|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From La Rabida Children's Hospital and Research Center (Drs. Wolf and Berry, Ms. O'Connor, and Ms. Coover), Chicago, IL; University of Chicago (Dr. Wolf), Chicago, IL; and Northwestern University (Dr. Berry), Chicago, IL.
Correspondence to: Raoul L. Wolf, MD, FCCP, La Rabida Children's Hospital and Research Center, E 65th St at Lake Michigan, Chicago, IL 60649; e-mail: rlwolf{at}midway.uchicago.edu
| Abstract |
|---|
|
|
|---|
Background: Asthma is the most common chronic disease of childhood, yet the frequency with which this condition is recognized among school-aged children varies widely. Several methods are used to increase the accuracy of detection of asthma, but many are cumbersome and difficult to apply on a large scale.
Design: We elected to validate a five-question instrument, the Brief Pediatric Asthma Screen (BPAS), to screen for the presence of asthma among children attending school in Region 5 of the Chicago school district, where the schools report a 2.7% frequency of asthma. The questionnaire was distributed to the parents of grade-school children at the time of report-card pick-up.
Setting: A clinical assessment was performed on a selected group of children whose parents completed the questionnaire in a school and in a hospital outpatient clinic.
Participants: Of 4,147 questionnaires that we distributed, 1,796 (43%) were returned. We excluded 341 children (19% of the total sample) whose parents reported that they had been diagnosed with asthma. The remaining pool indicated that the children of 183 responders (10%) had symptoms suggestive of asthma, while 1,272 parents (71%) indicated that their children did not have symptoms of asthma.
Measurements and results: We selected 90 of the respondents who did not indicate that their children had a diagnosis of asthma. Of this group, 81 completed the validation, in which their responses suggested symptoms of asthma (n = 34) or no asthma symptoms (n = 47). The children of these respondents were given a blinded clinical evaluation consisting of history, physical examination, and spirometry. The survey demonstrated a sensitivity of 75% and a specificity of 81.2% for the presence of asthma among those who were unaware of the diagnosis.
Conclusions: The BPAS is brief, can be filled out by parents, and appears accurate in detecting asthma.
| Introduction |
|---|
|
|
|---|
It is evident from these studies that there are problems inherent in evaluating asthma on a large scale in schools. Many authors have shown that there are significant consequences to missing school days on a large scale. To address this problem, we undertook to validate a short, five-point asthma questionnaire that called only for "yes" or "no" answers.
The instrument, the Brief Pediatric Asthma Screen (BPAS), was designed so that parents could complete it in a short time without supervision. The best currently available "gold standard" to validate the survey consisted of history, physical examination, and spirometry performed by an expert in the diagnosis of pediatric asthma. We therefore elected to validate the survey against this "gold standard". We distributed the instrument to elementary school children in Region 5 of the Chicago school district, comprising 108 elementary schools where the schools report a 2.7% frequency of asthma. The schools are located in impoverished neighborhoods with predominantly African-American populations. The school personnel in this region appreciated the need for greater awareness of asthma. The most frequent request from the elementary schools where we are involved had been for a simple method to determine the number of asthmatics in school. This request was a strong motivating factor in pursuing the validation. Current reporting methods in Chicago seriously underestimate the number of affected children, leading to many of the problems inherent in underdiagnosed asthma.
In this paper, we describe the results of testing the validity of the questionnaire against a "gold standard."
| Materials and Methods |
|---|
|
|
|---|
Questionnaire
We used an abbreviated questionnaire excerpted from the
International Study of Asthma and Allergies in
Childhood.13
This modified tool was initially developed
for use in clinical screening by Dr. Evalyn Grant (personal
communication; February 1998).
The BPAS contained five questions:
Parents completed the BPAS without supervision.
Cohort
As a first phase of the validation, copies of the questionnaire
were distributed in eight elementary schools in the target region,
Region 5.
School personnel gave forms to the parents of 4,147 children at the time of report-card pick-up, and completed forms were returned to the school. The return rate was 43% (1,796/4,147). For this aspect of the study, no attempt was made to increase the return rate by follow-up of nonrespondents or by adding incentives for completing the surveys. The surveys were classified into three categories: (1) negative; (2) positive; and (3) possible asthma, based on a prior classification scheme developed by the authors.
The negative group answered in the negative to questions 1 (prior diagnosis of asthma) and 5 (emergency department [ED] visit) and to two of the three symptom questions. We made a decision that one symptom alone was insufficient evidence of undiagnosed asthma, but were able to test this decision empirically in the analyses.
The positive group had answered "yes" to question 1, indicating that a physician had diagnosed asthma. Responses to the other four questions were irrelevant for this group.
The possible asthma group had answered "no" to question 1, but "yes" to question 5 or two or more of the remaining symptom questions.
