Design, statistical and methodological considerations: Comorbidity

Research of disease comorbidity and symptom co-occurrence raises several issues relating to study design and analytical techniques that require careful consideration. In this chapter, we first address methodological issues that are of particular relevance in comorbidity research, including symptom overlap and the resultant double counting of symptoms; the pitfalls and advantages of removing overlapping scale items; and the utility of creating latent variables or ‘symptom groups’. We then discuss the advantages and limitations of employing various study designs in the context of comorbidity research and make recommendations for maximising the scientific rigour of statistical analyses whilst ensuring that ethical standards are met. Finally, we highlight analytical techniques that are relatively novel and/or less commonly utilised in studies of comorbidity, and how these techniques might advance research in this field.