Archives of Pediatrics and Adolescent Medicine

2003 Archives of Pediatrics and Adolescent Medicine

doi: 10.1001/archpedi.157.10.951

Effects and Costs of Requiring Child-Restraint Systems for Young Children Traveling on Commercial Airplanes The Federal Aviation Administration has been planning a new regulation that would require children younger than 2 years to ride in an approved child-restraint seat on airplanes. This would mean that parents would have to buy separate seats for their young children; parents could no longer carry children on their laps in planes. Newman and colleagues examined the implications of this proposed policy if parents chose to drive to their destinations instead of fly because of the increased cost. Increased deaths from motor vehicle crashes would exceed deaths prevented by restraint use on planes if more than 5% to 10% of families switched from plane to car travel. Moreover, for each $1 increase in the cost of plane travel per family, the cost per death prevented would increase by $64 million. View LargeDownload Net effect of mandated child-restraint system use as a function of the proportion of families who choose to drive rather than fly and the average number of miles driven per diverted enplanement. The labeled isocontour lines show the specified net change in the annual number of deaths. See Article A Meta-analysis of Randomized Controlled Trials Evaluating the Efficacy of Epinephrine for the Treatment of Acute Viral Bronchiolitis This systematic review examined 14 randomized trials comparing epinephrine with a placebo or other bronchodilators. Among outpatients, epinephrine appeared to be better than the placebo and albuterol for short-term effects. However, there was inadequate evidence to make any conclusions about use of epinephrine in inpatient settings. The frequency of the problem and the costs for care demand that large-scale multicenter randomized trials be conducted. View LargeDownload See Article Abusive Head Injury as a Cause of Apparent Life-Threatening Events in Infancy Episodes of apparent life-threatening events involving changes in an infant's breathing pattern, color, and muscle tone are frightening to parents and often perplexing to pediatricians. In this series of 243 infants admitted for these episodes, Altman and colleagues found that the most common causes were reflux and seizures. In 6 infants (2.5%), an abusive head injury was the cause, and 2 of these infants died. While a wide spectrum of mostly benign disorders can cause these apparent life-threatening events, child abuse must be considered, and an appropriate evaluation should be done early in the hospital admission. See Article Trends in Psychotropic Medication Costs for Children and Adolescents, 1997-2000 The use of psychotropic medications in children has become a highly visible issue among both the lay and medical press. The cost implications are substantial, since medication costs are the fastest growing part of the health care budget. To provide more information on the factors underlying changes in psychotropic medication costs in children, Martin and Leslie examined insurance claims data from a national, 4-year sample. The largest changes in use were in atypical antipsychotics, atypical antidepressants, and selective serotonin reuptake inhibitors. Costs increased by 18%, more than half of which was due to a change in costlier medications within the same class of drugs. Almost half of the increase in cost was accounted for by 3 drugs. View LargeDownload See Article

McAnarney, Elizabeth

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.952

Bishai, David

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.953pmid: 14557152

2003 Archives of Pediatrics and Adolescent Medicine

doi: 10.1001/archpedi.157.10.954

The Archives of Pediatrics & Adolescent Medicine will publish a theme issue on mental health in August 2004. Our call for papers includes studies that provide new information on office screening for mental health problems, and guidance to practitioners on appropriate care and referral of patients with mental illness. We are especially interested in randomized trials that test new interventions, particularly those that can be provided by primary care physicians. For the best chance of consideration for this theme issue, papers should be received by January 1, 2004. Please consult our Web site (http://www.archpediatrics.com) for detailed instructions for authors.

Brosco, Jeffrey P.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.956pmid: 14557154

Hartling, Lisa; Wiebe, Natasha; Russell, Kelly; Patel, Hema; Klassen, Terry P.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.957pmid: 14557155

