[Skip to Navigation]
Sign In
May 2005

A Randomized Trial of Primary Care Provider Prompting to Enhance Preventive Asthma Therapy

Author Affiliations

Author Affiliations: Department of Pediatrics, University of Rochester School of Medicine and Dentistry and the Golisano Children’s Hospital at Strong (Drs Halterman, McConnochie, and Szilagyi, Mss Conn and Neely, and Mr Callahan); and School of Nursing, University of Rochester School of Medicine and Dentistry (Dr Yoos), Rochester, NY.

Arch Pediatr Adolesc Med. 2005;159(5):422-427. doi:10.1001/archpedi.159.5.422

Background  Guidelines recommend preventive medications for all children with persistent asthma, yet young urban children often receive inadequate therapy. This may occur in part because primary care providers are unaware of the severity of their patients’ symptoms.

Objective  To determine whether systematic school-based asthma screening, coupled with primary care provider notification of asthma severity, will prompt providers to take preventive medication action (prescribe a new preventive medication or change a current dose).

Design  Children aged 3 to 7 years with mild persistent to severe persistent asthma were identified at the start of the 2002-2003 school year in Rochester. Children were assigned randomly to a provider notification group (child’s primary care provider notified of asthma severity) or a control group (provider not notified of severity). Primary care providers of children in the provider notification group were sent a facsimile indicating the child’s symptoms and recommending medication action based on national criteria. Interviewers blinded to the child’s group assignment called parents 3 to 6 months later to determine if preventive actions were taken.

Results  Of 164 eligible children with mild persistent or more severe asthma, 151 (92.1%) were enrolled. Children in the provider notification group were not more likely to receive a preventive medication action than were children in the control group (21.9% vs 26.0%; P = .57). Additional preventive measures, including encouraging compliance with medications (33.3% vs 31.3%; P = .85), recommending environmental modifications (39.3% vs 42.4%; P = .86), and referrals for specialty care (6.6% vs 6.0%; > .99), also did not differ between the provider notification and control groups. At the end of the study, 52.4% of children in both groups with no medication changes were still experiencing persistent symptoms.

Conclusions  School-based asthma screening identified many symptomatic children in need of medication modification. Provider notification, however, did not improve preventive care. Findings suggest that more powerful interventions are needed to make systematic asthma screening effective.

Asthma-related morbidity and mortality continue to increase despite available treatment options.1,2 National guidelines recommend daily inhaled anti-inflammatory medications for all children with persistent asthma.3 Such medications reduce asthma severity4,5 and prevent exacerbations leading to hospitalizations.6-8 However, studies indicate that many children in the United States who should receive preventive medications are not receiving them.9-11 The greatest underuse of preventive medications occurs among young children living in the inner city.12-14

Practice guidelines for the management of asthma rely heavily on a stepwise approach in which medications are adjusted to maintain adequate asthma control.3,15 The effectiveness of this approach requires families to report significant symptoms to a health care provider and for the provider to respond appropriately. Therefore, adequate therapy for asthma requires accurate and timely communication and a partnership between families and providers.16,17 In a prior study, however, most primary care providers (PCPs) were unaware of the severity of their patients’ symptoms, even when the patients had been recently seen in the office. Consequently, these patients did not receive appropriate preventive asthma therapy.18 Among the children who were accurately identified by their provider as having mild persistent or more severe asthma, however, almost all received prescriptions for preventive medications. These findings suggest that when providers are aware of the frequency and severity of their patients’ symptoms, they correctly follow care guidelines. Enhanced communication between patients and providers regarding symptom frequency likely would improve the prescription of preventive therapy.

One method to assist PCPs in identifying gaps in care is to provide them with prompts to stimulate action regarding a particular clinical issue. Provider prompts have been effective in improving preventive care for various conditions. For example, provider prompts led to higher rates of breast cancer screening,19 alcohol counseling,20 and immunizations.21 To our knowledge, no study has tested provider prompts as a method to improve preventive asthma care for urban children.

Schools provide an ideal site for asthma symptom screening in young children. The objectives of this study were: (1) to identify a group of children with persistent asthma by symptom screening through schools and (2) to test the effectiveness of a provider prompt to improve the prescription of preventive medications for these children. We hypothesized that systematic school-based asthma symptom screening, coupled with PCP notification of severity, would prompt PCPs to take preventive medication action (prescribe a new preventive medication or change the current preventive medication dose) and subsequently increase the proportion of children receiving adequate preventive care. If effective, this relatively simple intervention might be sustained through existing resources of school health systems.


