Despite extensive searches we identified only 18 published reports of paediatric trials including standardised QOL measures. This undoubtedly represents a very small percentage of paediatric trials and supports previous findings that QOL data is seldom reported in paediatric clinical trials . Asthma and rhinitis were most frequently studied, perhaps because there is higher incidence for these conditions in children compared to other conditions such as cancer and cystic fibrosis . Further explanations include the non-life threatening nature of these conditions as well as the availability of disease specific measures compared to rarer illnesses.
In considering why there are relatively few trials including QOL measures, it is important to take into account the aims and purpose of the trial [40, 41]. The aim of most trials is to assess the impact of treatment on clinical variables, with QOL viewed to be of secondary importance if at all. It is not necessarily appropriate that QOL measures are included in all trials. Where QOL assessment is appropriate however, inclusion of a QOL measure must be hypothesis driven and an integral part of the clinical development programme rather than an added afterthought .
Quality of measures
Where QOL was measured, disease specific measures were most often used (N = 12) as is normally recommended for use in clinical trials. Only two trials included measures that satisfactorily fulfilled accepted criteria . Typically, information about measures included some reliability data although a third of studies failed to provide information about the validity of the scale. Most measures were brief and contained less than 30 items but many lacked age appropriate versions or parallel versions for child and proxy raters.
Selection of a measure of QOL is dependent on the psychometric properties of the instrument, as well as clinical and demographic variables characteristic of the sample. However, psychometric properties depend upon samples for which the scale has been validated. Hence it is important to ensure measures are used with clinical populations where psychometric data are available.
There are some grounds for assuming that QOL changes during childhood, and therefore satisfactory measures target specific age groups . There are difficulties identifying single measures that are appropriate across a wide age range and only half of measures identified in this review included age appropriate versions.
It is also generally recommended that ratings of QOL should be made by children themselves whenever possible . In cases of younger children proxy reports are necessary but there are questions about the relationship between child and parent report . It is therefore positive that most (73.3%) studies obtained ratings from children with only four relying on parents alone to provide proxy ratings.
CONSORT  guidelines recommend methods of reporting RCTs, but do not adequately deal with the issues concerning QOL assessment and psychometric validity. It is essential that trial developers select appropriate measures and are aware of the problems associated with QOL assessment.
Barriers to inclusion of QOL measures
Objections to inclusion of QOL measures in trials involve anticipated increased costs, extra time needed to gain patient and parent consent, and lack of sophistication of currently available measures . A major restriction to inclusion of QOL assessment in clinical trials remains limitations in currently available measures, especially for less prevalent chronic conditions. However, it is only through including measures that we will learn more and be able to develop a second generation of measures that do show more sophisticated properties.
A second problem is that disease specific measures may simply not be available for rare conditions. Attempts to develop such measures are promising and in this review instruments for ambylopia  and agranulocytosis  had been developed. In order to facilitate collection of QOL data from children with chronic illness, reliable and valid measures are increasingly required .
Other methodological limitations in current work include the lack of power calculations. Where the aim of the trial includes QOL assessment, power calculations must be performed and are an essential element of clinical trial design. In cases where measurement of QOL is a secondary endpoint, sample size calculations are rare and difficult to establish. However attempts should be made to hypothesise expected changes in QOL scores in relation to the agreed sample size prior to the trial .