All participants who completed a survey were sent a letter informing them that the survey was negative or that there was a possibility that their child had asthma. In this instance, the parents were advised to follow up with their physician, or to visit our clinic if they did not have an established relationship with a doctor.
Subjects for validation were selected among responding children from three elementary schools. Potential subjects were selected randomly by one of us who did not participate in the clinical evaluation and was the only investigator aware of the parents' responses in the surveys. Children in the positive group were excluded from the study as they already had a diagnosis of asthma. The purpose of this validation was not to evaluate a reported prior diagnosis.
Ninety children and their primary caregivers were enrolled in the validation study. All of the enrollees were invited to participate in a medical assessment. Eighty-one children completed the assessment, 47 of whom were in the negative group and 34 of whom were in the possible asthma group according to the BPAS. The parents who completed the assessment were each given a gift certificate to a local supermarket. The responses to questions 2 to 5 were correlated with whether or not the medical assessment indicated that the child had asthma.
Clinical Assessment
We used a medical assessment by a specialist in asthma as a
"gold standard" for the detection of the presence of asthma. The
medical assessment consisted of a medical history, physical
examination, and pulmonary function test.
An allergist and a specialist nurse who performed the medical assessment were blinded to the subjects' responses in the BPAS. The history and physical examination were based on usual clinical criteria for suspecting asthma in a child. Special attention was paid to symptoms of cough and breathlessness; nocturnal cough or wheeze; difficulty speaking, especially with exercise; difficulty with exercise; and missed school for "respiratory difficulties." The parents were also asked if the child regularly saw a pediatrician and how frequently. If the parents indicated that the ED was used, we asked for the reason. Evidence of atopy and triggering events for respiratory symptoms were other criteria used to evaluate the history.
Physical examination focused on upper and lower airway. We noted evidence of atopy and the presence and degree of airflow obstruction.
Patients were classified according to a schema that categorized the children as (1) asthma (at least two symptoms and clinical evidence of airway obstruction, such as reduced airflow, prolonged exhalation); or (2) probably asthma (two symptoms or clinical evidence of airway obstruction).
Pulmonary function testing consisted of spirometry, which was performed on a portable spirometer (Jones Medical Instrument Co; Oak Brook, IL) that was calibrated before each session. A certified pediatric pulmonary function technician who is skilled at obtaining reproducible flow/volume loops from young children performed the test. The flow parameters were assessed for reduced flow rates and a flow-loop contour that was indicative of airflow obstruction. Reversibility was not assessed. The spirometry was graded as (1) asthma (reduced peak flow, FEV1 < 75%, FEV1/FEV < 75%, concave volume-flow loop); or (2) suggestive of asthma (reduced midexpiratory flow, mild reduction in peak flow, normal to flattened flow loop); or (3) normal.
We combined the history, clinical assessment, and spirometry, and categorized each child as having definite asthma, no asthma, or probable asthma. For purposes of analysis, we collapsed the definite and probable categories into a single positive category.
Evaluation of Data/Validation of the BPAS
We compared the BPAS results with the assessment from the
composite clinical and spirometric evaluation by two criteria:
sensitivity (the ability to detect positives) and specificity (the
ability to exclude negatives). Sensitivity is calculated by dividing
the number of true-positives by the sum of the true-positives and
false-negatives, and then multiplying by 100. Specificity is computed
by dividing the number of true-negatives by the sum of the
true-negatives and false-positives, and then multiplying by 100.
| Results |
|---|
|
|
|---|
|
|
The a priori scheme was, in a sense, a weighted scoring. The ED question was given greater weight than the three symptom questions. We tried unweighted combinations, such as any one affirmative response to the four questions, any two affirmative responses, etc. Any one item provided the most sensitive scoring (84%), but was not specific enough (65%). Three affirmative responses was specific (94%), but not sensitive (60%). Two affirmative responses was slightly less sensitive than our weighted scheme (73%), but slightly more specific (83%).
Analysis of Results for Individual Schools
A breakdown of the results of applying the validated screen
to 4,147 students from eight Region 5 elementary schools, varying in
size from 115 to 850 students, is given in Table 3
. The response rate ranged from 36 to 55% (mean, 43%). Between 63 and
78% (mean, 71%) of the respondents did not have evidence of asthma.