BackgroundControversy exists surrounding the use of bronchodilators for bronchiolitis. Epinephrine hydrochloride is being used with increasing frequency in this group; however, its efficacy has not been systematically reviewed.ObjectiveTo systematically review randomized controlled trials comparing inhaled or systemic epinephrine vs placebo or other bronchodilators.Data SourcesMEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, primary authors, and reference lists.Study SelectionStudies were included if they (1) were randomized, controlled trials; (2) involved children 2 years or younger with bronchiolitis; and (3) presented quantitative outcomes.Data ExtractionTwo reviewers independently extracted data and assessed study quality.Data SynthesisWe included 14 studies (7 inpatient, 6 outpatient, and 1 patient status unknown). Thirteen of forty-five comparisons were significant. Among outpatients, results favored epinephrine compared with placebo for clinical score at 60 minutes (standardized mean difference [SMD], −0.81; 95% confidence interval [CI], −1.56 to −0.07), oxygen saturation at 30 minutes (weighted mean difference [WMD], 2.79; 95% CI, 1.50-4.08), respiratory rate at 30 minutes (WMD, −4.54; 95% CI, −8.89 to −0.19), and improvement (odds ratio, 25.06; 95% CI, 4.95-126.91); among inpatients, for clinical score at 60 minutes (SMD, −0.52; 95% CI, −1.00 to −0.03). Among outpatients, results favored epinephrine compared with albuterol sulfate (salbutamol) for oxygen saturation at 60 minutes (WMD, 1.91; 95% CI, 0.38-3.44), heart rate at 90 minutes (WMD, −14.00; 95% CI, −22.95 to −5.05), respiratory rate at 60 minutes (WMD, −7.76; 95% CI, −11.35 to −4.17), and improvement (odds ratio, 4.51; 95% CI, 1.93-10.53); among inpatients, respiratory rate at 30 minutes (WMD, −5.12; 95% CI, −6.83 to −3.41).ConclusionsEpinephrine may be favorable compared with placebo and albuterol for short-term benefits among outpatients. There is insufficient evidence to support the use of epinephrine among inpatients. Large, multicentered trials are required before routine use among outpatients can be strongly recommended.BRONCHIOLITIS, THE most common lower respiratory tract infection in infants, is characterized by fever, coryza, cough, expiratory wheezing, and respiratory distress (ie, increased respiratory rate, chest wall indrawing, thoracoabdominal asynchrony).It is most commonly associated with viruses, with the leading cause being the respiratory syncytial virus.Overall, it is estimated that 11% to 12% of all infants are affected in the first year of life, with 1% to 2% requiring hospitalization.Because of the prevalence and morbidity associated with bronchiolitis, the economic burden placed on health care services is substantial.Despite the frequency of the condition, considerable controversy remains regarding its management. Historically, children have been offered supportive care, including oxygen and supplemental fluids.Recent clinical trials have provided conflicting evidence regarding the benefit of pharmacological interventions. Much of the debate involves the role of bronchodilators.Although in common use, the efficacy of bronchodilators in this patient group has not been universally accepted. Two recent systematic reviews have found insufficient empirical support for widespread use of bronchodilators.Flores and Horwitzreviewed 8 randomized controlled trials (RCTs) to evaluate the efficacy of β2-agonists in bronchiolitis. Among 3 inpatient studies, the results were contradictory with respect to improved clinical scores, duration of hospital stay, and oxygen saturation. Among the 5 outpatient studies, there was no benefit in terms of hospitalization rate or respiratory rate. The reviewers found a statistically significant improvement in oxygen saturation and heart rate, but the results were deemed to be not clinically significant.Kellner and colleaguesreviewed 20 RCTs, 18 of which examined β2-agonists and 2, epinephrine hydrochloride. The review grouped all bronchodilators and compared these with placebo; they did not examine the relative efficacy of different bronchodilators. The reviewers found modest short-term improvements in clinical score among children with mild and moderate bronchiolitis. The results for oxygen saturation were inconclusive owing to heterogeneity between studies. They found no significant improvement in rate or duration of hospitalization. These authors concluded that bronchodilators could not be recommended for routine management in first-time wheezers.Although different nebulized bronchodilators such as albuterol sulfate (salbutamol), ipratroprium bromide, and epinephrine are being used in the treatment of bronchiolitis, research to date supports epinephrine as the bronchodilator of choice.Along with the β-adrenergic effects of bronchodilation, epinephrine adds α-adrenergic properties and is believed to offer the supplemental benefits of vasoconstriction in the bronchiolar vasculature. Along with others, Wohl and Chernickhave suggested that this vasoconstriction may reduce edema and mucous production, hallmarks in the pathology of acute viral bronchiolitis. Because of the unique mechanism of action of epinephrine and its increasing use in infants with bronchiolitis, we chose to specifically investigate the efficacy of this drug in the treatment of bronchiolitis. Thus, the objective of this study was to review RCTs that compared the effects of inhaled or systemic epinephrine vs placebo or other bronchodilators in infants and young children (age, ≤2 years) with bronchiolitis.METHODSCRITERIA FOR INCLUDING STUDIESAll RCTs evaluating the efficacy of epinephrine vs placebo or of epinephrine vs other bronchodilators in the treatment of bronchiolitis were considered for inclusion, regardless of language or publication status. All studies involving infants and young children 2 years or younger were eligible for inclusion. Bronchiolitis was defined as wheezing (with or without cough, tachypnea, and increased respiratory effort) associated with clinical evidence of a viral infection (eg, coryza and fever). Studies of inpatients and outpatients were included. Studies were included if they reported on at least 1 of the following outcome measures: clinical score, oxygen saturation (oximetry), admission to the hospital (rate of hospitalization), length of hospital stay, respiratory rate, heart rate, and results of pulmonary function tests.SEARCH STRATEGY FOR IDENTIFICATION OF STUDIESSearches of MEDLINE (January 1966 through December 2000), EMBASE (January 1988 through December 2000), and The Cochrane Central Register of Controlled Trials were conducted using the following terms: epinephrine, bronchiolitis, respiratory syncytial viruses, respiratory syncytialPneumovirus, respiratory syncytial virus, and adrenalin.The complete search strategies are available from the authors on request. We examined the reference lists of all selected articles for relevant studies. Primary authors of relevant trials were contacted for information on additional trials. We searched PubMed at the end of the project (September 2002) to identify any recent studies.STUDY SELECTIONThe initial search of all of the databases and reference lists was screened independently by 2 of us (L.H. and K.R.) to identify citations with potential relevance. The full text of selected articles was obtained. The 2 reviewers independently decided on trial inclusion using a standard form with predetermined eligibility criteria. Disagreements were resolved by consensus reached after discussion.QUALITY ASSESSMENTStudy quality for English-language studies was assessed independently by 2 of us (reviewers K.R., T.P.K., or L.H.); study quality of the Turkish reports was assessed by an independent reviewer. Quality was assessed on the basis of published reports in peer-reviewed journals when available; for 1 trial,quality assessment was based on the abstract and unpublished information from the author, as the manuscript was not yet available. Each study was evaluated using the Jadad 5-point scale to assess randomization (0-2 points), double blinding (0-2 points), and withdrawals and dropouts (0-1 point).The Jadad scale was chosen because it is the only quality assessment tool, to the best of our knowledge, that has been validated, and it incorporates components that are directly related to the control of bias. Concealment of allocation was assessed as adequate, inadequate, or unclear.Differences in quality ratings were resolved through discussion.DATA EXTRACTIONData from the English-language studies were extracted independently by 2 of us (L.H. and K.R.); data were extracted from the Turkish reports by a single individual. Additional data were requested from authors as necessary. A standard form was used that described the following: characteristics of the study (design, method of randomization, and withdrawals/dropouts), participants (age and sex), intervention (type, dose, route of administration, timing and duration of therapy, and cointerventions), control (agent and dose), outcomes (types of outcome measures, timing of outcomes, complications, and adverse events), whether the study used an intention-to-treat protocol, funding source, and results. Data were entered into RevMan 4.1 (The Cochrane Collaboration, Oxford, England, 2000) by one reviewer (K.R.) and checked for accuracy by a second reviewer (L.H.).Individual patient clinical score data were extracted from graphs for 1 study.Means were extracted from graphs for 4 studies,SDs for 1,and 95% confidence limits for 1.One trial used a crossover design; therefore, only data from the first phase were used in the meta-analysis.This same study included 2 placebo groups; data from both groups were pooled. In some cases, variance was imputed from confidence intervals (CIs)and SEs.To calculate the variance of change in oximetry in 1 study, the end time-point SDs were substituted with the baseline SDs.Finally, for 1 study,the mean SDs from other studies were substituted for missing SDs for the clinical score outcomes.DATA ANALYSISAnalyses were performed using RevMan 4.1 (The Cochrane Collaboration, Oxford, England, 2000) and Splus 2000 (Insightful Corporation, Seattle, Wash, 1999). Separate analyses were conducted for the 2 types of control groups (ie, placebo and nonepinephrine bronchodilators) and patient status (ie, inpatient or outpatient). Dichotomous data (eg, improvement) were expressed as Mantel-Haenszel odds ratios, and a common Mantel-Haenszel odds ratio with 95% CIs was calculated. The number needed to treat was derived for significant results to help clarify the degree of possible benefit for the averaged (inverse-variance method) baseline rates. There were too few studies to check whether the relative risk was constant across different baseline rates; therefore the numbers needed to treat were not provided for a range of baseline rates. The changes in clinical score and oximetry were calculated using baseline and time-point data; a correlation of 0.5 was assumed. The clinical scores were converted to a standardized mean difference because the 14 studies used a total of 6 different clinical scores. An overall standardized mean difference with 95% CIs was calculated. A standardized mean difference is "the difference between 2 means divided by an estimate of the within-group standard deviation."Other continuous data (eg, oximetry, heart rate, respiratory rate, and length of stay) were converted to the mean difference, and an overall weighted mean difference (with 95% CIs) was calculated. When mean differences (difference between treatment group means) are pooled by the inverse variance method, each mean difference is weighted by the inverse of the estimate's variance, giving more weight to studies with more precise estimates. Only 1 study included results of pulmonary function tests as an outcome; these results are presented separately.Results were calculated using random-effects models, regardless of heterogeneity. In particular, all clinical scores were calculated using random effects owing to the supposition that these clinical scores measured different clinical features of bronchiolitis or weighted these differently. Fixed effects were also calculated in a sensitivity analysis. Possible sources of heterogeneity were not assessed by subgroup and sensitivity analyses because of the small numbers of studies for each outcome. Publication bias was not assessed because of the small number of trials in each outcome, comparison, and patient status group included in the review. Post hoc power calculations were assessed using independent 2-sample ttests and Pearson χ2tests in PASS 2002 (Number Cruncher Statistical Systems, Kaysville, Utah, 2002).RESULTSThe results reported in this article differ slightly from those in a previously published abstract,as 5 trials were subsequently added.DESCRIPTION OF STUDIESSeventy-six unique references were identified (the full list of references is available from the authors). Twenty-five studies were identified as being potentially relevant. Fourteen studies met the inclusion criteria; there was 100% agreement between the 2 reviewers with respect to study inclusion. The included studies are described in Table 1. The studies were generally small, but ranged in sample size from 30 to 194. Most studies (n = 12) were published in English, with 2 published in Turkish.The studies were conducted in a variety of primarily high-income countries.Table 1. Characteristics of Included StudiesStudyCountryNo. of Participants per Study Group*Inpatient vs OutpatientClinical ScoreWheezing HistoryPrimary OutcomeType of Epinephrine†Abul-Ainine and Luyt(2002)England19 Epinephrine hydrochloride; 19 placeboInRDAIFirst timeRR, HRL-epinephrineBarlas et al(1998)Turkey15 Epinephrine; 15 placebo; 15 albuterol sulfateOutBarlas et alUnknownClinical scoreRacemicBertrand et al(2001)Chile16 Epinephrine; 14 albuterolInTal et alFirst timeClinical scoreL-epinephrineKristánsson et al(1993)Sweden/Norway15 Epinephrine; 14 placeboInKristáansson et alMixedClinical scoreRacemicHariprakash et al(2003)England38 Epinephrine; 37 placeboOutRDAIMixedAdmission rateL-epinephrineLowell et al(1987)United States16 Epinephrine; 14 placeboOutRACS (RDAI)MixedClinical scoreL-epinephrineMenon et al(1995)Canada20 Epinephrine; 21 albuterolOutRDAIFirst timeSaO2L-epinephrineMull et al(2002)United States34 Epinephrine; 32 albuterolOutRDAIFirst timeClinical scoreRacemicOkutan et al(1998)Turkey16 Epinephrine; 19 albuterol; 19 placeboUnknownRDIFirst timeClinical scoreUnknownPatel et al(2002)Canada50 Epinephrine; 51 albuterol; 48 placeboInRDAIFirst timeLOSRacemicRay and Singh(2002)India45 Epinephrine; 46 albuterolOutRDAI, Yale Observation ScaleMixedSaO2L-epinephrineReijonen et al(1995)Finland24 Epinephrine; 27 albuterol; 49 placeboInRDAIMixedClinical scoreRacemicSanchez et al(1993)Canada12 Epinephrine; 12 albuterolInTal et alFirst timeClinical scoreRacemicWainwright et al(in press)Australia99 Epinephrine; 95 placeboInNoneFirst timeLOS, ready for dischargeL-epinephrineAbbreviations: HR, heart rate; LOS, length of stay; RACS, respiratory assessment change score; RDAI, Respiratory Distress Assessment Instrument; RDI, Respiratory Distress Index; RR, respiratory rate; SaO2, arterial oxygen percent saturation.*Refers to study groups used in meta-analysis.†Delivered via nebulizer in all studies except Lowell et al,in which delivery was subcutaneous.A wide range of outcomes was reported. Table 1describes the primary outcomes studied in each trial. Secondary outcomes included clinical score; oxygen saturation; respiratory rate; heart rate; blood pressure; activity status; time receiving oxygen; highest oxygen flow rates; need for supplemental parenteral fluids; transcutaneous oxygen and carbon dioxide tensions; time from admission to normal oxygenation, adequate intake, and minimal respiratory distress; pulmonary mechanics; duration of hospitalization; rate of hospitalization; and improvement as defined by the individual trials.Although most studies measured clinical scores, a number of different scoring systems were used, and the scores were reported in different ways (Table 1and Table 2). The Respiratory Distress Assessment Instrument was the scoring system most commonly used. It was used in 7 studies; however, in 2 of these studies, scores were not reported. In 1 study,the authors reported the mean time to a Respiratory Distress Assessment Instrument score of no greater than 4, and in another studythe authors reported only a Pvalue for the mean change in score. The remaining studies used a variety of partially validated or unvalidated scales that measured different clinical features of bronchiolitis.Table 2. Description of Clinical Scores Used in the Included StudiesStudyValidatedMaximum PointsNo. of Points for Different Clinical FeaturesRetractions and IndrawingRespiratory RateWheezingCyanosisAuscultatory Breath SoundsGeneral ConditionNostril MovementRDAIPartially170-9NA0-8NANANANARACSPartially**NA*NANANANARDIUnclear90-30-30-3NANANANAKristánsson et alNo100-20-2NA0-20-20-2NABarlas et alNo150-30-30-3NANA0-30-3Tal et al(modified from Bierman and Pierson)Partially120-30-30-30-3NANANAAbbreviations: NA, not measured by the score; RACS, respiratory assessment change score; RDAI, Respiratory Distress Assessment Instrument; RDI, Respiratory Distress Index.*Based on the RDAI; it is the sum of the change in each variable.Most studies conducted short-term follow-up of up to 4 hours, whereas 3 studies followed up inpatients during their hospital stay (herein referred to as longer-term outcomes).In addition, 1 outpatient study evaluated 72-hour relapse rates,and 1 inpatient study asked general physicians to notify the study personnel of any deterioration in the patients' condition during the 48-hour postdischarge period (no data presented).METHODOLOGICAL QUALITY OF INCLUDED STUDIESThe methodological quality of studies is reported in Table 3. Three studies received pharmaceutical sponsorship; funding was received from other external sources in 6 trials; the source of funding was not mentioned in 4 trials; and 2 studies received no funding.Two studies conducted an intention-to-treat analysis.Four studies reported withdrawals and excluded these from the analysis.Eight studies did not report any withdrawals.Table 3. Methodological Quality of Included StudiesStudyRandomizationDouble-BlindingDescription of Withdrawals/Dropouts*Jadad Score†Allocation ConcealmentStatedMethod DescribedStatedMethod DescribedAbul-Ainine and Luyt(2002)✓Adequate✓UnclearInadequate3UnclearBarlas et al(1998)✓UnclearXNAInadequate1UnclearBertrand et al(2001)✓Unclear✓UnclearAdequate3UnclearHariprakash et al(2003)✓Adequate✓AdequateAdequate5AdequateKristánsson et al(1993)✓Unclear✓AdequateInadequate3UnclearLowell et al(1987)✓Unclear✓UnclearInadequate2AdequateMenon et al(1995)✓Adequate✓AdequateInadequate4AdequateMull et al(2002)‡✓Adequate✓UnclearInadequate3AdequateOkutan et al(1998)✓Inadequate✓InadequateInadequate1UnclearPatel et al(2002)✓Adequate✓AdequateAdequate5AdequateRay and Singh(2002)✓UnclearXNAInadequate1UnclearReijonen et al(1995)✓Unclear✓UnclearInadequate2UnclearSanchez et al(1993)✓Unclear✓UnclearInadequate2UnclearWainwright et al(in press)✓Unclear✓AdequateAdequate4AdequateAbbreviations: NA, not applicable; check mark, yes; X, no.*To be graded as "adequate," the description must include the number and reasons for withdrawal in each group; if there were no withdrawals it must be stated in the article.†Described in Jadad et al.‡Quality assessment based on abstract and unpublished information.EPINEPHRINE VS PLACEBOResults were stratified by inpatient vs outpatient status. Table 4presents the results for the epinephrine vs placebo comparison. Five inpatient studies compared epinephrine and placebo.Only 1 of 10 inpatient outcomes demonstrated a significant difference between treatment groups; change in clinical score at 60 minutes favored epinephrine.Table 4. Comparison of Epinephrine vs Placebo by Inpatient/Outpatient StatusOutcomeNo. of StudiesNo. of SubjectsSummary MeasureOverall Effect Measure (95% CI)*InpatientsChange in clinical score 30 minutes after treatment3140SMD−0.24 (−0.78 to 0.30)Change in clinical score 60 minutes after treatment267SMD−0.52 (−1.00 to −0.03)Change in oxygen saturation 30 minutes after treatment3140WMD−0.05 (−0.94 to 0.85)Change in oxygen saturation 60 minutes after treatment267WMD0.11 (−0.98 to 1.21)Length of stay, h2292WMD−5.90 (−16.23 to 4.43)Incidence of pallor 30 minutes after treatment129OR4.73 (0.46 to 48.77)OutpatientsChange in clinical score 30 minutes after treatment2105SMD−0.55 (−1.11 to 0.02)Change in clinical score 60 minutes after treatment130SMD−0.81 (−1.56 to −0.07)Change in oxygen saturation 30 minutes after treatment175WMD2.79 (1.50 to 4.08)Change in oxygen saturation 60 minutes after treatment130WMD1.20 (−0.13 to 2.53)Improvement†260OR25.06 (4.95 to 126.91)NNT‡1.69 (1.30 to 2.50)Admission rates2105OR0.51 (0.18 to 1.42)Inpatients/outpatient status unknownChange in clinical score 30 minutes after treatment135SMD−0.29 (−0.96 to 0.38)Change in clinical score 60 minutes after treatment135SMD−0.70 (−1.39 to −0.01)Abbreviations: CI, confidence interval; NNT, number needed to treat; OR, odds ratio; SMD, standardized mean difference; WMD, weighted mean difference.