Geographic Variation of Asthma Quality Measures Within and Between Health Plans

Published on: 
, , , ,
The American Journal of Managed Care, December 2005, Volume 11, Issue 12

Objective: To contrast health plan performance in treating pediatricasthma within and between different geographic regions.

Study Design: Retrospective analysis of administrative claimsdata for 18 health plans serving Medicaid enrollees.

Methods: The study population was 3970 children 5-18 yearsold with persistent asthma who were continuously enrolled in thesame Michigan Medicaid health plan for 2002 and 2003, with noother source of health insurance. Outcome measures were assessedbased on national guidelines for asthma management: at least 1asthma controller medication prescription, at least 1 outpatientvisit, 1 or more asthma emergency department visits, and an annualinfluenza vaccination.


Results: Adherence to national guidelines varied significantly(≤ .05) between plans. The proportion of children with at least 1asthma controller prescription ranged from 66% to 88%; the proportionof children with influenza vaccination ranged from 3% to46%. Plan ranking varied depending on the guideline measureused. The plan with the lowest aggregate proportion of childrenwho had asthma controller prescriptions (66%) had regional proportionsthat ranged widely, from 44% to 72%. Some plans wereobserved to rank highly in performance in 1 region and substantiallylower in other regions; similar with-plan regional variation wasfound for each outcome measure.

Conclusions: Assessments of adherence to Medicaid pediatricasthma management guidelines at the plan level may be insufficientto identify opportunities for improvement. Administrativeclaims-based profiles of plan performance that are sensitive toregional variations in plan characteristics may be particularly usefulin isolating and prioritizing quality-improvement opportunities.

(Am J Manag Care. 2005;11:765-772)

Over the past decade, quality of care has becomea central consideration in the selection ofhealth plans by employers. Performance measuressuch as the National Committee on QualityAssurance (NCQA) Health Plan Employer Data andInformation Set (HEDIS) measure of appropriate medicationuse are at the forefront for assessing quality ofcare among those with persistent asthma.1-3 Purchasersof healthcare services are interested in plan-to-plan performancevariation in asthma management, given thepotential economic benefits of effectively controllingthis common chronic disease.4-7

Guidelines for the

Diagnosis and Management of Asthma

Since 1991, the National Asthma Education andPrevention Program (NAEPP) has provided recommendationsfor effective asthma treatment, including guidelineshighlighting the components of care that areregarded as essential to asthma management, includingasthma assessment and monitoring, control of factorscontributing to asthma severity, pharmacotherapy, andpatient education. The NAEPP's 8 and the 2002update9 represent the most comprehensive, sciencebasedconsensus on asthma diagnosis and management;key clinical activities have been extracted from theseguidelines.10 Following the NAEPP guidelines has thepotential to reduce asthma-related deaths, hospitalizations,and school absences among children, which arenotably high among minority children.4

Despite the longstanding availability of NAEPP guidelines,wide variations in adherence to these asthmamanagement recommendations have been reported.11-19Prior studies have found ample opportunities forimprovement among Medicaid-enrolled children withasthma, with underuse of asthma controller medications,frequent emergency department (ED) utilization,and high rates of hospitalizations being reported.13,14,18,20,21 There is evidence suggesting the beneficialeffect of managed care on Medicaid enrollees with conditionssuch as asthma,22 but it is not clear that healthplans exhibit consistent performance across geographicareas. Although performance measures can be obtainedfor health plans operating in a given geographic area,these assessments are based on aggregate, plan-widedata.23 Understanding the degree to which aggregateplan-level assessments (eg, HEDIS) reflect plan performancein a specific locale is important to purchasers andhealth plans alike. Medicaid and health plan officials areespecially interested in comparative data on health planperformance in areas that represent particular challenges,because they can use these data to establish reasonabletargets for improvement.

The objective of this study was to characterize adherenceto guidelines for management of pediatric asthmaamong health plans serving Medicaid beneficiaries. Weexamined adherence from 2 perspectives: (1) variationbetween plans, contrasting the performance of plansoperating in similar geographic areas; and (2) variationwithin plans, comparing outcomes for enrollees withinthe same plan, but living in different geographic areas.In doing so, we explored the degree to which aggregatemeasures of health plan quality based on NAEPP guidelinesreflected plan performance within a given geographicarea.


