Improving the measurement of functional somatic symptoms with Item Response Theory

Abstract

More than 40 questionnaires are currently available to assess Functional Somatic Symptoms (FSS). Differences among them, as well as other methodological issues hamper the comparison of results between studies and the measurement of FSS. Sophisticated measurement techniques such as Item Response Theory (IRT) could help to deal with such issues. Aim: To determine which items of the somatization subscale of the SCL-90 are more informative and more able to discriminate between persons at different levels of severity of FSS. Methods: A sample of 82,740 adult participants, without somatic conditions, from the Lifelines Cohort Study was selected for analysis. The Graded Response Model (GRM) was applied to the twelve items of the somatization subscale of the SCL-90. Sensitivity analyses were performed using the data from all the participants who completed the somatization subscale in order to assess generalizability. Results: Mean threshold parameters were large ($\beta=2.28-5.10$) given that high levels of severity of FSS were necessary to report the highest answer options on the items. The items with the best ability to discriminate between levels of severity of FSS were item 11 ‘feeling weak physically’ ($\alpha=2.33$) and 12 ‘heavy feelings in arms or legs’ ($\alpha=2.50$). These also provided the most information. Sensitivity analyses provided similar results. Conclusions: The items ‘Feeling weak physically’, and ‘Heavy feelings in arms or legs’ are highly discriminative and informative symptoms to measure severity levels of FSS. Clinicians and researches may pay extra attention to these symptoms to augment the assessment of FSS.

Publication
Manuscript under review