*Boldface results significantly favor epinephrine hydrochloride except where indicated.†Defined within the individual studies.‡Given a baseline risk of 89% (95% CI, 78%-99%) for not improving.Three studies compared epinephrine and placebo among outpatients.Five of 10 outcomes were significant. Change in clinical score at 60 minutes after treatment, change in oxygen saturation at 30 minutes after treatment, respiratory rate at 30 minutes after treatment (weighted mean difference [WMD], −4.54; 95% CI, −8.89 to −0.19), and improvement favored epinephrine. In 1 study, improvement was defined as a positive change in the respiratory assessment change score of at least 4 U,and in the other study it was not defined.Heart rate at 60 minutes after treatment favored placebo. Admission rates (Figure 1), change in clinical score at 30 minutes after treatment, change in oxygen saturation at 60 minutes after treatment, and heart rate at 30 minutes after treatment were not significantly different between the treatment arms. Sensitivity analyses using fixed-effects models found 1 significant difference favoring epinephrine in change in clinical score at 30 minutes. One Turkish studydid not indicate its inpatient/outpatient status. This study reported a significant change in clinical score at 60 minutes favoring epinephrine compared with placebo.Figure 1.Metagraph of admissions to the hospital among outpatients. CI indicates confidence interval; OR, odds ratio.EPINEPHRINE VS ALBUTEROLTable 5presents the results of epinephrine vs albuterol. Four studies compared epinephrine with albuterol among inpatients.Only 1 of the 7 outcomes was statistically significant: respiratory rate at 30 minutes favored epinephrine compared with albuterol (WMD, −5.12; 95% CI, −6.83 to −3.41). The clinical scores, oxygen saturation, heart rate, and length of stay (Figure 2) outcomes showed no significant difference.Table 5. Comparison of Epinephrine vs Albuterol by Inpatient/Outpatient StatusOutcomeNo. of StudiesNo. of SubjectsSummary MeasureOverall Effect Measure* (95% CI)InpatientsChange in clinical score 30 minutes after treatment3105SMD−0.43 (−1.01 to 0.16)Change in clinical score 30 minutes after treatment after 24 h130SMD0.11 (−0.61 to 0.83)Change in clinical score 30 minutes after treatment after 36 h130SMD0.55 (−0.18 to 1.29)Change in oxygen saturation 30 minutes after treatment275WMD0.21 (−0.73 to 1.14)Length of stay, h2131WMD−3.96 (−25.55 to 17.62)OutpatientsChange in clinical score 30 minutes after treatment2107SMD−0.08 (−0.84 to 0.69)Change in clinical score 60 minutes after treatment4228SMD−0.21 (−0.74 to 0.32)†Change in clinical score 90 minutes after treatment2107SMD−0.32 (−0.82 to 0.19)Change in oxygen saturation 30 minutes after treatment2132WMD−1.31 (−3.15 to 5.76)Change in oxygen saturation 60 minutes after treatment3162WMD1.91 (0.38 to 3.44)Change in oxygen saturation 90 minutes after treatment141WMD−0.68 (−2.39 to 1.03)Improvement‡2120OR4.51 (1.93 to 10.53)NNT§4.55 (1.82 to infinity)Admission rates4228OR0.40 (0.12 to 1.33)†Incidence of pallor 30 minutes after treatment141OR6.00 (1.33 to 27.00)NNT∥2.78 (1.61 to 11.11)Inpatients/outpatient status unknownChange in clinical score 30 minutes after treatment135SMD0.16 (−0.51 to 0.83)Change in clinical score 60 minutes after treatment135SMD0.07 (−0.59 to 0.74)Abbreviations: CI, confidence interval; NNT, number needed to treat; OR, odds ratio; SMD, standardized mean difference; WMD, weighted mean difference.*Boldface results significantly favor epinephrine hydrochloride except where indicated. †Significant with fixed-effect estimates and favors epinephrine.‡Defined within the individual studies.§Given an albuterol sulfate risk of 22% (95% CI, 11%-32%) for not improving.∥Given an albuterol risk of 50% (95% CI, 38%-62%) for not being admitted.Figure 2.Metagraph of length of hospital stay (LOS) among inpatients. CI indicates confidence interval; WMD, weighted mean difference.Four outpatient studies reported on the epinephrine-albuterol comparison.Four of 16 outcomes showed the following statistically significant differences between treatment groups: change in oxygen saturation at 60 minutes, change in heart rate at 90 minutes (WMD, −14.00; 95% CI, −22.95 to −5.05), respiratory rate at 60 minutes (WMD, −7.76; 95% CI, −11.35 to −4.17), and improvement after treatment significantly favored epinephrine. Improvement in 1 study referred to patients in whom moderate and severe distress was converted to normal or mild distress after intervention; the other study did not define improvement.One outcome, the incidence of pallor at 30 minutes after treatment, favored albuterol. Sensitivity analyses using fixed-effects models found significant differences favoring epinephrine for change in clinical score at 60 minutes and admissions. In addition, fixed-effects analyses for heart rate at 60 minutes favored albuterol.One Turkish studydid not indicate its patient status (inpatients vs outpatients); neither of its 2 change-in-clinical-score outcomes was significant.OTHER OUTCOMESOnly 1 study evaluated pulmonary mechanics among 24 patients randomized to receive epinephrine or albuterol.Significant differences between pretreatment and posttreatment values were noted in inspiratory, expiratory, and total pulmonary resistance in the epinephrine group, but not the albuterol group. There were no significant differences compared with baseline values in either group with respect to tidal volume, minute ventilation, dynamic compliance, or duration of inspiration as a fraction of total breath duration.Because of the small number of studies that evaluated longer-term outcomes, some of these outcomes were not included in the meta-analysis. The largest trial, conducted by Wainwright et al,randomized 194 inpatients to epinephrine or placebo and found no differences between groups in length of stay or time ready for discharge. The second largest trial involved 149 inpatients randomized to epinephrine, albuterol, or placebo and found no significant difference between groups in length of stay or any secondary outcomes.Bertrand et alfollowed up 30 inpatients randomized to epinephrine or albuterol and found no statistically significant differences in length of stay or duration of oxygen therapy, although the trend favored epinephrine. Mull et alassessed the relapse rate at 72 hours after treatment among 66 outpatients randomized to epinephrine or albuterol and found no significant difference.Three studies reported on patient return to the hospital or emergency department after the study. Sanchez et alfound that only 3 of 24 patients (treatment group not specified) were readmitted to the hospital for acute wheezing during a 6- to 10-month follow-up period; Bertrand et alfound that no patients were readmitted in the 2 weeks after discharge from the hospital; and Patel et alreported that 93 of 149 infants (21 receiving epinephrine; 21, albuterol; and 25, placebo) had a medical visit in the week after discharge, that 8 of these visits (1 patient receiving epinephrine; 3, albuterol; and 4, placebo) were to the emergency department, and that 3 patients (receiving placebo) were readmitted. One study noted that children were sent home receiving oral medication but did not specify the type.COMMENTThe objective of this study was to provide some resolution to the uncertainty in the literature regarding the use of epinephrine in the treatment of bronchiolitis. Some evidence supports the use of epinephrine among outpatients. The combined results of the outpatient studies favored epinephrine compared with albuterol in terms of oxygen saturation at 60 minutes, heart rate at 90 minutes, respiratory rate at 60 minutes, and improvement. These results are based on a small number of studies of varying quality. Some evidence also suggests that epinephrine is favorable compared with placebo among outpatients in terms of clinical score at 60 minutes after treatment, oxygen saturation at 30 minutes after treatment, heart rate at 60 minutes after treatment, and overall improvement. None of the studies reported any significant adverse effects resulting from the administration of epinephrine, although 1 study reported significantly less pallor at 30 minutes after treatment in the albuterol group.Because of the small number of studies for each comparison, we did not have the ability to examine the relative efficacy of epinephrine among other potentially important subgroups such as first-time vs recurrent wheezers, severity of illness, specific viral etiology, age, and stage of the disease.We also did not have the ability to assess different forms of delivery such as type of epinephrine, route of delivery, number of administrations, and dosage. We used a more liberal definition of bronchiolitis, as is common in North America and parts of Europe.The results should be interpreted in light of this.Several factors may contribute to the lack of consistency in the findings. First, there may be no difference between treatment with epinephrine vs treatment with albuterol or placebo, and any significant findings may have been spurious associations resulting from multiple comparisons.The efficacy of the drug may be different for various subgroups (eg, outpatients vs inpatients). The subgrouping of outpatients vs inpatients may be a proxy for severity of illness, as those admitted may be more severely affected, later in the course of the disease, or more resistant to treatment. Continued focused evaluation within these subgroups is warranted.Six different scoring systems were used across the component studies, which resulted in statistically significant heterogeneity between studies. Multiple comparisons between clinical scores at different time points, among different subgroups (outpatients and inpatients), and for the different controls (albuterol and placebo) were performed, and only 3 of 14 comparisons resulted in statistically significant results. It is possible that these were spurious findings. Alternatively, the scoring systems may not be sensitive to clinically important differences. They may not measure, or may measure differentially, the clinical improvement in bronchiolitis. There is clearly a need to evaluate the clinical scores currently in use. Validation and checking sensitivity of the scores used in individual trials would facilitate comparisons between studies.More than a dozen different outcome measures were evaluated within the component trials. Because of the lack of consistency in the outcomes reported, there were few studies within each comparison. In primary studies, as in meta-analyses, care needs to be taken to specify the outcomes a priori to avoid bias that can arise if only those outcomes with significant results are reported.The quality of the trials was moderate, with a median Jadad score of 3. All studies were described as random, but only 4 studies described an appropriate method of randomization. Twelve of the 14 studies were described as double-blind, but only 5 studies described an appropriate method of double blinding. Inadequate blinding can overestimate the effect,which could skew the results in favor of either treatment, depending on the biases of the investigators. Investigators should be aware that adequate blinding is of particular concern in a study of epinephrine for 2 reasons. First, some investigators have noted reddish nasal discharge after administration of epinephrine. However, in a large trial by Patel et al,no instances of red nasal discharge were reported; the investigators suggested that this may be related to the age of the medication. Second, perioral pallor results with nebulized epinephrine. This is a concern in studies where a postmask assessment does not allow sufficient time for the pallor to dissipate (eg, 30 minutes). This issue is of most importance in studies that compare epinephrine with placebo vs those that compare epinephrine with albuterol, since many of the short-term adverse effects of albuterol are similar to those of epinephrine. Only 4 studies provided an adequate description of withdrawals and dropouts. Six studies reported adequate allocation concealment. Studies that do not properly conceal treatment allocation can overestimate treatment effects by as much as 40%.Finally, the meta-analysis may not have sufficient power to detect statistically significant differences between treatment groups. We calculated the power that the combined studies had to detect a simply pooled difference in the outcomes with largest combined sample size per comparison and patient status group. In a single trial with the same number of patients, there would have been only 7% to 57% power to detect a difference in these various comparisons. There would be less power for the other clinical score outcomes for which there were fewer studies and patients. The implication of this finding is that a number of large trials is needed to substantiate the relative efficacy of epinephrine in the treatment of bronchiolitis.CONCLUSIONSSome evidence suggests that epinephrine may be favorable compared with albuterol and placebo among outpatients. There is insufficient evidence to support the use of epinephrine for the treatment of bronchiolitis among inpatients. A validated, reliable scoring system is needed that is sensitive to important clinical changes in patients. The appropriateness of a scoring system may vary depending on the context in which it is used; eg, for acute changes, a clinical scoring system may be adequate, but for longer-term changes, inclusion of quality-of-life measures may be more appropriate (impact on feeding, family life, anxiety, difficulty breathing, etc). The use of a validated, reliable, and responsive scoring system would facilitate comparison of results across studies. A number of large, multicentered trials are required to examine the effectiveness of epinephrine compared with placebo and albuterol for infants presenting to the emergency department.What This Study AddsMuch controversy has surrounded the use of bronchodilators for bronchiolitis. Recent evidence has suggested that epinephrine hydrochloride may offer some clinical benefit. Epinephrine is being used with increasing frequency in this group; however, its efficacy has not been systematically reviewed.There is insufficient evidence to support the use of epinephrine for the treatment of bronchiolitis among inpatients. Some evidence suggests that epinephrine may be favorable compared with albuterol sulfate and placebo among outpatients. Further research needs include (1) a number of large, multicentered trials to examine the effectiveness of epinephrine compared with placebo and albuterol for infants presenting to the emergency department; and (2) development and validation of a reliable scoring system that is sensitive to important clinical changes in patients with bronchiolitis.TPKlassenRecent advances in the treatment of bronchiolitis and laryngitis.Pediatr Clin North Am.1997;44:249-261.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9057793MLEverardBronchiolitis: origins and optimal management.Drugs.1995;49:885-896.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7641603GFloresRIHorwitzEfficacy of β2-agonists in bronchiolitis: a reappraisal and meta-analysis.Pediatrics.1997;100:233-239.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9240805JDKellnerAOhlssonAMGadomskiEELWangBronchodilators for bronchiolitis [Cochrane Review on CD-ROM].Issue 2. Oxford, England: Cochrane Library, Update Software; 2000.MEBWohlVChernickState of the art: bronchiolitis.Am Rev Respir Dis.1978;118:759-781.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=212970CCMullRJScarfoneLRFerriNebulized epinephrine in the emergency department treatment of bronchiolitis [abstract].Pediatr Res.2002;51:100.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11756647ARJadadRAMooreDCarrollAssessing the quality of reports of randomized clinical trials: is blinding necessary?Control Clin Trials.1996;17:1-12.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8721797KFSchulzIChalmersRJHayesDGAltmanEmpirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials.JAMA.1995;273:408-412.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7823387DILowellGListerKoss HVonPMcCarthyWheezing in infants: the response to epinephrine.Pediatrics.1987;79:939-945.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=3295741VOkutanRAkinAEKurekciAYanikOOzcanEGokcayEffectiveness of nebulised adrenaline and salbutamol in the treatment of infants with bronchiolitis.Bull Gulhane Milit Med Acad.1998;40:199-204.AAbul-AinineDLuytShort term effects of adrenaline in bronchiolitis [in Turkish]: a randomised controlled trial.Arch Dis Child.2002;86:276-279.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11919104SHariprakashJAlexanderWCarrollRandomized controlled trial of nebulized adrenaline in acute bronchiolitis.Pediatr Allergy Immunol.2003;14:134-139.TReijonenMKorppiSPitkäkangasSTenholaKRemesThe clinical efficacy of nebulized racemic epinephrine and albuterol in acute bronchiolitis.Arch Pediatr Adolesc Med.1995;149:686-692.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7767427ISanchezKoster JDeREPowellRWolsteinVChernickEffect of racemic epinephrine and salbutamol on clinical score and pulmonary mechanics in infants with bronchiolitis.J Pediatr.1993;122:145-151.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8419602SKristjánssonCarlsen KCLødrupGWennergrenI-LStrannegårdK-HCarlsenNebulised racemic adrenaline in the treatment of acute bronchiolitis in infants and toddlers.Arch Dis Child.1993;69:650-654.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8285776CWainwrightLAltamiranoMCheneyA multi-centre randomised controlled double blind trial of nebulised adrenaline in infants with bronchiolitis.N Engl J Med.2003;349:27-35.PBertrandHAraníbarECastroISánchezEfficacy of nebulized epinephrine versus salbutamol in hospitalised infants with bronchiolitis.Pediatr Pulmonol.2001;31:284-288.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11288210S-LTNormandTutorial in biostatistics meta-analysis: formulating, evaluating, combining, and reporting.Stat Med.1999;18:321-359.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10070677MClarkeADOxmanedsCochrane reviewers' handbook 4.1 [updated June 2000].In: Review Manager[computer program]. Version 4.1. Oxford, England: Cochrane Collaboration; 2000.LHartlingNWiebeKFRussellHPatelTPKlassenEpinephrine for the treatment of acute viral bronchiolitis: a systematic review [abstract].Pediatr Res.2002;51:86.CBarlasNKiperAGocmenRacemic adrenaline and other treatment regiments in mild and moderate bronchiolitis [in Turkish].Cocuk Sagligi Ve Hastaliklari Dergisi.1998;41:155-165.KMenonTSutcliffeTPKlassenA randomized trial comparing the efficacy of epinephrine with salbutamol in the treatment of acute bronchiolitis.J Pediatr.1995;126:1004-1007.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7776075MSRayVSinghComparison of nebulized adrenaline versus salbutamol in wheeze associated respiratory tract infection in infants.Indian Pediatr.2002;39:12-22.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11805349HPatelRWPlattGSPekelesFDucharmeA randomized, controlled trial of the effectiveness of nebulized therapy with epinephrine compared with albuterol and saline in infants hospitalized for acute viral bronchiolitis.J Pediatr.2002;141:818-824.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=12461499ATalCBavilskiDYohaiJEBearmanRGorodischerSWMosesDexamethasone and salbutamol in the treatment of acute wheezing in infants.Pediatrics.1983;71:13-18.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=6129609LHAllenThe role of nutrition in the onset and treatment of metabolic bone disease.Nutr Update.1983;1:263-281.CWBiermanWEPiersonThe pharmacologic management of status asthmaticus in children.Pediatrics.1974;54:245-247.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=4847861GAColditzJNMillerFMostellerHow study design affects outcomes in comparisons of therapy, I: medical.Stat Med.1989;8:441-454.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=2727468Corresponding author: Terry P. Klassen, MD, MSc, FRCPC, Department of Pediatrics, University of Alberta, 2C3.67 Walter C. Mackenzie Health Sciences Centre, Edmonton, Alberta, Canada T6G 2R7 (e-mail: [email protected]).Accepted for publication March 6, 2003.The Alberta Research Centre for Child Health Evidence, Edmonton, is supported by an establishment grant from the Alberta Heritage Foundation for Medical Research, Edmonton.We thank Marlene Dorgan, MLIS, and Ellen Crumley, MLIS, for their assistance with searching, and Metin Gulmezoglum, MD, for his assistance with translation and data extraction of the Turkish studies.This review has been registered with the Cochrane Collaboration. Regular updates will be available in the Cochrane Library.