This study is based on a retrospective analysis ofadministrative claims data for the Michigan Medicaidprogram and was approved by the University ofMichigan institutional review board.

Study Population

A total of 5792 children between 5 and 18 years ofage in the Michigan Medicaid program were classified ashaving persistent asthma by HEDIS criteria during 2years (2002 and 2003), allowing up to a 1-month enrollmentgap each year. The HEDIS criteria for persistentasthma require administrative claims evidence of a childhaving any of the following within a calendar year: at least4 asthma medication-dispensing events, at least 1 inpatientor 1 ED claim with a primary diagnosis of asthma, orat least 4 outpatient visits with an asthma diagnosis andat least 2 asthma medication-dispensing events.3 HEDISclassifies an asthma medication-dispensing event as 1prescription for an asthma medication for an amount lasting30 days; these medications are identified by a comprehensivelist from NCQA.24 Although HEDIS criteriacurrently require that health plans assess asthma outcomesusing administrative claims data for services furnishedduring 1 year (ie, the "measurement year") topersons identified as having asthma during the prior year,we used a more stringent case definition. We requiredclaims evidence of asthma in both years because there isevidence suggesting that the specificity of the HEDIS criterionfor identifying asthma cases can be improved byrequiring claims evidence of asthma in the measurementyear in addition to the prior year.25,26

We excluded children not enrolled in the same managedcare plan throughout the study period (n = 168),children enrolled in fee-for-service Medicaid (n = 1345),and children with other health insurance (n = 309) toensure that claims for all services were available inMedicaid administrative data files. Our resulting sampleconsisted of 3970 children enrolled in 18 plans, representing69% of Medicaid children with persistent asthma.Demographic characteristics and location of residencewere obtained from Medicaid eligibility files. The countyof beneficiary residence was classified as being urban ornonurban based on the presence of a metropolitan statisticalarea, as defined by the US Census Bureau.27

Children were classified a priori into 1 of 4 geographicareas based on general similarities in population andhealth-services characteristics. Region 1 was comprisedof children residing in 3 urban centers that are broadlyperceived by Medicaid and health plan officials as areasin which plans consistently have difficulty meetingMedicaid program performance objectives. The largelyrural counties in the northern portion of the state,including the Upper Peninsula, were included in region2. Region 3 included the suburban southeastern counties,and region 4 was comprised of the southwesterncounties. Of the 18 plans in our sample, 10 had studysubjects in 2 or more of these regions.

Outcomes Measured

We used a variety of performance measures for pediatricasthma management to better understand thedegree to which these measures are consistent with thesingle HEDIS measure for appropriate asthma medication,which is widely used by commercial and Medicaidhealth plans. Our 4 indicators of asthma managementwere based on NAEPP recommendations10: (1) the proportionof children with persistent asthma who had atleast 1 asthma controller medication dispensing event in2003; (2) the proportion of children with persistentasthma who had at least 1 outpatient visit in 2003; (3)the proportion of children with persistent asthma whohad 1 or more asthma ED visits in 2003; and (4) the proportionof children with persistent asthma who receivedan influenza vaccination in 2003. We included an indicatorof long-term asthma controller medication usebecause these medications are essential in the preventionof exacerbations and chronic symptoms amongthose with persistent asthma.10 Our outpatient visitmeasure was based on the recommendation that asthmapatients be seen by a physician once every 1-6months, depending on the severity of the asthma.10Because asthma issues can potentially be reviewed by aphysician at any office visit, we used a less stringentrequirement and determined those children who had atleast 1 outpatient visit (acute or routine) for any condition.We included an indicator of ED use because thesevisits are an indicator of poorly controlled asthma orinadequate planning for asthma exacerbations.8 Ourmeasure for influenza vaccination is based on the recommendationby the Centers for Disease Control andPrevention that all persons with chronic disease, includingasthma, receive an influenza vaccination annually.28