Setting and participants

During the 2002-2003 school year, parents of 3- to 7-year-old children entering the Rochester City School District completed a routine school health and development form that inquired about chronic diseases, including asthma. Parents who indicated asthma on their child’s school health form were contacted by telephone to assess their child’s eligibility for the study.

To be eligible for participation, the child’s parent needed to report that a physician had diagnosed their child as having asthma. In addition, children’s symptoms needed to be consistent with mild persistent to severe persistent asthma according to the National Heart, Lung, and Blood Institute (NHLBI) guidelines.3 Therefore, children were required to meet at least 1 of the following 2 symptom criteria: (1) an average of more than 2 days per week with asthma symptoms or (2) more than 2 nights per month with symptoms. Parents of eligible children who wanted to participate provided informed consent over the telephone, and a copy of the consent form was mailed to the parent. The University of Rochester Institutional Review Board and the Rochester City School District’s Department of Research, Evaluation, and Testing approved the study protocol.

Enrollment and randomization

Following the consent process, parents completed a baseline telephone interview that inquired about sociodemographic characteristics, their child’s asthma severity, and medication use. We assessed asthma severity by asking the parent how many days a week during the past 4 weeks their child had daytime and nighttime asthma symptoms. Acute visits were assessed by asking the parent how many times their child needed emergency medical visits for asthma during the past 12 months. If any emergency medical visits occurred, parents were asked to specify how many were to the physician’s office and how many were to the emergency department. In addition, parents were asked how many times their child had been hospitalized for asthma in the past 12 months, and what medications their child was taking for asthma.

At the conclusion of the baseline interview, children were randomized into either the provider notification group or the control group. Randomization was stratified by use of preventive medications at baseline and was blocked in groups of 6. Randomization cards were made from a table of random numbers and were kept in sealed, opaque, sequentially numbered envelopes until after the baseline assessment was completed.

Provider Notification Group

Primary care providers of children in the provider notification group were sent a letter by facsimile, signed by the principal investigator (J.S.H.), indicating the number of days the child experienced daytime and nighttime symptoms during the past 4 weeks and the number of emergency medical visits for asthma during the past 12 months. The letter also included a copy of the 2002 NHLBI guidelines for asthma management and a recommendation to consider medication action based on the child’s current therapy. In addition, providers of children in the provider notification group were asked to complete a 2-page survey about the child’s asthma history and current medication use based on their knowledge of the child and review of the child’s medical records. To confirm that providers actually received and read the prompt, we sent the survey in the same transmission with the prompt and asked them to return this survey to our study team by facsimile. If a return facsimile was not received within 2 weeks of sending the original prompt, the study team placed a reminder telephone call to the office and sent a new copy of the prompt if needed.

Control Group

For children assigned to the control group, PCPs were not contacted. Parents of children in both groups were notified of their child’s asthma severity at the telephone interview, and recommendations were made for them to speak with their child’s provider about their child’s asthma symptoms.

Assessment of outcome measures

Interviewers blinded to the child’s group assignment called parents 3 to 6 months after randomization to assess the outcome measures for this study. Parents were asked whether they had made contact with their PCP, and whether a new preventive medication was prescribed or the dose of a previously prescribed preventive medication was increased since the date of the child’s enrollment into the program. Primary care providers of children in both groups also were surveyed during this same period. These surveys inquired about contacts with the patient, prescriptions or changes in medications, and other preventive measures, including encouraging compliance with preventive medications, discussion of environmental triggers, or referrals to an asthma or allergy specialist. Providers of children in the provider notification group also were asked to respond to an open-ended question inquiring whether they thought the information provided in the prompt was helpful.

Because all of the children enrolled in this study had persistent asthma symptoms, each should have been using a daily preventive medication according to national guidelines. For those children already using a preventive medication, an increase in dose or an additional preventive medication was indicated because persistent symptoms continued to occur in these children. Therefore, we defined a preventive medication action as either a new prescription of a preventive medication or an increase in dose of a current preventive medication. If either the parent or the provider described a new medication or a change in medication, then we assumed for study purposes that a preventive medication action had been taken.