King, Tracy M.; Alperovitz-Bichell, Kari; Rowe, Peter C.; Lehmann, Harold P.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.965pmid: 14557156

Newman, Thomas B.; Johnston, Brian D.; Grossman, David C.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.969pmid: 14557157

ContextThe US Federal Aviation Administration is planning a new regulation requiring children younger than 2 years to ride in approved child-restraint seats on airplanes.ObjectivesTo estimate the annual number of child air crash deaths that might be prevented by the proposed regulation, the threshold proportion of families switching from air to car travel above which the risks of the policy would exceed its benefits, and the cost per death prevented.DesignRisk and economic analyses.ResultsChild-restraint seat use could prevent about 0.4 child air crash deaths per year in the United States. Increased deaths as a result of car travel could exceed deaths prevented by restraint seat use if the proportion of families switching from air to car travel exceeded about 5% to 10%. The estimate for this proportion varied with assumptions about trip distance, driver characteristics, and the effectiveness of child-restraint seats but is unlikely to exceed 15%. Assuming no increase in car travel, for each dollar increase in the cost of implementing the regulation per round trip per family, the cost per death prevented would increase by about $6.4 million.ConclusionsUnless space for young children in restraint seats can be provided at low cost to families, with little or no diversion to automobile travel, a policy requiring restraint seat use could cause a net increase in deaths. Even excluding this possibility, the cost of the proposed policy per death prevented is high.THE US FEDERAL Aviation Administration (FAA) will soon propose a regulation mandating use of child-restraint systems (CRSs) for all children flying in aircraft.As a consequence, children younger than 2 years would no longer be able to travel on their parents' laps but would require a seat of their own, with the costs passed on to their parents, other passengers, or the airlines' shareholders. This proposed regulation represents a policy change on the part of the FAA, which argued in a 1995 report to Congress that CRSs on aircraft would prevent a maximum of 5 child plane crash deaths per 10 years and would result in a net increase of 82 deaths per 10 years because of families shifting to other, less safe modes of travel.The FAA's 1995 analysis was not accepted by the US National Transportation Safety Board, which argued that the FAA's estimates of diversion of travel from airplanes to cars were too high.Since their 2000 meeting, a regulation mandating CRS use for young children on airplanes has been on the National Transportation Safety Board's list of "Most Wanted Transportation Safety Improvements."Recently, the American Academy of Pediatrics Committee on Injury and Poison Prevention released a policy statement recommending a federal requirement for CRS use on airplanes for all children younger than 2 years.The American Academy of Pediatrics committee acknowledged the FAA's previous concern that parents might opt to drive rather than pay for tickets for their young children but dismissed this possibility, stating "no data support this argument."The committee also did not present any analyses of the numbers of lives that might be saved by the policy or the policy's costs.Because we disagree with the American Academy of Pediatrics committee's assertion that no data support the FAA's concern about travel substitutionand believe that analyses of benefits and costs can inform policy decisions like this one, we estimated the possible benefits, risks, and costs of the proposed policy by using a range of values for key unknown variables. We were specifically interested in how many young child air crash deaths might be prevented, the threshold proportion of families switching from air to car travel above which the projected harms of the policy would exceed its projected benefits, and how little the extra seats for young children would have to cost for the policy to approach the cost-effectiveness of other available injury prevention interventions.METHODSThe key inputs for our calculations, sources, and ranges for sensitivity analyses are summarized in Table 1. We obtained the risk of death per 100 million enplanements according to the US National Transportation Safety Board databaseby averaging data for years 1982 through 2001, the period for which data were available. (An enplanement is defined as "a revenue passenger boarding an aircraft.") We estimated the total number of enplanements per year at 650 million, the approximate average of the last 5 years. Air travel decreased substantially after September 11, 2001, but was rising steadily before that; we used the average because we have no way of predicting future changes. We used the FAA's estimate that about 1% of enplanements are by children younger than 2 years for a total of 6.5 million enplanements in that age group.Table 1. Inputs for Calculations of Potential Numbers of Deaths Prevented by Reduced Air Fatalities and Caused by Diversion to Ground TravelVariableBase Case ValueRangeReference or SourceAir travelOverall child and adult death rate per 100 million enplanements*27.7NANTSBNo. of enplanements per round trip2.5NAEstimateNo. of enplanements per year650 000 000NANTSBPercentage of round trips by children younger than 2 years1.0NAFAA2Proportion of deaths occurring in survivable crashes0.3NANTSBRelative risk of death for unrestrained young children in survivable crashes, as compared with risk for restrained young children2.51.5-4.0FAA,2Berg et al,and Corneli et alRisk of death for restrained young children in survivable crashes, as compared with overall passenger risk0.50.2-1.0FAA,2Berg et al,and Corneli et alGround travelAverage family size traveling, including child3NAEstimateAdjusted average family size traveling, including child†2.4NAEstimateAverage US motor vehicle occupancy1.6NABureau of Transportation StatisticsRisk of car death for families with young children, as compared with that for average drivers0.30.2-1.0Evans et alAverage motor vehicle death rate per 100 million vehicle-miles traveled1.5NAUS Department of TransportationAverage net number of miles driven per diverted enplanement300200-400EstimateAbbreviations: FAA, US Federal Aviation Administration; NA, not applicable; NTSB, US National Transportation Safety Board.*An enplanement is defined as "a revenue passenger boarding an aircraft."†See "Methods" section.We estimated in 2 steps the number of deaths of children younger than 2 years that might be prevented by CRS use. First we estimated the proportion of fatalities that occur in survivable crashes. For simplicity, we defined a survivable crash as any crash in which there were survivors. Of the 2784 deaths on US air carriers from 1982 through 2001, 832 (30%) occurred as a result of crashes in which there were survivors,so we used 30% as our estimate. We assumed CRS use would reduce fatalities only for these survivable crashes. We then estimated the number of lives that could be saved by CRS use as a function of 2 relative risks: the risk of death for restrained children younger than 2 years, as compared with that of other passengers, and the risk of death for unrestrained young children, as compared with that for restrained young children. The relative risk for unrestrained young children, as compared with that for other passengers, is thus the product of these 2 relative risks. Because only 1% of enplanements are by children younger than 2 years, we used the risk for all passengers in place of the risk for passengers older than 2 years to simplify calculations.For our base case estimate, we used 2 sources of data: extrapolations from car crash data and the FAA's 1995 report to Congress,which included detailed analyses of children in survivable aircraft crashes from 1978 through 1994. In fatal car crashes, restrained infants and toddlers have about 20% to 60% lower risk of death or serious injury than do restrained adults, even when their tendency to ride in the back seat, which is safer, is taken into account.The FAA's analysis used a 50% lower risk of death or serious injury in restrained young children, as compared with that in other passengers.We used a 50% lower risk for restrained young children, as compared with that in other passengers, as our base case, and we varied this between 0.2 and 1.0 in sensitivity analyses. In car crashes, unrestrained young children have about 2 to 3 times the risk of death or serious injury, as compared with that in restrained young children; in the FAA analysis of airplane crashes, the risk ratio for unrestrained vs restrained young children was 2.25. We used 2.5 as our base case, varying it in sensitivity analyses from 1.5 to 4.0.The number of motor vehicle fatalities that might be caused by diversion of travelers from planes to cars depends not only on the proportion of young children whose parents make this decision but also on the average distance they would drive and the average risk per mile driven. On the other hand, fatalities from plane crashes are related to the number of enplanements rather than to the number of miles flown. We estimated that the average net increase in car travel (ie, driving distance to the destination minus driving distance to the airport) per enplanement for families switching from planes to cars would be 300 miles, with a range of 200 to 400 miles for sensitivity analyses. Assuming 2.5 enplanements per round trip, this would translate into 750 miles per round trip, with a range of 500 to 1000 miles for sensitivity analyses.Because the risk of motor vehicle fatality is provided per 100 million vehicle-miles traveled,we adjusted for ways in which the risk per vehicle-mile traveled by these children and their families might differ from the averages reflected in the statistics. This risk depends on the average vehicle occupancy for these trips, because the higher the occupancy rate the greater the potential number of fatalities per mile traveled. However, the number of fatalities per mile traveled does not increase proportionately to the number of passengers because passengers have lower death rates than drivers do and because death rates per mile traveled also include deaths in nonpassengers (eg, pedestrians).We estimated the average vehicle occupancy for the extra trips at 3 persons, as compared with the national average of 1.6 persons,but adjusted this occupancy downward to 2.4 (ie, only 1.5 rather than 1.9 times the national average) to account for the lower death rates per mile for passengers vs drivers.In addition to the factors noted, the risk of motor vehicle fatalities depends on risk characteristics of the drivers. Evans et alestimated coefficients by which the average risk can be multiplied to take into account differences in risk of drivers, cars, and types of driving. We used their coefficients and assumed drivers would be about 30 years old (0.7) and not drink any alcohol when driving (0.6). The coefficient for rural interstate driving was 0.53; we used 0.7 to reflect higher than average but not exclusive use of rural interstate highways. The product of these coefficients (0.7 × 0.6 × 0.7) is 0.3; we used this as our base case, and we used estimates from 0.2 to 1.0 in sensitivity analyses.We used commercially available software for calculations (Excel 97; Microsoft Corp, Redmond, Wash) and for confirming all calculations and producing sensitivity analysis graphs (DATA 3.5; TreeAge Corp, Williamstown, Mass). We rounded numbers displayed in tables but carried through all calculations with full precision. Cost estimates are based on July 2002 US dollars; comparisons with dollars from other years were adjusted by using the Consumer Price Index.To convert costs per enplanement to costs per round trip, we estimated that the average round trip included 2.5 enplanements, which is equivalent to estimating that 75% of round trips are nonstop and 25% require a single plane change in each direction. Spreadsheets with all calculations, the downloaded FAA crash data, and the decision trees we used are available on request.RESULTSAccording to our base case assumptions, the estimated number of young child air travel deaths that would be prevented by CRS use is about 0.4 deaths per year. This number could be as low as 0.05 deaths per year if unrestrained young children had only 1.5 times the risk of restrained young children and if restrained young children had only 20% of the risk of all passengers. Alternatively, it could be as high as 1.6 deaths per year if unrestrained young children had 4 times the risk of restrained young children and if restrained young children had no lower risk than did other passengers (Table 2). In this last case, the risk of death of unrestrained young children would be 4 times the risk of other passengers. Because the risk of death among the 3360 passengers in survivable crashes from 1982 through 2001 was 25%and the risk of death in young children in such crashes cannot exceed 100%, this last relative risk cannot exceed 4.Table 2. Potential Number of Child Fatalities From US Air Travel Prevented per Year by Use of a Child-Restraint System*Relative Risk for Unrestrained Young Children, as Compared With That for Restrained Young ChildrenRelative Risk for Restrained Young Children, as Compared With That for All Passengers0.200.350.500.751.001.50.050.090.140.200.272.00.110.190.270.410.542.50.160.280.410.610.813.00.220.380.540.811.084.00.320.570.811.221.62*Assumes 6.5 million young child enplanements (defined as "revenue passenger boarding an aircraft") per year, 27.7 total deaths per 100 million enplanements, and 30% of deaths occurring in survivable crashes. The relative risk for restrained young children, as compared with that for all passengers is the product of the row and column headings. Base case value is in boldface.We estimated the threshold proportion of families switching from air to car travel at which the projected increase in motor vehicle deaths would exceed the projected reduction in plane crash deaths from CRS use. For our base case, this proportion was 5.4%. As the number of miles per enplanement decreases, the safety advantage of air travel compared with that of car travel decreases and the proportion of families that can switch to car travel without causing a net increase in deaths increases. About 13% of families could switch from air to car travel without a net increase in deaths if their average enplanement were for only 200 miles (Figure 1). For enplanements for fewer than 135 miles, driving is estimated to be safer. On the other hand, as the number of miles driven per diverted enplanement and the proportion of families choosing to drive increase, we project a small increase in deaths (Figure 1).Figure 1.Net effect of mandated child-restraint system use as a function of the proportion of families who choose to drive rather than fly and the average number of miles driven per diverted enplanement. The labeled isocontour lines show the specified net change in the annual number of deaths.The break-even proportion for switching from air to car travel is also sensitive to assumptions about the relative risk of death for the families choosing to drive, particularly as that relative risk declines below the base case estimate of about 0.3. If the motor vehicle death rate per vehicle-mile traveled for families switching from air to car travel were one fifth of the national average (ie, a relative risk of 0.20), about 12% could switch from air to car travel without a net increase in deaths (Figure 2). If the risk per vehicle-mile traveled of families switching to car travel were the same as the national average, car deaths caused would exceed air crash deaths prevented if only about 1% of families chose to drive (Figure 2).Figure 2.Net effect of mandated child-restraint system use as a function of the proportion of families who choose to drive rather than fly and the relative risk of automotive fatality for families choosing to drive, as compared with that of average drivers. The labeled isocontour lines show the specified net change in the annual number of deaths.Because the policy would lead to a net increase in deaths if more than about 5% to 10% of families switched from air to car travel, we assumed that no families would switch when we calculated the cost of the policy per round trip per death prevented. The cost per year is simply the number of enplanements by young children per year (about 6.5 million) times the cost per round trip divided by 2.5 (the estimated number of enplanements per round trip). To calculate the cost per death prevented, we divide this cost by the number of deaths prevented per year (0.4). Thus, ignoring the possibility of increased deaths due to diversion to car travel, the base case estimate for cost of mandatory CRS use is about $6.4 million per death prevented for each dollar cost of the policy per round trip per young child (Table 3). For example, if the additional cost per round trip were $200 per young child, the cost per death prevented, ignoring car crash deaths, would be about $1.3 billion.Table 3. Cost of Mandatory Child-Restraint System Use Policy per Life Saved and per Life-Year Saved, Assuming No Increase in Deaths From Ground TravelCost per Round Trip, $Cost per Life Saved,* $Cost per Life-Year Saved (3% Discounting),† $200.001 283 594 06342 786 469100.00641 797 03221 393 23450.00320 898 51610 696 61720.00128 359 4064 278 64710.0064 179 7032 139 3235.0032 089 8521 069 6621.006 417 970213 9320.251 604 49353 483*Lives saved by families forgoing travel altogether are not included. Assumes 0.4 lives saved per 6.5 million child enplanements (defined as "a revenue passenger boarding an aircraft") and 2.5 enplanements per round trip.