PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Pediatrics. Author manuscript; available in PMC Mar 1, 2010.
Published in final edited form as:
PMCID: PMC2723164
NIHMSID: NIHMS65088
Socioeconomic, Family, and Pediatric Practice Factors Affecting the Level of Asthma Control
Gordon R Bloomberg, MD,a Christina Banister, BA, CCRP,b Randall Sterkel, MD,c Jay Epstein, MD,c Julie Bruns, MA,d Lisa Swerczek, RN,e Suzanne Wells, RN,e Yan Yan, MD, PhD,f and Jane M Garbutt, MDag
aDepartment of Pediatrics, Division of Allergy and Pulmonary Medicine, Washington University School of Medicine, St. Louis, MO.
bDepartment of Medicine, Washington University School of Medicine, St. Louis, MO.
cDepartment of Pediatrics, Washington University School of Medicine, St. Louis, MO
dBJC Healthcare, St. Louis, MO
eSt Louis Children’s Hospital, St. Louis, MO
fDepartment of Surgery, Division of Biostatistics, Washington University, School of Medicine, St Louis Children’s Hospital, St. Louis, MO
gDepartment of Pediatrics, Department of Medicine, Washington University School of Medicine, St. Louis, MO, St. Louis Children’s Hospital, St. Louis, MO
Corresponding Author: GR Bloomberg, MD, Department of Pediatrics, Division of Allergy and Pulmonary Medicine, Washington University School of Medicine, St Louis Children’s Hospital, One Children’s Place, St. Louis, MO, 63110, E-mail bloomberg/at/kids.wustl.edu
Background
Multiple issues bear on effective control of childhood asthma.
Objective
To identify factors related to the level of asthma control in children receiving asthma care from community pediatricians.
Patients and Methods
Data for 362 children participating in an intervention study to reduce asthma morbidity were collected by telephone administered questionnaire. Level of asthma control (“well controlled,” partially controlled,” or “poorly controlled”) was derived from measures of recent impairment (symptoms, activity limitations, albuterol use) and the number of exacerbations in a 12 month period. Data also included demographic characteristics, asthma-related quality of life, pediatric management practices, and medication usage. Univariable and multivariable analyses were used to identify factors associated with poor asthma control and to explore the relationship between control and use of daily controller medications.
Results
Asthma was “well controlled” for 24% of children, “partially controlled” for 20%, and “poorly controlled” for 56%. Medicaid insurance (p=0.016), the presence of another family member with asthma (p=0.0168), and outside the home maternal employment, (p=0.025), were significant univariable factors associated with poor asthma control. Medicaid insurance had an independent association with poor control (OR 0.49, 95% CI 0.28-0.9). Seventy-six percent of children were reported by parents as receiving a daily controller medication. Comparison of guidelines recommended controller medication with level of control indicated that a higher step level of medication would have been appropriate for 74% of these children. Significantly lower overall quality of life scores were observed in both parents and children with poor control. (ANOVA, p<0.05)
Conclusion
Despite substantial use of daily controller medication, children with asthma continue to experience poorly controlled asthma and reduced quality of life. While Medicaid insurance and aspects of family structure are significant factors associated with poorly controlled asthma, attention to medication use and quality of life indicators may further reduce morbidity.
Keywords: Childhood asthma, asthma control, asthma outcomes
Asthma is a highly visible and demanding part of primary care practice. It is both a chronic and acute episodic disease that has become a major cause of childhood disability, (1-3) with pediatricians becoming increasingly responsible for diagnosing and managing its care. (4) Ambulatory care for asthma in terms of visit rates has steadily increased since the early 1990’s and has continued to increase despite a plateau in both childhood asthma prevalence and ambulatory care visits over recent years. (3, 5) However, reports from national telephone surveys indicate that asthma morbidity is not being adequately addressed.(6-9).
During enrollment for an intervention study, interviews with parents generated data about asthma morbidity and care. These data provided an opportunity to use parent reported information to characterize multiple attributes of school-aged children with persistent asthma and identify factors that influence morbidity.
