INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As the field of health-related, quality of life research continues to grow, the development of measures for children continues to lag behind those for adults (1). Although the SF-36, in particular, and several other generic and condition-specific adult measures have become widely used tools in research and clinical practice, measures for the pediatric population are only beginning to emerge. The Child Health Questionnaire (CHQ-PF50), a generic health status tool released for general public use in 1996 (4), is one of several recent efforts to gauge pediatric, health-related quality of life from the patient's (or in the case of young children, parent's) perspective (5, 6). Prior to its public release, the underlying structural integrity of the CHQ was carefully evaluated in children 5-18 yr of age using traditional item scaling analysis and other psychometric methods. Scaling tests were also performed for children with and without chronic health conditionsincluding asthma (4). Early evaluations of the CHQ in children with asthma were obtained from the clinical trial context and from data provided from a managed care organization. The current study extends these findings by examining the performance of the CHQ at two points in time among moderate and low-income inner-city children receiving asthma care.

Several pediatric, asthma-specific health status measures are now available and in use (1, 7). Although condition-specific measures might be more sensitive to changes in the health status of children with asthma, for some applications, generic measures might be desired. In particular, generic measures are needed to compare the health status of children with asthma to that of other children (4, 12). Because generic health status instruments typically employ a basic core set of concepts, these concepts should be applicable across diseases and may be most suitable in some circumstances to assess health-related quality of life or functional status. Several generic instruments have been used in studies of pediatric asthma (3, 4, 6, 12, 13).

Ideally, generic measures might be used in concert with disease-specific instruments, providing both breadth and depth (14). A generic instrument provides information about the relative burden of the disease, whereas a specific tool can potentially provide a more discrete picture. In some instances, however, a generic tool may be sufficient.

In this study, we utilize a test-retest study design with a sample of children with asthma and their caregivers known to be at high risk for poor outcomes to examine several questions regarding CHQ-PF50 performance: Do items within each of the 12 scales of the CHQ-PF50 fit together as single constructs when administered to a sample of parents of children with asthma? (Internal consistency reliability.) Under conditions where their child's asthma appears to be clinically stable, do CHQ-PF50 respondents give similar answers from one time to another? (Short-term test-retest reliability.) Does the CHQ differentiate illness severity among children with asthma? (Construct validity.)

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The CHQ-PF50 is a paper-and-pencil measure completed by parents of children ages 5 through 12 yr. The instrument includes a broad spectrum of child- and family-focused health areas divided into 12 concepts (Physical Functioningsix items; Role/Social Limitations- Physicaltwo items; General Health Perceptionsfive items; Bodily Pain/Discomforttwo items; Family Activitiessix items; Role/Social Limitations-Emotional/Behavioral (counts as two concepts) three items; Parent Impact Timethree items; Parent Impact Emotionthree items; Self-Esteemsix items; Mental Healthfive items; Behaviorfive items; Family Cohesionone item; Change in Healthone item). Individual profile scores for each of the concepts can be computed. It is also possible to derive two summary scores (Physical and Psychosocial). Individual items require participants to respond on a Likert-type scale with higher scores indicating better or more positive health status.

For this project, CHQ-PF50 data were collected from February 1997 to February 1998. Adult caregivers of children, ages 5-12 yr, with asthma (n = 74) were recruited for participation through patient lists from a local managed care organization and asthma clinics at two Chicago medical centers, one public and one private. Eligibility required that the child be previously diagnosed with asthma, free from major comorbidities (e.g., cystic fibrosis, congential heart disease, bronchopulmonary dysplasia), enrolled in no other asthma project at recruitment or for the duration of the study, and accompanied by a parent/ guardian who could provide informed consent and communicate in English.

Four times over the 2-mo study period (baseline, 48 h, 4 wk, and 8 wk postbaseline) parents completed, in person, the self-report CHQ-PF50 and a series of other questionnaires on asthma history and social demographics. Findings from baseline and 48-h follow-up data collection will be reported here. Children performed spirometry at each data collection point according to American Thoracic Society protocol (15) using DX spirometry system software. Asthma severity was assessed using three measures: a symptom-free day instrument, medication use form, and pulmonary function testing (see below). Adult respondents who completed the entire study were paid $80 for their participation, and children received gift certificates.

The study was approved by the institutional review boards of the American Academy of Pediatrics, Rush-Presbyterian-St. Luke's Medical Center, and Cook County Hospital.