†Based on life expectancy of 75 years, which is equivalent to 30 years with 3% discounting.To obtain the estimated cost per life-year saved, we assumed a 75-year additional life expectancy and 3% discounting, so the cost per death prevented can be divided by 30. That is, the cost per discounted life-year is about $214 000 for each dollar cost per round trip per young child. Thus, if the average round-trip cost per young child were $200, the cost per discounted life-year saved would be about $43 million. Put another way, for the cost per discounted life-year to be in the range of $50 000, the cost of complying with the regulation in our base case would need to be about $0.25 per round trip per young child and there could be no diversion to cars. These costs can easily be adjusted to account for different assumptions about the possible number of young child air crash deaths prevented per year. For example, at the extreme of the sensitivity analysis, where the number of lives saved by requiring CRS use is 1.6 per year and there is no increase in deaths from motor vehicle crashes, the cost per life-year saved would be about $53 000 for each dollar cost per round trip per young child.COMMENTUsing available data on the risk of fatalities from air travel and the survivability of crashes and reasonable assumptions for relative risks of death for restrained and unrestrained young children involved in crashes, we found that the number of deaths that could be prevented in the United States with mandatory CRS use in commercial aircraft is small–probably less than 1 and almost certainly less than 2 per year. The number of deaths that could be prevented by mandatory CRS use is limited because the number of deaths of unrestrained young children in survivable crashes is already low.Because of the small projected absolute benefit of the policy, it is important to examine its risks—not because the risks are large but because they could be small and still exceed the projected benefits. We found that a policy of requiring CRS use for airplane travel is likely to lead to a net increase in deaths caused by increased motor vehicle travel if the proportion of families switching to automobile travel exceeds about 5% to 10%. This threshold varied with the estimated number of lives saved by CRS use on airplanes, the average length of the added round trips by car, and the risk profile of the drivers but is unlikely to exceed 15%.The small magnitude of potential benefit per young child also makes the cost per life saved high unless the cost per round trip per young child is close to zero. Even if the policy led to no increase in car travel and cost only $20 per round trip per young child, the cost per life saved would be about $4.3 million per discounted life-year. After we adjusted the cost-effectiveness estimates in a review by Tengs et alupward to 2002 dollars, the cost is still more than 33 000 times the cost per life-year of mandatory seat belts and child restraints for cars and more than 60 times the median cost per life-year saved of 132 other fatal-injury prevention interventions.Our estimates for young child deaths that might be prevented by CRS use are based partly on extrapolations from car crash data. Unlike car crashes, in survivable airplane crashes, only about 60% of deaths are caused by the impact; most of the remainder are caused by heat, smoke, and toxic fumes. Use of CRSs presumably would provide less protection from some of these injuries than from injuries related to impact. On the other hand, in at least 1 widely publicized crash, lack of a CRS might have interfered with locating and evacuating a young child.Because of lack of data for airplane crashes and these 2 differences from car crashes acting in opposite directions, we extrapolated from car crash data and acknowledged uncertainties in sensitivity analyses.Our results for possibly preventable young child air crash deaths are in general agreement with those of others who have examined these data. Our base case estimate of about 0.4 air crash deaths per year prevented by CRS use is virtually the same as the FAA's estimate of 5 deaths in 10 years. Fife et alalso estimated about 0.6 lives per year could be saved by CRS use. Their estimate for the relative risk for unrestrained young children, as compared with that for adults, was about 6, which is higher than ours, but they excluded crashes in which injuries were not caused by deceleration or in which all deaths occurred in compartments with no survivors. In addition, their analysis was based on only 5 young child deaths in 14 crashes from 1976 through 1979.The number of additional deaths due to car travel depends largely on the proportion of families choosing to drive rather than fly, which we did not attempt to estimate in this study. The FAA used economic modeling to estimate that about 20% of families would opt for car rather than air travel.Their model was more comprehensive than the one reported here because it took into account other modes of ground travel besides cars and the likelihood that, depending on the distance of the trip and the increase in the fare, many families would forgo travel entirely, which would lead to a small reduction in air fatalities in older passengers and in children.Although full details of the calculations are not provided, the reason that the FAA estimate of a net increase of 8.2 deaths per year is higher than ours appears to be primarily because of higher estimates for the number of families choosing to drive and a smaller downward adjustment in the risk per mile driven.McKenzie and Leeestimated an increase of 5 deaths and 175 disabling injuries per year from car travel as a result of mandatory CRS use—a number higher than ours, probably for the same reasons. Our model shows results similar to those of the FAAand McKenzie and Leeif the proportion of families choosing to drive is about 20% and the relative risk of death from these car crashes is about the national average (Figure 2).There are many factors we did not consider in this analysis. We did not consider possible benefits of CRS use in reducing nonfatal injuries. However, the net number of serious nonfatal injuries prevented is likely to be small because serious nonfatal injuries from air travel are less common than are fatalities.In contrast, serious nonfatal injuries from motor vehicle travel are about 80 times as common as are fatalities.Thus, the reduction of nonfatal injuries from CRS use is likely to be small, and the possible increase from diversion to car travel would be much larger. We also did not consider a variety of other outcomes that might be affected by a policy of mandatory CRS use. These outcomes include convenience of families and comfort of young children and of surrounding passengers, which could be enhanced by the extra room available for the child in a CRS or diminished if children were kept in CRSs when they wanted to be held or breastfed. Other possible benefits of CRS use we did not consider include decreased anxiety of parents and airline personnel and reduction of injuries to young children during turbulence and to surrounding passengers from unrestrained young children during crashes.Our cost estimates were all expressed as cost per round trip per child. The main cost would be the cost of a ticket, but there would also be the cost of buying a CRS approved for airplane use and additional costs to airlines and inconvenience to passengers of ensuring adherence to the rule. For example, airlines would need to ensure that the model of safety seat brought by parents was approved for air travel and would need a supply of approved safety seats to loan or rent to passengers who arrive at the airport without an approved seat.We also did not estimate costs and benefits of alternatives to mandatory CRS use on aircraft. One scenario modeled by the FAA and found far superior to mandatory CRS use(p2-10)is to allow a family member to reserve an adjacent seat for a young child's use on a space-available basis. If this were done, except when flights were full, parents could put their child in a CRS in the seat next to them, without needing to buy a ticket and without revenue loss to the airline. An approach that might reduce injuries due to turbulence is a safety harness that attaches to the parent's seat belt. However, because of the risk to the child of being crushed between the adult and the seat in front in the event of a crash, these harnesses are not approved by the FAA for use during takeoff and landing.Additional research could allow us to estimate some of the parameters of our model with greater confidence. Such parameters include the average family size and average number of enplanements per round trip of families traveling with young children, the performance of different brands of CRSs in airplane crash simulations, and the average net number of vehicle-miles that families electing to drive rather than fly would travel. In addition, families could be surveyed about the price sensitivity of their decision to drive rather than to buy a ticket for their child. The trouble with this last approach, however, is that the answers to such questions on a survey might not match what people actually would do, and the answers would probably strongly depend on whether and how the projected benefits of CRS use were quantified for parents. We do not advocate additional research into any of these areas because we believe the probability that the results would substantively affect the lopsided cost-benefit balance of a mandatory CRS policy is close to zero.We conclude that a policy of requiring CRS use on aircraft is likely to lead to a small net increase in deaths and injuries unless the cost of complying with the policy is low enough that fewer than 5% to 10% of families with young children switch from air to car travel. Even if young children in CRSs were allowed to fly at no cost to their families, the associated societal cost would need to be less than $1 per child per round trip for the cost-effectiveness of CRS use on aircraft to approach that of other available injury prevention interventions.What This Study AddsUnrestrained children younger than 2 years have died in potentially survivable airplane crashes. However, the effects, costs, and risks of attempting to prevent such deaths by requiring CRSs for all children flying on airplanes are not known.Results of this study show that requiring CRSs on airplanes would prevent few airplane crash deaths and might cause an increase in motor vehicle deaths if many families switched to travel by car rather than paying additional fares for their young children. Irrespective of that possibility, the cost of the regulation per death prevented would be high–about $1.3 billion if the price of the round-trip ticket for the young child were $200. These findings suggest that a policy requiring CRSs on airplanes would be a poor use of societal resources.Federal Aviation Administration National Aviation Safety Data Analysis CenterNTSB safety recommendations to the FAA with FAA responses database recommendation A-95-51.Available at: http://www.nasdac.faa.gov. Accessed March 12, 2002.Federal Aviation AdministrationReport to Congress: Child Restraint Systems: Report of the Secretary of Transportation to the United States Congress Pursuant to Section 522 of the Federal Aviation Administration Authorization Act of 1994, Pub L No.103-305. Washington, DC: US Dept of Transportation; 1995.BSweedlerTestimony of Barry Sweedler, Director, Office of Safety Recommendations, National Transportation Safety Board, before the Committee on Transportation and Infrastructure Subcommittee on Aviation, House of Representatives, regarding legislation to require the use of child safety restraint systems aboard aircraft.August 1, 1996. Available at: http://www.ntsb.gov/speeches/S960801.htm. Accessed March 12, 2002.American Academy of Pediatrics Committee on Injury and Poison PreventionRestraint use on aircraft.Pediatrics.2001;108:1218-1222.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11694707PLAndersonRDMcLellanJPOvertonGLWolframPrice elasticity of demand.Available at: http://www.mackinac.org/1247. Accessed March 13, 2003.Air Transport AssociationEconomics glossary: common aviation terms and references.Available at: http://www.airlines.org/public/industry/display1.asp?nid=5244. Accessed July 23, 2001.US National Transportation Safety BoardTable 3: passenger injuries and injury rates, 1983 through 2002, for US air carriers operating under 14 CFR 121.Available at: http://www.ntsb.gov/aviation/Table3.htm.Accessed April 14, 2002.US National Transportation Safety BoardAccidents involving passenger fatalities US airlines (part 121), 1982present.March 2002. Available at: http://www.ntsb.gov/aviation/Paxfatal.htm. Accessed April 15, 2002.MDBergLCookHMCorneliDDVernonJMDeanEffect of seating position and restraint use on injuries to children in motor vehicle crashes.Pediatrics.2000;105:831-835.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10742328HMCorneliLJCookJMDeanAdults and children in severe motor vehicle crashes: a matched-pairs study.Ann Emerg Med.2000;36:340-345.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11020681Bureau of Transportation StatisticsNational transportation statistics 2000.Available at: http://www.bts.gov/publications/nts/2000/index.html. Accessed March 20, 2002.LEvansMCFrickRCSchwingIs it safer to fly or drive?Risk Anal.1990;10:239-246.US Department of Transportation National Highway Traffic Safety AdministrationTraffic Safety Facts 1999: A Compilation of Motor Vehicle Crash Data From the Fatality Analysis Reporting System and the General Estimates System.Washington, DC: US Dept of Transportation, National Highway Traffic Safety Administration; 2000:15.CJohnstonFPRivaraRSoderbergChildren in car crashes: analysis of data for injury and use of restraints.Pediatrics.1994;93:960-965.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8190584US Department of Labor, Bureau of Labor StatisticsConsumer Price Index, all urban consumers (CPI-U).Available at: ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt. Accessed September 6, 2002.TOTengsMEAdamsJSPliskinFive-hundred life-saving interventions and their cost-effectiveness.Risk Anal.1995;15:369-390.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7604170European Transport Safety CouncilIncreasing the surival rate in aircraft accidents: impact protection, fire survivability and evacuation.Available at: http://www.etsc.be/rep_air2.htm. Accessed April 1, 2003.JBrown-LohrTestimony before the US House of Representatives Subcommittee on Aviation: hearing on HR 1309: child safety restraint systems requirement on commercial aircraft.Available at: http://commdocs.house.gov/committees/Trans/hpw104-63.000/hpw104-63_0.htm. Accessed April 1, 2003.DFifeBRosnerWMcKibbenRelative mortality of unbelted infant passengers and belted non-infant passengers in air accidents with survivors.Am J Public Health.1981;71:1242-1246.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7294268RBMcKenzieDRLeeEnding the free airplane rides of infants: a myopic method of saving lives: CATO Institute Briefing Paper, No. 11.Available at: http://www.cato.org/pubs/briefs/bp-011.html. Accessed April 15, 2002.Corresponding author: Thomas B. Newman, MD, MPH, Dept of Epidemiology and Biostatistics, University of California, San Francisco, School of Medicine, Box 0560, San Francisco, CA 94143 (e-mail: [email protected]).Accepted for publication December 3, 2002.

