 Albert Einstein College of Medicine, Department of Epidemiology and Population Health, Bronx, New York, USA Factors associated with early asthma readmission have not been fully studied. To identify predictors of early readmission, we performed a matched case-control study of children discharged with primary diagnosis of asthma. Cases were readmitted with asthma within 30 days of discharge. Controls were not readmitted. Conditional logistic regression analysis was used. History of asthma hospitalization within the past 12 months was an independent predictor of early readmission (OR 1.89, p = 0.021). Modifiable factors such as medical treatment and management during and upon discharge from the index admission did not predict early asthma readmission. Keywords asthma, modifiable predictors, early readmission, inner-city, children INTRODUCTION Asthma is one of the most frequent causes of hospital ad- missions among children (1). It accounts for almost 200,000 childhood hospitalizations per year (2) and 1.6 billion dol- lars in annual medical costs for children in the United States (3). Asthma readmissions contribute substantially to this bur- den. Children residing in New York City have the highest rates in the nation for hospitalization related to asthma, and the Bronx, in particular, has the highest rates of asthma-related hospitalization and death (4). Twenty to fifty percent of children hospitalized for asthma will be readmitted with the same diagnosis within the fol- lowing year (5–8). This late readmission has been related to demographic factors (7, 9), asthma severity at index admis- sion (9), and a history of previous hospitalizations for asthma (7). A number of children discharged with asthma will be readmitted early, within 30 days of discharge. Factors associ- ated with early asthma readmission have not been fully stud- ied. Identifying these factors, especially if they are amendable to change, would allow physicians to recognize patients at risk for early readmission and enhance their asthma manage- ment. Thus, the aim of this study was to identify modifiable predictors of early readmission in inner-city children with asthma at one children’s hospital. METHODS Design and Setting We conducted a matched case-control study of a cohort of children hospitalized for asthma at the Children’s Hos- pital at Montefiore (CHAM), Bronx, New York, between Presented in part at the Eastern Society for Pediatric Research Annual Meeting, Old Greenwich, Connecticut, USA, March 2005 and Pediatric Academic Societies’ Annual Meeting, Washington, DC, USA, May 2005. Corresponding author: Philip O. Ozuah, M.D., Ph.D., Children’s Hospital at Montefiore, 3415 Bainbridge Avenue, Bronx, NY 10467; E-mail: pozuah@montefiore.org January 1998 and December 2004, where the prevalence of early asthma readmission is three to four percent. Cases were defined as children hospitalized with asthma exacerbation and readmitted with the same diagnosis within 30 days of discharge from the index admission. Controls were children hospitalized for asthma but not readmitted within 30 days of discharge. Index admission was defined as the subject’s first hospitalization for asthma during the study period. Subjects We used computerized health records to identify all pedi- atric patients (0–21 years of age) discharged with a primary diagnosis of asthma (International Classification of Diseases, Ninth Revision, ICD-9 493.0) during the study period. The electronic records allowed us to retrieve data on subject’s gender, race, date of birth, medical record number, discharge date of the index admission, and occurrence of asthma read- mission within a 30-day period. If a case had more than one early readmission during the study period, the first readmis- sion within 30 days of discharge from the index admission was selected; thus, each observation in the data represented an individual child. Of those children discharged with a primary diagnosis of asthma during the study period (n = 5, 104), 173 subjects met the eligibility criteria for early readmission and were identified as cases. Of these, medical charts were unavailable for 7 cases, and 14 cases had more than one early readmission during the study period and were excluded. Analysis included 152 cases that were eligible for study entry. We selected up to two controls from the cohort of remain- ing 4,931 children discharged from CHAM with a primary diagnosis of asthma but not readmitted within 30 days of in- dex admission. We identified 293 controls to match to our 152 cases. Cases and controls were matched on age (±2 years), gender, race, season, and year of index admission. Season was defined as winter (December, January, and February), spring (March, April, and May), summer (June, July, and August) 37 J Asthma Downloaded from informahealthcare.com by Michigan University on 10/30/14 For personal use only.

