End-digit preference of readings of grip strength measured with analogue-display devices.

Grip strength is frequently measured with analogue dynamometers graded to the nearest 2 kg. The end digits of such measurements should be even numbers or zero with frequency of 20% each. The accuracy of the grip strength measurements was never assessed for the possible end digit preference. We aimed to assess the end digit preference in the measurements of grip strength and the possible differences between examiners experienced and inexperienced in geriatrics. We did a retrospective analysis of grip strength data previously gathered. Data were recorded by two teams of examiners who performed grip strength measurements, using analogue devices, in older hospitalized patients. Team A consisted of geriatricians or physiotherapists working in the geriatric care setting. Team B consisted of physiotherapists without special experience in geriatrics who received a short, standardized instruction. We analysed a total of 251 (37.8% by team experienced in geriatrics) measurements. Grip strength was recorded with significant end digit preference leading to biased results. This was especially true of non-geriatric as opposed to geriatric trained examiners. Overall, values measured by Team A were 8.3 times more likely to end, correctly, on even digit or zero than end digits recorded by Team B (OR 8.3, 95%CI 2.9-25.0, p < 0.0001). We demonstrated a significant end digit preference, predominantly in examiners without experience in geriatrics. Standardized training and constant quality checks but also training in the importance of accuracy and the importance of muscular assessment in older patients are needed to ensure correct recording of grip strength values.