Powell, Elizabeth C.; Hampers, Louis C.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.978pmid: 14557158

ObjectivesTo describe the variation among physicians in test ordering when caring for children with gastroenteritis and to explore the effect of hospital charge information on such variation.DesignProspective, nonmasked, observational study and controlled trial of price information.SettingUrban, university-affiliated pediatric emergency department.ParticipantsPediatric emergency medicine faculty (n = 10) and fellows (n = 5).MethodsTest-ordering practices were reviewed during 3 periods: control, intervention, and washout. During the intervention period, test charge information was placed on patients' emergency department records. Telephone contact with families was initiated 7 days after care.ResultsWe included 3198 visits. Individual physician mean test charges varied more than 2-fold during the control period (mean, $127; range, $82-$185). Based on their test charges (control period), physicians were assigned to the "high" (n = 8) or "low" (n = 7) test user group. Differences in mean charges in high vs low test users during the control period ($144 vs $112) persisted in the intervention period ($80 vs $52; Mann-Whitney P= .01), as did rates of intravenous fluid use (20% vs 14% in both periods). Among the lowest-acuity patients, low test users exhibited greater price sensitivity (vs high users). Patients treated by low test users did not differ in improved condition (82% vs 86%) or family satisfaction (93% vs 92%); they had more unscheduled follow-up (25% vs 17%; P<.01), but were no more often admitted (5% vs 3%; P= .11).ConclusionsPhysicians varied in resource use when treating children with gastroenteritis. High and low test users were sensitive to price information. This intervention did not seem to compromise patient outcome.PRACTICE VARIATION among physicians is well documented.This variation in part reflects uncertainty in medicine: although the number of clinical guidelines is increasing, in many specific clinical situations physicians determine the tests or interventions needed for their patients.Thus, clinicians develop individual styles of patient management based on their experiences, their training, and their temperament. Some physicians are comfortable ordering fewer tests, whereas others prefer the additional information made available by using more studies.Inefficient use of diagnostic tests at academic institutions has been observed.This seems in part to be a result of the academic environment, where the variety of payer relationships and hospitalwide billing procedures effectively insulate the individual physician from the relative costs of diagnostic testing.In contrast, physicians in community practices face logistical and direct financial disincentives to ordering tests.Test-ordering behavior may also reflect resident training, most of which occurs in the academic setting, and the career track of many academic pediatric emergency medicine physicians, who remain in an academic setting after completing their training.Although it is difficult to know the specific tests needed to provide high-quality medical care, it is clear that inefficient use of diagnostic studies contributes to health care costs.Providing test price information has been demonstrated to result in fewer tests being ordered.In a previously published work, Hampers et aldescribed the general effect of price information in the emergency department (ED), without detailing interphysician variations. In this study, we describe the variation in test-ordering behavior in a single setting (the pediatric ED) among physicians with similar training (pediatric emergency medicine faculty and fellows) in managing a single disease (gastroenteritis). We also looked at the effects of the provision of test charge information on individual physician test-ordering practices and patient outcomes.METHODSThis study was conducted in the ED of an urban, tertiary care pediatric hospital with an annual volume of 39 000 patients. Board-certified pediatric emergency medicine physicians or fellows, working independently, supervised the care of all patients studied. Although pediatric, emergency medicine, and family practice house staff participated in the care of many patients, they had limited autonomous decision-making authority.On arrival at the ED, patients were triaged by acuity of illness to 1 of 4 categories: emergent, urgent high, urgent low, and nonurgent. Between September 1, 1997, and March 31, 1998, a data form was attached to each patient chart at triage. The form asked physicians to identify patients aged 2 months to 10 years who had a complaint of vomiting, diarrhea, decreased oral intake, or fever. We defined gastroenteritis as vomiting or diarrhea, and we included all children with those complaints. Children with chronic illness (a history of immunosuppression or immunodeficiency, an inborn error of metabolism, or a ventriculoperitoneal shunt) were excluded.Data were collected during 3 periods. During the control period (September 1, 1997, to December 31, 1997), physicians were asked to circle the tests they ordered on the data form. During the intervention period (January 1, 1998, to March 31, 1998), physicians were asked to circle the tests they ordered on a similar data form that included the hospital charge for each test. During the washout period (April 1-30, 1998), the forms used in the control period were again used. The control period was 1 month longer than the intervention period because of seasonal variation in the incidence of gastroenteritis; this allowed enrollment of a similar number of cases in each of these 2 periods. Hospital charges for individual tests did not change during the study.One of the investigators (L.C.H.) or a research assistant reviewed the ED records after the visit to collect the following information: age, race or ethnicity (as recorded by registration personnel), insurance status, initial vital signs, diagnostic testing, use of intravenous fluids, disposition (admitted to the hospital or discharged from the ED), and attending or fellow supervising care. We calculated years in practice for supervising physicians as years since completion of residency.During the control and intervention periods, patient families were contacted by telephone approximately 1 week after the ED visit. Using a structured interview, caretakers were asked to describe the child's overall condition (better, same, or worse), whether they had taken the child to see a health care provider since the ED visit (scheduled or unscheduled and office or urgent care/ED), and whether the child had been admitted to the hospital. If the caretaker had been with the child at the time of the ED visit, they were asked to describe their overall satisfaction with their child's care (very unsatisfied, somewhat unsatisfied, somewhat satisfied, very satisfied). Caretakers who spoke only Spanish were interviewed in Spanish. Because the washout period was designed only to estimate the effect of the intervention on test charges, telephone interviews were not conducted during that time.Data were analyzed using statistical software (SPSS for Windows; SPSS Inc, Chicago, Ill). For comparisons between groups of categorical data we used the χ2test; we used a 2-tailed ttest for continuous variables. Test charges, which were not normally distributed, were compared between physicians and between periods using the Mann-Whitney test; 95% confidence intervals (CIs) for means are reported. Statistical significance was set at P<.05. The study was approved by the institutional review board of Children's Memorial Hospital (Chicago).RESULTSData were available for 1415 visits during the control period and 1429 visits during the intervention period. Review of daily ED records suggested that the data we report represented an estimated 90% of eligible cases; reasons for inappropriate exclusion included clerical error (no form was attached to the ED record) and physician error (the physician did not complete the form). The demographic and clinical characteristics of eligible patients who were not included in this study did not differ from those of patients who were included.Each physician (n = 15) treated a mean (SD) of 94 (29) patients (range, 40-136) during the control period and 95 (31) patients (range, 44-159) during the intervention period.The 15 participating physicians exhibited wide variation in their clinical approaches. Physician-specific mean test charges during the control period ranged from $82.40 to $185.29 (overall mean, $126.66; median, $74.00). Based on this variation in test charges, physicians were divided into 2 groups: "high" test users (6 faculty physicians and 2 fellows) and "low" test users (4 faculty physicians and 3 fellows). Mean (SD) duration of practice was 7 (5) years for high test users and 4 (1) years for low test users (Mann-Whitney P= .20). Mean test charges associated with high vs low test users during the control period were $143.84 (95% CI, $130.39-$157.28) vs $112.43 (95% CI, $101.32-$123.54) (Figure 1).Test charges associated with high and low test users during the control and intervention periods.The demographic and clinical characteristics of children cared for by high and low test users are given in Table 1. There were no significant differences between groups in race or ethnicity, insurance status, triage category, or clinical characteristics. We also compared the case-mix of patients within each period; there were no significant differences between high and low test users in patient distribution by triage category during either period (χ2, control period: P= .40; and intervention period, P= .98).Table 1. Demographic and Clinical Characteristics of Children Cared for by 8 High and 7 Low Test Users*CharacteristicHigh Users (n = 1351)Low Users (n = 1493)PValueRace or ethnicity, No. (%)White219 (16)251 (17).02Black307 (23)343 (23)Hispanic701 (52)775 (52)Other74 (6)71 (5)Not available50 (4)53 (4)Insurance, No. (%)Commercial388 (29)433 (29).46Medicaid801 (59)903 (60)Uninsured162 (12)157 (10)Triage category, No. (%)Emergent3 (0.2)5 (0.3).64Urgent high266 (20)277 (19)Urgent low617 (46)707 (47)Nonurgent465 (34)504 (34)Clinical characteristics, mean (SD)Age, mo33 (30)34 (30).34Temperature, °C37.8 (1.2)37.8 (1.2).40Heart rate, beats/min137 (25)136 (24).12Respiratory rate, breaths/min34 (11)34 (11).13*Because of rounding, percentages may not all sum to 100.Table 2gives test charges for patients cared for by high and low users during the control and intervention periods stratified by triage category. Because emergent patients accounted for less than 0.5% of each group, they were combined with urgent high patients for ease of data analysis and presentation. Within triage categories, the charge differences between high and low test users were statistically significant for each category and within each period except for urgent high patients during the intervention period. Overall, the mean charges for tests were 48% (95% CI, 39%-56%) lower during the period when physicians were provided with price information. The biggest declines were observed among low test users (–53% overall) and for patients who were triaged as nonurgent (–58% overall and –67% for low test users).Table 2. Test Charges per Patient in the Control and Intervention Periods for Children Cared for by High and Low Test UsersVariableControl PeriodIntervention Period% Change*Patients, No.Charges, Mean, $Patients, No.Charges, Mean, $OverallHigh users641143.8471080.42−44Low users774112.4371952.43−53Subtotal1415122.66142966.33−48Triage categoryUrgent highHigh users148283.09121178.98−36Low users159228.16123154.85−32Subtotal307254.64244167.31−34Pvalue.02.37Urgent lowHigh users283136.9033485.71−37Low users366111.8434144.98−60Subtotal649122.7767565.13−47Pvalue†.03.01NonurgentHigh users21055.0425526.25−52Low users24939.3925512.98−67Subtotal<45946.5551019.61−58Pvalue†.05.01*P= .01 for all.†By Mann-Whitney test.Among patients admitted to the hospital (17% during the control period and 11% during the intervention period), there were no significant differences in test charges between high and low test users during either period (control and intervention, Mann-Whitney, P= .56 for both), and there were no significant decreases in charges between periods (Mann-Whitney, P= .09). Among patients not admitted to the hospital, there were significant mean charge differences between periods (control period, $85.58; intervention period, $36.88; Mann-Whitney P<.01) and within periods between high and low test users.More patients had no tests ordered during the intervention period (66% vs 41%; χ2P<.01) (Table 3). The differences between high and low test users were significant for each test and for no tests during the control period. These differences were less marked during the intervention period.Table 3. Tests Obtained by High and Low Users During the Control and Intervention Periods*TestPatients, No. (%)Control Period (n = 1415)Intervention Period (n = 1429)NoneHigh users (n = 641)232 (36)435 (61)Low users (n = 774)347 (45)513 (71)Subtotal579 (41)948 (66)Pvalue†.01.01Complete blood cell countHigh users162 (25)97 (14)Low users149 (19)72 (10)Subtotal311 (22)169 (12)Pvalue†.01.03Serum electrolytesHigh users96 (15)67 (9)Low users78 (10)43 (6)Subtotal174 (12)110 (8)Pvalue†.01.01UrinalysisHigh users133 (21)74 (10)Low users123 (16)56 (8)Subtotal256 (18)130 (9)Pvalue†.02.08*P<.01 for all.†By χ2test.High test users administered intravenous fluids to 20% of patients during the control and intervention periods; low test users gave intravenous fluids to 14% of their patients during each period. Patients treated vs not treated with intravenous fluids had higher test charges during the control period (chemistry charges, $72.80 vs $5.34; and total charges, $91.57 vs $303.76). Among patients treated with intravenous fluids by high test users, charges declined significantly between the control and intervention periods (chemistry charges, $77.00 vs $47.80; and total charges, $314.56 vs $223.96; Mann-Whitney P<.01 for each). Among patients treated with intravenous fluids by low test users, charges also declined significantly between periods (chemistry charges, $67.91 vs $43.86; and total charges, $291.17 vs $176.50; Mann-Whitney P<.01 for each).We successfully obtained telephone follow-up information from 556 (39%) of 1415 families during the control period and 599 (42%) of 1429 families during the intervention period; 555 (48%) of these 1155 had been cared for by high test users and 600 (52%) had been treated by low test users. The percentage of children who were "better" was similar in practice style groups (86% for high users vs 82% for low users; χ2P= .20), as was the proportion of caretakers who reported that they were "somewhat satisfied" or "very satisfied" with their ED visit (92% vs 93%; P= .40). Among children treated by low test users, there was more unscheduled follow-up (25% vs 17%; P<.01), mainly with primary care providers. Unscheduled returns to the ED occurred in 12% of patients cared for by low test users and 9% cared for by high test users (P= .08). The proportion of children who were admitted to the hospital during follow-up was similar (3% of patients treated by high test users and 5% of those treated by low test users; P= .11).A total of 354 patients were included in the brief washout period. Patients of high users (n = 135) had charges similar to those observed during the intervention period (mean, $80.79, Mann-Whitney P= .99); low users (n = 219) had a modest increase in charges (+11%), which was not statistically significant (mean, $58.71; Mann-Whitney P= .35).COMMENTThere was enormous variation among physicians in tests ordered for pediatric patients with gastroenteritis, suggesting that some practiced in a resource-intensive manner and that others provided more efficient care. Variation in patient acuity does not explain these practice differences, and neither does family insurance status. As a result, average patient charges within triage categories vary 2-fold or greater. Providing price information was associated with a significant decrease in the number of tests ordered, reduced overall charges, and reduced some of the variation among physicians in test-ordering charges, particularly among visits triaged as urgent high.Hampers et alpreviously described the general effect of price information in the ED, without detailing interphysician variations. The focus of the present study was to explore the implications of this intervention on the practice styles of faculty and fellows. In addition to reporting striking individual practice variation, we believe that this analysis provides information about the receptiveness of members of a group of academic practitioners to educational interventions designed to alter their behaviors.Results of the present analysis suggest that high and low test users are sensitive to price information. Among higher-acuity patients, high test users had greater declines than low test users, suggesting greater discretionary test use among high test users during the control period. During the intervention period, the mean charge difference between the high and low test users was less, suggesting a more consistent approach to test use in higher-acuity patients.Overall test charges declined the greatest among less acute patients (–58% for nonurgent patients), suggesting that price information had the greatest impact on discretionary test ordering when physicians were treating children who were the least ill. Despite their lower utilization rates in the control period, low test users seemed to be even more sensitive to price information than were high users, further widening the gap between these 2 groups in the intervention period. It may be that the test-ordering practices of the low test users in part reflected a predisposition to price sensitivity. The high test users may have personality traits that made them less price sensitive, resulting in higher resource use in the control period and less sensitivity to charge information. This finding suggests an interesting paradox: price information may have the greatest impact on providers who are already practicing the most efficiently. Some of the variation in test ordering may reflect training environment: high test users, on average, had been in practice longer than low test users. Low test users may have been trained during a more "cost-conscious" time.We do not believe that the resource-efficient physicians we observed did not obtain needed tests. More specifically, the data for the low test users suggested that rates of family satisfaction, reports that the child was better, and rates of hospital admission were similar to those observed among their higher-spending peers.Although mean charges declined for patients within each stratum, there were no significant differences between the control and intervention periods in charges for patients who were admitted to the hospital. Although the methods used to estimate dehydration in children are imprecise and were not standardized for this study, we observed consistency among physicians across periods in the percentage of children treated with intravenous fluids. Thus, test use among children admitted to the hospital and treatment with intravenous fluids were not sensitive to test price information. Specific charge information was provided only for tests; it is possible that additional charge information for other interventions (intravenous fluid administration and hospital admission) would alter the results we present.The washout period was brief and involved a limited number of patients. However, among high test users, overall charges during the washout period were similar to those in the intervention period, and low test users showed a modest increase. This suggests that short-term effects of the intervention on high and low test users persisted.Other pediatric studies have observed differences in test ordering specific to practice settings and to training. There is clear variation among care providers in the tests obtained on febrile infants.Differences in the care of children with febrile seizures treated in community EDs vs university-affiliated children's hospitals have also been reported.Practice variation between physicians trained in pediatric emergency medicine and those trained in emergency medicine in the management of croup has been demonstrated.The present study, performed at a single institution among physicians who had completed or were in training to complete pediatric emergency medicine fellowships, suggests that individual behaviors also account for the significant variation in test-ordering practices.Although gastroenteritis is not a high-cost disease, it is a high-volume disease, and the variation in charges becomes significant over time. There were clear differences in the use of intravenous fluids among high and low test users that was unexplained by case-mix differences. During the intervention period, fewer patients had "routine" tests ordered (serum electrolytes, complete blood cell count, urinalysis, and urine culture). A higher fraction of patients had no tests ordered. The challenge is that the "best" test-ordering strategy, which allows for only needed tests and maximizes efficiency for children with gastroenteritis, is not known.High and low test users were sensitive to price information. Although these data suggest that information that caused efficient physicians to order even fewer tests resulted in satisfactory outcomes, additional study of this issue is warranted. Because most cases of gastroenteritis in infants and toddlers have a good prognosis, it is not surprising that the variation in testing we report did not seem to impact outcome or satisfaction.It is difficult to completely separate the effect of attending physician test-ordering behavior from that of the house staff they supervised. Some of the house staff (postgraduate years 1 and 2) participated during both periods; others spent a single month during either the control or the intervention period in the ED. Because each physician (attending or fellow) supervised all house staff, it is unlikely that systematic differences between house staff explain the variation among supervising physicians. Although the fellows supervised house staff independent of the attending, it is possible that the attending had an impact on the tests that the fellow ordered. However, because each attending was present with each fellow at various times, it is unlikely that this explains the variation we report within each stratum and between periods.It is also difficult to completely separate the effect of the incidence of gastroenteritis, which has seasonal variation, on attending physician test-ordering behavior. It is possible that physicians ordered fewer tests during the intervention period because gastroenteritis was more common then. Although this may account for some of the differences between periods, it is unlikely to explain the variation in charges between high and low test users during each period. Although our case-mix seems consistent in control and intervention periods, there may have been unrecognized differences in acuity, as the proportion of children admitted to the hospital was lower in the intervention period. However, as we reported in the results, significant differences in test charges persisted in a separate analysis of patients who were not admitted to the hospital.Because there are no generally accepted outcome measures for children with gastroenteritis treated and then discharged from the ED, we chose to obtain information about family perception that the child was better, family satisfaction with the ED encounter, rates of hospital admission during follow-up, and rates of unscheduled follow-up with primary care providers or in the ED. Although we observed a higher rate of unscheduled follow-up when fewer tests were ordered, there were no differences in hospital admission rates or in parent satisfaction. The reasons for the higher rates of unscheduled follow-up are not clear but may include unmet parental expectations for tests or physician failure to obtain needed studies. We do not know whether diagnostic studies were obtained during follow-up visits. Follow-up visits contribute to overall health care costs; thus, further study is needed to address this observation. Nevertheless, we believe it is unlikely that failure to obtain studies contributed to significant morbidity, as parent satisfaction and hospital admission rates were similar for both groups.Our study is limited by failure to obtain follow-up information on many patients. We used a fairly narrow time window, and many families provided incomplete contact information. Although this limits the generalizability of our results, there seem to be no systematic biases in follow-up rates between high and low test users, suggesting that the data we present are valid.Gastroenteritis represents a narrow range of complaints, and we limited this study to healthy children. In addition, we defined gastroenteritis broadly as vomiting, diarrhea, or both, and it is possible that children with other diagnoses were included in the data we present. For most children, this illness is self-limited, and thus we were unlikely to observe poor outcomes. For this reason, much test ordering in the management of gastroenteritis is discretionary. It is clear that within a single ED there was enormous variation among physicians in test ordering, contributing to significant variation in patient charges. Moreover, all physicians were sensitive to price information, and it reduced some of the variation among physicians, particularly when caring for higher-acuity patients. Further study is needed to determine the best methods to allow physicians to provide cost-sensitive, efficient, high-quality care to their patients.What This Study AddsData about physician-specific variation in test-ordering behavior and the effects of charge information are limited. We describe the variation in test use in a single ED among physicians with similar training in managing children with gastroenteritis. The data suggest that there was significant variation and that high and low test users were sensitive to price information. Low test use did not seem to be associated with poor outcomes.JEWennbergUnderstanding geographic variations in health care delivery.N Engl J Med.1999;340:32-39.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9878643EMPetrackNCChristopherJKriwinskyPain management in the emergency department: patterns of analgesic utilization.Pediatrics.1997;99:711-714.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9113948LCHampersJLTrainorRLListernickSetting-based practice variation in the management of simple febrile seizure.Acad Emerg Med.2000;7:21-27.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10894238Not AvailableNational Guideline Clearinghouse Web site Vol 2001.Rockville, Md: Agency for Healthcare Research and Quality; 2001. Available: at http://www.guideline.gov. Accessed July 3, 2003.JMCameronThe indirect costs of graduate medical education.N Engl J Med.1985;312:1233-1238.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=3921836PJGrecoJMEisenbergChanging physicians' practices.N Engl J Med.1993;329:1271-1275.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8413397ETWongMMMcCarronSTShawOrdering of laboratory tests in a teaching hospital: can it be improved?JAMA.1983;249:3076-3080.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=6854832KGrumbachTBodenheimerMechanisms for controlling costs.JAMA.1995;273:1223-1230.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7707633DMBerwickKLColtinFeedback reduces test use in a health maintenance organization.JAMA.1986;255:1450-1454.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=3951079GMGrossmanA review of physician cost-containment strategies for laboratory testing.Med Care.1983;21:783-802.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=6888030KGrumbachTBodenheimerPainful vs painless cost control.JAMA.1994;272:1458-1464.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7933431RCSachderaLSJeffersonJCoss-BuEffects of availability of patient-related charges on practice patterns and cost containment in the pediatric intensive care unit.Crit Care Med.1996;24:501-506.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8625641LCHampersSChaDJGutglassSEKrugHJBinnsThe effect of price information on test-ordering behavior and patient outcomes in a pediatric emergency department.Pediatrics.1999;103:877-882.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10103325MSKramerEDShapiroManagement of the young febrile child: a commentary on recent practice guidelines.Pediatrics.1997;100:128-134.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9200370JAFinkelsteinCLChristiansenRPlattFever in pediatric primary care: occurrence, management, and outcomes.Pediatrics.2000;105:260-266.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10617733LCHampersSGFariesPractice variation in the emergency management of croup.Pediatrics.2002;109:505-508.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11875148Corresponding author: Elizabeth C. Powell, MD, MPH, Division of Pediatric Emergency Medicine, Children's Memorial Hospital, Box 62, Chicago, IL 60614 (e-mail: [email protected]).Accepted for publication March 13, 2003.This study was funded in part by a special project grant from the Ambulatory Pediatric Association, McLean, Va.We thank Robert R. Tanz, MD, for his thoughtful review of this manuscript.

