Background: The screening process for systematic reviews is resource-intensive. Although previous machine learning solutions have reported reductions in workload, they risked excluding relevant papers. Objective: We evaluated the performance of a 3-layer screening method using GPT-3.5 and GPT-4 to streamline the title and abstract-screening process for systematic reviews. Our goal is to develop a screening method that maximizes sensitivity for identifying relevant records. Methods: We conducted screenings on 2 of our previous systematic reviews related to the treatment of bipolar disorder, with 1381 records from the first review and 3146 from the second. Screenings were conducted using GPT-3.5 (gpt-3.5-turbo-0125) and GPT-4 (gpt-4-0125-preview) across three layers: (1) research design, (2) target patients, and (3) interventions and controls. The 3-layer screening was conducted using prompts tailored to each study. During this process, information extraction according to each study’s inclusion criteria and optimization for screening were carried out using a GPT-4–based flow without manual adjustments. Records were evaluated at each layer, and those meeting the inclusion criteria at all layers were subsequently judged as included. Results: On each layer, both GPT-3.5 and GPT-4 were able to process about 110 records per minute, and the total time required for screening the first and second studies was approximately 1 hour and 2 hours, respectively. In the first study, the sensitivities/specificities of the GPT-3.5 and GPT-4 were 0.900/0.709 and 0.806/0.996, respectively. Both screenings by GPT-3.5 and GPT-4 judged all 6 records used for the meta-analysis as included. In the second study, the sensitivities/specificities of the GPT-3.5 and GPT-4 were 0.958/0.116 and 0.875/0.855, respectively. The sensitivities for the relevant records align with those of human evaluators: 0.867-1.000 for the first study and 0.776-0.979 for the second study. Both screenings by GPT-3.5 and GPT-4 judged all 9 records used for the meta-analysis as included. After accounting for justifiably excluded records by GPT-4, the sensitivities/specificities of the GPT-4 screening were 0.962/0.996 in the first study and 0.943/0.855 in the second study. Further investigation indicated that the cases incorrectly excluded by GPT-3.5 were due to a lack of domain knowledge, while the cases incorrectly excluded by GPT-4 were due to misinterpretations of the inclusion criteria. Conclusions: Our 3-layer screening method with GPT-4 demonstrated acceptable level of sensitivity and specificity that supports its practical application in systematic review screenings. Future research should aim to generalize this approach and explore its effectiveness in diverse settings, both medical and nonmedical, to fully establish its use and operational feasibility.

Long-term benzodiazepine (BZD) use is not recommended in the treatment of chronic insomnia, and psychosocial interventions, particularly cognitive behavioral therapy for insomnia (CBT-I), are a potential treatment option for discontinuing BZDs. This systematic review and meta-analysis aimed to clarify whether psychosocial interventions are effective for discontinuing BZD hypnotics in patients with chronic insomnia. A literature search of major electronic databases was conducted up to July 2018. Two researchers independently selected relevant publications, extracted data, and evaluated methodological quality according to the Cochrane criteria. Eight randomized-controlled trials, all of which evaluated CBT-I, were included in this review, and meta-analyses were performed. The results indicated that short-term (≤3 mo) CBT-I plus gradual tapering was more effective than gradual tapering alone for discontinuing BZDs hypnotics (risk ratio: 1.68, 95% confidence interval [CI]: 1.19-2.39, p = 0.003) and for improving insomnia symptoms (g: -0.69, 95% CI: -1.09 - -0.28, p = 0.0009). However, the long-term (12 mo) efficacy of CBT-I for discontinuing BZDs was not significant (risk ratio: 1.67, 95% CI: 0.91-3.07, p = 0.10). Thus, CBT-I is effective for discontinuing BZD hypnotics for ≤3 mo. Further studies are needed to clarify the long-term efficacy of psychosocial interventions for discontinuing BZD hypnotics.

Purpose: There is a lack of evidence regarding answers for clinical questions about treating insomnia disorder. This study aimed to answer the following clinical questions: (1) how to use each hypnotic and non-pharmacological treatment differently depending on clinical situations and (2) how to reduce or stop benzodiazepine hypnotics using alternative pharmacological and non-pharmacological treatments.Methods: Experts were asked to evaluate treatment choices based on 10 clinical questions about insomnia disorder using a nine-point Likert scale (1 = “disagree” to 9 = “agree”). The responses of 196 experts were collected, and the answers were categorized into first-, second-, and third-line recommendations.Results: The primary pharmacological treatment, lemborexant (7.3 ± 2.0), was categorized as a first-line recommendation for sleep initiation insomnia, and lemborexant (7.3 ± 1.8) and suvorexant (6.8 ± 1.8) were categorized as the first-line recommendations for sleep maintenance insomnia. Regarding non-pharmacological treatments for primary treatment, sleep hygiene education was categorized as the first-line recommendation for both sleep initiation (8.4 ± 1.1) and maintenance insomnia (8.1 ± 1.5), while multicomponent cognitive behavioral therapy for insomnia was categorized as the second-line treatment for both sleep initiation (5.6 ± 2.3) and maintenance insomnia (5.7 ± 2.4). When reducing or discontinuing benzodiazepine hypnotics by switching to other medications, lemborexant (7.5 ± 1.8) and suvorexant (6.9 ± 1.9) were categorized as first-line recommendations.Conclusion: Expert consensus indicates that orexin receptor antagonists and sleep hygiene education are recommended as first-line treatments in most clinical situations to treat insomnia disorder.

Aim: This systematic review and meta‐analysis evaluated whether bright light therapy (BLT) is an effective and safe treatment for manic/depressive symptoms and a preventive measure for recurrent mood episodes in patients with bipolar disorder.Methods: A literature search of major electronic databases was conducted in June 2019, including all published articles up to that date. Two researchers independently selected relevant publications, extracted data, and evaluated methodological quality according to the Cochrane criteria.Results: Six randomized controlled trials (RCT) evaluated the efficacy of BLT for bipolar depression. A meta‐analysis found no significant differences between BLT and placebo for the following outcomes: (i) rates of remission from depressive episodes (risk ratio [RR]: 1.81, 95% confidence interval [CI]: 0.43 to 7.64, P = 0.42); (ii) depressive symptom scores (standardized mean difference: −0.25, 95%CI: −0.74 to 0.23, P = 0.30); and (iii) rates of manic switching (RR: 1.00, 95%CI: 0.28 to 3.59, P = 0.26). The sensitivity analysis for studies with low overall indirectness did show a significant antidepressant effect for BLT (RR: 3.09, 95%CI: 1.62 to 5.90, P = 0.006). No RCT investigated the effect of BLT in preventing the recurrence of mood episodes in the euthymic state or in improving manic symptoms in the manic state. No severe adverse events were reported.Conclusion: While a meta‐analysis was unable to demonstrate the efficacy of BLT for bipolar depression, a sensitivity analysis did show a significant effect. Further well‐designed studies are needed to clarify the effectiveness of BLT, not only for the depressive state but also for other states, in the treatment of bipolar disorder.

Click on the article title to read more.