Asthma severity was assessed at follow-up by asking the parents the number of daytime and nighttime symptoms their child had during the past 4 weeks. The follow-up survey also inquired about acute visits (to the physician’s office and emergency department) and hospitalizations since the baseline interview.

Assessment of covariates

Independent variables included the child’s age, sex, race (defined as white, black, or other), and ethnicity (Hispanic or not Hispanic), Medicaid insurance, parent education (<high school or ≥high school degree), and parent age (<30 or ≥30 years). Provider variables included the length of time the child had been a patient at the office (<6 or ≥6 months) and the provider’s specialty (pediatrics or other). We defined children as using a preventive medication if they described using a preventive anti-inflammatory medication (inhaled corticosteroid, cromolyn sodium or montelukast sodium).


All subjects were kept in their originally assigned treatment groups for analysis. Subjects missing both a parent and provider follow-up survey were excluded from analysis (n = 1). Analysis was performed using a commercially available software program (SPSS, version 11.5; SPSS Inc, Chicago, Ill). t tests and χ2 tests were used to compare outcomes between the 2 groups. Multiple regression analyses were performed to predict preventive medication action, controlling for potentially confounding variables. < .05 was considered statistically significant.


Seven hundred children entering the school district were identified as having symptoms of asthma on their school health forms, and 593 were assessed for eligibility. We were unable to contact 107 families. Four hundred twenty-nine children were ineligible, and the remaining 164 children had mild persistent to severe persistent asthma and were eligible to participate (Figure). Thirteen children were not enrolled because of parent refusal (n = 9), lack of a primary care physician (n = 2), or lack of parental response (n = 2). One hundred fifty-one children were enrolled in the program, for a response rate of 92.1%. Seventy-four children were randomized into the provider notification group, and 77 were randomized into the control group. Seventy-two different PCPs were identified for the group of children, and 19 PCPs had patients in each treatment arm. Of providers in the provider notification group, 90.0% confirmed receipt of their prompt with a returned facsimile.

Participant flow diagram.

Participant flow diagram.

One hundred fifty children had data available for analysis for the primary outcome (73 in the provider notification group and 77 in the control group). We were able to include data for a particular subject if a follow-up was completed with either the parent or the provider. The overall response rate for parental follow-up interviews was 92.6% (93.2% of children in the provider notification group and 92.2% of children in the control group). The response rate for the follow-up provider surveys was 91.3% (89.0% of children in the provider notification group and 93.5% of children in the control group). For subjects with provider and parent follow-up data available, responses for the primary outcome agreed 89.8% of the time. Of the follow-up surveys, 94.3% occurred between 4 and 5 months after the baseline survey, and the amount of time until follow-up was the same for children in both groups.

Table 1 describes the demographic characteristics for the children and parents in the study. Of the 150 children, 58.0% were boys, 59.3% were black, and 66.7% received Medicaid insurance. At baseline, children in the provider notification group and children in the control group were similar with respect to demographic characteristics and asthma severity. In addition, there was an equal proportion of children in both groups who were receiving daily preventive medications at baseline.

Table 1. 
Prognostic Indicators and Baseline Values*
Prognostic Indicators and Baseline Values*

Table 2 shows the proportion of children with preventive actions taken during the study period for the 2 groups. The children in the provider notification group were not more likely to have received a preventive medication action (either a new prescription of preventive medication or a change in medication dose) compared with the children in the control group. Similarly, there were no differences between the 2 groups of children on other preventive measures. Specifically, when compared with PCPs of children in the control group, providers of children in the provider notification group were not more likely to encourage compliance with medications, recommend environmental modifications, request a follow-up visit for asthma, or refer the children for specialty care for their asthma. At follow-up, 52.4% of children in both groups with no medication changes were still experiencing persistent symptoms.

Table 2. 
Preventive Actions*
Preventive Actions*

Because we assumed that health care contacts would be necessary for a preventive medication action to occur, we considered whether the intervention prompted contacts with the PCP. Table 3 shows the parents’ report of health care use for the children in the provider notification group compared with the children in the control group. Overall, most (82.1%) children had contact with their PCP during the study period; the children in the provider notification group were not more likely to have contact compared with the children in the control group. Approximately one third of the children in each group had an acute visit for asthma, and there were no differences between the 2 groups on these measures. If children did not have contact with a health care provider during the study period, they were unlikely to receive an alteration in their preventive medication. Among children with no contact reported, only 8.0% had a preventive medication action taken.