International Classification of Diseases, Ninth

Revision, Clinical Modification (ICD-9-CM)


Outcomes were determined using Medicaid administrativeclaims data reported by health plans for healthservicesencounters and pharmacy claims. Our outcomemeasure for asthma controller medication reflectsNCQA HEDIS criteria, which include several classes ofasthma medications: inhaled corticosteroids, cromolyn/nedocromil, leukotriene modifiers, and methylxanthines.3 Outpatient visits were identified based onCurrent Procedural Terminology (CPT) procedure ordiagnosiscodes indicating evaluation and management, preventive,or general medical examination visits. We classifiedED visits based on CPT procedure codes (99281-99285)and the Uniform Bill-92 (UB-92) revenue code (0450) forED services. We determined the proportion of subjectswith an influenza vaccination based on CPT procedurecodes (90657-90660, 90724) and diagnosiscodes (V048 and V066).

Statistical Analysis

Chi-square tests were performed to assess differencesin demographic characteristics between regions.The mean proportion and95% confidence intervalswere estimated for eachplan's overall outcomesbased on 2003 claims.Regional outcomes wereestimated when a plan had≥30 subjects in a region.To aid in comparison, weranked each plan on eachmeasure (ED visits wereranked in reverse order).Reduction and summarizationof claims data weredone with SAS version 8software (SAS InstituteInc, Cary, NC).



Table 1 illustrates thedemographic characteristicsof the 3970 childrenwith persistent asthma in our study sample. The asthmahealth services utilized by the majority of children(65%) during the study period met multiple HEDIS criteriafor persistent asthma; 34.5% had only asthma prescriptionclaims, and 0.5% had only asthma ED visits.Most of our subjects were younger than 15 years of age(82%), predominantly male (59%), and black (59%). Themajority of subjects (90%) lived in urban areas. Region 1(urban centers) had a significantly higher proportion ofchildren who were between 10 and 14 years of age andwere black, and both region 1 and region 3 were entirelyurban or suburban (≤ .001).






Table 2 summarizes each plan's outcome measuresand the corresponding rank order relative to otherplans. Wide variations were observed for each outcomemeasure compared with statewide mean values. Forexample, prescriptions of controller medications rangedbetween plans from 88% to 66% (≤ .05). Four plans (E,R, M, and A) had controller medication use that was significantlyhigher than the statewide mean (74.8%), and 1plan (D) had controller medication use that was significantlylower (≤ .05). Similar variation was observedbetween plans for each remaining outcome measure. Sixplans (R, K, E, M, N, and F) had proportions of childrenwith at least 1 outpatient visit that were significantlyhigher than the statewide mean (82.6%); however, in 1plan (B) that proportion was significantly lower (≤.05). The proportions of children with asthma ED visitswere significantly higher than the statewide mean(27.7%) for 4 plans (N, M, L, and K), but were significantlylower for plans B and Q (≤ .05). Plans H, K, and Ehad influenza vaccination rates significantly higher thanthe statewide average (16.8%), whereas plans G, C, andB each had vaccination rates that were significantlylower (≤ .05).

Taking each of the 4 outcome measures intoaccount, 1 plan (E) had consistently favorable rankingsand 2 (B and C) had consistently unfavorable rankings.However, plan performance often was inconsistentacross the 4 outcome measures. For example, plan Awas ranked 4th based on use of long-term controllermedications, but had a less favorable ranking for outpatientvisits (9th), ED visits (15th), and influenza vaccinations(9th). Similarly, plan D was ranked last (18th)for use of long-term controller medications, but hadmore favorable rankings for outpatient visits (11th), EDvisits (7th), and influenza vaccinations (12th).



