Retrospective judgements on health and quality of life in patients and the general population: Response shift and recall bias

点击翻译按钮获取中文摘要

Health-related quality of life (QoL) has become an important outcome criterion in clinical practice and research. Changes in QoL as evidence of therapeutic benefit are often calculated as the difference between a post-value and a pre-value. The problem that arises here is that the assessment criteria may have changed between pretest and posttest. One way to analyze and to correct this response shift effect is the thentest. The respondents are asked at the post-time to assess the situation retrospectively as it was at the pre-time. The difference between the original pretest and the thentest is then used to infer such response shift effects. However, it is also possible that such differences are effects of recall bias, i.e., that in the memory systematically higher or lower levels of QoL are stored. While various theoretically possible explanations of such effects are discussed in the literature, it remains unclear how these can be empirically separated and tested. One approach to analyze such effects is to ask patients about their assessment of QoL before the onset of the disease. Studies in the literature show that these retrospective judgements generally reflect a better QL than the current judgements of the general population. The relevant studies are mostly based on persons who have suffered an accident; corresponding studies on other patient collectives are sparse. In addition, it remains unclear whether this is a general better assessment of the past in the sense of recall bias or a shift in the assessment standard due to the disease. Furthermore, the findings on specific conditions for the occurrence of response shift effects are inconsistent. In the project applied for, a large sample of the general population (n > 2000) will be used to analyze the extent to which retrospective assessments of subjective health systematically exceed or fall short of the assessment made at that time. On the basis of four clinical samples (cancer patients and patients in cardiologic rehabilitation, n about 300 each), the retrospective QoL judgements (evaluation of the situation before the disease) are compared with the values given by the general population. It will be examined whether a systematic overestimation of the retrospectively assessed QoL can be proven, whether and to what extent it exceeds the retrospective effects of the general population, and which (sociodemographic and clinical) conditions are predictive factors for such effects. The empirical data sets are already available, they were collected in connection with three other projects, but have not yet been analyzed. Therefore, we do not apply for funds for data collection, but only for the analysis of the existing data and their publication.