• Disability;
  • Outcome measurement;
  • Childhood Health Assessment Questionnaire;
  • Juvenile arthritis


  1. Top of page
  2. Abstract


To determine the relative discriminant validity of 3 new versions of the Childhood Health Assessment Questionnaire (CHAQ) and to determine the relative concordance between children and their parents.


The parents of 48 children with musculoskeletal disability and 101 nondisabled controls were given the CHAQ and the 3 revised versions in random order. Children older than 5 years also completed the questionnaires.


All 3 new versions of the CHAQ were more sensitive at differentiating patients from controls (relative efficiency 1.35–1.65); the versions with 8 added items were even more sensitive (relative efficiency 1.79–2.32). The new versions of the CHAQ suffered less from a ceiling effect and were more normally distributed. Concordance between children and parents was moderate to high for all versions.


Researchers and clinicians should consider using a revision of the CHAQ. By being more sensitive at differentiating patients, the revised versions of the CHAQ will allow fewer subjects to be studied and will be more able to detect differences between patients in the clinic.


  1. Top of page
  2. Abstract

Although chronic childhood musculoskeletal disorders are rarely fatal, they frequently result in physical disabilities that have significant health impacts. The burden of illness associated with such disorders is frequently underestimated (1). Such disorders as juvenile rheumatoid arthritis may induce substantial disease-related chronic pain (2) and affect children's independence and their ability to perform daily or desired activities to varying degrees. Such disorders may also have implications for a child's or adolescent's developmental progress (3); these in turn may predispose or lead to both psychosocial (4) and physical (5) manifestations beyond the childhood years. Chronic childhood musculoskeletal disorders are associated with increased utilization of health care services and lead to tremendous personal and societal economic costs (6). Thus, further investigation into effective means of evaluating and treating chronic childhood musculoskeletal disorders is warranted.

It is important to have accurate self-report measures of physical ability for patients with chronic conditions, such as musculoskeletal disorders. Self-report measures are known to yield information that may not be attainable from clinician-based scales focusing on disease activity (7, 8). It is vital for the delivery of patient-oriented care to recognize patients' self perceptions of multiple dimensions of their function—including aspects of self care, day-to-day and long-term variability in functional well-being, and disease impact on daily activities. Self-report health status measures may be helpful in optimizing the design and monitoring of patients' routine care plan and have become very important in optimizing therapeutic protocols, clinical trials, and other research.

In the pediatric population, it is valuable to have measurement tools that allow parents to act as effective proxies in those circumstances when child self report is not possible. Questionnaires designed or later validated for parental proxies have been utilized for various chronic conditions in children, including asthma (9), diabetes (10), end-stage renal disease (11), and musculoskeletal conditions (8, 12–14). More recently, the validity and utility of involving adolescents and children themselves in the evaluation of chronic diseases have also been demonstrated. For instance, self report by children aged 5–15 years regarding their musculoskeletal condition during the development of the Activities Scale for Kids uncovered additional items that were not recognized as being important by parents, clinicians, or other researchers (15). Children as young as 5 years have been shown to be capable of understanding and using simple rating scales, particularly where questions relate to physical tasks (16, 17).

The most widely utilized functional status tool for childhood musculoskeletal disorders is the disability index of the Childhood Health Assessment Questionnaire (CHAQ) (4). Adapted from the adult Health Assessment Questionnaire (HAQ), the CHAQ was initially developed to assess childhood arthritis and has been shown to be an effective instrument with high validity, reliability, and responsiveness to change over time (18). Over the years, the use of the CHAQ has broadened to other pediatric rheumatologic conditions, such as juvenile dermatomyositis (19) and chronic musculoskeletal pain (20), as well as other musculoskeletal conditions such as spina bifida (21). The CHAQ has been designed for use either as a self report or a parental proxy questionnaire. This 30-item questionnaire is thought to have advantages over other measures of physical function due to such aspects as its multidimensionality (it assesses 8 domains of physical function) (22). The CHAQ is in use internationally; cross-cultural adaptations were recently validated in 32 countries (23).

Despite its advantages and pervasive use, the CHAQ is similar to many other measures of physical function in that it focuses only on the negative aspects of ability and does not measure an individual's strengths; that is, it measures only disability and not the whole spectrum of ability. As with many physical function survey tools, the recognition that there is a continuum of function (from very poor health and function to excellent health and function) is not considered with the CHAQ; health has been narrowly defined as the lack of limitations (24).