Upload: philip-o

Post on 02-Mar-2017

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Predictors of Early Hospital Readmission for Asthma Among Inner-City Children

Journal of Asthma, 43:37–40, 2006Copyright C© 2006 Taylor & Francis Group, LLCISSN: 0277-0903 print / 1532-4303 onlineDOI: 10.1080/02770900500446997

ORIGINAL ARTICLE

Predictors of Early Hospital Readmission for Asthma AmongInner-City Children

MARINA REZNIK, M.D., M.S.,1 SUSAN M. HAILPERN, B.S.N., M.P.H., M.S.,2 AND PHILIP O. OZUAH, M.D., PH.D.1,∗

1Children’s Hospital at Montefiore, Department of Pediatrics, Albert Einstein College of Medicine, Bronx, New York, USA2Albert Einstein College of Medicine, Department of Epidemiology and Population Health, Bronx, New York, USA

Factors associated with early asthma readmission have not been fully studied. To identify predictors of early readmission, we performed a matched

case-control study of children discharged with primary diagnosis of asthma. Cases were readmitted with asthma within 30 days of discharge. Controls

were not readmitted. Conditional logistic regression analysis was used. History of asthma hospitalization within the past 12 months was an independent

predictor of early readmission (OR 1.89, p = 0.021). Modifiable factors such as medical treatment and management during and upon discharge from

the index admission did not predict early asthma readmission.

Keywords asthma, modifiable predictors, early readmission, inner-city, children

INTRODUCTION

Asthma is one of the most frequent causes of hospital ad-missions among children (1). It accounts for almost 200,000childhood hospitalizations per year (2) and 1.6 billion dol-lars in annual medical costs for children in the United States(3). Asthma readmissions contribute substantially to this bur-den. Children residing in New York City have the highest ratesin the nation for hospitalization related to asthma, and theBronx, in particular, has the highest rates of asthma-relatedhospitalization and death (4).

Twenty to fifty percent of children hospitalized for asthmawill be readmitted with the same diagnosis within the fol-lowing year (5–8). This late readmission has been related todemographic factors (7, 9), asthma severity at index admis-sion (9), and a history of previous hospitalizations for asthma(7). A number of children discharged with asthma will bereadmitted early, within 30 days of discharge. Factors associ-ated with early asthma readmission have not been fully stud-ied. Identifying these factors, especially if they are amendableto change, would allow physicians to recognize patients atrisk for early readmission and enhance their asthma manage-ment. Thus, the aim of this study was to identify modifiablepredictors of early readmission in inner-city children withasthma at one children’s hospital.

METHODS

Design and SettingWe conducted a matched case-control study of a cohort

of children hospitalized for asthma at the Children’s Hos-pital at Montefiore (CHAM), Bronx, New York, between

Presented in part at the Eastern Society for Pediatric Research AnnualMeeting, Old Greenwich, Connecticut, USA, March 2005 and PediatricAcademic Societies’ Annual Meeting, Washington, DC, USA, May 2005.

∗Corresponding author: Philip O. Ozuah, M.D., Ph.D., Children’sHospital at Montefiore, 3415 Bainbridge Avenue, Bronx, NY 10467;E-mail: [email protected]

January 1998 and December 2004, where the prevalence ofearly asthma readmission is three to four percent. Cases weredefined as children hospitalized with asthma exacerbationand readmitted with the same diagnosis within 30 days ofdischarge from the index admission. Controls were childrenhospitalized for asthma but not readmitted within 30 days ofdischarge. Index admission was defined as the subject’s firsthospitalization for asthma during the study period.

SubjectsWe used computerized health records to identify all pedi-

atric patients (0–21 years of age) discharged with a primarydiagnosis of asthma (International Classification of Diseases,Ninth Revision, ICD-9 493.0) during the study period. Theelectronic records allowed us to retrieve data on subject’sgender, race, date of birth, medical record number, dischargedate of the index admission, and occurrence of asthma read-mission within a 30-day period. If a case had more than oneearly readmission during the study period, the first readmis-sion within 30 days of discharge from the index admissionwas selected; thus, each observation in the data representedan individual child.

