Although the APACHE II score is the most widely used scoring system in intensive care units worldwide, its reliability and variability have not been extensively studied. Differences in case-mix may complicate comparison and interpretation of results. We hypothesised that a degree of variability might be inherent to use of the APACHE II scoring system, and decided to assess intra-observer variability in APACHE II scoring as a potential indicator of inherent score variability. APACHE II scores were assessed twice from the charts of 11 patients by 14 physicians, with a time interval of 4 (range 3.5-4.5) months between the two assessments. Intra-observer was found to be approximately 15%. These findings are in agreement with previous observations regarding inter-observer variability in APACHE II scoring, and strongly suggest that there is an inherent score variability of about 15%.