Importance Large language models (LLMs) may facilitate the labor-intensive process of systematic reviews. However, the exact methods and reliability remain uncertain. Objective To explore the feasibility and reliability of using LLMs to assess risk of bias (ROB) in randomized clinical trials (RCTs). Design, Setting, and Participants A survey study was conducted between August 10, 2023, and October 30, 2023. Thirty RCTs were selected from published systematic reviews. Main Outcomes and Measures A structured prompt was developed to guide ChatGPT (LLM 1) and Claude (LLM 2) in assessing the ROB in these RCTs using a modified version of the Cochrane ROB tool developed by the CLARITY group at McMaster University. Each RCT was assessed twice by both models, and the results were documented. The results were compared with an assessment by 3 experts, which was considered a criterion standard. Correct assessment rates, sensitivity, specificity, and F1 scores were calculated to reflect accuracy, both overall and for each domain of the Cochrane ROB tool; consistent assessment rates and Cohen kappa were calculated to gauge consistency; and assessment time was calculated to measure efficiency. Performance between the 2 models was compared using risk differences. Results Both models demonstrated high correct assessment rates. LLM 1 reached a mean correct assessment rate of 84.5% (95% CI, 81.5%-87.3%), and LLM 2 reached a significantly higher rate of 89.5% (95% CI, 87.0%-91.8%). The risk difference between the 2 models was 0.05 (95% CI, 0.01-0.09). In most domains, domain-specific correct rates were around 80% to 90%; however, sensitivity below 0.80 was observed in domains 1 (random sequence generation), 2 (allocation concealment), and 6 (other concerns). Domains 4 (missing outcome data), 5 (selective outcome reporting), and 6 had F1 scores below 0.50. The consistent rates between the 2 assessments were 84.0% for LLM 1 and 87.3% for LLM 2. LLM 1's kappa exceeded 0.80 in 7 and LLM 2's in 8 domains. The mean (SD) time needed for assessment was 77 (16) seconds for LLM 1 and 53 (12) seconds for LLM 2. Conclusions In this survey study of applying LLMs for ROB assessment, LLM 1 and LLM 2 demonstrated substantial accuracy and consistency in evaluating RCTs, suggesting their potential as supportive tools in systematic review processes.

Objective: Non-invasive brain stimulation (NIBS) is beneficial to adult patients with depression, but its safety and efficacy in combination with antidepressants in children and adolescents with depression are not clear. We conducted a preliminary meta-analysis to objectively evaluate its clinical effect and provide information for future research and clinical practice. Methods: PubMed, Cochrane Library, Embase, and Web of Science were searched systematically to find clinical trials published in English before April 11, 2023. Stata software was used for meta-analysis, and random or fixed effect models were used to combine effect sizes. Results: Nine studies were eligible and included (n = 393). No articles about children were included in the analysis. The results showed that the remission rate was 40% (95% confidence interval [CI]: 13% to 71%). The scores of Children's Depression Rating Scale (CRDS) and Hamilton's depression scale (HAMD) significantly decreased compared to baseline value (MD = -27.04, 95% CI: -30.95, -23.12 and MD = -12.78, 95% CI: -19.55 to -6.01). In addition, the incidence of all adverse events was 13% (95% CI: 5%, 23%), and all were minor pain-related events. Conclusion: The combination of NIBS and antidepressants has been shown to notably alleviate depressive symptoms in adolescents, offering a considerable level of safety. This therapeutic synergy is particularly effective in patients with major depressive disorder, where repetitive transcranial magnetic stimulation augmented with antidepressants can enhance the amelioration of depressive symptoms. Systematic review registration: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023442215, PROSPERO CRD42023442215.