Of those children discharged with a primary diagnosis ofasthma during the study period (n = 5, 104), 173 subjectsmet the eligibility criteria for early readmission and wereidentified as cases. Of these, medical charts were unavailablefor 7 cases, and 14 cases had more than one early readmissionduring the study period and were excluded. Analysis included152 cases that were eligible for study entry.

We selected up to two controls from the cohort of remain-ing 4,931 children discharged from CHAM with a primarydiagnosis of asthma but not readmitted within 30 days of in-dex admission. We identified 293 controls to match to our 152cases. Cases and controls were matched on age (±2 years),gender, race, season, and year of index admission. Season wasdefined as winter (December, January, and February), spring(March, April, and May), summer (June, July, and August)

37

J A

sthm

a D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/30/

14Fo

r pe

rson

al u

se o

nly.

Page 2: Predictors of Early Hospital Readmission for Asthma Among Inner-City Children

38 M. REZNIK ET AL.

and fall (September, October, and November). If more thantwo matched controls were identified per case, controls withthe discharge date of the index admission closest to that of thecase were chosen. The institutional review board at Monte-fiore Medical Center, Bronx, New York, approved this study.

Data CollectionOne of the authors (M.R.) performed medical chart reviews

to abstract clinical and demographic data. Written recordsfrom the Emergency Department (ED), index hospital ad-mission, and discharge were also used. The variables of in-terest were chosen on the basis of literature review (5–10).Information was collected in the following four categories:demographics, variables related to past asthma history, vari-ables related to the index admission, and variables related todischarge from the index admission.

Demographic data, including age at index admission, gen-der, race, and medical insurance (Medicaid-recipients, otherinsurance, or no insurance) were collected. Variables relatedto past asthma history included a history of asthma-relatedED visits and admissions to the hospital in the past 12 months,a prior history of an intensive care unit (ICU) admission forasthma, being premature at birth, prescription of inhaled cor-ticosteroids (ICS) before the index admission as per parentalreport of medications taken at home, and exposure to envi-ronmental triggers such as roaches and tobacco smoke.

Variables related to the index admission included oxy-gen saturation level on presentation to the ED as mea-sured by pulse oximetry, length of hospital stay (LOS) indays, therapeutic management of asthma, including sub-cutaneous injection of epinephrine or intravenous (IV) in-jection of magnesium sulfate, and number and frequencyof aerosolized beta-agonist treatments, oxygen requirementduring the index admission, and receipt of a pulmonaryconsultation as initiated by the hospital physicians or sub-ject’s primary medical doctor (PMD). Variables related todischarge from the index admission included last recordedoxygen saturation level in the medical chart before dis-charge, presence of wheezing, and prescription of ICS as re-ported in the physician’s discharge note and nurse’s dischargeinstructions.

Statistical AnalysisData were maintained in SPSS version 11.5 (SPSS Inc,

Chicago, IL) and STATA version 8.2 (STATA, College Sta-tion, TX) statistical softwares. Chi-square tests were per-formed to test for differences in categorical matching char-acteristics. For continuous variables, we used generalizedestimating equations (GEE) statistics to test for differ-ences in matched characteristics. GEE are methods of pa-rameter estimation well suited for the analysis of corre-lated data, as is the case in matched case-control studies(11, 12).

We performed conditional logistic regression analyses formatched data to test for the association of independent vari-ables with early readmission status. Regression analyses fol-lowed model building strategies suggested by Hosmer andLemeshow (13). Univariate analysis was performed for eachvariable under consideration. Variables with a p value <0.25in univariate analysis were initially entered into a multivari-

ate model. The final model included all clinically importantand/or statistically significant variables. Conditional logisticregression models were run for only those cases and controlswithout missing data (n = 324).