Plan enrollees were stratified into the 4 geographicregions; in total, there were 30 combinations of plans byregion, yielding a subsample of 3780 children (95%) ofour initial sample. Table 3 presents the 4 asthma outcomemeasures by plan for each of the 4 regions. Plansin urban centers were observed to have outcomes thatvaried significantly between plans. For example, thetable illustrates that among the 9 plans in region 1, planA had significantly higher rates of asthma controllermedication use, office visits, and influenza vaccinationscompared with most other plans in the region (≤ .05).Rates varied widely across measures in each region; forexample, use of long-term asthma controller medicationvaried by as little as 12% in region 2 and as much as 44%in region 4. Figure 1 shows an example of regional variation,using asthma controller medications. Plan A,where the aggregate use of these medications was 79%,had outcomes ranging from 82% in region 1 to 67% inregion 4 (≤ .05). Similarly, plan G had regional resultsthat varied substantially from its aggregate outcome of74%, ranging from 80% (region 1) to 66% (region 4).




Plans with a relatively favorable performance rankingin 1 region did not necessarily have a comparable rankingin another region. In addition, plans with favorablerankings based on 1 outcome did not necessarily haveconcordant rankings using another measure. Figure 2illustrates these results for plan D, showing regional differencesfor each of the 4 outcomes measured. The figure shows that asthmacontroller medicationuse in regions 2 and 3was significantly higher(≤ .05) than use inregion 4. The proportionof children with atleast 1 outpatient visitwas significantly higherin region 2 than region1 (≤ .05), and alsowas higher than that inregion 3. Asthma EDuse in plan D varied significantlybetween regions,ranging from16% in region 3 to 39%in region 1 (≤ .05).Influenza vaccinationrates ranged from alow of none reported(region 4) to 20% (region3).


Quality-improvement(QI) initiatives arecommonplace amonghealth plans, and theresources devoted tothese efforts continuesto increase.29-31 Ourfindings suggest theimportance of recognizingthat individualhealth plans may have adiverse profile for performancein pediatricasthma management bygeographic region; thus,aggregate quality indicatorsmay not adequatelydescribe planexperiences. We alsofound that plan performanceoften is inconsistentacross multiplemeasures of quality of asthma care based on NAEPPguidelines. Some plans had notably low proportions ofannual outpatient visits—a fundamental indicator ofaccess to care—yet also had relatively low rates of asthmaED visits. Our findings also indicate that relying on asingle HEDIS-type aggregate measure of asthma care willfail to capture within-plan variations that may reflect differencesin (1) the underlying severity of illness amongenrollees, (2) characteristics of the physician network inspecific locales, (3) access to local and regional healthcareservices (eg, EDs, public health services), and (4)community norms.


In addition, the observed within-plan variation mayreflect underlying differences by region with respect tothe demographic composition of enrollees, which priorstudies have found to influence patient compliance withrecommended asthma management strategies and relatedoutcome measures.13,32-34 Other findings from a priorstudy of Michigan Medicaid plans illustrate that pediatricasthma QI programs varied substantially betweenplans in terms of comprehensiveness, use of providerincentives, and the inclusion of allied health professionals.35 However, we also observed in that study that fewplans used QI programs with interventions that areknown to be effective in changing physician practices oroutcomes for asthma patients. The findings from thisstudy suggest that plans can use regional stratification ofasthma outcomes to identify and prioritize QI initiativestoward the areas in greatest need.

The finding that asthma outcomes varied withinplans is consistent with other studies, both for asthmaoutcomes measures17 and other measures of quality.36Though some have argued for HEDIS-type measures atthe physician-practice level,37 the feasibility of thisapproach is limited due to small sample sizes.38Moreover, variability in individual physician performanceexists from 1 patient to another,39 as well as acrossdifferent aspects of quality measures.40 In addition, ourassessment of pediatric asthma management outcomessuggests that despite the many challenges, some plansoperating in urban centers are able to perform at substantiallyhigher levels than other plans.

This study has several limitations. We used a HEDISbasedapproach with administrative claims data, becausethis methodology is used widely in the United States andis familiar to Medicaid and health plan officials. Althoughthe completeness of the data reported by MichiganMedicaid plans is generally good, some incompletenessmay be present (eg, influenza vaccinations obtainedthrough health fairs or free clinics). Despite variations inthe timeliness, accuracy, and completeness of plans'claims data, findings from a recent study suggest thatpediatric asthma outcome measures are robust to dataloss and accurately reflect plan quality even when datamay be incomplete.41 We used 2 years of eligibility datato establish our study population. Plans considering theapplication of this approach to asthma and other conditionsmay improve their specificity by applying HEDISeligibility criteria to 2 years as the basis for identifyingcases.25,26 Another limitation is that our findings arebased on administrative claims data and severity of asthmacould not be determined; consequently, potential differencesin severity of asthma between and within plansare not reflected in these results.

