From the authors:
L. van den Bemt and co-workers raise an important issue concerning the use of both self- and interviewer-administered versions for the validation of the Asthma Control Questionnaire (ACQ) in children and one we considered carefully when designing the study. We wanted to ensure that the ACQ could be used in clinical practice (individual children followed over time), paediatric research (6–16 yrs) and adult research (≥12 yrs). Therefore, two separate questionnaires was not an option. The development of the ACQ had ensured that it contained the questions that are important for assessing asthma control in children 6–16 yrs of age (content validity) but we knew the self-administered adult version could not be completed unaided by younger children. We considered validating an interviewer version for all ages (6–16 yrs) but realised this would be less practical for older children and also for adult clinical trials that enrol patients ≥12 yrs of age. The decision to develop an interviewer version for younger children (and for older children with inadequate literacy or numeracy skills) that could be combined with data from the self-administered version was based on strong evidence [1, 2]. Nevertheless, to minimise further any risk of bias, which might have affected the estimation of the measurement properties, children were consistent in the version they completed. L. van den Bemt and co-workers cited our validation of the Paediatric Asthma Quality of Life Questionnaire as evidence that within- and between-subject variance differs across age groups [3]. In that study, within-subject variance was very consistent across all age groups. Between-subject variance was consistent in the 7–10-yr and 15–17-yr groups but higher in the 11–14-yr group, suggesting that there was a wider range of impairment in the middle group. This does not mean that reliability, responsiveness or validity of the instrument were different in the different age groups; they were actually very similar.
Addressing the concern about ACQ interviewer instructions and parental help, we could have asked interviewers to read the self-administered version to younger children and given no guidance about parental involvement (which happens with many other questionnaires). This approach is open to errors. Although children as young as 6 yrs old understand the 7-point scale and can give very reliable responses [4, 5], they need the responses on a card and the concept explained in a standardised manner. For children <8 yrs of age, the concept of “during the last week” is also difficult and interviewers have to know how to check this. Interviewers first read each question to all children exactly as it is written in the original ACQ. However, we found during cognitive debriefing that a few of the youngest children did not understand some of the questions very accurately. We could have left it to interviewers to explain these questions. This is the usual approach but interviewers sometimes do not fully understand the concepts themselves and give an erroneous interpretation. To minimise this source of error, we give standardised alternative wording (interestingly, this novel approach has raised a new methodological challenge as the interviewer version is adapted for new languages: the questions that are difficult for young children to understand vary between languages). Ideally, one would like children to respond to all the questions but they sometimes need a bit of help. Although parents and primary caregivers often have a poor perception of their child’s asthma-related quality of life, their perception of their clinical asthma status is better [6]. We tried to use their input as little as possible but it was sometimes required for question 6 concerning the number of puffs of rescue bronchodilator used each day.
For patient-reported outcomes (PROs) there is rarely a “gold standard” with which to compare the new instrument for evidence of validity. Construct validity is the building up of evidence that the instrument is measuring what it purports to measure. In this study, we have provided the first evidence of the validity of the ACQ in children. Not only did the instrument give good measurement properties that were similar to those observed in adults (important for the use of the ACQ in adult (≥12 yrs of age) studies), the fact that these were observed in a relatively small sample size strengthens the evidence of validity rather than weakens it. Nevertheless, we agree that, like all other PROs, construct validity grows with additional studies and we would welcome further evaluation of this instrument. In addition, for adults, we have determined the cut-point on the ACQ score between adequate and inadequate asthma control, and this must now be established for children.
Footnotes
Statement of Interest
Statements of interest for E.F. Juniper and K. Gruffydd-Jones can be found at www.erj.ersjournals.com/site/misc/statements.xhtml
- ©ERS 2011