Setting and study population
This cross-sectional study used data from the 362 patients enrolled in an intervention study evaluating the Telephone Asthma Program (TAP). Families eligible to participate in the TAP study required: the child was 5-12 years old; a physician diagnosis of asthma for at least a year; had at least one of the following, an acute exacerbation requiring an unscheduled office visit, a course of oral steroids, an Emergency Department (ED) visit or hospitalization within the prior year. The child also had to be using daily controller medications and/or have symptoms or interference with normal activity > 3 or more days per week, or short acting beta-2 agonist use more than 8 times during the past 2 weeks. Families were ineligible to participate if they did not have a phone or could not speak English, if a sibling was already a study subject, or if the child with asthma had another disease that required regular follow-up by the pediatrician (such as cystic fibrosis or sickle cell disease) or was participating in another asthma study. The study was approved by the Washington University Human Studies Committee, and all parents signed an informed consent.
Potential subjects were recruited from families who called the Telephone Triage Service (TTS) at St Louis Children’s Hospital (SLCH) for asthma care between January, 2003 and December, 2005. Upon consent of the child’s pediatrician, a member of the study team called the parent to invite them to participate in the study. We identified 483 eligible participants of whom 362 participated in the study and completed the 20-minute baseline interview. For 98% (355/362) of the study population, the time gap between the index call to the TTS and the baseline interview was at least 2 weeks, and for 84% (303/362) it was at least a month.
We obtained demographic information, and asked about use of an Asthma Action Plan, (AAP), the frequency of asthma check-ups, and use of controller medications. Parents reported the frequency of asthma symptoms (wheeze, cough, shortness of breath and chest tightness), use of reliever medications in the past week, the number of days missed from school or physical education class restrictions in the past 2 weeks, the number of courses of oral steroids in the past 3 months, and the number of ED visits or hospitalizations in the past year. These varied time frames were used to capture asthma events that occur over irregular intervals and need to be incorporated into the global assessment of asthma control and, at the same time, minimize recall bias. Although data about night-time symptoms were collected, an error in instrument development precluded use of these data. Parents were asked about their level of satisfaction with the asthma care received from the child’s pediatrician and responded using a Likert scale, (poor, fair, good, very good, excellent).
Disease-specific quality of life (QOL) was measured for children and parents using the Pediatric Asthma Quality of Life Questionnaire (PAQLQ)(10) and Pediatric Asthma Caregiver’s Quality of Life Questionnaire (PACQLQ)(11) respectively. (With permission) For both instruments, answers are expressed on a 7-point scale, with a higher score indicating a better quality of life.
Definitions
Use of Controller Medications
To assess use of controller medications, parents were asked “Apart from albuterol and oral steroids, which medications has your child used for asthma in the past 7 days?” and they provided the name, dose and dosing frequency used. The step level of controller treatment was categorized using the comparative daily doses for ICS and the definitions for stepped care provided in the NAEPP Guidelines, 2002, current at the time of data collection. (12) Treatment was determined to be less than optimal if a child not well controlled was eligible to be increased one step level of controller medications. Consequently, children currently receiving maximal levels of ICS, but still poorly controlled were excluded from this subgroup analysis. Observations were excluded from the assessment of the stepped-level of care if data about dose or dosing frequency were unavailable (n=25), or the child’s therapeutic regimen could not be classified within the guidelines (n=30) (Most often, this was because the child was receiving both fluticasone/salmeterol and montelukast, a combination not recognized in the 2002 asthma guidelines.) (12) Adherence was elicited by directly asking about the number of days in the past week the child was receiving each medication, and if the child sometimes missed a dose or a day of medication.
Asthma Control
Asthma control was assessed in terms of short term morbidity and history of exacerbations (Table 1).(13) Short-term morbidity included the frequency of daytime symptoms and albuterol use in the past week and activity limitations in the past two weeks. History of exacerbations included oral steroid bursts in the past 3 months, and ED visits or hospitalizations in the past 12 months. For each item, a 3-point scale was used to define children that were “well controlled,” “partially controlled,” and “poorly controlled.” The maximal score for any one of the five items was the child’s overall level of control. During the period of data analysis, a draft of the 2007 NAEPP guidelines became available. This guideline introduced the terms “impairment” for short-term morbidity and “risk” for long term morbidity. (13)
Table 1
Table 1
Definitions Used To Assess Level of Asthma Control
Data Analysis
Demographic, family, and pediatric practice characteristics were evaluated as predictors of partially and poorly controlled asthma. Continuous variables are reported as mean (standard deviation, SD) or median (range), and categorical data as proportions. In subgroup analyses, differences were compared using the students t-test, ANOVA, Wilcoxon rank sum test, Chi-squared test or Fisher’s exact test as appropriate.