Our findings have several implications for healthplans and Medicaid program managers. We found thatmany plans had quality-of-care measures indicatingthat improvements focused on care practices for asthmapatients are warranted. Several indicators of asthmamanagement based on national guidelines illustrate agenerally high reliance on ED care and notably lowinfluenza vaccination rates among children with persistentasthma. Results from several recent studies offerperspectives on potential interventions that focus onreducing ED use15,42-45 and improving rates of influenzavaccination46 among children with asthma.

Our findings also suggest that aggregate assessmentsof Medicaid pediatric asthma management practices atthe overall plan level are not sufficient to identify targetedopportunities for improvement. Profiles of plan performancethat are sensitive to regional variations in plancharacteristics may be particularly useful in isolatingand prioritizing QI opportunities. Medicaid plans canuse this information to gauge the adequacy of existingprovider networks in selected geographic areas and totarget areas that would benefit most from more intensiveintervention.

Although our approach to establishing geographicregions was based on county (or city) of residence, applicationof this approach in other states requires anunderstanding of the geographic distribution of planenrollees, health plan participation, and health-servicesavailability. Consequently, Medicaid and health planofficials may wish to characterize geographic areas thatdo not necessarily follow existing boundaries establishedfor local health departments or health-planningpurposes.


Although there is some additional administrative burden,these findings demonstrate that claims data can beused to expand the scope of performance measurementbeyond what is currently considered by the HEDIS program.The method described here offers a more broadbasedassessment of plan quality, with a more diverseselection of outcome measures that may more accuratelyportray asthma management among health planenrollees. Health plans may find this approach particularlyvaluable when assessing the effects of asthma QIefforts in selected areas of operation; it also may be usefulfor evaluating other outcomes and health plan performancecriteria.


The authors thank Susan Moran, BSN, MPH, of the MichiganDepartment of Community Health for her insights and comments regardingthis study.

From the Child Health Evaluation and Research (CHEAR) Unit, Division of GeneralPediatrics, University of Michigan, Ann Arbor, Mich.

This study was supported by the Michigan Department of Community Health.

Address correspondence to: Kevin J. Dombkowski, DrPH, MS, University of Michigan,Division of General Pediatrics, 300 N Ingalls, Ann Arbor, MI 48109-0456.

J Natl Med Assoc.

1. Ohldin A, Mims A. The search for value in health care: a review of the NationalCommittee for Quality Assurance efforts. 2002;94(5):344-350.

J Healthc Qual.

2. O'Kane ME. Margaret E. O'Kane on healthcare accreditation. Interview byWanda Bishop and Kevin C. Park. 2001;23(3):21-25.


2003, Volume 2: Technical Specifications.

3. National Committee for Quality Assurance. Use of appropriate medications forpeople with asthma. In: ® Washington, DC: National Committee for Quality Assurance; 2003.

4. National Center for Health Statistics. Asthma prevalence, health care use andmortality, 2002. February 10, 2005. Available at: Accessed February 16. 2005.

Manag Care Interface.

5. Bootman JL, Crown WH, Luskin AT. Clinical and economic effects of suboptimallycontrolled asthma. 2004;17(1):31-36.

Am J Manag


6. Skrepnek GH, Skrepnek SV. Epidemiology, clinical and economic burden, andnatural history of chronic obstructive pulmonary disease and asthma. 2004;10(5 suppl):S129-S138.

J Allergy Clin Immunol.

7. Weiss KB, Sullivan SD, Lyttle CS. Trends in the cost of illness for asthma in theUnited States, 1985-1994. 2000;106(3):493-499.

Guidelines for the

Diagnosis and Management of Asthma.