Table 3. 
Parent Report of Health Care Use*
Parent Report of Health Care Use*

A logistic regression analysis was performed to identify factors independently predicting a preventive medication action. All covariates included in the model are shown in Table 4. The only factors that significantly predicted the occurrence of a preventive action were acute visits for asthma and baseline use of preventive medications. Those children with 3 or more acute visits for asthma during the past year were more likely than children with fewer visits to have a preventive medication action taken. In addition, those children who were using a preventive medication at baseline (at a suboptimal dose) were more than 3 times as likely to have a preventive medication action taken compared with children who were not using a preventive medication at baseline. Sociodemographic characteristics, provider factors, and treatment group allocation were not associated with the occurrence of preventive medication actions.

Table 4. 
Predictors of Preventive Medication Action in Logistic Regression
Predictors of Preventive Medication Action in Logistic Regression

Last, in an attempt to better understand the reason why the prompt was ineffective, we compiled the open-ended responses from the providers in the provider notification group regarding the following question, “Was the information in the original notification helpful?” Of the 73 providers in this group, 27 (37.0%) responded. The most common comment, offered by 10 providers, was that the prompt helped by increasing their awareness of the child’s symptoms. Of these 10 providers, 7 made a preventive medication change. Another 9 providers, who believed that their own assessment revealed mild intermittent rather than persistent symptoms, made no medication changes. The remaining 8 respondents mentioned that the child was already using a preventive medication that did not warrant change (n = 4) or that they were unable to contact the family to schedule a follow-up visit (n = 4).


This study evaluated school-based asthma symptom screening with provider prompting as a method to improve the prescription of preventive medications for urban children with asthma. Our screening method proved highly effective in identifying many children with persistent asthma symptoms who warranted more intensive management of their disease according to NHLBI guidelines. However, we found no evidence that provider notification increased the likelihood of improving preventive care. More than half of the children continued to have mild persistent or more severe asthma symptoms at the follow-up assessment, and most had no changes to their medication regimen.

A number of studies have evaluated provider prompting as a means to improve preventive care. Several, which involved prompts regarding issues such as smoking status,22 vaccination needs,21 or mammography screening,19 had positive outcomes. However, few studies have evaluated prompting as a means to improve care for asthma. Shiffman and colleagues tested the usefulness of a handheld computer to prompt physicians about the American Academy of Pediatrics’ guidelines for office management of acute asthma.23 Physicians who received this prompt followed care guidelines better than unprompted physicians, suggesting that immediate access to national guidelines could increase the use of recommended care practices by physicians.

The present study was designed to test a system that used a preexisting screening mechanism used throughout an entire city school district, combined with a simple prompt to the PCP. We ensured that the facsimile was brought to the attention of the PCP by requesting a return facsimile, and received confirmation of receipt from 90.0% of providers. In addition, the widely endorsed NHLBI guidelines for the management of children with asthma were included with the prompt.

There may be several reasons why this prompt was ineffective. First, some providers related that they did not agree with our assessment of the child’s asthma severity, or believed that the symptoms elicited in our survey were transient. Our study team was trained to inquire about asthma symptoms with systematic questions based on NHLBI guidelines, and providers may not have used the same method of inquiry.

Second, given other known barriers to adherence to asthma preventive care guidelines, the prompt may not be sufficient to elicit a change in physician behavior. Contributing factors include a potential lack of familiarity with the guidelines, concerns about adverse effects of corticosteroids, and confusion regarding guideline implementation.24-28 In this study, some of the providers seemed to misunderstand the guidelines because they suggested in their comments that changes in preventive medications were not warranted if the child already was using a preventive medication.

Third, in contrast to most of the successful studies of provider prompting, our prompt was delivered at a time when the patient was not physically present in the office. Providers may need contact with the family and the realization that preventive medication action is warranted at the time of the contact. While most children in both groups had contacts with their care providers at some point during the study period, there were no differences between the groups in the proportion of children having contacts with their PCPs. These contacts likely did not result from prompting, and recognition that preventive action was warranted at these contacts did not result from study procedures. Because asthma symptoms vary over time, it is possible that the children’s persistent symptoms that were described at baseline had diminished by the time contact occurred. The validity of this explanation is brought into question, however, by the fact that most parents at follow-up interviews still described persistent symptoms.