In order to identify variables that are independently associated with the outcomes, we first performed univariate analysis, and then performed multivariate analysis. Due to the ordinality of our outcome (well controlled=1, partially controlled=2, and poorly controlled=3), we used the mean score statistic to test the null hypothesis of no association versus the alternative hypothesis of a location difference among the different levels of nominal categorical variables. The statistical significance was determined by Chi-square distribution with (s-1) degree of freedom, where s is the number of nominal categorical levels. For ordinal categorical variables, we used the correlation statistic to test the null hypothesis of no association versus the alternative hypothesis of linear association between these variables and outcomes. The statistical significance was determined by Chi-square distribution with 1 degree of freedom. For the multivariate analyses, we selected the variables with a P value <=0.2 in the univariate analysis and the variables thought to be clinical important as the candidates for stepwise procedure. Finally, interactions were tested among those variables selected through the stepwise procedure. The proportional odd assumption for variables retained in the model was examined using the score statistic.
A probability of P < 0.05 (two-tailed) was used to establish statistical significance with all tests. All statistical analyses were done using STATA 9.2 (Stata Corp.2001. Stata Statistical Software: Release 9.2 College Station, TX: Stata Corporation) or SAS version 9.2, (SAS Institute Inc, Cary, NC.)
Patient and Family Characteristics
Of 362 survey respondents, 94% were mothers. The children were school age (5-12 years), with 62% males, and 61% Caucasian race/ethnicity (Table 2). All had a history of symptoms consistent with persistent asthma ((12, 14); with a median duration of diagnosis of 5 years (Range, 1-12 years).
Table 2
Table 2
Baseline Patient and Family Characteristics of 362 Participants
The families generally resided in suburban areas and all had access to pediatric care with 73% having commercial health insurance and 22% Medicaid (Table 2). 5% had “other” forms of insurance or were “self-pay”. The majority of families (71%) were headed by two parents. Eighty-two percent of respondents had some college education with 41% holding a college/post graduate degree. 72% of respondents were employed outside the home, either part-time or full-time. Asthma was also a common occurrence within the family, with 53% of families having another member with asthma.
Asthma Morbidity
Impairment
The degree of short-term asthma morbidity was substantial. In the past week, daily or continuous symptoms were present in 20% of children: 18% were using albuterol for symptoms 3-6 days per week and 17% were using albuterol daily; 19% were limited in their normal daily activities for at least 5 days. For 326 children who attended school, 26% of mothers reported their children were absent from school within the previous 2 weeks (Median 2 days, range 0.5-9.0 days) and 27% reported restrictions in physical education (Median 2 days, range 1-14 days). In the past 2 weeks, 19% of caregivers missed work because of the child’s asthma symptoms (Median 2 days, Range-1-8 days).
Risk
Many children had history of a recent exacerbation. 183 children (51%) had received one or more courses of oral corticosteroids in the past 3 months, 59% had 1 ED visits, and 10% had been hospitalized for an asthma exacerbation in the past 12 months (Table 3).
Table 3
Table 3
Asthma Morbidity and Level of Asthma Control Assessed from Parent Reported Data for 362 Children
Level of Control
87 (24%) of the children were “well-controlled,” 71 (20%) were “partially controlled,” and 204 (56%) were “poorly-controlled” (Table 3).
Quality of Life
Asthma-related quality of life scores for both parents and children were high. The mean (SD) overall score for parents was 5.8 (1.2) and for children was 5.7 (1.0). Overall QOL scores varied significantly by level of asthma control, for both parents and children, (p<0.05, ANOVA). For children categorized as well-controlled, partially-controlled, and poorly-controlled, the mean QOL scores (SD), were 6.5 (0.5), 6.2 (0.9), 5.4 (1.3) for parents and 6.1 (0.9), 5.7 (1.0), 5.5 (1.1) for children, respectively.