8. National Asthma Education and Prevention Program. Bethesda, Md: National Institutes of Health,National Heart, Lung, and Blood Institute; July 1997. Publication 97-4051.

Guidelines for the

Diagnosis and Mangement of Asthma— Update on Selected Topics, 2002.

9. National Asthma Education and Prevention Program. Bethesda, Md: National Institutes of Health, National Heart, Lung, and BloodInstitute; June 2003. Publication 02-5074.

MMWR Morb Mortal Wkly Rep.

10. Centers for Disease Control and Prevention. Key clinical activities for qualityasthma care: recommendations of the National Asthma Education and PreventionProgram. 2003;52(RR-6):1-8.

J Pediatr.

11. Walders N, Kopel SJ, Koinis-Mitchell D, McQuaid EL. Patterns of quick-reliefand long-term controller medication use in pediatric asthma. 2005;146(2):177-182.

Med Care.

12. Fuhlbrigge A, Carey VJ, Adams RJ, et al. Evaluation of asthma prescriptionmeasures and health system performance based on emergency department utilization.2004;42(5):465-471.


13. Shields AE, Comstock C, Weiss KB. Variations in asthma care by race/ethnicityamong children enrolled in a state Medicaid program. 2004;113(3):496-504.

Arch Pediatr

Adolesc Med.

14. Finkelstein JA, Lozano P, Farber HJ, Miroshnik I, Lieu TA. Underuse of controllermedications among Medicaid-insured children with asthma. 2002;156(6):562-567.


15. Stempel DA, Roberts CS, Stanford RH. Treatment patterns in the months priorto and after asthma-related emergency department visit. 2004;126(1):75-80.

Ann Epidemiol.

16. Allen-Ramey FC, Samet JM, Rand CS, Joseph CL. Trends in use of inhaledcorticosteroids for asthma management: 1994-1998. 2004;14(3):161-167.

Am J Manag Care.

17. Legorreta AP, Liu X, Zaher CA, Jatulis DE. Variation in managing asthma: experienceat the medical group level in California. 2000;6(4):445-453.

Ann Allergy Asthma Immunol.

18. Apter AJ, Van Hoof TJ, Sherwin TE, Casey BA, Petrillo MK, Meehan TP.Assessing the quality of asthma care provided to Medicaid patients enrolled in managedcare organizations in Connecticut. 2001;86(2):211-218.

Arch Pediatr

Adolesc Med.

19. Lozano P, Finkelstein JA, Hecht J, Shulruff R, Weiss KB. Asthma medicationuse and disease burden in children in a primary care population. 2003;157(1):81-88.

Ann Allergy Asthma Immunol.

20. Smith MJ, Rascati KL, McWilliams BC. Inhaled anti-inflammatory pharmacotherapyand subsequent hospitalizations and emergency department visits amongpatients with asthma in the Texas Medicaid program. 2004;92(1):40-46.


21. Anarella J, Roohan P, Balistreri E, Gesten F. A survey of Medicaid recipientswith asthma: perceptions of self-management, access, and care. 2004;125(4):1359-1367.

Health Serv Res.

22. Bindman AB, Chattopadhyay A, Osmond DH, Huen W, Bacchetti P. Theimpact of Medicaid managed care on hospitalizations for ambulatory care sensitiveconditions. 2005;40(1):19-38.

23. National Committee for Quality Assurance. Health plan report card. Availableat: Accessed April 1, 2005.

24. National Committee for Quality Assurance. Use of appropriate medications forpeople with asthma denominator. Available at: Accessed December 5, 2003.

Ann Allergy Asthma Immunol.

25. Berger WE, Legorreta AP, Blaiss MS, et al. The utility of the Health PlanEmployer Data and Information Set (HEDIS) asthma measure to predict asthma-relatedoutcomes. 2004;93(6):538-545.

Public Health Rep.

26. Dombkowski KJ, Wasilevich E, Lyon-Callo S. Pediatric asthma surveillanceusing Medicaid claims. 2005;120(5):515-524.

27. US Department of Commerce Census Bureau. Michigan Metropolitan Areas,Counties, and Central Cities. Available at: Accessed February 25, 2004.