LOS and receipt of subcutaneous epinephrine or IV mag-nesium sulfate were used as surrogate measures of asthmaseverity during the index admission. Prescription of ICS,presence of wheezing, and oxygen saturation level on dis-charge from the index admission served as markers of asthmaseverity on discharge. We retained LOS, receipt of subcuta-neous epinephrine or IV magnesium sulfate, prescription ofICS, presence of wheezing, and oxygen saturation level ondischarge in the final model to adjust for asthma severityduring and on discharge from the index admission.

RESULTS

Medical records were available and reviewed for 152 casesand 293 matched controls. Cases and controls were success-fully matched (age p = 0.89; gender p = 0.75; race p =0.96; season of index admission p = 0.99). Demographiccharacteristics of the entire cohort included: 61.0% male,62.2% Hispanic, 34.2% African American, 3.6% other races;53.7% were Medicaid-recipients and 4.7% had no insurance.Mean age of the cohort was 5.99 years ± 5.18 (range .20–20.30). The majority of the index admissions occurred duringthe fall (39.6%). The mean duration between admission dateof a readmission and discharge date of the index admissionfor cases was 15.8 ± 8.5 days (range 0.0–30.0), and the meanLOS for the index admission was 3.1 ± 1.8 days (range .5–12.0).

Individual predictors of early readmission revealed thatcases were more likely to have been hospitalized for asthmain the past 12 months (OR 2.22, 95% CI 1.40–3.50), tohave visited the ED for asthma in the past 12 months (OR3.28, 95% CI 1.55–6.94), to have a history of an ICU ad-mission for asthma (OR 2.18, 95% CI 1.26–3.78), to havereceived a pulmonary consultation during the index admis-sion (OR 1.87, 95% CI 1.12–3.10), and to have been pre-scribed ICS before the index admission (OR 1.61, 95% CI1.02–2.52). We found no significant difference between casesand controls for history of prematurity at birth, LOS, re-ceipt of subcutaneous epinephrine or IV magnesium sulfateduring the index admission, exposure to environmental trig-gers, prescription of ICS on discharge from the index admis-sion, oxygen saturation level on presentation for the indexadmission or discharge, the need for oxygen supplementa-tion during the index admission, and presence of wheezingat discharge (Table 1). After adjusting for asthma severityduring and upon discharge from the index admission, a his-tory of an asthma-related hospital admission within the past12 months was found to be an independent predictor ofearly readmission in a conditional logistic regression model(Table 2).

DISCUSSION

This is the first study to evaluate the predictors of earlyasthma readmission within 30 days of discharge in inner-city Bronx children. In our study population, history ofasthma hospitalization in the past 12 months was associatedwith early readmission. Modifiable factors, such as medical

J A

sthm

a D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/30/

14Fo

r pe

rson

al u

se o

nly.

Page 3: Predictors of Early Hospital Readmission for Asthma Among Inner-City Children

PREDICTORS OF EARLY READMISSION FOR PEDIATRIC ASTHMA 39

TABLE 1.—Odds ratios for individual predictors of early asthma readmission.∗

Predictors OR ( 95% CI) P value

Hospital admission in the past 12 months 2.22 (1.40–3.50) 0.001ED visit in the past 12 months 3.28 (1.55–6.94) 0.002Prior ICU admission 2.18 (1.26–3.78) 0.005Pulmonary consultation during index admission 1.87 (1.12–3.10) 0.016Prescription of ICS before index admission 1.61 (1.02–2.52) 0.039Prematurity at birth 1.72 (.85–3.47) 0.129LOS 1.08 (.97–1.21) 0.221†

Roaches at home 1.36 (.85–2.17) 0.195Cigarette exposure at home 0.81 (.54–1.21) 0.306Receipt of subcutaneous epinephrine or IV

magnesium sulfate on index admission1.49 (.85–2.60) 0.162

Prescription of ICS on discharge 1.35 (.85–2.14) 0.200Presence of wheezing on discharge 0.98 (.64–1.51) 0.942Oxygen saturation on presentation for the index

admission0.99 (.93–1.05) 0.727†

Oxygen saturation on discharge 1.00 (.87–1.15) 0.917†

Oxygen requirement during index admission 1.06 (.67–1.66) 0.816

OR, odds ratio; CI, confidence interval; ED, Emergency Department; ICU, intensive careunit; ICS, inhaled corticosteroids; LOS, length of hospital stay in days; IV, intravenous.

