Researchers developed a machine learning classification model to identify restless legs syndrome in patients with end-stage renal disease who are on dialysis. The study included 396 individuals from a single cohort. The team compared a Support Vector Machine model against other common algorithms like random forests and decision trees to see which worked best for detecting the condition.
The Support Vector Machine model showed the best performance in testing. It achieved an area under the curve of 0.791, with an accuracy of 0.761. The analysis also found that beta-2 microglobulin levels and anemia were the most influential factors in the model's predictions.
The study notes that this tool might help screen for prevalent cases and inform clinical decisions. However, the researchers state that the model requires prospective external validation in multi-center cohorts before it can be used in clinical practice. Safety data were not reported in this study.
readers should understand that this is a development stage tool. It is not yet ready for routine use without further testing in larger, diverse groups of patients.