Agreement of clinical measurements of liver size with ultrasound when performed by medical students.

AIM

The accuracy of physical examination techniques in detecting liver disease is unclear. We sought to determine the perceived location of the lower liver border via scratch, percussion, palpation and ballottement with novice medical student examiners, compared with ultrasound localisation.

METHOD

Five novice medical students learnt four liver examination techniques and measured the lower liver border in 19 healthy volunteers. The difference between the examination technique and ultrasound (bias) and limits of agreement of each method were estimated by mixed linear models and shown using Bland-Altman- like plots.

RESULTS

All techniques had similar bias for the lower liver border, between 1.6 and 1.9 cm superior to the ultrasound measurement. Percussion had the smallest variability with increasing liver size. Limits of agreement were wide for all techniques, smallest for palpation (6.04 cm) and largest for scratch (7.2 cm) compared to the mean liver distance of 8 cm. Conclusion There was no difference in bias between the scratch, percussion, palpation and ballottement techniques regarding the lower liver border. All techniques had very wide limits of agreement, although palpation had the smallest. Liver size examination by novice medical students in healthy subjects is unreliable. Further research is needed using examiners with a different level of expertise and participants of varying body habitus and confirmed liver disease.