∗OR, 95% CI, and p values obtained by conditional logistic regression, unless specifiedotherwise.

† p values obtained from GEE statistics.

treatment and asthma management during and on dischargefrom the index admission, were not associated with earlyreadmission.

To date, most studies of asthma readmissions have focusedon predictors of late readmissions for asthma, within one yearof discharge from the index admission (5–10). In a nestedcase-control study, Minkovitz et al. (8) found that childrenreadmitted for asthma within one year of index admissionwere more likely to receive a pulmonary consultation duringthe index admission and to have preventive asthma medica-tions at home. Our findings support this report.

A retrospective study of predictors of late asthma read-mission in 1,034 New Zealand children revealed that latereadmission was associated with severity of index admis-sion as measured by the physician’s subjective assessmentof severity and receipt of intravenous therapy (9). IV magne-sium sulfate and subcutaneous epinephrine have been used inthe management of moderate to severe acute asthma attacksin children with improvement in clinical status (14–16). Wedid not find a significant association between early readmis-sion and severity of the index admission as measured by thereceipt of subcutaneous epinephrine or IV magnesium sul-fate. Although, given the size of the point estimate, such anassociation could not be ruled out.

TABLE 2.—Multivariate analysis of predictors of early asthma readmission.∗

Predictors OR (95% CI) P value

Hospital admission in the past 12 months 1.89 (1.10 – 3.25) 0.021Pulmonary consultation during admission 2.05 (.93 – 4.52) 0.075Prior ICU admission 1.99 (.93 – 4.27) 0.075Prescription of ICS before index admission 1.32 (.64 – 2.70) 0.449LOS 1.06 (.91 – 1.24) 0.440Receipt of subcutaneous epinephrine or IVmagnesium sulfate on index admission

1.47 (.73 – 2.94) 0.280

Prescription of ICS on discharge 0.66 (.31 – 1.38) 0.271Presence of wheezing on discharge 1.22 (.73 – 2.03) 0.456Oxygen saturation on discharge 0.88 (.75 – 1.04) 0.131

OR, odds ratio; CI, confidence interval; ICU, intensive care unit; ICS, inhaled corticos-teroids; LOS, length of hospital stay in days; IV, intravenous.

∗OR, adjusted, 95% CI and p values for the final model were obtained by conditionallogistic regression (n = 324).

The finding that cases were more likely to have a prior his-tory of an ICU admission for asthma is intuitive. An admis-sion to an ICU for asthma has been shown to be a marker ofasthma severity and a risk factor for asthma death (17). Caseswere also more likely to have visited the ED for asthma in thepast 12 months. Although statistically significant in univari-ate analysis, a visit to the ED within the past 12 months wasnot significant in multivariate analysis and was eliminated inthe final regression model. Given that inner-city children fre-quently use the ED for asthma as their primary source of care(18), an ED visit for asthma may not be the best individualpredictor of early readmission for this inner-city population.

We found that children with early readmission were sig-nificantly more likely to have been prescribed ICS beforethe index admission. An earlier study of predictors of lateasthma readmission found similar results (8). One possibleexplanation of our counterintuitive finding may be that be-ing prescribed ICS merely suggests a presence of persistentasthma and greater disease severity. In fact, National AsthmaEducation and Prevention Program (19) recommends inhaledanti-inflammatory medications for the treatment of persistentasthma, supporting our conclusion that these children mayhave had more severe disease.