Medications
Table 4
Table 4
Medication Use and Relation to Control and Adherence A. Use of Asthma Controller Medications By Children Reported By 362 Families
A. Medications
(Table 4, A) 275 (76%) of children reported using a controller/maintenance medication in the week prior to the interview. 219 (60%) were using an ICS either alone (66), or in combination with salmeterol (52), montelukast (50) or both (48), or with other controller medications (3). 55 (15%) used montelukast alone and 1 used sodium cromolyn.
B. Stepped Management
Table 4, B presents the controller medication, reported by the parent, categorized by the step level outlined in the 2002 guidelines. (12) Of 307 children included in this subgroup analysis, 87 (28%) were not taking controller medications, 65 (75%) of whom had partially or poorly controlled asthma. Among the 29 children receiving maximal therapy, 24 (83%) were still only partially or poorly controlled. These 24 children were excluded from the analysis assessing the adequacy of the child’s stepped-level of treatment. Of the remaining 283 children, 209 (74%) were not well-controlled although 196 (69%) reported daily use of controller medications.
C. Adherence
We asked parents, “In the past seven days, what medications apart from oral steroids and albuterol has your child taken for asthma?” This was followed by questions related to missed doses and days for each individual medication. The overall parent-reported usage of controller medications in the seven days preceding the interview was high. 89% of parents reported that their child was using an ICS and received the medication for > 5 days/week, as did 94% of those treated with montelukast.
Office and Home Management Practices
Access to primary care was readily available. In the 2 weeks prior to the telephone interview, 20% of parents had called their pediatrician’s office because of asthma symptoms with calls ranging from one to ten calls within that period (Table 2). 17% of parents telephoned for advice on multiple occasions after hours. In the previous 2 weeks, 41 (11%) of the children visited a health care provider for emergent asthma care because symptoms were not controlled (16 at the primary care office only, 8 at the office and ED, 13 ED only, 4 ED with a hospitalization). When respondents were asked to rate the overall quality of asthma care by the pediatrician, 70% responded that it was “very good” or “excellent.”
A routine asthma care visit in the past 6 months occurred for 173 (48%) children. (Table 2) Within this group, 58% were seen at the pediatrician’s office, and 41% were seen by an asthma specialist. Although 235 (65%) had at one time received an AAP, only 49% (115/235) of this group had the AAP updated in the past 6 months. 66% of parents reported they had talked to their pediatrician regarding the child’s early warning signs of asthma and 55% that their child had ever been observed using their inhaled medications by their physician or nurse.
Predictors of Poor Control
The results of the univariable analyses are provided in Table 5. Three variables were significant in the association with poorly controlled asthma, Medicaid insurance, (X27.32, DF 2, p=0.0166), homes where the mother worked full-or part-time outside of the home, (X25.15, DF 2, p=0.025), and the presence of another family member in the home with asthma, (X29.35, DF 2, p= 0.017). In the multivariate analysis, Medicaid insurance was the only independent factor associated with poorly controlled asthma. Children with Medicaid insurance were less likely to be well controlled (OR 0.489; 95% CI: 0.276, 0.865). Removing Medicaid insurance from the multivariate models failed to identify any other statistically significant independent predictors of poorly controlled asthma.
Table 5
Table 5
Univariate Analysis of the Relationship between Demographic and Asthma Related Factors with Level of Control in 362 Children
Asthma morbidity was significant in this cohort of children from the general asthma population. Although the cohort was identified through contact with a telephone triage service for asthma care, the time between the index call and the baseline interview was adequate to assume that these reported data represented the child’s usual state of control. It is disconcerting to see the degree of activity limitations and missed school and work days that persist for families with asthma, despite widespread availability of effective therapy. Activity limitations, commonly used by patients and their parents to minimize asthma symptoms, (7) accounted for the largest group of children classified as “poorly controlled” in this population. These limitations may not be identified unless specific questions are asked, causing asthma control to be overestimated and the prescribed step-level of medication sub-optimal. (15)
We identified several factors associated with poorly controlled asthma. Medicaid health insurance, another family member with asthma, and the mother working part-time or full-time outside the home were associated with poorly controlled asthma in the univariate analyses, and Medicaid health insurance was an independent predictor. That family issues may be related to poorly controlled asthma has been noted previously. (16) Someone else in the family with asthma, single parenting and mothers working outside the home suggest competing priorities that interfere with parental knowledge of their child’s level of asthma control, daily use of controller medications, and opportunities for asthma monitoring visits. These factors may be used to identify families that may benefit from more frequent monitoring contacts, and additional support and education to augment effective home management of their child’s asthma.