With 0 as the best possible score (representing no functional limitations), the CHAQ has been demonstrated to suffer from a ceiling effect whereby scores are clustered at the normal end of the scale, or near 0 (19, 25, 26). Specific limitations in the research and clinical utility of the CHAQ occur as a result of the ceiling effect. False-negative outcomes may ensue as the ceiling effect makes the scale intrinsically less sensitive to milder levels of disability (24). As with the adult version of the HAQ, a marked ceiling effect also decreases the ability to assess improvement in health for those patients who are close to or at the ceiling (24, 27, 28).

We are interested in 2 potential ways of reducing the ceiling effect and improving the sensitivity of a functional measure, such as the CHAQ. One method is to use different response options in answering the same questions (items) posed by the original questionnaire. These different response options allow for answers that include strengths as well as limitations (see below). The other method is to add more challenging items that will allow respondents who have close to normal function to score below the ceiling.

Over the past few years, there has been an increasing emphasis on recognizing the variability in function and strengths that can be shown by individuals who had classically been given designations of disability (29). This new outlook is demonstrated by the reorientation of the World Health Organization's International Classification framework regarding functioning and disablement from that entitled “Impairment, Disabilities and Handicaps” to that of “Impairment, Activities and Participation” in 1997 (30). If we apply this outlook to functional status measures, we might develop tools that identify patients' strengths and abilities rather than their disabilities alone; such tools would likely be more clinically sensitive and responsive. We hypothesized that by adapting the CHAQ so that an individual's strengths as well as weaknesses could be measured (i.e., by utilizing new response scales as well as by adding more challenging questions), the limitations of the CHAQ—namely the ceiling effect and poor sensitivity for children with relatively good function—may be reduced.

The goal of this study was to evaluate if a modified version of the CHAQ that measures strengths as well as weaknesses offers an increased ability to 1) distinguish between patient and control populations, 2) exhibit less of a ceiling effect (i.e., assess near-perfect function), and 3) show a more normal distribution in both patient and control populations compared with the original questionnaire. To study this question, we used the CHAQ and 3 modified versions. We also sought to evaluate the interobserver reliability of the modified versions when parents act as proxy respondents. Our specific hypotheses were that the modified versions of the CHAQ, when compared with the original, would be more evenly distributed around an average value, would differentiate more between healthy children and those with musculoskeletal disorders, and would exhibit minimal or no ceiling effects.


  1. Top of page
  2. Abstract


Patients with musculoskeletal conditions were consecutively recruited over a 3-month period from the rheumatology, hemophilia, and orthopedic clinics at the Hospital for Sick Children (HSC), a pediatric tertiary care center, and the Bloorview MacMillan Children's Centre (BMCC), a referral pediatric rehabilitation center. Patients who presented to the rheumatology service but who did not have a musculoskeletal disorder (e.g., isolated uveitis) were excluded; similarly, patients who had been given a diagnosis of systemic lupus erythematosus but who had never had a musculoskeletal problem (e.g., never had arthritis nor avascular necrosis, etc.) were also excluded. All patients from the hemophilia service were screened by a hemophilia clinic nurse and considered for the study if they had had at least 1 major or 2 minor bleeds into their joints. Patients at the orthopedic services at the HSC and the BMCC that self identified as having had a musculoskeletal problem—short-term (e.g., fractures), congenital (e.g., spina bifida), or chronic—were invited to participate.

During the same time period, 2 groups of control subjects who did not have musculoskeletal problems were consecutively recruited. The first group was enrolled from the dermatology and ears, nose, and throat (ENT) clinics at the HSC. The second group was enrolled from municipally run summer day camps. None of the control children had a history of chronic physical limitation or musculoskeletal disability. Using 2 sets of controls allowed us to include children with nonmusculoskeletal illnesses as well as those who were healthy and highly functioning with respect to physical activity.

A previous study assessing the variability of CHAQ scores in children from the HSC with musculoskeletal disease determined a mean population CHAQ score of 0.325 with a standard deviation of 0.56 (31). Assuming similar variability in the CHAQ scores of control children without musculoskeletal disease with a presumed mean score of 0, a minimum sample size of 47 subjects and 47 controls was required to obtain a power of 0.80 with an α error of 0.05.