Richardson, Laura P.; DiGiuseppe, David; Garrison, Michelle; Christakis, Dimitri A.

2003 JAMA Pediatrics

doi: 10.1001/archpedi.157.10.984pmid: 14557159

BackgroundRacial disparities have previously been noted in antidepressant use among Medicaid-covered youth.ObjectiveTo determine if racial and ethnic differences are due to dissimilarity in the prevalence of diagnosed depression or disparate treatment patterns.MethodsClaims were examined for 192 441 youth (5-18 years old) who had been continuously enrolled in Medicaid from July 1, 1997, to December 31, 1998. Diagnosed depression was defined as having been assigned an International Classification of Diseases, Ninth Revisioncode for a depressive disorder in a medical claim. Logistic regression methods were used to evaluate the association between race/ethnicity and (1) depression diagnosis and (2) depression treatment in the 6 months following a new episode of diagnosed depression. All analyses were controlled for age, sex, and rural or urban residence.ResultsTwo percent of the total sample had a depression diagnosis, 25% of which were new episodes of depression. Compared with white youth, Hispanic (odds ratio [OR], 0.51; 95% confidence interval [CI], 0.46-0.57), Asian/Pacific Islander (OR, 0.16; 95% CI, 0.12-0.21), and black (OR, 0.31; 95% CI, 0.26-0.37) youth were less likely to have a depression diagnosis. Following a new diagnosis, Native American (OR, 0.29; 95% CI, 0.18-0.46) and Hispanic (OR, 0.42; 95% CI, 0.30-0.61) youth were less likely than white youth to have received an antidepressant or a mental health specialty visit.ConclusionsRacial and ethnic disparities exist in both the prevalence and treatment of diagnosed depression. Future studies should examine underlying reasons for these disparities and how they affect the quality of care for depressed Medicaid-covered youth.BY THE age of 18 years, it is estimated that 20% of youth will have experienced at least one episode of major depression.Depressed youth are at increased risk for suicide, school failure, substance abuse, nicotine dependence, early pregnancy, and social isolation.Although psychotherapyand selective serotonin reuptake inhibitor antidepressantsare thought to be effective treatments for adolescent depression, it is not known what proportion of depressed youth receive these treatments. In the United States, less than half of youth with diagnosable mental health disorders report having received treatment for these disorders; youth with depressive disorders are less likely to receive mental health services than those with attention-deficit/hyperactivity disorder or other disruptive disorders.Another concern is the possibility that racial and ethnic disparities exist in depression diagnosis and treatment. Although community-based studies suggest that minority youth are at least as likely as white youth to have a depressive disorder or symptoms of depression,minority youth have been found to be less likely than white youth to have received antidepressant medicationsor professional help for depression.Possible explanations for these disparities include underdiagnosis and undertreatment of depression in minority youth. It is also possible that there are cultural differences in family preferences for depression treatment.Youth insured by Medicaid are an especially important population in which to examine these issues. Although Medicaid pays approximately a quarter of all mental health costs for children and adolescents in the United States,no studies have documented the prevalence of diagnosed depressive disorders in this population. Additionally, because of their social situation, youth covered by Medicaid may represent a vulnerable population for whom underdiagnosis or undertreatment may pose especially grave risks.We conducted a study with 2 primary objectives: (1) to determine the prevalence of depressive disorder diagnoses in a statewide Medicaid population and (2) to explore whether racial or ethnic disparities exist with respect to diagnosis and treatment of depressive disorders in this sample. We hypothesized that minority youth would be less likely to be diagnosed as having a depressive disorder and that once their condition was diagnosed they would be less likely to receive antidepressant medications or mental health services.METHODSSETTINGThis study was conducted using Washington State Medicaid claims data collected from July 1, 1997, through December 31, 1998. During these years, the Washington State Medicaid program covered youth up to the age of 19 years in families with incomes up to 200% of the federal poverty level. Most youth were enrolled in a Medicaid managed care plan unless they had complex chronic medical conditions or were placed in foster care, in which case they were covered by a fee-for-service plan. Administrative data captured in claims format were available for both managed care and fee-for-service plan enrollees. Services covered by Medicaid included outpatient and inpatient health care and mental health care, as well as prescription medications.SUBJECTSYouth (aged 5-18 years) were eligible to be included in this study if they were continuously enrolled in Medicaid from July 1, 1997, through December 31, 1998. This 18-month window was chosen to allow for the assessment of a full year of "new episodes of depression" (defined as having a coded visit for depression after a 6-month period without antidepressant use or depression diagnosis).DEPRESSIVE DISORDER DIAGNOSESYouth were considered to have a depressive disorder diagnosis if they had been assigned an International Classification of Diseases, Ninth Revision (ICD-9)code for a depressive disorder at any time during the 18-month study period. The depression ICD-9codes used are as follows. Depressive disorder diagnosis codes were selected based on prior claims-based studies of adult depression.ICD-9Diagnosis Category296.2Major depressive disorder, single episode296.3Major depressive disorder, recurrent episode298.0Depressive-type psychosis300.4Neurotic depression or dysthymic disorder309.0Brief depressive reaction309.1Prolonged depressive reaction309.28Adjustment disorder with mixed emotional features311Depressive disorder, not otherwise specified313.1Misery/unhappiness—childhood disturbancesBecause of concerns about stigma, we hypothesized that clinicians might be more likely to use codes of lower perceived severity for coding depression in children. Thus, codes for adjustment disorder with depressive and anxiety features (309.28), brief depressive reaction (309.0), and misery/unhappiness of childhood (313.1) were also included in our definition of depressive disorders. There were 9 ICD-9code fields available for each claim, although few records had more than 2 or 3 coded diagnoses. A youth was considered to have depression if an included code was present in any ICD-9code field. Youth with a bipolar disorder diagnosis at any time during the study period (n = 1336) were excluded from analysis, because their needs and treatment were anticipated to differ from youth with depression alone.Treatment in the 6 months following diagnosis was evaluated only in youth who presented with a "new episode of depression." Consistent with methods used by others,youth were classified as having a "new episode of depression" if they had an ICD-9code for depression after a 6-month window with no depressive disorder diagnosis and had not filled an antidepressant medication prescription.TREATMENT VARIABLESAntidepressant use was defined as having at least one pharmacy claim for a selective serotonin reuptake inhibitor, tricyclic, or other antidepressant in the 6 months following presentation with a new episode of depression. Antidepressant drugs were identified using National Drug Codes for antidepressant medications.Mental health specialty visits were also assessed in the 6 months following presentation with a new episode of depression. Youth were defined as having a mental health specialty visit if they had (1) an evaluation and management visit coded by a psychiatrist or psychologist, (2) a visit coded with a Current Procedural Terminology code for psychotherapy or psychiatric assessment by any mental health care provider, or (3) a state-specific code for a mental health visit in the capitated mental health system.RACE/ETHNICITYData on subject race and ethnicity were obtained from Medicaid eligibility data. As part of the Medicaid eligibility paperwork, parents were asked to categorize the race or ethnicity of their child on a single question on the enrollment questionnaire. Based on these responses, youth were categorized as being white, Hispanic, black, Native American, Asian/Pacific Islander, other, or unknown. All racial and ethnic groups were mutually exclusive. The "other" category consisted of youth who were identified by their parents as not fitting into one of these categories or as being of more than one racial and/or ethnic group. Youth without an identified racial or ethnic category were dropped from the analysis.COVARIATESCovariates included sex, subject age, and rural or urban residence, since each of these covariates has been shown to be associated with the likelihood of receiving a diagnosis for depressionor treatment for depression.Age categories (5-10 years, 11-14 years, and 15-18 years) were selected to assess for differences by developmental stage. They were also selected to capitalize on the known increase in prevalence of depression diagnoses for girls that occurs around the age of 15 years.Rural or urban residence status was determined using the Rural and Urban Commuting Area coding system developed for the Washington State Area in conjunction with the Federal Office of Rural Health Policy and the Department of Agriculture's Economic Research Service.We hypothesized that youth who were eligible for Medicaid based on the presence of a disabling condition or enrollment in foster care might be more likely to have depression and that reasons for Medicaid enrollment might differ by race or ethnicity. Therefore, we also controlled for category of Medicaid eligibility, with categories including low income (the TANF program [Temporary Assistance for Needy Families] and CHIP [State Children's Health Insurance Program], or Washington State–specific expanded financial eligibility criteria), disabled, in foster care, or other (eg, pregnancy and blindness). Differences between managed care and fee-for-service coverage were not examined due to complexities related to an individual's ability to change coverage status during the 18-month study period and the possibility of a managed care beneficiary receiving certain services in the fee-for-service sector.STATISTICAL ANALYSISData preparation was performed using SAS statistical software, version 8.2,and all analyses were performed using STATA 7 statistical software.Descriptive analyses were performed to assess the characteristics of depressed and nondepressed youth in our sample. The χ2methods were used to compare demographic information by race/ethnicity and by presence of depression diagnoses. Subsequently, estimates of depression diagnosis period prevalence were calculated using all eligible youth as a denominator. Given known differences in depression prevalence by age and sex,period prevalence analyses were stratified by both of these variables.To test the hypothesis that the period prevalence of depression diagnoses would vary by race and ethnicity, logistic regression analyses were performed to test race and ethnicity categories as predictors and presence of a depression diagnosis as an outcome. Logistic regression methods were also used to test the hypothesis that youth of minority racial and ethnic groups would be less likely than white youth to receive antidepressants or to have a mental health specialty visit in the 6 months following presentation with a new episode of depression. Three separate analyses were run for the treatment outcome: the first for antidepressant use, a second for having received mental health specialty care, and a third for having received either an antidepressant or mental health specialty care. All regression analyses were adjusted for age, sex, rural or urban residence, Medicaid eligibility category, and an interaction term for age and sex.RESULTSThere were 369 006 youth covered by Washington State Medicaid at any time from July 1, 1997, through December 31, 1998, and 192 441 youth who were continuously enrolled (Figure 1). There were no significant racial or ethnic differences in youth who were continuously enrolled vs those who were excluded for noncontinuous enrollment. Among continuously enrolled youth, 4048 had a depression diagnosis at any time during the 18-month study period (diagnosis sample), and 1459 met criteria for having a new episode of depression (treatment sample). The demographics of the total sample are given in Table 1.Figure 1.Study sample flow diagram.Table 1. Demographics of Total SampleCharacteristicNo. (%) (n = 192 441)Male sex97 810 (51)Age category, y5-10107 801 (56)11-1456 123 (29)15-1828 517 (15)Race or ethnicityWhite114 955 (60)Hispanic29 137 (15)Black14 478 (7)Native American5729 (3)Asian/Pacific Islander11 541 (6)Other42 621 (22)Unknown3121 (2)Residence settingRural143 978 (75)Urban46 969 (25)Medicaid eligibility categoryLow income178 431 (93)Disabled7855 (4)Foster care5810 (3)Other345 (0.2)MAIN SAMPLE DEMOGRAPHICSMales and females were evenly distributed in the main sample. More than half of the youth were 5 to 10 years old, and an additional 29% were 11 to 14 years old. Consistent with the population in Washington State, youth in the sample were predominantly white, with Hispanic youth and black youth as the 2 next largest self-identified racial and ethnic groups. Twenty-two percent of our sample was not identified as belonging solely to any one of the available racial or ethnic categories. A small number of youth (2% of the original sample) were coded as having an unknown race or ethnicity; these youth were dropped from further analyses.Approximately 75% of study youth lived in urban settings and 25% lived in rural settings. Most youth qualified for Medicaid due to low income status (93%) vs disability (4%), foster care enrollment (3%), or other reason for eligibility (0.2%). Compared with white youth, black and Native American youth were less likely to be enrolled for financial reasons and more likely to be enrolled due to foster care placement; Asian and Hispanic youth were more likely than white youth to be enrolled based on low income status (P<.001). Hispanic and Native American youth were more likely than white youth to live in rural settings, whereas Asian/Pacific Islander youth and black youth were more likely to live in urban settings (P<.001). Asian/Pacific Islander youth in our sample tended to be older than white youth (P<.001), whereas Hispanic youth tended to be younger (P<.001). Black and Native American youth had similar age distributions to white youth. There were no differences in sex distribution by race or ethnicity.DIAGNOSESOverall, 4084 youth (2% of study sample) had a diagnosis of depression on a claim at some time during the study period. Consistent with community-based depression studies, the period prevalence of depression diagnoses increased with age (P<.001). The male-female ratio of youth with depression diagnoses shifted from 2:1 in the 5- to 10-year-old group to 1:2 in the 15- to 18-year-old group. Although the magnitude of depression period prevalence varied with race and ethnicity, similar patterns of the age and sex distribution of depression diagnosis were seen within all racial and ethnic groups. The period prevalence ranged from a low of 0.3 per 1000 population in 5- to 10-year-old male Asian/Pacific Islanders to a high of 94 per 1000 population in 15- to 18-year-old female Native Americans (Figure 2and Figure 3). In white youth, the age- and sex-adjusted period prevalence of depression diagnoses was 36 per 1000 population compared with 18 per 1000 population in Hispanics, 12 per 1000 population in blacks, 5 per 1000 population in Asian/Pacific Islanders, 40 per 1000 population in Native Americans, and 17 per 1000 population for youth in the "other" category.Figure 2.Prevalence of depression diagnoses per 1000 males.Figure 3.Prevalence of depression diagnoses per 1000 females.With the exception of Native Americans, youth from racial and ethnic minority groups were less likely to have a depression diagnosis present when compared with white youth (Hispanic: odds ratio [OR], 0.51; 95% confidence interval [CI], 0.46-0.57; black: OR, 0.31; 95% CI, 0.26-0.37; Asian/Pacific Islander: OR, 0.16; 95% CI, 0.12-0.21; and other: OR, 0.50; 95% CI, = 0.43-0.58). Native American youth had a similar rate of depression diagnoses compared with white youth (OR, 1.14; 95% CI, 0.98-1.34). These differences were present after adjustment for sex, age, Medicaid eligibility category, and rural or urban residence.TREATMENTThe treatment analysis was conducted among 1459 youth (0.8% of the total sample) who met criteria for having a new episode of depression. The