The key as to which questionnaire should be chosen should be governed by the QL hypothesis to be tested. Once the key QL aspects and timepoints have been fixed the questionnaire can be chosen. The difference between questionnaires often relates to the degree to which they emphasize objective compared with subjective dimensions . It is particularly important that the questionnaire addresses all the components that are important to the patient population and that are susceptible to being affected (either positively or negatively) by the intervention.
Most of the standard questionnaires cover much the same ground, and the choice may then be influenced by continuity (the trials group may have used the questionnaire in the previous trial and are comfortable with it), the need to make comparisons across trials, or the fact that one questionnaire may concentrate in greater detail on the symptoms or side-effects of interest.
Several papers have asked patients to complete more than one questionnaire and have then compared the results. The questionnaires compared in this way include the FACT-G and the EORTC QLQ-C30 , the FLIC and the EORTC QLQ-C30  and three lung-cancer-specific questionnaires, the EORTC, FACT, and LCSS .
Despite apparently addressing similar domains the studies revealed a number of differences. It was found that only one of the five subscales on the FACT-G was adequately covered by the QLQ-C30, and only three of the eight QLQ-C30 subscales were adequately covered by the FACT-G. The QLQ-C30 does not cover social well being and relationship with the doctor, whereas the FACT-G does not cover cognitive functioning and the LCSS captures the impact of toxicity rather than the detail of side-effects. Although there is general agreement in terms of the domains that questionnaires cover, the actual definitions used appear to differ. Thus social functioning is defined within the FLIC as an 'inclination for companionship with friends and family,' in the EORTC it is defined as 'the impact of disease on social activity and family life.'
Modification of these standard questionnaires is generally not advised and sometimes not permitted, as this would breach copyright. Indeed the questionnaires should be used in exactly the recommended format for the following good reasons:
♦ the psychometric performances of individual scales and items when used alone are not known,
♦ only a small proportion of patients find any questions upsetting,
♦ patients are generally not bothered by questions about symptoms they do not have.
Was this article helpful?