The Statistical Analyses of Questionnaires: Validity, Reliability, Responsiveness, Range and Bias

摘要

Injuries and disease commonly affect the ability of an individual to perform activities of daily living. The use of regional outcome measures or scoring systems is important as it allows comparison between these injuries and disease, and allows clinicians to assess progression and the effects of different treatment modalities. A patient-completed questionnaire is efficient in terms of time and resources, and allows the assessment of outcome without the need to attend an outpatient clinic. It is important that the scoring system allows satisfactory regional outcome measurements specific for the hand. The validity, reliability, responsiveness and bias are used for the assessment of various questionnaires. The validity establishes whether the outcome measure actually measures what it was designed to. The criterion, construct, and content or face validity provide various assessment parameters that allow a complete picture of the validity of a questionnaire to be established. An outcome tool has test-retest reproducibility if the same result is obtained when tested at different time points once the condition has stabilised. The responsiveness is the ability to detect clinically important changes in disability at various intervals. The effect size and the standardised response mean are used to evaluate the responsiveness. Bias can occur when assumed independent variables such as age, gender, hand dominance or dominant side injured affect responses. There are also many practical issues concerning the use of questionnaires including feasibility of use for the patient and the clinician. This paper looks at the statistical analyses of questionnaires.