Prescription of ICS on discharge from the index admissionwas not associated with readmission status. While it might beexpected that being prescribed ICS on discharge would havereduced the likelihood of early asthma readmission, we didnot find this association in our study. It is possible that thosewith early readmission did not fill the prescription for ICS orwere not adherent to the discharge treatment plan.

We found that a history of hospital admission for asthmain the past 12 months independently predicted early readmis-sion. A study of people with asthma, 5 to 45 years, of age inNew Zealand revealed that an asthma admission in the past12 months was associated with late readmission and death(20). Asthma-related hospitalization in the past 12 months,in addition to being a measure of disease severity, may alsorepresent one’s ability to access health care.

Several studies have looked at disparities of asthma admis-sions and readmissions by gender, age, ethnicity, and seasonalvariation (10, 21–23). In our study, subject selection was notlimited to any specific age group. Although factors associ-ated with readmission for preschool and school-aged childrenwith asthma may differ from those associated with adoles-cents and young adults, we controlled for these differences bymatching cases and controls on age. We also matched casesand controls on gender, race, season, and year of index admis-sion since these were nonmodifiable variables of no interestto our study.

Some limitations to our study should be noted. Clinicaldata were abstracted from medical charts and were limitedto the information documented in the charts. For example,parental education, baseline asthma severity, and socioeco-nomic status (with the exception of insurance status) werenot consistently recorded and could not be studied. Factorsother than asthma severity on presentation that may have in-fluenced the decision to admit or readmit were not studied.These factors may have included family psychosocial status,time of day on presentation to the hospital, and living dis-tance from the hospital. We did not measure adherence to theprescribed preventive asthma medications before and after

J A

sthm

a D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/30/

14Fo

r pe

rson

al u

se o

nly.

Page 4: Predictors of Early Hospital Readmission for Asthma Among Inner-City Children

40 M. REZNIK ET AL.

the discharge from the index admission due to inconsistencyof reporting of this factor in the medical records. In addition,medical chart review and unavailability of pharmacy recordsdid not allow us to determine if cases had their prescriptionsfilled. A prospective study measuring adherence to ICS wouldlikely reveal an inverse association between medication com-pliance and early asthma readmission. Finally, our study pop-ulation was composed of primarily inner-city, low-income,and minority children hospitalized at a single children’s hos-pital. Thus, our results may not be generalizable to popu-lations of different ethnic composition and socioeconomicstatus.

Despite these limitations, our findings have implicationsfor the care of inner-city asthmatic children. Early asthmareadmission occurred among a cohort of children with greaterdisease severity and can therefore be predicted. History ofasthma-related admission or ED visits in the past 12 months,prior ICU admission, receipt of pulmonary consultation, andbeing prescribed ICS before the index admission may iden-tify children who are more likely to be readmitted. Recogniz-ing this population at risk for early readmission would allowphysicians to enhance the discharge planning with close out-patient follow-up and develop effective preventive interven-tions. Shifting asthma management to the ambulatory caresetting may improve asthma-related health outcomes by de-creasing admissions and preventing early readmissions.

ACKNOWLEDGMENTS

The authors would like to thank Arthur E. Blank, PhD(Albert Einstein College of Medicine), Hillel W. Cohen,DrPH (Albert Einstein College of Medicine), and Peter A.Selwyn, MD, MPH (Montefiore Medical Center) for theirreview and thoughtful feedback.

REFERENCES

1. Akinbami LJ, Schoendorf KC. Trends in childhood asthma: prevalence,

health care utilization, and mortality. Pediatrics 2002; 110:315–322.

2. Mannino DM, Homa DM, Akinbami LJ, Moorman JE, Gwynn C, Redd

SC. Surveillance for asthma–United States, 1980–1999. MMWR Surveill

Summ 2002; 29:1–13.

3. Smith DH, Malone DC, Lawson KA, Okamoto LJ, Battista C, Saunders

WB. A national estimate of the economic costs of asthma. Am J Respir Crit

Care Med 1997; 156:787–793.