Inhaled-corticosteroids and montelukast were more widely prescribed in the study cohort than reported in earlier studies. (6, 9, 17) indicating the increased awareness of the need for daily controller therapy, rather than SABAs, to reduce exacerbations. However, many children remained symptomatic despite use of effective controller medications. Although self-reported adherence for these medications was high, it is well recognized that there are serious discrepancies between reported and actual medication use. (18-20) It is also possible, that after initiating daily controller use, there may not be sufficient monitoring and follow-up to ensure the child’s step-level of controller medication was adequate for optimal control. (6, 12, 16, 17, 21, 22)
Asthma maintenance visits were infrequent in the study population. Parents may not recognize the need for maintenance visits and resist follow-up when their child is “doing well.” In addition, they may have low expectations for asthma control,(15, 21, 23) accepting the child’s limitations as “unavoidable consequences of the their disease.” (9) Parents may also not see the value of regular checkups if they are hesitant to accept a diagnosis of asthma as a chronic disease, and may also underreport the asthma status of their child. (24) These factors may explain why most parents reported a high degree of satisfaction with their child’s asthma care despite high levels of asthma morbidity.
We acknowledge several limitations to our study. We used self-reported data and did not confirm data accuracy. We used a short time frame to limit recall bias. (13) Acute exacerbations serious enough to require an ED visit or hospitalization are likely to be memorable events. We did not use office charts to confirm these data, but these records are often incomplete. (25) Our method to assess the level of asthma control has not been validated, but is similar to the methodology proposed by Fuhlbrigge (7) and is consistent with the approach used in the 2007 NAEPP guidelines. (13) Unfortunately, we were unable to include night-time symptoms in our control assessment due to an error in the response scale. This omission is likely to result in underestimation of the extent of partially or poorly controlled asthma. (26)
CONCLUSIONS
Our findings have implications for maintenance asthma care provided by community pediatricians. Despite widespread use of ICS and montelukast as controller medications, many patients are not using the appropriate step-level of controller medications to fully control symptoms. Economic factors, family priorities, parent undervaluation of their child’s symptoms, and lack of adequate monitoring and follow-up by the parent and the PCP may all play a role in not meeting the expectation of adequate treatment and reduced morbidity. Interventions for parents and physicians are needed to enable the goals of effective therapy to be met.
Abbreviations
NAEPPNational Asthma Education and Prevention Program
EPRExpert Panel Report
EDEmergency Department
ICSInhaled Corticosteroid
AAPAsthma Action Plan
TTSTelephone Triage Service
SABAShort Acting β-Agonist

Footnotes
Funding: AHRQ (#HS 15378), Dr Yan Yan funding: R01 HL07919-05
The authors have indicated that they have no financial relationships related to 18this article to disclose.
None of the authors have any potential, perceived, or real conflict of interest.
1. Adams P, Hendershot G. 1996 Vital Health Stat National Center for Health Statistics, 1999. 200 Vol. 10. 1999. Current estimates from the National Health Interview Survey. [PubMed]
2. Newacheck PW, Halfon N. Prevalence, Impact, and Trends in Childhood Disability Due to Asthma. Arch Pediatr Adolesc Med. 2000;154(March):287–293. [PubMed]
3. Akinbami LJ The State of childhood asthma, United States, 1980-2005. National Center for Health Statistics; Hyattsville, MD: 2006. Report No.: 381.