It is not surprising that the children with multiple acute visits for asthma were the most likely to have a preventive medication action taken, because the acute visits themselves may attract the attention of the care provider and elicit a change in their care regimen. In addition, the children already using a preventive medication at baseline were more likely to have a change in medication occur during the study. Neither of these consequences was influenced by provider prompting.


Limitations include the fact that blinding of parents and PCPs to the allocated treatment group was not possible. Therefore, self-reported responses could have been biased by parents or providers, assuming that they were receiving or providing enhanced preventive care. However, because no differences were found between the 2 groups, it is unlikely that this bias played a significant role in the findings. Second, the children included in this study were young, and it is possible that providers are more reluctant to prescribe preventive medications for young children. Third, some providers had children in both treatment arms of the study. The prompts for children in the intervention group possibly could have altered management for patients in the control group. However, most subjects had no change in their treatment regimen, so this potential conservative bias did not lead to heightened treatment in both groups. Last, only urban children in Rochester were included in this study, therefore limiting the ability to generalize to children and providers in similar settings.


Many children continue to be undertreated for asthma,9-11 especially young children living in the inner city.12-14 This study tested the effectiveness of information transfer to narrow the communication gap between patients and providers about asthma symptoms. We were able to use an existing school health screening program to identify many children with persistent asthma symptoms, and to relay that information to each child’s PCP. It seems, however, that solving the problem of undertreatment requires more than the transfer of information about severity classification. Findings suggest that more powerful interventions are needed to make systematic asthma symptom screening effective in changing treatment. These might include more aggressive case management (eg, contacting families on multiple occasions to assess ongoing symptoms and to encourage physician contact), more timely prompting of physicians (eg, a reminder about asthma screening results at the next office contact), and ongoing education in primary care practices regarding implementation of the guidelines for asthma care. Further study is needed to determine whether an effective prompting method can be devised to complement the efficiency of the school-based screening.

Correspondence: Jill S. Halterman, MD, MPH, Department of Pediatrics, Golisano Children’s Hospital at Strong, Box 777, 601 Elmwood Ave, Rochester, NY 14642 (jill_halterman@urmc.rochester.edu).

Back to top
Article Information

Accepted for Publication: June 23, 2004.

Funding/Support: This study was supported by grants from the Halcyon Hill Foundation, Rochester; the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program; and the Strong Children’s Research Center, Rochester.

Acknowledgment: We thank Andrew MacGowan and the Rochester City School District for their support with this project; George B. Segel, MD, for his insightful review of the manuscript; and Crista Crittenden for her assistance in manuscript preparation.