ObjectiveTo compare the outcomes associated with the use of lasmiditan, rimegepant, ubrogepant, and zavegepant for the acute management of migraine headaches.MethodsWe searched four electronic databases from database inception to August 31, 2023, to identify randomized controlled trials (RCTs) that report efficacy and safety for the acute treatment of migraine. The risk of bias in the included RCTs was evaluated according to the Cochrane tool, and the certainty of evidence using the CINeMA approach. We conducted frequentist network meta-analyses (NMA) to summarise the evidence. Data were analyzed using R-4.3.1.ResultsA total of 18 eligible studies including 10 different types of interventions with 22,429 migraine patients were included. NMA results showed that compared to ubrogepant (25 mg and 50 mg) and zavegepant, lasmiditan (100 mg and 200 mg) exhibits an elevated probability of achieving pain relief within a 2-hour interval. Similarly, relative to zavegepant, rimegepant (75 mg) and ubrogepant (50 mg and 100 mg) demonstrate an enhanced likelihood of sustaining pain relief over a 24-hour period. Furthermore, in contrast to ubrogepant (25 mg) and lasmiditan (50 mg), rimegepant (75 mg) presents a heightened probability of achieving freedom from photophobia within 2 h. Regarding safety, lasmiditan carries the highest risk of adverse events, which are associated with an increased incidence of adverse effects, including dizziness, somnolence, asthenia, paresthesia, and fatigue.ConclusionsIn this NMA, a spectrum of evidence ranging from very low to high levels underscores the favorable efficacy and tolerability of rimegepant 75 mg and ubrogepant 100 mg, positioning them as potential candidates for the acute management of migraine. Concurrently, lasmiditan (100 mg and 200 mg) exhibits notable efficacy, albeit accompanied by an increased susceptibility to adverse events. These findings should still be approached with caution, primarily due to the intrinsic limitations associated with indirect comparisons.

Background: The clinical efficacy and safety of intravenous immunoglobulin (IVIg) treatment for COVID-19 remain controversial. This study aimed to map the current status and gaps of available evidence, and conduct a meta-analysis to further investigate the benefit of IVIg in COVID-19 patients. Methods: Electronic databases were searched for systematic reviews/meta-analyses (SR/MAs), primary studies with control groups, reporting on the use of IVIg in patients with COVID-19. A random-effects meta-analysis with subgroup analyses regarding study design and patient disease severity was performed. Our outcomes of interest determined by the evidence mapping, were mortality, length of hospitalization (days), length of intensive care unit (ICU) stay (days), number of patients requiring mechanical ventilation, and adverse events. Results: We included 34 studies (12 SR/MAs, 8 prospective and 14 retrospective studies). A total of 5571 hospitalized patients were involved in 22 primary studies. Random-effects meta-analyses of very low to moderate evidence showed that there was little or no difference between IVIg and standard care or placebo in reducing mortality (relative risk [RR] 0.91; 95% CI 0.78-1.06; risk difference [RD] 3.3% fewer), length of hospital (mean difference [MD] 0.37; 95% CI - 2.56, 3.31) and ICU (MD 0.36; 95% CI - 0.81, 1.53) stays, mechanical ventilation use (RR 0.92; 95% CI 0.68-1.24; RD 2.8% fewer), and adverse events (RR 0.98; 95% CI 0.84-1.14; RD 0.5% fewer) of patients with COVID-19. Sensitivity analysis using a fixed-effects model indicated that IVIg may reduce mortality (RR 0.76; 95% CI 0.60-0.97), and increase length of hospital stay (MD 0.68; 95% CI 0.09-1.28). Conclusion: Very low to moderate certainty of evidence indicated IVIg may not improve the clinical outcomes of hospitalized patients with COVID-19. Given the discrepancy between the random- and fixed-effects model results, further large-scale and well-designed RCTs are warranted

Background: Based on randomized controlled trials, a network meta-analysis was conducted to compare treatment effects across varenicline and related smoking interventions. Methods: English databases were screened for randomized controlled trials reporting the effect of varenicline as treatment for smoking. The risk of bias in included trials was assessed using the Cochrane Handbook tool. Stata 15.1 software was used to perform network meta-analysis, and the GRADE approach was used to assess the evidence credibility on the tobacco treatment effects of different interventions. Results: Thirty-four studies involving 26,130 smokers were included in the network meta-analysis. Varenicline and 11 other interventions were reported, yielding 66 pairs of comparisons. Network meta-analysis showed that varenicline monotherapy or its combination with other interventions were superior in achieving smoking cessation compared to bupropion, nicotine replacement therapy, counselling, and placebo. Furthermore, compared to the varenicline, evident abstinence superiority was found in varenicline + bupropion (odds ratio = 1.49, 95% confidence interval [1.02, 2.18]). Finally, the surface under the cumulative ranking curve value indicated that varenicline + bupropion has the highest probability to become the best intervention. Conclusions: Varenicline monotherapy increased the odds of smoking cessation further than bupropion monotherapy, nicotine replacement therapy, counselling, and placebo, while varenicline combined with other interventions may even achieve a better abstinence effect. More credible evidence has been reported indicating that the combination of varenicline and bupropion is a superior treatment for smoking.

