OR WAIT null SECS
Investigators designed an algorithm with high diagnostic accuracy to reduce unnecessary polysomnography examinations and optimize the identification of participants with iRBD.
According to a recent study, an algorithm could reduce the amount of unnecessary polysomnography (PSG) examinations. The investigators designed an algorithm that demonstrated high diagnostic accuracy of PSG-proven isolated rapid eye movement (REM) sleep behavior disorder (iRBD).
This approach can also facilitate more efficient recruitment for research studies by avoiding unnecessary examinations. The study aimed to optimize the identification of participants with iRBD from the general population.
Aline Seger, MD, Department of Neurology, Faculty of Medicine and University Hospital Cologne, University of Cologne, and the team of investigators compared the performance of the RBD screening questionnaire (RBDSQ), which is commonly used, with sleep expert ratings. To improve classification accuracy, they selected the most specific items of the RBDSQ and included information from additional sleep-related questionnaires.
A total of 185 participants were screened, of whom 124 received PSG after expert selection, and 78 (62.9%) were diagnosed with iRBD. Findings revealed that the RBDSQ total score was a significant predictor of being chosen for PSG by the sleep expert. However, the RBDSQ total score had low specificity.
The RBDSQ total score showed a sensitivity of 95.2%, a specificity of 26.2%, an accuracy of 72.4%, and an area under the curve (AUC) of 0.68. Among participants receiving PSG, the RBDSQ total score showed a sensitivity of 96.2% and a specificity of 6.5% at a cutoff score of >5 points and an AUC of 0.64 (accuracy of 62.9%).
When comparing the algorithm to the sleep expert decision, 77 instead of 124 PSGs (62.1%) would have been carried out, and 63 (80.8%) iRBD patients would have been identified. 32 of 46 (69.6%) unnecessary PSG examinations could have been avoided.
In the stepwise multiple regression analysis of step 1 (selection of participants based on the expert's rating), the final model consisted of an RBDSQ subscore of items 6.1+6.2+6.3-10, the Pittsburgh Sleep Quality Index (PSQI) component score daytime dysfunction, and the STOP-Bang questionnaire.
The final step 1 model reached 81.1% classification accuracy (sensitivity, 91.9%; specificity, 59%; AUC, 0.84; P < 0.001).
In step 2 (identification of iRBD participants upon PSG), the final model consists of the RBDSQ subscore based on items 6.1+6.2+6.4-9 (notably, partly different items than in step 1), age, and the PSQI component score sleep disturbances. With this second step, classification accuracy reached 76.6% (sensitivity, 83.3%; specificity, 65.2%; AUC, 0.82; P < 0.001).
Overall, the investigators reported that the proposed algorithm displayed high diagnostic accuracy for PSG-proven iRBD cost-effectively and may be a convenient tool for research and clinical settings. The ROC curves for both step 1 and step 2 models are provided in the Supporting Information.
The AUCs of both final models significantly differed from the AUC using the RBDSQ total score alone, indicating a significantly better classification accuracy of the resulting step 1 and step 2 algorithms.
“This study presents a novel screening algorithm to optimize the identification of participants with iRBD from the general population,” investigators wrote. “Our approach does not require a selection step for PSG by a sleep expert. The algorithm allowed us to identify 80% of iRBD patients who were identified by a sleep expert at a 40% reduction in PSG numbers.”
Seger, A., Ophey, A., Heitzmann, W., Doppler, C.E.J., Lindner, M.-S., Brune, C., Kickartz, J., Dafsari, H.S., Oertel, W.H., Fink, G.R., Jost, S.T. and Sommerauer, M. (2023), Evaluation of a Structured Screening Assessment to Detect Isolated Rapid Eye Movement Sleep Behavior Disorder. Mov Disord. https://doi.org/10.1002/mds.29389