Adams  PFMarano  MA Current estimates from the National Health Interview Survey, 1994.  Vital Health Stat 10 1995;(193, pt 1)1- 260PubMedGoogle Scholar
 Asthma mortality and hospitalization among children and adults: United States, 1980-1993.  MMWR Morb Mortal Wkly Rep 1996;45350- 353Google Scholar
National Asthma Education Program, Expert Panel Report II: Guidelines for the Diagnosis and Management of Asthma: Update on Selected Topics.  Bethesda, Md US Dept of Health and Human Services2002;NIH publication 02-5075
Barnes  PJPederson  S Efficacy and safety of inhaled corticosteroids in asthma.  Am Rev Respir Dis 1993;148(4, pt 2)S1- S26PubMedGoogle ScholarCrossref
Kamada  AKSzefler  SJMartin  RJAsthma Clinical Research Network, Issues in the use of inhaled glucocorticoids.  Am J Respir Crit Care Med 1996;1531739- 1748PubMedGoogle ScholarCrossref
Donahue  JGWeiss  STLivingston  JMGoetsch  MAGreineder  DKPlatt  R Inhaled steroids and the risk of hospitalization for asthma.  JAMA 1997;277887- 891PubMedGoogle ScholarCrossref
Greenberger  P Preventing hospitalizations for asthma by improving ambulatory management.  Am J Med 1996;100381- 382PubMedGoogle ScholarCrossref
Wennergren  GKristjansson  SStrannegard  IL Decrease in hospitalization for treatment of childhood asthma with increased use of anti-inflammatory treatment, despite an increase in the prevalence of asthma.  J Allergy Clin Immunol 1996;97742- 748PubMedGoogle ScholarCrossref
Ordonez  GAPhelan  PDOlinsky  ARobertson  CF Preventable factors in hospital admission for asthma.  Arch Dis Child 1998;78143- 147PubMedGoogle ScholarCrossref
Homer  CJSzilagyi  PGRodewald  L  et al.  Does quality of care affect rates of hospitalization for childhood asthma?  Pediatrics 1996;9818- 23PubMedGoogle Scholar
Halterman  JSAligne  CAAuinger  PMcBride  JSzilagyi  PG Inadequate therapy for asthma among children in the United States.  Pediatrics 2000;105272- 276PubMedGoogle Scholar
Kattan  MMitchell  HEggleston  P  et al.  Characteristics of inner-city children with asthma.  Pediatr Pulmonol 1997;24253- 262PubMedGoogle ScholarCrossref
Eggleston  PAMalveaux  FJButz  AM  et al.  Medications used by children with asthma living in the inner city.  Pediatrics 1998;101349- 354PubMedGoogle ScholarCrossref
Warman  KLSilver  EJMcCourt  MPStein  REK How does home management of asthma exacerbation by parents of inner-city children differ from NHLBI guideline recommendations?  Pediatrics 1999;103422- 427PubMedGoogle ScholarCrossref
Yoos  HLMcMullen  A Symptom monitoring in childhood asthma: how to use a peak flow meter.  Pediatr Ann 1999;2831- 39PubMedGoogle ScholarCrossref
Warman  KL Management of asthma exacerbations: home treatment.  J Asthma 2000;37461- 468PubMedGoogle ScholarCrossref
Gavin  LAWamboldt  MZSorokin  NLevy  SYWamboldt  FS Treatment alliance and its association with family functioning, adherence, and medical outcome in adolescents with severe, chronic asthma.  J Pediatr Psychol 1999;24355- 365PubMedGoogle ScholarCrossref
Halterman  JSYoos  HLKaczorowski  JM  et al.  Providers underestimate symptom severity among urban children with asthma.  Arch Pediatr Adolesc Med 2002;156141- 146PubMedGoogle ScholarCrossref
Harris  RPO’Malley  MSFletcher  SWKnight  BP Prompting physicians for preventive procedures.  Am J Prev Med 1990;6145- 152PubMedGoogle Scholar
Saitz  RHorton  NJSullivan  LMMoskowitz  MASamet  JH Addressing alcohol problems in primary care.  Ann Intern Med 2003;138372- 382Google ScholarCrossref
Minkovitz  CSBelote  ADHigman  SMSerwint  JRWeiner  JP Effectiveness of practice-based intervention to increase vaccination rates and reduce missed opportunities.  Arch Pediatr Adolesc Med 2001;155382- 386PubMedGoogle ScholarCrossref
Chang  HCZimmerman  LHBeck  JM Impact of chart reminders on smoking cessation practices of pulmonary physicians.  Am J Respir Crit Care Med 1995;152984- 987PubMedGoogle ScholarCrossref
Shiffman  RNFreudigman  KABrandt  CALiaw  YNavedo  DD A guideline implementation system using handheld computers for office management of asthma: effects on adherence and patient outcomes.  Pediatrics 2000;105767- 773PubMedGoogle ScholarCrossref
Doerschug  KCPeterson  MWDayton  CSKline  JN Asthma guidelines.  Am J Respir Crit Care Med 1999;1591735- 1741PubMedGoogle ScholarCrossref
Picken  HAGreenfield  STeres  DHirway  PSLandis  JN Effect of local standards on the implementation of national guidelines for asthma.  J Gen Intern Med 1998;13659- 663PubMedGoogle ScholarCrossref
Cabana  MDEbel  BECooper-Patrick  LPowe  NRRubin  HRRand  CS Barriers pediatricians face when using asthma practice guidelines.  Arch Pediatr Adolesc Med 2000;154685- 693PubMedGoogle ScholarCrossref
Cabana  MDRand  CSPowe  NR  et al.  Why don’t physicians follow clinical practice guidelines? a framework for improvement.  JAMA 1999;2821458- 1465PubMedGoogle ScholarCrossref
Finkelstein  JALozano  PShulruff  R  et al.  Self-reported physician practices for children with asthma.  Pediatrics 2000;106886- 896PubMedGoogle Scholar