Between 5 and 13% (mean, 10%) had possibly undiagnosed asthma, while
16 to 24% (mean, 19%) reported a previous diagnosis of asthma by a
physician. These data are representative of the respondents only, and
are not indicative of prevalence data.
|
The odds ratio for wheezing after play was 3.9. The odds ratios for coughing and ED visit were both close to 1. The odds ratio for episodes of wheezing was 2.5, but it was not statistically significant (p > 0.05).
| Discussion |
|---|
|
|
|---|
The BPAS correlated well with the clinical standard. Very similar sensitivity and specificity were achieved by using any three "yes" responses to questions 2 through 5 in an unweighted score or by using a scoring method that weighted the ED question more heavily. Further validation studies may determine if either the weighted or unweighted method is superior to the other. In this sample, the single item "wheezing after play," the unweighted "two or more affirmative responses," and the weighted scheme (ED visit or any two symptoms) performed similarly well. Taken in isolation, a visit to the ED for wheezing is very specifiic, but has poor sensitivity. This reflects the parents' interpretation of "wheeze," as this symptom may originate in the upper airway or be attributable to a viral infection independent of asthma. In any case, these data indicate that a simple five-question instrument is valid in detecting children with asthma and suspicious symptoms in school. In this sample, the single item "wheezing after play" and the full four items of BPAS taken together performed equally well. In cases where brevity is of the utmost importance, the item "wheezing after play" can be used as a reasonable screen for potential undiagnosed asthma. In other cases, especially if the screen is being used by a clinician for clinical rather than pure research purposes, the additional questions provide valuable information in and of themselves. Furthermore, we believe that the full screen and the single item need further testing in diverse samples to establish whether the single item is consistently as good a predictor of undiagnosed asthma as the five-item screen. In general, multi-item scales are superior to single-item scales in terms of consistency, stability, and robustness across samples and settings. On the other hand, a single-item screen is considered less stable than a multi-item index, and so we are recommending use of all five items in the BPAS with either weighted or unweighted scoring.
When we simultaneously assessed the significance of each component of the survey in a logistic regression analysis, the only question that retained predictive value was wheezing after play or other exercise. Nocturnal cough is often used as a guide to control of asthma. However, this symptom did not demonstrate unique predictive value when the regression equation was adjusted for (or included) wheezing after exercise. Likewise, wheezing alone or an ED visit did not predict asthma.
The correlation between asthma symptoms and objective signs is not always clear. Chang et al15 indicated that patients with asthma who have a predominant cough presentation were more sensitive to capsaicin challenge. On the other hand, Faniran et al,16 17 using a validated survey of chronic cough, questioned whether persistent cough is synonymous with asthma. Bai et al18 examined the grouping of asthma symptoms and the predictive power of symptom questions. They noted that questions that predicted asthma "syndrome" differed from those that predicted cough. The significance of reduced pulmonary function measures in predicting asthma control or severity is not clear. Teeter and Bleecker19 did not demonstrate a correlation among six asthma symptoms and the degree of obstruction as measured by FEV1, and Marabini et al20 noted that FEV1 was not a reliable guide for reducing steroid dose. One third of children who have asthma-like symptoms but no objective signs will develop asthma.21 From a broader perspective, McCowan et al22 also noted that school absence alone was insufficient as a marker of asthma severity. Thus, it seems that there is a need for an objective detection tool for undiagnosed asthma.7
The data collection was not designed to determine the prevalence of asthma in the schools. The purpose of the screening we did in public schools was to provide screening as a direct service to schools to help them identify children with asthma. We made no systematic attempts to boost response rates through follow-up efforts, unless the individual school requested that we do so or helped us in follow-up. We have no reason to believe there was any systematic bias from nonresponse. We conducted the validation study with a random sample of children whose parents had completed the screen. Thus, the overall rate of 19% of parents who returned the surveys and noted that their children carried a diagnosis of asthma should not be taken as a prevalence rate. This number is probably high, but is in keeping with data from other investigators.22 23 24 25 Anderson et al5 demonstrated a rise in prevalence of wheezing among school children in London, England. It is of interest that 10% of children met criteria in the validated survey for possible asthma. Others have noted similar numbers from the United States and other countries.1 26 27 These children have problems in school, and the source of their difficulty may be unrecognized. As a result, the seriousness of their condition is often not appreciated.24 28 29 30 A simple tool such as the BPAS is helpful in identifying these children.
| Conclusion |
|---|
|
|
|---|
| Footnotes |
|---|
Abbreviations: BPAS = Brief Pediatric Asthma Screen; ED = emergency department
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
D. N. Homnick and S. R. DeJong Parent-Reported Physician Diagnosis Is an Important Factor in Asthma Management: An Elementary School Survey Clinical Pediatrics, June 1, 2007; 46(5): 431 - 436. [Abstract] [PDF] |
||||
![]() |
L. Coover, C. Vega, V. Persky, E. Russell, R. Blase, R. Wolf, M. Garcia, E. Grant, and M. Dublin A Collaborative Model to Enhance the Functioning of the School Child With Asthma Chest, October 1, 1999; 116(suppl_2): 193S - 195S. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |