BackgroundSerious mental illness (SMI) is associated with elevated mortality compared to the general population; the majority of this excess is attributable to co-occurring common physical health conditions. There may be variation within the SMI group in the distribution of physical co/multi-morbidity. This study aims to a) compare the pattern of physical co- and multi-morbidity between patients with and without SMI within a South London primary care population; and, b) to explore socio-demographic and health risk factors associated with excess physical morbidity among the SMI group.MethodsData were obtained from Lambeth DataNet, a database of electronic patient records derived from general practices in the London borough of Lambeth. The pattern of 12 co-morbid common physical conditions was compared by SMI status. Multivariate ordinal and logistic regression analyses were conducted to assess the strength of association between each condition and SMI status; adjustments were made for potentially confounding socio-demographic characteristics and for potentially mediating health risk factors.ResultsWhile SMI patients were more frequently recorded with all 12 physical conditions than non-SMI patients, the pattern of co-/multi-morbidity was similar between the two groups. Adjustment for socio-demographic characteristics – in particular age and, to a lesser extent ethnicity, considerably reduced effect sizes and accounted for some of the associations, though several conditions remained strongly associated with SMI status. Evidence for mediation by health risk factors, in particular BMI, was supported.ConclusionsSMI patients are at an elevated risk of a range of physical health conditions than non-SMI patients but they do not appear to experience a different pattern of co-/multimorbidity among those conditions considered. Socio-demographic differences between the two groups account for some of the excess in morbidity and known health risk factors are likely to mediate the association. Further work to examine a wider range of conditions and health risk factors would help determine the extent of excess mortality attributable to these factors.