Background. There has been increasing interest in the development of m
easures to quantify baseline severity of illness and thus provide a mo
re meaningful interpretation of health outcomes. Objective. We aimed t
o assess the inter-rater reliability of a generic measure of illness s
everity, the Duke University Severity of Illness (DUSOI) checklist. Me
thods. Selected general practices and hospital outpatient departments
across Tyneside and Humberside in the UK were used as the setting. Thi
rty-three clinicians were posted copies of the same set of 14 patient
records and asked to rate the severity of illness of each patient usin
g the DUSOI. The subjects were 27 GPs, three consultant chest physicia
ns and three diabetologists. The main outcome measures were: (i) intra
class correlation coefficients used to express clinician agreement upo
n severity scores; and (ii) a 'gold standard', constructed in order to
assess clinician agreement upon the health problems identified in the
patient record to be considered as constituents of overall severity.
Results. It was found that the degree of inter-rater reliability of se
verity scoring was satisfactory, with an intraclass correlation coeffi
cient of 0.43. In terms of identification of problems to be rated, the
re was considerable agreement between raters and a specially compiled
'gold standard' of health problems. Conclusion. The DUSOI has potentia
l use in routine clinical practice, but strategies should be developed
in order to maximize the reliability of rating illness severity on su
ch a generic measure.