Objective: The aim of this study was to examine the most effective delivery format of cognitive behavioral therapy for insomnia (CBT-I) on insomnia in cancer patients. Methods: We searched five databases up to February 2021 for randomized clinical trials that compared CBT-I with inactive or active controls for insomnia in cancer patients. Outcomes were insomnia severity, sleep efficiency, sleep onset latency (SOL), wake after sleep onset (WASO), and total sleep time (TST). Pairwise meta-analyses and frequentist network meta-analyses with the random-effects model were applied for data analyses. Results: Sixteen unique trials including 1523 participants met inclusion criteria. Compared with inactive control, CBT-I could significantly reduce insomnia severity (mean differences [MD] = -4.98 points, 95% confidence interval [CI]: -5.82 to -4.14), SOL (MD = -12.29 min, 95%CI: -16.48 to -8.09), and WASO (MD = -16.58 min, 95%CI: -22.00 to -11.15), while increasing sleep efficiency (MD = 7.62%, 95%CI: 5.82% to 9.41%) at postintervention. Compared with active control, CBT-I could significantly reduce insomnia severity (MD = -2.75 points, 95%CI: -4.28 to -1.21), SOL (MD = -13.56 min, 95%CI: -18.93 to -8.18), and WASO (MD = -6.99 min, 95%CI: -11.65 to -2.32) at postintervention. These effects diminished in short-term follow-up and almost disappeared in long-term follow-up. Most of the results were rated as "moderate" to "low" certainty of evidence. Network meta-analysis showed that group CBT-I had an increase in sleep efficiency of 10.61%, an increase in TST of 21.98 min, a reduction in SOL of 14.65 min, and a reduction in WASO of 24.30 min, compared with inactive control at postintervention, with effects sustained at short-term follow-up. Conclusions: CBT-I is effective for the management of insomnia in cancer patients postintervention, with diminished effects in short-term follow-up. Group CBT-I is the preferred choice based on postintervention and short-term effects. The low quality of evidence and limited sample size demonstrate the need for robust evidence from high-quality, large-scale trials providing long-term follow-up data.

This review compared the efficacy and acceptability of different delivery formats for cognitive behavioral therapy for insomnia (CBT-I) in insomnia. We searched five databases for randomized clinical trials that compared one CBT-I delivery format against another format or control conditions for insomnia in adults. We used pairwise meta-analyses and frequentist network meta-analyses with the random-effects model to synthesize data. A total of 61 unique trials including 11,571 participants compared six CBT-I delivery formats with four control conditions. At post-intervention, with low to high certainty evidence, individual, group, guided self-help, digital assisted, and unguided self-help CBT-I could significantly increase sleep efficiency and total sleep time (TST) and reduce sleep onset latency (SOL), wake after sleep onset (WASO), and insomnia severity compared with treatment as usual (MD range for sleep efficiency: 7.81%-12.45%; MD range for TST: 16.14-33.96 min; MD range for SOL:-22.42 to-13.81 min; MD range for WASO:-40.84 to-19.48 min; MD range for insomnia severity:-6.40 to-3.93) and waitlist (MD range for sleep efficiency: 7.68%-12.32%; MD range for TST: 12.67-30.49 min; MD range for SOL:-19.07 to-10.46 min; MD range for WASO:-47.10 to-19.15 min; MD range for insomnia severity:-7.59 to-5.07). The effects of different CBT-I formats per-sisted at short-term follow-up (4 wk-6 mo). Individual, group, and digital assisted CBT-I delivery formats would be the more appropriate choices for insomnia in adults, based on post-intervention and short-term effects. Further trials are needed to investigate the long-term effects of different CBT-I formats. (c) 2022 Elsevier Ltd. All rights reserved.

