Abstract
Background: Semantic differential scales, in which respondents rate their relative agreement between two opposing statements, are rarely used in assessment of health related quality of life (QOL). Here we present a use-case study describing our experiences constructing such a scale, part of a larger project to generate a questionnaire, the VALVQ, to measure QOL in heart valve disease (HVD).
Methods and results: Individuals with HVD were identified from echocardiographic databases. Initial questionnaire content was generated from semi-structured interviews with 34 individuals with HVD as well as three family members and five clinical experts. Four methods were used to generate semantic differential items. Firstly, two participants could give opposing points, allowing an item to be generated directly. Secondly, a participant could report a single experience at one extreme, with the research team generating the opposing statement. Thirdly, a participant could report an experience in the middle of two extremes, with two opposing statement generated by the research team. Finally, subtly different items were generated where an obvious corresponding item was not clear, allowing for future psychometric testing to examine item robustness. 64 semantic differential items were generated for pilot testing. 46 valid completed questionnaires were returned. Multiple items required changes due to phrasing of the item content, but no difficulties were reported with the semantic differential format.
Conclusion: The semantic differential scale can be considered for health related QOL research. Further head-to-head comparisons with the more common Likert format will allow assessment of differences in respondent biases between these questionnaire formats.