4. New York City Department of Health Childhood Asthma Initiative. Asthma

Facts, Second Edition, 2003. http://www.nyc.gov/html/doh/downloads/

pdf/asthma/sfacts.pdf. Accessed July 8, 2005.

5. Macarthur C, Calpin C, Parkin PC, Feldman W. Factors associated with

pediatric asthma readmissions. J Allergy Clin Immunol 1996; 98:992–993.

6. Bloomberg GR, Trinkaus KM, Fisher EB, Musick JR, Strunk RC. Hospital

readmissions for childhood asthma: a 10-Year Metropolitan Study. Am J

Respir Crit Care Med 2003; 167:1068–1076.

7. Farber HJ. Risk of readmission to hospital for pediatric asthma. J Asthma

1998; 35:95–99.

8. Minkovitz CS, Andrews JS, Serwint JR. Rehospitalization of children with

asthma. Arch Pediatr Adolesc Med 1999; 153:727–730.

9. Mitchell EA, Bland JM, Thompson JMD. Risk factors for readmission to

hospital for asthma in childhood. Thorax 1994; 49:33–36.

10. Chen Y, Dales R, Stewart P, Johansen H, Scott G, Taylor G. Hospital read-

missions for asthma in children and young adults in Canada. Pediatr Pul-

monol 2003; 36:22–26.

11. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous

outcomes. Biometrics 1986; 42:121–130.

12. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear

models. Biometrika 1986; 73:13–22.

13. Hosmer D, Lemeshow S. Applied Survival Analysis. New York: John Wiley

& Sons, 1999.

14. Lin YZ, Hsieh KH, Chang LF, Chu CY. Terbutaline nebulization and

epinephrine injection in treating acute asthmatic children. Pediatr Allergy

Immunol 1996; 7:95–99.

15. Gurkan F, Haspolat K, Bosnak M, Dikici B, Derman O, Ece A. Intravenous

magnesium sulphate in the management of moderate to severe acute asth-

matic children nonresponding to conventional therapy. Eur J Emerg Med

1999; 6:201–205.

16. Ciarallo L, Brousseau D, Reinert S. Higher-dose intravenous magnesium

therapy for children with moderate to severe acute asthma. Arch Pediatr

Adolesc Med 2000; 154:979–983.

17. Belessis Y, Dixon S, Thomsen A, Duffy B, Rawlinson W, Henry R, Morton

J. Risk factors for an intensive care unit admission in children with asthma.

Pediatr Pulmonol 2004; 37:201–209.

18. Farber HJ, Johnson C, Beckerman RC. Young inner-city children visiting the

emergency room (ER) for asthma: risk factors and chronic care behaviors.

J Asthma 1998; 35:547–552.

19. NIH, National Asthma Education and Prevention Program (National Heart,

Lung, and Blood Institute). Second Expert Panel on the Management

of Asthma. Expert Panel Report 2: Guidelines for the Diagnosis and

Management of Asthma. National Institutes of Health: Bethesda, MD,

1997.

20. Crane J, Pearce N, Burgess C, Woodman K, Robson B, Beasley R.

Markers of risk of asthma death or readmission in the 12 months fol-

lowing a hospital admission for asthma. Int J Epidemiol 1992; 21:737–

744.

21. Wallace JC, Denk CE, Kruse LK. Pediatric hospitalizations for asthma: use

of a linked file to separate person-level risk and readmission. Prev Chronic

Dis 2004; 1:A07.

22. Rushworth RL, Rob MI. Readmissions to hospital: the contribution of mor-

bidity data to the evaluation of asthma management. Aust J Public Health

1995; 19:363–367.

23. Kimes D, Levine E, Timmins S, Weiss SR, Bollinger ME, Blaisdell C.

Temporal dynamics of emergency department and hospital admissions of

pediatric asthmatics. Environ Res 2004; 94:7–17.

J A

sthm

a D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y M

ichi

gan

Uni

vers

ity o

n 10

/30/

14Fo

r pe

rson

al u

se o

nly.