BackgroundDespite strong evidence for cognitive behavioral therapy for insomnia (CBT-I), uptake remains constrained by poorly understood cognitive, attitudinal, and practical barriers. This study examined determinants of willingness to enroll in sleep improvement programs among adults at risk of sleep disorders—including insomnia, obstructive sleep apnea, and comorbid psychological distress—attending a tertiary sleep clinic in China.MethodsA cross-sectional knowledge-attitudes-practices survey was conducted among 2,661 adults attending the sleep and behavioral medicine outpatient clinic at Ganzhou Hospital-Nanfang Hospital, Southern Medical University, Ganzhou, Jiangxi, China, between February 2022 and June 2025. Willingness to enroll in a structured sleep improvement program was assessed alongside sleep health knowledge, perceived need, CBT-I versus medication effectiveness beliefs, telehealth acceptability, clinical severity (Insomnia Severity Index, Epworth Sleepiness Scale, STOP-Bang), psychological symptoms (PHQ-2, GAD-2), perceived barriers, and sociodemographic characteristics. Multivariable logistic regression identified independent predictors of willingness to enrollment, with secondary analyses evaluating model discrimination and testing prespecified interactions.ResultsAmong 2,661 participants (median age 45 years, 56.5% female, median ISI = 13), 1,386 (52.1%) expressed willingness to enroll. Univariable comparisons showed no significant differences between willing and not-willing groups across demographics, clinical characteristics, or barriers (all p>0.05). However, multivariable modeling revealed that when considered simultaneously, cognitive-attitudinal factors emerged as significant independent predictors, suggesting complex interactions rather than simple bivariate associations. In multivariable models, perceived need (OR = 1.20, 95% CI: 1.16–1.25, p<0.001), beliefs that CBT-I is more effective and durable than sleep medication (OR = 1.12, 95% CI: 1.08–1.16, p<0.001), sleep health and treatment knowledge (assessed by a six-item knowledge score) (OR = 1.09, 95% CI: 1.05–1.13, p<0.001), and anxiety symptoms (OR = 1.07, p=0.005) positively predicted willingness. Paradoxically, depression symptoms (OR = 0.94, p<0.001) and insomnia severity (OR = 0.93, p<0.001) inversely predicted willingness. Model discrimination was modest (AUC = 0.543, 95% CI: 0.504–0.590). Time (mean 3.54) and cost (3.44) were most severe barriers but showed no independent association with willingness (p>0.05).ConclusionCognitive-attitudinal factors (perceived need, CBT-I beliefs, knowledge) independently predicted enrollment willingness, whereas demographics and practical barriers did not. Depression and insomnia severity paradoxically reduced willingness, creating an inverse care law. However, poor model discrimination and measurement of stated willingness rather than actual enrollment limit conclusions. Prospective validation and motivational enhancement strategies for patients with depression are needed.
Radically different