Abstract
INTRODUCTION: Insomnia and obstructive sleep apnea (OSA) are each linked to elevated risks of depression. When comorbid (COMISA), these risks increase further, highlighting the need for effective depression screening. This study evaluated the screening accuracy of a novel software, MEB-001, for detecting current major depressive episode (cMDE) in individuals identified as with and without suspected COMISA (sCOMISA). METHODS: We conducted a retrospective sub-analysis from a prospective multicenter study at U.S. sleep clinics, including 296 adults who underwent routine polysomnography (PSG). Electrocardiogram and electroencephalogram signals, along with items 1 and 2 of the self-report depression screener Patient Health Questionnaire, 9 items (PHQ-9), were used to generate MEB-001 screening output. The Mini International Neuropsychiatric Interview (MINI) served as the diagnostic reference for cMDE. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for MEB-001 and PHQ-9 (cut-off ≥10), with subgroup comparisons conducted using Fisher's exact and McNemar's tests (p < 0.05). RESULTS: MINI identified cMDE in 15.5% of participants (16.9% in sCOMISA; 14.2% in non-COMISA). Against the MINI, MEB-001 achieved 84.8% sensitivity, 72.0% specificity, 35.8% PPV, and 96.3% NPV in the full cohort; PHQ-9 ≥10 showed similar performance (89.1, 68.4, 34.2, and 97.2%, respectively). MEB-01's performance did not differ between sCOMISA and non-sCOMISA (all p ≥ 0.68), and no significant McNemar differences were found in the subgroups (p > 0.19). DISCUSSION: MEB-001 demonstrated consistent cMDE screening performance comparable to the PHQ-9, supporting its potential utility in sleep clinic settings.