0592 Sponsorship Bias and Methodological Quality in Randomized Controlled Trials About Digital CBT for Insomnia

Abstract Introduction Industry sponsorship bias refers to distortions that can occur at any point in a study to favor its sponsor. Digital cognitive behavioral therapy for insomnia (CBT-I) is a nonpharmacological treatment whose studies are often sponsored by the developer companies. The aim of this study was to assess whether sponsorship bias in digital CBT-I studies predicts more favorable outcomes on randomized controlled trials (RCTs). Methods A systematic review of PubMed and Web of Science databases was conducted to identify RCTs on digital CBT-I in adults. Information was collected on the type of publication, methodological quality, and involvement of the authors with the company developing the digital CBT-I studied. Industry sponsorship bias was measured on a 0 to 4 scale, based on the number of authors primarily affiliated to the companies developing the digital CBT-I interventions, or to mentions of sponsorship or funding received from these companies. The methodological quality of the RCTs was evaluated using the van Tulder scale Results Twenty-eight RCT analyses were included. The most frequent intervention was SHUTi (39.28%), followed by non-specified digital interventions for insomnia. The mean insomnia severity index (ISI) scores were 15.04±4.89 in the control group, 10.75±5.06 in the experimental group, and the mean effect size was 1.06±1.66. There was no correlation between the sponsorship bias level and the ISI effect size (ρ=1.000; p=0.198). The ISI effect size was equivalent among studies with and without sponsorship bias (p=0.405). The presence of sponsorship bias was associated with both open access publication (p=0.022) and lower methodological quality (p=0.027). Conclusion Sponsorship bias did not favor the results of digital CBT-I interventions, as no correlation was found between industry sponsorship bias and the ISI score. However, sponsorship bias was significantly associated with both open access publication and with lower methodological quality. Support (if any) This work was supported by the Associação Fundo de Incentivo à Pesquisa (AFIP), São Paulo, Brazil. IPAL and VAK receive scholarship grants from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). ST is a grant recipient from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).