This study describes part of the development process of the seven DISABKIDS cross-national condition-specific modules. The DISABKIDS instrument for children and adolescents is the first to be developed cross-nationally in collaboration with several European countries and to include a chronic generic and condition-specific module.
The DISABKIDS instrument has several advantages. First the construction of the chronic generic and condition-specific modules allows for a comprehensive assessment of HRQoL. The chronic generic module can be used in conjunction with any of the condition-specific modules. Combining these modules gives the clinician and investigator the unique opportunity to compare between countries and between different conditions.
The second advantage is the simultaneous cross-national patient-derived development of the DISABKIDS instrument. Children and adolescents from each DISABKIDS country were included in the developmental process of the instrument. HRQoL statements were collected from the cross-national focus groups and interviews. Investigators from the DISABKIDS centres were involved in the item selection process, assuring that all items where relevant in each country. This was again tested in the cognitive interview in the pilot study. This simultaneous setup in different countries supported the developmental process by taking into account cross-national consensus on important HRQoL issues.
In addition, the construction of the DISABKIDS instrument has been a reflective one, combining subjective and statistical procedures. Item selection and reduction was not carried out solely through the use of statistical methods, but also through the inclusion of qualitative factors, such as the views of children and adolescents (gained from cognitive interview) and clinical judgement. The domain structure that resulted from the pilot test was to a great extent successfully replicated after the field test. The reliability of each domain was satisfactory in each condition-specific module.
However, some limitations should be given consideration. The number of respondents in some condition groups in both the pilot and the field test was relatively small, CP (n = 21 and 43) and atopic dermatitis groups (n = 29 and 65) in particular (Table and ). It was therefore not possible to solely use statistical methods to develop these modules. It is important to carry out further data collection and to test the reliability and validity in larger patient groups for these conditions. It will also be necessary to carry out large cross-national studies in the future in order to use modern psychometric methods based on Item Response Theory (IRT), which will permit the testing of differential item functioning across cultures and inform the degree to which cross-national comparisons can be validly made. The use of such IRT-based tests was not possible at this stage of the development of the measure because IRT methods require very large sample sizes.
A second limitation is that the condition-specific modules were not tested in every country. Only asthma was tested in all the participating DISABKIDS countries. The Cronbach's alphas were adequate for each asthma domain in each country. The lower alphas in Greece might not only be due to lower numbers of tested participants but also to the fact that the researched population included mostly exercise-induced asthma, which might result in a different impact on their HRQoL. As the number of participants in the other chronic conditions was generally low the reliability per country will still need to be explored in more detail.
Future studies will be necessary to provide more details on the reliability and validity of the DISABKIDS modules, especially in larger groups and in different countries. Evidence also needs to be supplied on the value of the instrument in clinical practice. Further possibilities include testing the chronic generic module for applicability in other chronic medical conditions (e.g. haemophilia, heart disease or obesity).
The developmental steps within the DISABKIDS project have included a combination of qualitative and quantitative methods. The two methods were used in succession in order to complement each other, as has been the case throughout the DISABKIDS project. The qualitative data (cognitive interview and clinical judgement) collected in the pilot study was first used to disregard irrelevant items. This was followed by the psychometric calculations. In some cases the project members found removed items to be clinically relevant. These were therefore added as the two extra items in the field study.
Although the process of item reduction for each of the condition-specific modules was similar and included well know procedures [28
], it remains difficult to describe the developmental process. As the value of each test depended on the size of the group and the number of items in the domain, and common sense judgements were also included, the taken steps may not always seem transparent. The number of countries included in the study meant that there were more national factors and individual opinions to include. Several processes within the DISABKIDS project (team meetings, group discussions) have influenced decisions. An example was the post-hoc decision to add extra items based on clinical relevance.