4. Stafford R, Ma J, Finkelstein SN, Haver K, Cockburn I. National trends in asthma visits and asthma pharmacotherapy, 1978-2002. J Allergy Clin Immunol. 2003;111:729–735. [PubMed]
5. Hing E, Cherry D, Woodwell D Advance data from vital and health statistics. National Center for Health Statistics; Hyattsville, Maryland, 20782: 2006. National Ambulatory Medical Care Survey: 2004. [PubMed]
6. Adams R, Fuhlbrigge A, Guilbert T, Lozano P, Martinez F. Inadequate use of asthma medication in the United States: Results of the Asthma in America national population survey. J Allergy Clin Immunol. 2002;110:58–64. [PubMed]
7. Fuhlbrigge AL, Guilbert T, Spahn J, Peden D, Davis K. The Influence of Variation in Type and Pattern of Symptoms on Assessment in Pediatric Asthma. Pediatrics. 2006;118(2):619–625. [PubMed]
8. Rabe K, Vermeire PA, Soriano JB, Maier WC. Clinical management of asthma in 1999: the Asthma Insights and Reality in Europe (AIRE) study. Eur Respir J. 2000;16:802–807. [PubMed]
9. Rabe K, Adachi M, Lai CKW, et al. Worldwide severity and control of asthma in children and adults: The global Asthma Insights and Reality surveys. J Allergy Clin Immunol. 2004;114:40–47. [PubMed]
10. Juniper E, Guyatt G, Feeny D, Ferrie P, Griffith L, Townsend M. Measuring quality of life in children with asthma. Quality of Life Research. 1996;5(1):35–46. [PubMed]
11. Juniper E, Guyatt G, Feeny D, Ferrie P, Griffith L, Townsend M. Measuring quality of life in the parents of children with asthma. Quality of Life Research. 1996;5(1):27–34. [PubMed]
12. JACI. 5. Vol. 110. 2002. Expert Panel Report 2: Guidelines for the Diagnosis and Management of Asthma Guidelines for the Diagnosis and Management of Asthma; pp. S141–219. [PubMed]
13. Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma. Summary Report JACI. 2007;120(5):S93–138. [PubMed]
14. Anonymous The case for managing asthma care. State of Health Care in America; 1995. Data watch; pp. 4–5. [PubMed]
15. Cabana MD, Slish KK, Nan , Lin X, Clark NM. Asking the Correct Questions to Assess Asthma Symptoms. Clin Pediatr. 2005;44:319–325. [PubMed]
16. Halterman J, Aligne C, Auinger P, McBride J, Szilagyi P. Inadequate Therapy for Asthma Among Children in the United States. Pediatrics. 2000;105(1):272–276. [PubMed]
17. Goodman DC, Lozano P, Stukel T, Chang C, Hecht J. Has Asthma Medication Use in Children Become More Frequent, More Appropriate, or Both? Pediatrics. 1999;104:187–194. [PubMed]
18. Bender B, Wamboldt FS, O’Connor S, Rand C, Szefler SJ, Milgrom H. Measurement of children’s asthma medication adherence by self-report, mother report, canister weight, and Doser CT. Annals of Allergy, Asthma, & Immunology. 2000;85:416–421. [PubMed]
19. Bauman LJ, Wright E, Leickly FE, et al. Relationship of Adherence to Pediatric Asthma Morbidity Among Inner-City Children. Pediatrics. 2002. http://www.pediatrics.org/cgi/content/full/110/1/e6 [PubMed]
20. Bender B, Milgrom H, Apter A. Adherence intervention research: What have we learned and what do we do next? J Allergy Clin Immunol. 2003;112:489–494. [PubMed]
21. Dozier A, Aligne CA, Schlabach MB. What is Asthma Control? Discrepancies between Parents’ Perceptions and Official Definitions. Journal of School Health. 2006;76(6):215–218. [PubMed]
22. Jones C, Clement LT, Morphew T, et al. Achieving and maintaining asthma control in an urban pediatric disease management program: The Breathmobile Program. Journal of Allergy and Clinical Immunology. 2007;119:1445–1453. [PubMed]
23. Halterman JSM, Conn KM, Yoos LH, et al. A Potential Pitfall in Provider Assessments of the Quality of Asthma Control. Ambulatory Pediatrics. 2003;3(2):102–105. [PubMed]
24. Yoo H, Johnson S, Voight R, Campeau L, Yawn BP, Juhn Y. Characterization of asthma status by parent report and medical record review. Journal Allergy Clin Immunol. 2007;120(6):1468–1469. [PubMed]
25. Cabana MD, Bruckman D, Meister K, Bradley JF, Clark N. Documentation of Asthma Severity in Pediatric Outpatient Clinics. Clinical Pediatrics. 2003;42(2):121–125. [PubMed]
26. Colice G, Burgt JV, Song J, Stampone P, Thompson PJ. Categorizing Asthma Severtiy. Am J Respir Crit Care Med. 1999;160:1962–1967. [PubMed]