MMWR Morb Mortal Wkly Rep.

28. Centers for Disease Control and Prevention. Recommendations of the AdvisoryCommittee on Immunization Practices (ACIP). 2004;53(RR-6):1-44.


29. Landon BE, Epstein AM. Quality management practices in Medicaid managedcare: a national survey of Medicaid and commercial health plans participating inthe Medicaid program. 1999;282(18):1769-1775.

Jt Comm J Qual Improv.

30. Landon BE, Huskamp HA, Tobias C, Epstein AM. The evolution of quality managementin state Medicaid agencies: a national survey of states with comprehensivemanaged care programs. 2002;28(2):72-82.

Am J Manag Care.

31. Seidman JJ, Weiss KB. Health plans' use of asthma quality improvement projectsto meet NCQA accreditation standards. 2001;7(6):567-572.

J Health Care Poor


32. Ali S, Osberg JS. Differences in follow-up visits between African American andwhite Medicaid children hospitalized with asthma. 1997;8(1):83-98.


33. Blais L, Beauchesne M-F. Use of inhaled corticosteroids following dischargefrom an emergency department for an acute exacerbation of asthma. 2004;59(11):943-947.


34. Boudreaux ED, Emond SD, Clark S, Camargo CA Jr. Race/ethnicity and asthmaamong children presenting to the emergency department: differences in diseaseseverity and management. 2003;111(5 pt 1):e615-e621.

Am J Med Qual.

35. Cabana MD, Dombkowski KJ, Yoon EY, Clark SJ. Variation in pediatric asthmaquality improvement programs by managed care plans. 2005;20(4):204-209.

Med Care.

36. Brook RH, Kamberg CJ, Lohr KN, Goldberg GA, Keeler EB, Newhouse JP.Quality of ambulatory care. Epidemiology and comparison by insurance status andincome. 1990;28(5):392-433.

Am J Prev Med.

37. Shih SC, Bost JE, Pawlson LG. Standardized health plan reporting in four areasof preventive health care. 2003;24(4):293-300.

Am J Manag Care.

38. Huang IC, Diette GB, Dominici F, Frangakis C, Wu AW. Variations of physiciangroup profiling indicators for asthma care. 2005;11(1):38-44.

Med Care.

39. Palmer RH, Wright EA, Orav EJ, Hargraves JL, Louis TA. Consistency in performanceamong primary care practitioners. 1996;34(9 suppl):S52-66.

Med Care.

40. Gandhi TK, Francis EC, Puopolo AL, Burstin HR, Haas JS, Brennan TA.Inconsistent report cards: assessing the comparability of various measures of thequality of ambulatory care. 2002;40(2):155-165.

Health Serv Res.

41. Stuart B, Singhal PK, Magder LS, Zuckerman IH. How robust are health planquality indicators to data loss? A Monte Carlo simulation study of pediatric asthmatreatment. 2003;38(6 pt 1):1547-1561.

J Am Board Fam Pract.

42. Fredrickson DD, Molgaard CA, Dismuke SE, Schukman JS, Walling A.Understanding frequent emergency room use by Medicaid-insured children withasthma: a combined quantitative and qualitative study. 2004;17(2):96-100.

J Pediatr.

43. Smith SR, Jaffe DM, Fisher EB, Trinkaus KM, Highstein G, Strunk RC.Improving follow-up for children with asthma after an acute emergency departmentvisit. 2004;145(6):772-777.



44. Lara M, Duan N, Sherbourne C, Halfon N, Leibowitz A, Brook RH. Children'suse of emergency departments for asthma: persistent barriers or acute need? 2003;40(3):289-299.


45. Zorc JJ, Scarfone RJ, Li Y, et al. Scheduled follow-up after a pediatric emergencydepartment visit for asthma: a randomized trial. 2003;111(3):495-502.


46. Daley MF, Barrow J, Pearson K, et al. Identification and recall of children withchronic medical conditions for influenza vaccination. 2004;113(1 pt1):e26-e33.