Analytic Plan

Descriptive statistics. Basic descriptive statistics (mean, standard deviation, range, percentage of responses at floor/ceiling) were examined for all individual items and scales. In accord with published instructions (4), computed scores were transformed giving each scale a possible range from 0 to 100. For all CHQ-PF50 scales, higher scores indicate more positive functioning or better health status.

Reliability testing. Tests of internal consistency reliability included item-total correlation, Cronbach's  for each scale overall, and  if the item was deleted for individual scale items. To examine test-retest reliability, we used correlation between forms and intraclass correlation coefficient among the subsample of children who were judged to be clinically stable over the designated analysis period. Two physicians (K.W. and E.G.), who had no clinical interaction with child participants, independently rated each child's clinical stability between baseline and the 48-h follow-up using data from (1) parent-completed medication use form, (2) parent-completed symptom-free day measure, and (3) researcher-administered pulmonary function test (PFT). Disagreements over individual patient scores (< 20% of all cases) were resolved through discussion until consensus was achieved. Test-retest reliability analyses presented in this study include only those children who received a clinical stability rating of "no change" (86.5%, n = 64).

Differences in scale scores by selected covariates. t-Tests were used to assess differences in scale scores by selected demographic variables. Demographic items with more than two response choices were dichotomized for analyses as follows: child age (5-9 yr versus 10-12 yr), child race (African American versus other), mother's education level (high school graduate or less versus some college or more), yearly family income (less than $30,000 versus $30,000+), health insurance status (private insurance versus other), and marital status (married versus unmarried).

Validity testing. To explore the measure's ability to distinguish levels of disease burden or severity, t-test and analysis of variance (ANOVA) were used. Variables to measure three dimensions of severity (recent symptom activity, medication use, and PFT) were constructed. The symptom activity variable was created from three items on a symptom-free day instrument (number of days child experienced wheezing or tightness in the chest or cough in the past 2 wk; number of days in the past 2 wk child had a cough that was not from a cold; and number of nights in the past 2 wk child woke up because of asthma, wheezing, tightness in the chest, or cough). Possible values for each item ranged from 0 to 14. The item with the highest number of days was used to determine the symptom activity score. Categories of symptom activity were defined as follows: low = 0 to 2 d, moderate = 3 to 10 d, high = 11+ d. For the current analyses, moderate and high categories were combined due to the small sample size in the high group (n = 7). The medication use variable was computed from two items on the medication use form ("In the past two weeks, how often did your child use an inhaler to take bronchodilator medicine?" and "In the past two weeks, how often did your child use a nebulizer?"). Values for each of the item ranged from 1 to 6 (1 = never and 6 = every day), and the item with the highest score was used to determine the medication severity score. Categories of medication use severity were defined as follows: low = 1 to 2, medium = 3 to 4, and high = 5 to 6. Lastly, the pulmonary function variable was determined by forced expiratory volume in 1 s (FEV₁) percentage predicted (pre--agonist) scores and categorized as follows: mild,  80% of predicted; moderate, 61 to 79% of predicted; and severe,  60% of predicted. Sixty-eight children (92%) provided usable PFT data. For the current analyses, the moderate and severe categories were collapsed due to the small size of the severe subsample (n = 6). All statistical analyses were performed using SPSS software (SPSS Inc., Chicago, IL).

Throughout the RESULTS section, data from several earlier asthma studies detailed in the CHQ reference manual (4) will be presented along with data from the current project; however, we make no statistical comparisons between these studies.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Demographic Characteristics

Study respondents were primarily mothers (87.8%) and the sample included slightly more male (51.4%) than female children. Mean child age was 9.1 yr (range = 5-12 yr). Additional demographic characteristics of the sample are reported in Table 1.

Descriptive Statistics for CHQ-PF50 Scales

Basic descriptive information about the 12 scales and 2 summary scores is presented in Table 2. Mean scale scores ranged from 56.7 to 88.4. Lowest scores were reported in the General Health Perceptions domain and highest scores in the Role/ Social-Emotional/Behavior domain. On several scales, more than 25% of respondents scored the maximum possible total of 100, and for two scales (Role/Social-Physical and Role/Social-Emotional/Behavior) more than 60%, indicating potential ceiling effects for some scales among children with asthma. In Table 3, current study scale means are compared to national norms for all children and to scores obtained in four prior studies of children with asthma published in the CHQ reference manual (4). For the majority of scales and both summary scores, current study means were lower (indicating worse health or functioning) than those observed in the other studies reported.

Internal Consistency Reliability Testing

Item-total correlations including Cronbach's  for overall scales and individual  levels for the scale if an item was deleted are detailed in Table 4. With acceptable to high item-total correlations for the majority of items, Cronbach's  levels are consistently high for all but the "General Health Perceptions" scale. Table 5 presents Cronbach's  from the current study and four published asthma studies (4). In 8 of the 11 scales examined, coefficient  levels obtained in the current project were higher than  levels observed in the earlier pediatric asthma samples.

Test-Retest Reliability

Data from baseline and the 48-h follow-up were used to assess test-retest reliability (see Table 6). For the 11* scales examined, correlation between forms ranged from a low of 0.17 (Bodily Pain) to a high of 0.74 (Parent ImpactEmotion); intraclass correlation coefficients ranged from a low of 0.37 (Parent ImpactTime) to a high of 0.84 (Behavior).

Differences in Scale Scores by Selected Covariates

In an assessment of potential differences in mean scale scores by selected covariates (e.g., mother's level of education, family income, health insurance status, study recruitment site, parental marital status, child race, child sex, and child age), no differences were noted by mother's educational level, child race, or child age. Only one significant difference was observed by child sex (mean score for the Parent ImpactEmotion scale was significantly lower for boys than girls; 57.2 versus 71.1, t = 2.17, p  0.03).

Several differences were found by family composition, family income, and insurance status. At baseline, mean scores for children in married parent families were significantly higher than those in nonmarried parent families for the Role/Social- Emotional/Behavior, Parent Impact-Time, Behavior, Mental Health, General Health Perceptions, and Family Activities scales and the Physical and Psychosocial summaries (p  0.05). Families with a yearly income of $30,000+ scored significantly higher on the Role Physical, Bodily Pain, Mental Health, General Health Perceptions scales, and the Physical Health Summary than those with annual household incomes of  $29,999 (p  0.05). Similarly, respondents with private health insurance (as opposed to public or no insurance) reported higher scores on 6 of the 12 scales (Physical Functioning, Role/Social-Physical, Role/Social-Emotional/Behavior, Parent-Impact, Time, Mental Health, General Health Perceptions) and both summary scores (p  0.05).

Validity Testing

We assessed validity by evaluating whether CHQ-PF50 scales were sensitive to differences in disease activity or severity using the three constructed severity variables described above (recent symptom activity, medication use, and FEV₁ [% predicted]). Figure 1 compares scale scores of children categorized as having "low" versus "moderate/high" recent symptom activity. Children whose recent symptom activity was categorized as "low" had consistently higher mean scale and summary scores (Physical Functioning, Role/Social-Physical, General Health Perceptions, Bodily Pain, Family Activity, Self-Esteem, Mental Health, Physical Summary, Psychosocial Summary; p  0.05).

View larger version (46K): [in this window] [in a new window]   Figure 1.   Differences in mean Child Health Questionnaire (CHQ) scale scores by severity (recent symptom activity) as reported by caregivers of children ages 5-12 with asthma. Open bars, low recent symptom activity; solid bars, moderate/high recent symptom activity. PF = Physical Functioning; RP = Role/Social-Physical; GH = General Health Perceptions; BP = Bodily Pain; FA = Family Activities; REB = Role/Social- Emotional/Behavioral; PT = Parent Impact-Time; PE = Parent Impact- Emotion; SE = Self-Esteem; MH = Mental Health; BE = Behavior; FC = Family Cohesion; PhS = Physical Summary; PsS = Psychosocial Summary. *p  0.05.

Differences in mean scale scores by medication use severity (classified as "low," "medium," and "high use") did not follow a similar pattern. As illustrated in Figure 2, no consistent trend emerged. ANOVA revealed no significant differences between mean scale or summary scores among the medication use groups.

View larger version (72K): [in this window] [in a new window]   Figure 2.   Differences in mean CHQ scale scores by severity (medication use) as reported by caregivers of children ages 5-12 yr with asthma. Open bars, low use; solid bars, medium use; dotted bars, high use. PF = Physical Functioning; RP = Role/Social-Physical; GH = General Health Perceptions; BP = Bodily Pain; FA = Family Activities; REB = Role/ Social-Emotional/Behavioral; PT = Parent Impact-Time; PE = Parent Impact-Emotion; SE = Self-Esteem; MH = Mental Health; BE = Behavior; FC = Family Cohesion; PhS = Physical Summary; PsS = Psychosocial Summary.

Using PFT scores (FEV₁% predicted) as an indicator of severity (see Figure 3) did not produce significant differences in scale scores. However, mean scores were in the expected direction for a number of CHQ scales.

View larger version (43K): [in this window] [in a new window]   Figure 3.   Differences in mean CHQ scale scores by severity (FEV1% predictedpre--agonist administration) as reported by caregivers of children ages 5-12 yr with asthma. Open bars, mild; solid bars, moderate/severe. PF = Physical Functioning; RP = Role/Social-Physical; GH = General Health Perceptions; BP = Bodily Pain; FA = Family Activities; REB = Role/Social-Emotional/Behavioral; PT = Parent Impact-Time; PE = Parent Impact-Emotion; SE = Self- Esteem; MH = Mental Health; BE = Behavior; FC = Family Cohesion; PhS = Physical Summary; PsS = Psychosocial Summary.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Performance characteristics of the CHQ-PF50 when used in a sample of low- to moderate income inner-city parents of children with asthma presented mixed results. Although the instrument addresses a broad range of health concepts, our findings suggest that some scales may be more salient than others in assessing health status of children at highest risk for asthma morbidity.

Descriptive statistics from the current study illustrate that CHQ-PF50 scales capture a wide range of experiences, and are similar to those observed in other samples of children with asthma, but lower than national norms for a general population. Some concern about ceiling effects for several scales, however, is warranted. Large ceiling effects may indicate that certain items lack relevancy or sensitivity for children with asthma, at least in the present sample, and will limit a scale's responsiveness to change in a child's health condition.

Findings do, however, indicate high internal consistency for all but one scale (General Health Perceptions), which is consonant with previously published results (4). Generally, levels of Cronbach's  observed in the current sample are higher than those found in other studies of asthma using the CHQ-PF50. This may be attributable to the homogeneity of our sample, characteristics of the individual items, use of a longer questionnaire version than used in some other studies, or unknown attributes of the early referent studies of children with asthma. Relatively low correlations for test-retest reliability raise questions about the measure's short-term stability and warrant further exploration in studies of both ill and well children.

Differences noted by several key demographic variables suggest that children with asthma may fare better or worse in some aspects of functional health or well being based on parental marital status, family income, and health insurance status. Similar differences based on parent marital status in a representative sample of U.S. children have been reported elsewhere (16).

In tests of validity, the majority of CHQ-PF50 scales were sensitive to differences in disease severity when quantified by recent symptom activity but not when severity was measured by medication use or premedicated FEV₁ (% predicted). The measure's inability to distinguish severity based on FEV₁ might be expected given that spirometry targets organ function at only one point in time whereas CHQ-PF50 items required respondents to report on a variety of experiences over the past 4 wk. The findings may also indicate that the CHQ-PF50 provides important information about a child's health status beyond that obtainable from a test of lung function. Other studies have noted a similar lack of association between FEV₁ and functional status (7, 17, 18).

Despite some limitations (e.g., ceiling effects, low short-term test-retest correlations) the findings of the current study are important for several reasons. First, this study provides new information about test-retest reliabilitya component absent in many studies of outcome measures. Further, the current study utilized a sample overrepresenting low-income, minority, and inner-city respondents in the United States (although even these results may have limited applicability to a higher risk inner-city population that has poor access to continuing care or use the Emergency Department as their primary source of care). It is critical that any quality of life instrument to be used with children with asthma, whether generic or condition specific, demonstrate validity and reliability for the population that may stand to benefit the most from a better understanding of the impact of asthma on daily life. Validity is not an inherent property of an instrument but is dependent on a measure's performance in a population (19); thus additional evidence to support the validity of the CHQ-PF50 (as well as other QOL instruments) with different populations of children is needed. Future studies can help determine whether the findings of the current study are generalizable or whether they are specific to a lower income, largely minority, urban population. Socioeconomic status and the social context in which children develop may play a role in quality of life measurement that is yet undefined.

If the CHQ-PF50 is to measure health status of children with asthma accurately and appropriately, additional data are also necessary to examine whether the instrument is responsive to changes in a child's condition over time. Such studies require the collection of longitudinal data to determine whether scale scores improve or worsen under conditions in which a child's asthma improves or worsens. Such information, although costly and difficult to collect, is vital to developing psychometrically sound measures and establishing clinical relevance.

To advance the science of functional health research for children with asthma and their families, future efforts must also include comparisons of condition-specific and generic instruments. Outcome measurement will be enhanced as instruments are used and critiqued by those beyond the developers themselves. Studies that directly compare two or more instruments are necessary to evaluate the relative strengths and weaknesses of different measures and different types of measures as well as examine the potential links between generic and condition-specific outcome instruments (20). Although many in health services research continue to search for a pediatric "gold standard," given the current state of the field, the "best" measure will likely depend on individual project aims, methods, resources available, and the population under study.