This study was approved by the Research Ethics Boards at the HSC and the BMCC. All families provided informed consent prior to participation. Parents of eligible children of any age, and the eligible children themselves if 5 years old and above, were invited to participate. Where both a child and his or her parent were available and capable, each was independently interviewed (to yield paired data to examine for interrater concordance between children and parents). For all other analyses, data from the parent were used where available. If 2 parents were present for a given child, parents themselves determined which one parent participated. Parents and children were guided through their respective randomly ordered questionnaires by 1 of 2 trained researchers, and remained blinded to each other's answers.

The 2 trained researchers worked from a pretested interview script. The interview script and the questionnaire booklets were pretested on a pilot group of 24 families from the HSC rheumatology clinic. In the course of the study, the researchers repeatedly observed one another to ensure that a similar interview style was maintained.


Each questionnaire booklet contained 4 versions of the CHAQ questionnaire with different rating scales: the original, visual analog scale (VAS), categorical, and choice scales (see below). Because previous research has indicated that participants' responses to the first of a series of rating scales may systematically influence responses to subsequent scales in a questionnaire (32), the 4 versions were placed in random order within each booklet. The assignment of random questionnaire order was done using computer-generated random numbers. In addition to the 4 different forms of the CHAQ questionnaire, each subject answered a demographic questionnaire, a 1-page pain description sheet, and the 1-page Quality of My Life questionnaire (33). At the end of each session, an interview quality-scoring sheet was completed by the researcher to reflect their assessment of the degree of difficulty the respondent had in completing the questionnaires.

Each modified version of the CHAQ consisted of items concerning the same activities of daily living over the same time frame as the original CHAQ. To measure perceived strengths, as well as deficits, the revised scales introduced the concept of a comparison between the capabilities of the subject and other children his or her age. The choice of wording for the modified questionnaires was done iteratively with pilot subjects recruited from the HSC rheumatology clinic.

The revision that we called the VAS used a 10-cm VAS for the response to each CHAQ question (see Figure 1). The VASCHAQ was doubleanchored with the anchors being: “much worse than other kids my age” and “much better than other kids my age.” The time period was “over the past week,” which is the same time period used in the original CHAQ (OrigCHAQ).

thumbnail image

Figure 1. Examples of the modified questionnaire responses: A, response options for the visual analog scale Childhood Health Assessment Questionnaire (CHAQ) questionnaire; B, response options for the categorical CHAQ; C, response options for the choice CHAQ.

Download figure to PowerPoint

The categorical scale (CatCHAQ) offered 5 response options for each CHAQ item. Subjects were asked to consider their abilities compared with most other kids their age and were asked to rate themselves (or their children as appropriate) as either “much worse,” “a little worse,” “the same,” “a little better,” or “much better” in their ability to perform the named activity in the past week (see Figure 1).

The choice scale (ChoiceCHAQ) consisted of 2 sentence stems for each question, one expressing that the activity was able to be performed “really easily” by some children, the other expressing that the activity was able to be performed “slowly or with difficulty” by some children. Respondents were asked to choose 1 of the 2 sentence stems that was most like the subject in the past week, and then subsequently to indicate whether the chosen statement was “really true” or “sort of true” for the subject (see Figure 1). (Full modified questionnaires are available from the authors.)

In addition to the 30 items asked in the original CHAQ, the 3 modified versions also included the addition of 8 new challenging questions. These 8 challenging questions were derived both from the investigators' experience and via key informant interviews with musculoskeletal patients at the HSC rheumatology clinic. Patients were asked to identify activities that they believed to be particularly challenging, either for themselves or their peers, but which were reasonably compatible with abilities of children their age. The authors then reviewed the activities and formulated the 8 new items; these items were pretested as part of the modified questionnaires administered to the 24 pilot families. These items are as follows:

  • 1
    I think I could have done climbing activities by myself (examples: climbing trees, rocks, or climbing over a fence).
  • 2
    I think I could have played team sports with others in my class (examples: basketball, baseball, soccer, hockey).
  • 3
    I think I could have played some sports by myself or with a few friends (examples: dribbling and shooting basketball).
  • 4
    I think I could have played team sports in competitive leagues (examples: local basketball, baseball, soccer, or hockey teams).
  • 5
    I think I could have kept my balance while playing rough games (examples: tag, wrestling, karate, judo).
  • 6
    I think I could have done activities I usually enjoy for a long time without getting tired out (examples: swimming, jogging, tennis, badminton, rowing, skiing).
  • 7
    I think I could have run in a race (example: 100-meter dash).
  • 8
    I think I could have worked carefully with my hands (examples: building Lego, making models, sewing, making bead necklaces).

The 3 new questionnaires were scored differently from the OrigCHAQ. We did not consider domain structures with the new questionnaires, but rather the summary score was determined by averaging all the answered items. Also for the new questionnaires, no consideration was made for the use of aids, devices, or outside help. In previous work, we found that changing the scoring system of the OrigCHAQ changed the sensitivity and distribution of its scores by only a very small degree (34); so for this study the OrigCHAQ was scored using its standard method (18).


Group differentiation.

Discriminant validity of the 4 questionnaire versions was determined on the basis of their ability to differentiate between control and subject populations. The relative ability of the original and modified questionnaires to differentiate between controls and patients was tested using the relative efficiency statistic (ratio of squared t-scores).

Score distribution.

To determine whether the modified questionnaires exhibited less of a ceiling effect and a more normal distribution when compared with the original questionnaire, normal probability plots, box-and-whisker plots, and the Kolmogorov-Smirnov (K-S) standard test of normality were used. These analyses were conducted separately for both patient and control populations.

Parent-child agreement.

The intraclass correlation coefficient (ICC) was used to compare concordance between parent and child responses for the different versions of the questionnaire. We adapted the benchmarks proposed by Landis and Koch (35) and considered the ICC to indicate poor agreement if the value was <0, slight agreement at 0.00–0.20, fair agreement at 0.21–0.40, moderate agreement at 0.41–0.60, substantial agreement at 0.61–0.80, and almost perfect agreement at 0.81–1.00 (35).

Additional challenging questions.

To assess the additional information conferred by the 8 added challenging questions to the modified versions, the calculation of relative efficiency was repeated including and excluding the added questions. For reporting purposes, the original CHAQ and its 30-question adaptations (VASCHAQ, CatCHAQ, and ChoiceCHAQ) are referred to below as the CHAQ-30 whereas versions with the additional questions are called CHAQ-38.


  1. Top of page
  2. Abstract


The patient and control groups were similar, with the exception that a significant age difference was found for the 3 sample groups, with the patients older than the controls; there was no difference between the 2 control groups (Table 1). Patients were recruited from rheumatology (n = 35), orthopedic (n = 9), and hemophilia (n = 3) clinics, with primary diagnoses including juvenile rheumatoid arthritis, autoimmune rheumatologic disorders, injuries, fractures, spina bifida, and hemophilia with a history of hemarthroses. Controls were recruited from municipal summer camps (n = 46), as well as ENT (n = 31) and dermatology (n = 24) clinics. One patient originally recruited from the ENT clinic was assigned to the patient group due to musculoskeletal paralysis; exclusion of this patient made no difference in the results.

Table 1. Demographics of subjects and controls: age, sex, and identity
 Subjects (n = 48)Controls (n = 101)P
Clinic (n = 55)Park (n = 46)
  • *

    NS = Not significant.

Age, mean (range)9.6 (2.1–17.5)6.15 (1.1–16.8)6.12 (2.4–11)< 0.0001
Sex of respondents, % female85.463.658.70.03
Sex of children, % female56.247.350.0NS*
No. of parent-child pairs26132 
No. of parents only (child > 5 years)71122 
No. of parents only (child < 5 years)122715 
No. of children only (age > 5 years)347 

Ability to differentiate groups.

All 4 versions of the CHAQ-30 differentiated between patients and controls (Table 2). The 2 control groups were treated as 1 because separate analyses found no difference between the results for the 2 control groups. Each of the modified CHAQ-30 versions demonstrated greater efficiency relative to the OrigCHAQ, with the VASCHAQ-30 scale having the best ability to discriminate.

Table 2. CHAQ-30 modified versions: comparison of summary scores for subjects and controls*
Questionnaire versionnMedian score for subjects (range)Median score for controls (range)2 Sample t-testMann-Whitney U testRelative efficiency
ClinicParkT statisticPU statisticZ scoreP
  • *

    CHAQ-30 = 30-item Childhood Health Assessment Questionnaire; df = degrees of freedom; VAS = visual analog scale.

  • P value refers to the comparison of the mean value between subjects and controls using the 2 sample (unpaired) t-test.

  • P value refers to the comparison of the median value between subjects and controls using the Mann-Whitney U test with ties omitted.

Original, 0 = best (0–3)1390.625 (0–2.875)0.125 (0.0–2.0)0.0 (0.0–0.875)−3.450 (63 df)0.001139−3.140.00171
VAS, 10 = best (0–10)1455.32 (0.83–9.38)5.92 (2.93–9.82)6.23 (4.65–9.88)−4.434 (95 df)≤ 0.0001708−4.16≤ 0.00011.65
Categorical, 2 = best (−2–2)141−0.03 (−1.8–2)0.11 (−0.93–2)0.31 (−0.04–1.86)−4.002 (87 df)≤ 0.0001352−4.37≤ 0.00011.35
Choice, 4 = best (0–4)1383.66 (1.6–4)3.95 (2.67–4)3.95 (3.0–4.0)−4.143 (57 df)≤ 0.000155−3.670.00021.44

Given the statistically significant difference between the ages of the subject and control populations, general linear models with age as a covariate were constructed to determine whether age was a potential confounder of group differences in summary scores; the inclusion of age as a covariate made no difference in any of the results.

Comparing the 38-question modified versions of the original CHAQ, all 3 modified versions better discriminated between patients and controls than the OrigCHAQ (Table 3). The VASCHAQ-38 was best at differentiating between patients and controls, followed by the ChoiceCHAQ-38, then by the CatCHAQ-38.

Table 3. CHAQ-38 modified versions: comparison of summary scores for subjects and controls*
Questionnaire versionNMedian score for subjects (range)Median score for controls (range)2 Sample t-testMann-Whitney U testRelative efficiency
ClinicParkT statisticPU statisticZ scoreP
  • *

    CHAQ-38 = 38-item Childhood Health Assessment Questionnaire; VAS = visual analog scale; df = degrees of freedom.

  • P value refers to the comparison of the mean value between subjects and controls using the 2 sample (unpaired) t-test.

  • P value refers to the comparison of the median value between subjects and controls using the Mann-Whitney U test with ties omitted.

VAS, 10 = best (0–10)1455.04 (0.76–9.4)5.97 (2.49–9.5)6.44 (6.60–9.85)−5.26 (94 df)≤ 0.00011,157−4.72≤ 0.00012.32
Categorical, 2 = best (−2–2)141−0.07 (−1.84–2)0.16 (−1.35–2)0.29 (−0.13–1.77)−4.62 (84 df)≤ 0.0001413−5.15≤ 0.00011.79
Choice, 4 = best (0–4)1383.53 (1.58–4)3.88 (2.39–4)3.88 (3–4)−4.70 (57 df)≤ 0.0001385−4.57≤ 0.00011.85

Distribution of scores.

Of the 4 versions of the CHAQ-30 used in patients, the VASCHAQ-30 differed the least from a normal distribution (K-S test statistic 0.082, P = 0.90) whereas the OrigCHAQ (K-S test statistic 0.2, P = 0.04) was statistically significantly not normal. Analysis of the modified versions with the inclusion of the 8 challenging questions broadened the score distribution for all 3 scales, resulting in smaller K-S test statistics and similar or larger P values, indicating a closer approximation to normality (Table 4). Similar results were seen for the control subjects (data not shown).

Table 4. CHAQ-30 and CHAQ-38 modified versions: Kolmogorov-Smirnov test in patients*
Questionnaire versionCHAQ-30CHAQ-38
K-S test statisticPK-S test statisticP
  • *

    CHAQ-30 = 30-item Childhood Health Assessment Questionnaire; CHAQ-38 = 38-item CHAQ; K-S = Kolmogorov-Smirnov; VAS = visual analog scale.

  • P value refers to probability of the observed distribution being sampled from the normal distribution.


As seen in the box-and-whisker plots (Figure 2), a strong ceiling effect was clearly observed in the OrigCHAQ, with scores for patients and the control populations clustering at the best achievable score of 0. To a slightly lesser extent, a strong ceiling effect clustering at the best achievable score of 4 was also noted in the ChoiceCHAQ for the control group, irrespective of whether the 30-question version or the 38-question version was used. In contrast, the distributions of the VASCHAQ and CatCHAQ for both the 30-question and 38-question versions were approximately normal.

thumbnail image

Figure 2. Box-and-whisker plots of the original and modified versions of the Childhood Health Assessment Questionnaire (CHAQ) for both patients and controls. The middle line in each box represents the median value. The shoulders of the box correspond to the 25th and 75th percentile scores. The whiskers enclose all the values, with the exception of outliers, which are plotted separately. A, The original CHAQ (OrigCHAQ) for both patients and controls. B, The 30-question and the 38-question visual analog scale CHAQs (VASCHAQ-30 and VASCHAQ-38) for both patients and controls. C, The 30-question and the 38-question categorical CHAQ (CatCHAQ-30 and CatCHAQ-38) for both patients and controls. D, The 30-question and the 38-question choice CHAQ (ChoiceCHAQ-30 and ChoiceCHAQ-38) for both patients and controls.

Download figure to PowerPoint

Agreement between parents and children.

The mean child-parent differences were smallest between parental proxies and their children with the VASCHAQ and ChoiceCHAQ scales, with similar results for both the CHAQ-30 and the CHAQ-38 modified versions (Table 5).

Table 5. Concordance between children and their parents*
Questionnaire versionCHAQ-30CHAQ-38
Mean differences (child − parent)t-testPIntraclass correlation coefficientMean differences (child − parent)t-testPIntraclass correlation coefficient
  • *

    CHAQ-30 = 30-item Childhood Health Assessment Questionnaire; CHAQ-38 = 38-item CHAQ; VAS = visual analog scale.

Original−0.19−2.3 (37 df)0.0240.463
VAS0.13−0.44 (38 df)0.660.3330.24−0.81 (38 df)0.420.413
Categorical0.263.02 (36 df)0.00460.6710.263.02 (36 df)0.00460.675
Choice0.0490.82 (36 df)0.420.5270.0631.032 (36 df)0.310.669

Using the ICC benchmarks discussed above, the VASCHAQ-30 exhibited fair agreement and the VASCHAQ-38 demonstrated moderate agreement between children and parents. Both the OrigCHAQ and the ChoiceCHAQ-30 also exhibited moderate agreement between parent and child data, whereas the ChoiceCHAQ-38 and both versions of the CatCHAQ demonstrated substantial agreement (Table 5). Thus, overall, parent-and-child pairs were concordant in their responses for all 4 versions of the CHAQ, supporting the use of these questionnaires whether they are completed by parental proxies or by children themselves.


  1. Top of page
  2. Abstract

We have shown that the modified versions of the CHAQ demonstrated better differentiation between controls and subjects (and therefore greater discriminant validity) when compared with the original CHAQ. Greatest differentiation was noted with the CHAQ-38 modified versions that included the new challenging questions and with the VASCHAQ. This greater discriminant validity allows the modified CHAQ questionnaires to better guide clinical research and management by allowing for more precise, meaningful comparisons of subject and control groups and by clarifying the distinct range of functional abilities of subject populations.

One advantage of the modified versions of the CHAQ is that, because they are more sensitive, studies that use these new versions will need fewer subjects. For example, to see the difference between our patients and our controls, a clinical study (in which the OrigCHAQ was the outcome) would have had to enroll ∼144 subjects (to achieve an α error level of 0.05 and a β error level of 0.90) based on standard calculations (36). However, with the same criteria using the VASCHAQ-38, one would only need to enroll 50 subjects. Trials using the CatCHAQ-38 and ChoiceCHAQ-38 would require 68 and 84 subjects, respectively. Clearly, using a more sensitive version of the CHAQ may lead to less expensive, easier to perform studies that require fewer patient volunteers.

All modified versions of the CHAQ, as well as the original CHAQ, demonstrated fair to substantial interrater reliability or concordance between parent-child pairs, thus supporting the use of the modified questionnaires as either self report or parental proxy measures.

As was initially hypothesized, the modified versions of CHAQ resulted in a more normal distribution, a diminished ceiling effect, and greater sensitivity compared with the origCHAQ. In particular, of the 3 modified response versions, the VASCHAQ appears to offer the best differentiation between subjects and controls and the most normal distribution. However, the CatCHAQ may be the easiest to complete, and may thus be preferred by some investigators; the CatCHAQ also showed the best concordance as a proxy measure. Given these advantages, it would likely be worthwhile for investigators to utilize the VASCHAQ or CatCHAQ in future studies.

Our study should be interpreted considering certain possible limitations. Some patients and control families did not have enough time to complete all the questionnaires. As a result, there was some missing data. However, 90% of the respondents completed all of the required questionnaires and the quality of the interviews—even when incomplete—was high as recorded by the interviewers. Given the strength of our findings, we do not believe that the missing data has substantially affected our conclusions. Also, this study was carried out at only a single center. It is possible that children and their families elsewhere might respond to the altered versions of the CHAQ differently due to different language nuances. Our results should be confirmed at other sites before a new version of the questionnaire is adopted widely. Finally, we did not measure the sensitivity of the new questionnaires over time; this should be a priority for future research.

By using alternate scaling methods in which the full spectrum of physical function is explored, and by adding questions involving more challenging activities, we have shown that the discriminant validity and sensitivity and the marked ceiling effect of the CHAQ can be improved. Of the 3 modified versions, the VASCHAQ was the best at differentiating between subjects and controls, and also yielded data that was the most normal in distribution, whereas the CatCHAQ showed the highest concordance between children and their parents. Future clinical trials should consider using the 38-question versions of the VASCHAQ or CatCHAQ because they are more sensitive at differentiating patients than the original CHAQ and because they do not exhibit a ceiling effect.


  1. Top of page
  2. Abstract
  • 1
    Badley EM, Rasooly I, Webster GK. Relative importance of musculoskeletal disorders as a cause of chronic health problems, disability, and health care utilization: findings from the 1990 Ontario Health Survey. J Rheumatol 1994; 21: 50514.
  • 2
    McGrath PA, Speechley KN, Seifert CE, Biehn JT, Cairney AE, Gorodzinsky FP, et al. A survey of children's acute, recurrent, and chronic pain: validation of the pain experience interview. Pain 2000; 87: 5973.
  • 3
    Schikler KN. Is it juvenile rheumatoid arthritis or fibromyalgia? Med Clin North Am 2000; 84: 96782.
  • 4
    Tucker LB. Whose life is it anyway: understanding quality of life in children with rheumatic diseases. J Rheumatol 2000; 27: 811.
  • 5
    Zitting P, Vanharanta H. Why do we need more information about the risk factors of the musculoskeletal pain disorders in childhood and adolescence? Int J Circumpolar Health 1998; 57: 14855.
  • 6
    Newacheck PW, Halfon N. Prevalence and impact of disabling chronic conditions in childhood. Am J Public Health 1998; 88: 6107.
  • 7
    Rider LG, Feldman BM, Perez MD, Rennebohm RM, Lindsley CB, Zemel LS, et al, Juvenile Dermatomyositis Disease Activity Collaborative Study Group. Development of validated disease activity and damage indices for the juvenile idiopathic inflammatory myopathies: I. Physician, parent, and patient global assessments. Arthritis Rheum 1997; 40: 197683.
  • 8
    Duffy CM, Duffy KN. Health assessment in the rheumatic diseases of childhood. Curr Opin Rheumatol 1997; 9: 4407.
  • 9
    Asmussen L, Olson LM, Grant EN, Fagan J, Weiss KB. Reliability and validity of the Children's Health Survey for Asthma. Pediatrics 1999; 104: e71.
  • 10
    Wake M, Hesketh K, Cameron F. The Child Health Questionnaire in children with diabetes: cross-sectional survey of parent and adolescent-reported functional health status. Diabet Med 2000; 17: 7007.
  • 11
    Kurtin PS, Landgraf JM, Abetz L. Patient-based health status measurements in pediatric dialysis: expanding the assessment of outcome. Am J Kidney Dis 1994; 24: 37682.
  • 12
    Young NL, Williams JI, Yoshida KK, Wright JG. Measurement properties of the activities scale for kids. J Clin Epidemiol 2000; 53: 12537.
  • 13
    Daltroy LH, Liang MH, Fossel AH, Goldberg MJ. The POSNA pediatric musculoskeletal functional health questionnaire: report on reliability, validity, and sensitivity to change: Pediatric Outcomes Instrument Development Group, Pediatric Orthopaedic Society of North America. J Pediatr Orthop 1998; 18: 56171.
  • 14
    Reid GJ, Lang BA, McGrath PJ. Primary juvenile fibromyalgia: psychological adjustment, family functioning, coping, and functional disability. Arthritis Rheum 1997; 40: 75260.
  • 15
    Young NL, Yoshida KK, Williams JI, Bombardier C, Wright JG. The role of children in reporting their physical disability. Arch Phys Med Rehabil 1995; 76: 9138.
  • 16
    Chambers CT, Johnston C. Developmental differences in children's use of rating scales. J Pediatr Psychol 2002; 27: 2736.
  • 17
    Young NL, Williams JI, Yoshida KK, Bombardier C, Wright JG. The context of measuring disability: does it matter whether capability or performance is measured? J Clin Epidemiol 1996; 49: 1097101.
  • 18
    Singh G, Athreya BH, Fries JF, Goldsmith DP. Measurement of health status in children with juvenile rheumatoid arthritis. Arthritis Rheum 1994; 37: 17619.
  • 19
    Huber AM, Hicks JE, Lachenbruch PA, Perez MD, Zemel LS, Rennebohm RM, et al, Juvenile Dermatomyositis Disease Activity Collaborative Study Group. Validation of the Childhood Health Assessment Questionnaire in the juvenile idiopathic myopathies. J Rheumatol 2001; 28: 110611.
  • 20
    Flato B, Aasland A, Vandvik IH, Forre O. Outcome and predictive factors in children with chronic idiopathic musculoskeletal pain. Clin Exp Rheumatol 1997; 15: 56977.
  • 21
    Alman BA, Bhandari M, Wright JG. Function of dislocated hips in children with lower level spina bifida. J Bone Joint Surg Br 1996; 78: 2948.
  • 22
    Tennant A, Kearns S, Turner F, Wyatt S, Haigh R, Chamberlain MA. Measuring the function of children with juvenile arthritis. Rheumatology (Oxford) 2001; 40: 12748.
  • 23
    Ruperto N, Ravelli A, Pistorio A, Malattia C, Cavuto S, Gado-West L, et al. Cross-cultural adaptation and psychometric evaluation of the Childhood Health Assessment Questionnaire (CHAQ) and the Child Health Questionnaire (CHQ) in 32 countries: review of the general methodology. Clin Exp Rheumatol 2001; 19: S19.
  • 24
    McHorney CA. Generic health measurement: past accomplishments and a measurement paradigm for the 21st century. Ann Intern Med 1997; 127: 74350.
  • 25
    Ruperto N, Ravelli A, Migliavacca D, Viola S, Pistorio A, Duarte C, et al. Responsiveness of clinical measures in children with oligoarticular juvenile chronic arthritis. J Rheumatol 1999; 26: 182730.
  • 26
    Bode RK, Klein-Gitelman S, Miller ML, Lechman TS, Pachman LM. Disease activity score for children with juvenile dermatomyositis: reliability and validity evidence. Arthritis Rheum 2003; 49: 715.
  • 27
    Stucki G, Stucki S, Bruhlmann P, Michel BA. Ceiling effects of the Health Assessment Questionnaire and its modified version in some ambulatory rheumatoid arthritis patients. Ann Rheum Dis 1995; 54: 4615.
  • 28
    Liang MH. Evaluating measurement responsiveness. J Rheumatol 1995; 22: 11912.
  • 29
    Simeonsson RJ, Chen J, Hu Y. Functional assessment of Chinese children with the ABILITIES index. Disabil Rehabil 1995; 17: 40010.
  • 30
    World Health Organization. ICIDH-2: International Classification of Impairments, Activities and Participation: a manual of dimensions of disablement and functioning. Beta 1 draft for field trials. Geneva: World Health Organization; 1997. p. 122.
  • 31
    Dempster H, Porepa M, Young N, Feldman BM. The clinical meaning of functional outcome scores in children with juvenile arthritis. Arthritis Rheum 2001; 44: 176874.
  • 32
    Peterson RA. Constructing effective questionnaires. Thousand Oaks (CA): Sage Publications; 2000.
  • 33
    Feldman BM, Grundland B, McCullough L, Wright V. Distinction of quality of life, health related quality of life, and health status in children referred for rheumatologic care. J Rheumatol 2000; 27: 22633.
  • 34
    Whitney-Mahoney KJ, Young NL, Feldman BM, Badley EM. Alternative scoring methods improve the responsiveness of the Childhood Health Assessment Questionnaire (CHAQ) [abstract]. Arthritis Rheum 2001; 44: S381.
  • 35
    Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 15974.
  • 36
    Buchner A, Faul F, Erdfelder E. G•Power: a priori, post-hoc, and compromise power analyses for the Macintosh. Trier, Germany: University of Trier; 1997; version 2.1.2.