DIABETIC foot sensory assessments showed good overall reliability between trained examiners, although individual measurements and exact scores were less consistent, according to a study of the Rotterdam Diabetic Foot (RDF) Test Battery.
The findings suggest that the tool can provide reproducible overall measures of sensory loss under standardised conditions. However, the researchers caution that its routine clinical use and ulcer risk classification require further evaluation.
The study included 138 patients with symptoms of diabetic sensorimotor polyneuropathy, recruited through the ongoing DeCompression trial. Participants had Type 1 or Type 2 diabetes and varying degrees of sensory loss.
Researchers assessed the full 39-item RDF battery, including continuous measurements and normal-or-abnormal scoring, alongside a shorter 13-item version. The assessments cover several aspects of foot sensation, with information on balance, numbness and previous foot complications also incorporated.
Two investigators administered the full battery, while five investigators with differing experience assessed the shorter version. All received training, were blinded to each other’s results and performed paired assessments on the same day after a short interval.
For the versions using normal-or-abnormal scoring, overall reliability was good. Intraclass correlation coefficients were 0.83 for the 39-item battery and 0.80 for the 13-item battery.
Nevertheless, examiners obtained identical total scores in only 42.9% and 47.1% of assessments, respectively. These findings distinguish consistency in measuring differences between patients from exact agreement on an individual’s score.
Continuous measurements showed more variable performance. Exact agreement across individual items ranged from 33.9% to 100%, while item-level reliability coefficients were poor. Agreement improved when small scoring differences were allowed, indicating that many discrepancies were limited in size.
Classification using previously established ulcer risk thresholds was more consistent than exact scoring. Examiners assigned the same category in 96.4% of paired assessments using the 39-item battery and 87.1% using the shorter version.
Importantly, the study did not collect subsequent ulcer outcomes. Agreement on a risk category therefore does not demonstrate that the classification accurately predicts future ulcers in this population.
The authors also note that the shorter battery may be more practical, although assessment duration and implementation were not directly evaluated. Small changes in an individual patient’s measurements should be interpreted cautiously because observer variation could contribute.
Further work in routine care and independent populations is needed. For now, the results support the reproducibility of overall sensory staging after training, while highlighting the limits of individual measurements and the need to distinguish reliability from predictive performance.
Reference
Boers N et al. Interobserver reliability of the Rotterdam Diabetic Foot Study Test Battery for assessing cutaneous sensibility. Diabetes/Metabolism Research and Reviews. 2026;42:e70251. doi:10.1002/dmrr.70251.
Featured image: Smeilov on AdobeStock
Each article is made available under the terms of the Creative Commons Attribution-Non Commercial 4.0 License.