age, sex, and racial and ethnic distributions of new episodes of depression were similar to those seen in the prevalence analyses of the entire sample. Overall, 37% of youth who had a new episode of depression had filled an antidepressant medication prescription, and 26% of youth had a mental health specialty visit in the subsequent 6 months. Fifty-five percent of youth received at least 1 of these 2 forms of depression treatment.Youth for whom race/ethnicity category was unknown were dropped from the sample, leaving 1435 youth for the regression analysis. After adjusting for age, sex, and rural or urban residence, Hispanic (OR, 0.42; 95% CI, 0.30-0.61), Native American (OR, 0.29; 95% CI, 0.18-0.46), and other (OR, 0.55; 95% CI, 0.34-0.88) youth were less likely than white youth to have the combined outcome of either a filled antidepressant prescription or mental health visit in the 6 months following presentation with a new episode of depression (Table 2). There were no significant differences between white youth and black or Asian/Pacific Islander youth. Our definition of depression was broad and included some diagnoses for which treatment might not be given, such as adjustment disorder. In a post hoc analysis, we narrowed our definition to include only diagnostic categories that are used to assess treatment in adult studies (ICD-9codes 296.2, 296.3, 300.4, and 311). Limiting analyses to youth who had one of these diagnostic codes did not significantly change the results of our analysis.Table 2. Likelihood of Filling an Antidepressant Prescription or Having a Mental Health Visit Within 6 Months of New Episode of Depression*Race/EthnicityNo. of Subjects (n = 1435)Odds Ratio (95% Confidence Interval)Any Antidepressant Prescription FilledAny Mental Health VisitAny Mental Health Visit or Antidepressant Prescription FilledWhite1048ReferentReferentReferentNative American1540.22 (0.12-0.41)0.63 (0.36-1.09)0.29 (0.18-0.46)Hispanic900.45 (0.30-0.69)0.61 (0.40-0.93)0.42 (0.30-0.61)Black480.68 (0.37-1.28)0.75 (0.37-1.50)0.65 (0.36-1.17)Asian/Pacific Islander190.76 (0.29-1.97)1.37 (0.50-3.71)0.88 (0.35-2.20)Other760.62 (0.37-1.02)1.15 (0.68-1.96)0.55 (0.35-0.88)*All analyses adjusted for sex, age, sex × age interaction, Medicaid eligibility category, and rural or urban residence.COMMENTThe findings of our study suggest that racial and ethnic disparities exist in both the prevalence of diagnosed depression and the treatment of depression once identified. Hispanic, black, and Asian/Pacific Islander youth in this Medicaid sample were less likely than white youth to have a medical claim with a diagnosis of depression. In addition, Native American and Hispanic youth were less likely than white youth to have mental health specialty care or to have filled an antidepressant medication prescription in the 6 months following presentation with a new episode of depression.Assuming that the underlying prevalence of depression is similar in all racial and ethnic groups,what are possible explanations for the observed racial and ethnic disparities in the depression diagnosis prevalence? First, physician identification of depressive disorders may differ by race and ethnicity. Mental health disorders may be particularly prone to diagnostic bias because patterns of depressive symptoms and the stigma associated with mental health disorders are culturally based and cultural experiences may differ by race and ethnicity.Primary care physicians have been shown to be less likely to detect depression in adult members of some racial minority groups.Additionally, the concordance of race and ethnicity between patient and physician has been shown to be associated with patient trust of their physician.Lack of trust in a physician might prevent patients and their families from discussing sensitive mental health issues.Second, differences in depression diagnosis prevalence in our data may be due to racial and ethnic differences in accessing the health care system for depression and mental health treatment preferences. Compared with white youth, minority youth report that they are more likely to receive mental health services outside the medical sector in school settings.Other settings where youth may receive mental health services that might not be captured in our claims data include churches, community centers, and social services.Third, continuity of medical care may also play an important role in the racial and ethnic disparities in diagnosis noted in our study. Minority youth have been shown to be less likely than white youth to have a regular source of care, to receive scheduled immunizations, and to obtain appropriate medications for asthma.Physician familiarity with pediatric patients has been associated both with identification of psychosocial problems and with treatment compliance.Finally, although racial/ethnic differences in depression prevalence have not been observed in community-based studies,Medicaid-covered youth are not a random sample from the community. It is possible that the factors that contribute to the need for Medicaid coverage and to the development of depression, such as parental depression or family disruption, may differ by race and ethnicity. However, even if the observed differences in diagnosis are due to racial and ethnic variation in underlying depression prevalence, the apparent disparities in depression treatment are still concerning.It is not clear why there are racial and ethnic differences in depression treatment once depression is diagnosed. Although community-based studies do not support lesser severity of depression in minority compared with white youth,it may be that racial and ethnic differences in the reporting of depression symptoms result in the perception of lower depression severity in minority youth. Another possibility is that there is physician bias in depression treatment; prior studieshave shown that racial and ethnic bias exists in treatment for other diseases, such as adult cardiac disease. Racial and ethnic differences in depression treatment may also reflect patient and family preferences for specific types of depression treatment. For example, alternative therapies are commonly used in both Native American and Hispanic cultures, and adult members of some minority groups report a high level of distrust regarding the use of psychotropic medications.In our data, lack of receipt of services may reflect either the lack of prescriptions or referrals by physicians or lack of follow-through by patients and families. A lack of follow-through may also account for differences between racial and ethnic disparities noted in claims data but not on physician self-report of prescribed treatments.Limitations in our design may have affected our results. First, this study was conducted in a Medicaid sample from a single state, and the results may not be generalizable to other settings. This may explain why racial and ethnic disparities in diagnosis and treatment were noted in our study but not in a prior studyconducted in private pediatric office settings. Second, it is possible that claims data do not completely represent all care that youth have received. This may particularly be a problem for patients in managed care systems where claims are not required for reimbursement. Although we do not anticipate that there would be large racial and ethnic differences in claims reporting, to the extent that claims data are incomplete, we may have underestimated the overall prevalence of treatment in our sample. Third, the categories of race and ethnicity were broad; there are likely to be differences within categories that are not accounted for in this analysis. Finally, few black and Asian youth were diagnosed as having depression in our sample, resulting in a small sample size for the treatment analysis. It is possible that the lack of difference between white youth and youth in these categories was due to the small sample size.Because these were administrative data, we did not have information regarding depression that was diagnosed but not coded. Although this lack of coding may have led to underestimation of prevalence estimates, it should only have contributed to racial and ethnic disparities if physicians have different depression coding practices by race and ethnicity. Since the prevalence of depression in white youth (3.5%) was similar to or less than community-based estimates (5%),it is unlikely that the differences observed in our study represent overdiagnosis of depression in Medicaid youth.Despite these limitations, this study raises important questions regarding depression treatment and detection in Medicaid youth. Racial and ethnic disparities were noted both in depression diagnoses and in treatment of new episodes of depression. Future studies should address the replicability of these findings, the underlying reasons for these disparities, racial and ethnic differences in care-seeking behaviors and treatment preferences, and how these differences affect quality of care for depressed youth.What This Study AddsPrior studies have shown that racial and ethnic disparities exist in the use of antidepressants and receipt of psychotherapy among youth. However, few studies have specifically studied diagnosis and treatment patterns for depression in this age group. In this study, we examined racial and ethnic disparities in both the presence of coded diagnoses of depression and treatment once depression was identified.We found that, with the exception of Native American youth, youth who were members of racial and ethnic minority groups were less likely to have a coded diagnosis of depression than white youth. We also found that Native American and Hispanic youth were less likely to have received antidepressants or a mental health visit in the 6 months after depression was diagnosed. Thus, racial and ethnic disparities were noted both in depression diagnoses and in treatment of new episodes of depression. Future studies regarding racial and ethnic differences in care-seeking behaviors and treatment preferences and how these differences affect quality of care for depressed youth may increase our understanding regarding why these disparities exist and may provide insight regarding how we can improve the treatment of depression for all youth.PMLewinsohnHHopsRERobertsJRSeeleyJAAndrewsAdolescent psychopathology, I: prevalence and incidence of depression and other DSM-III-R disorders in high school students.J Abnorm Psychol.1993;102:133-144.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8436689BBirmaherNDRyanDEWilliamsonDABrentJKaufmanChildhood and adolescent depression: a review of the past 10 years, part II.J Am Acad Child Adolesc Psychiatry.1996;35:1575-1583.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8973063MKovacsPresentation and course of major depressive disorder during childhood and later years of the life span.J Am Acad Child Adolesc Psychiatry.1996;35:705-715.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8682751KDMichaelSLCrowleyHow effective are treatments for child and adolescent depression? a meta-analytic review.Clin Psychol Rev.2002;22:247-269.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11806021GJEmslieJHHeiligensteinKDWagnerFluoxetine for acute treatment of depression in children and adolescents: a placebo-controlled, randomized clinical trial.J Am Acad Child Adolesc Psychiatry.2002;41:1205-1215.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=12364842MBKellerNDRyanMStroberEfficacy of paroxetine in the treatment of adolescent major depression: a randomized, controlled trial.J Am Acad Child Adolesc Psychiatry.2001;40:762-772.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11437014MStroberMDeAntonioSSchmidt-LacknerThe pharmacotherapy of depressive illness in adolescents: an open-label comparison of fluoxetine with imipramine-treated historical controls.J Clin Psychiatry.1999;60:164-169.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10192591PJLeafMAlegriaPCohenMental health service use in the community and schools: results from the four-community MECA study.J Am Acad Child Adolesc Psychiatry.1996;35:889-897.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8768348PWuCWHovenHRBirdDepressive and disruptive disorders and mental health service utilization in children and adolescents.J Am Acad Child Adolesc Psychiatry.1999;38:1081-1090; discussion 90-2.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10504806PCohenSKasenJSBrookELStrueningDiagnostic predictors of treatment patterns in a cohort of adolescents.J Am Acad Child Adolesc Psychiatry.1991;30:989-993.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=1757449US Department of Health and Human ServicesMental Health: Culture, Race and Ethnicity. A Supplement to Mental Health: A Report of the Surgeon General.Rockville, Md: US Dept of Health and Human Services; 2001.JMZitoDJSaferServices and prevention: pharmacoepidemiology of antidepressant use.Biol Psychiatry.2001;49:1121-117.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11430854PWuCWHovenPCohenFactors associated with use of mental health services for depression by children and adolescents.Psychiatr Serv.2001;52:189-195.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11157117RSturmJRingelCBaoNational Estimates of Mental Health Utilization and Expenditures for Children in 1998.Bethesda, Md: National Institute of Mental Health; 2000. NIMH Working Paper 205.CAMelfiTWCroghanUse of claims data for research on treatment and outcomes of depression care.Med Care.1999;37(4 suppl Lilly):AS77-AS80.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10217395KRostJHumphreyKKelleherPhysician management preferences and barriers to care for rural patients with depression.Arch Fam Med.1994;3:409-414.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8032501MMWeissmanVWarnerPWickramaratneDMoreauMOlfsonOffspring of depressed parents: 10 years later.Arch Gen Psychiatry.1997;54:932-940.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9337774Not AvailableRural Health Research Center Web page.Available at: http://www.fammed.washington.edu/wwamirhrc/rucas/RUCA_description.htm.Not AvailableSAS for Windows, Version 8.2.Cary, NC: SAS Institute Inc; 1999.StatacorpSTATA Statistical Software: release 7.0.College Station, Tex: Stata Corporation; 2001.US Public Health ServiceMental Health: A Report of the Surgeon General.Rockville, Md: US Dept of Health and Human Services, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health; 1999.LJKirmayerDGroleauAffective disorders in cultural context.Psychiatr Clin North Am.2001;24:465-478, vii.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=11593857SJBorowskyLVRubensteinLSMeredithPCampMJackson-TricheKBWellsWho is at risk of nondetection of mental health problems in primary care?J Gen Intern Med.2000;15:381-388.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10886472SSahaMKomaromyTDKoepsellABBindmanPatient-physician racial concordance and the perceived quality and use of health care.Arch Intern Med.1999;159:997-1004.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10326942PAEgglestonFJMalveauxAMButzMedications used by children with asthma living in the inner city.Pediatrics.1998;101(3 pt 1):349-354.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9480996BAHahnChildren's health: racial and ethnic differences in the use of prescription medications.Pediatrics.1995;95:727-732.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=7724312PWNewacheckDCHughesJJStoddardChildren's access to primary care: differences by race, income, and insurance status.Pediatrics.1996;97:26-32.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=8545220KJKelleherGEChildsRCWassermanTKMcInernyPANuttingWPGardnerInsurance status and recognition of psychosocial problems: a report from the Pediatric Research in Office Settings and the Ambulatory Sentinel Practice Networks.Arch Pediatr Adolesc Med.1997;151:1109-1115.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=9369872ECharneyRBynumDEldredgeHow well do patients take oral penicillin? a collaborative study in private practice.Pediatrics.1967;40:188-195.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=5006583KASchulmanJABerlinWHarlessThe effect of race and sex on physicians' recommendations for cardiac catheterization.N Engl J Med.1999;340:618-626.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10029647KJKelleherCDMooreGEChildsMLAngelilliDMComerPatient race and ethnicity in primary care management of child behavior problems: a report from PROS and ASPN.Med Care.1999;37:1092-1104.http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=10549612Corresponding author: Laura P. Richardson, MD, MPH, Child Health Institute, 6200 NE 74th St, Suite 210, Seattle, WA 98115 (e-mail: [email protected]).Accepted for publication March 28, 2002.This project was supported in part by funding from the Nesholm Family Foundation. Dr Richardson also receives funding through the Klingenstein Third Generation Foundation (New York, NY) Fellowship in Child and Adolescent Depression and the Poncin Family Scholarship (University of Washington School of Medicine, Seattle).