Objective: A network meta-analysis (NMA) was conducted to investigate the effect of varenicline (VAR), bupropion (BUP), and nicotine replacement therapy (NRT) on smoking cessation. Methods: Eight databases were searched in May 2021, and only randomized controlled trials (RCTs) using varenicline, bupropion, or NRT (single or combined) for smoking cessation were included. The risk of bias in the included RCTs was assessed using the Cochrane Handbook tool. Stata 15.1 software was used to perform NMA, and the quality of the evidence was evaluated using Confidence in Network Meta-analysis (CINeMA). Findings: Twenty RCTs involving 16,702 smokers were included. The risk of bias results showed that 10 RCTs were rated as high, three were low, and seven were unclear. A total of 21 pairs were compared based on seven interventions. The NMA showed that, compared to the placebo (PLA), the other six interventions had significant efficacy in smoking cessation, where VAR + BUP showed the best effect of all treatments (odds ratio (OR) = 6.08, 95% confidence interval (CI) [3.47, 10.66]). Moreover, VAR + BUP was superior to VAR + NRT (OR = 1.66, 95% CI [1.07, 2.59]) and the three monotherapies (VAR, BUP, and NRT). In the monotherapies, the results of pairwise comparisons of VAR, BUP, and NRT did not show significant differences. Finally, the surface under the cumulative ranking curve (SUCRA) value indicated that VAR + BUP had the greatest probability of becoming the best intervention. Conclusions: The efficacy of VAR, BUP, and NRT alone increased the odds of smoking abstinence better than the placebo, combined interventions were superior to monotherapy, and VAR combined with other interventions had a better smoking cessation effect.

Background: Many meta-analyses (MAs) on Chinese medicine (CM) as an adjunctive treatment for gastric cancer have been published in recent years. However, the pooled evidence reported in MAs and their methodological quality remain unknown. Therefore, we designed a study to comprehensively evaluate and summarize the current evidence of CMs for gastric cancer in published MAs. Methods: A systematic search on MAs published in English from inception to 1st September 2021 was conducted in PubMed and Embase. The AMSTAR-2 tool was used to evaluate the methodological quality of the included MAs, and the results of the quality assessment were visualized using the evidence mapping method. Stata 17/SE was used for statistical analysis (Registration number: INPLASY202190005). Results: A total of 20 MAs (16 pairwise and 4 network MAs) were included from 118 records. These MAs were published in 14 journals from 2013 to 2021, with the number of patients and trials ranging from 688 to 6,857, and from 10 to 85, respectively. A large number of CMs (e.g., AiDi, FuFangKuShen, and HuaChanSu) in combination with chemotherapy for gastric cancer were identified among the included MAs. According to the pooled results reported in MAs, when compared to chemotherapy alone, CMs in combination with chemotherapy not only improve various outcomes on efficacy (e.g., objective response rate, quality of life) but also reduce various adverse reactions (e.g., leucopenia, nausea and vomiting). Only 2 MAs were low in terms of the overall methodological quality, while the other 18 MAs were all critically low. The methodology was required to be advanced significantly, mainly involving: study protocol and registration, explanation for the inclusion of study design, list of excluded studies with justifications, adequate details of included studies, reporting on funding sources of primary studies, and evaluation of the potential impact of risk of bias. In addition, MAs that received funds support (β = 2.68; 95%CI: 0.40 to 4.96; p = 0.024) or were published in journals with higher impact factor (β = 2.81; 95%CI: 0.69 to 4.92; p = 0.012) had a higher score on the overall methodological quality in the univariate analysis, but the results were not statistically significant according to the multivariate analysis. Conclusion: Combining CMs with chemotherapy can potentially improve clinical outcomes and reduce the relevant adverse effects in patients with gastric cancer. However, the methodological quality of relevant MAs requires significant improvement, and the current evidence needs to be validated through multinational trials that are well-designed and have a large sample size.

Background & Carbon ion radiotherapy (CIRT) and proton beam therapy (PBT) are promising methods for prostate cancer, however, the consensus of an increasing number of studies has not been reached. We aimed to provide systematic evidence for evaluating the efficacy and safety of CIRT and PBT for prostate cancer by comparing photon radiotherapy. Materials and Methods: We searched for studies focusing on CIRT and PBT for prostate cancer in four online databases until July 2021. Two independent reviewers assessed the quality of included studies and used the GRADE approach to rate the quality of evidence. R 4.0.2 software was used to conduct the meta-analysis. A meta-regression test was performed based on the study design and tumor stage of each study. Results: A total of 33 studies including 13 CIRT- and 20 PBT-related publications, involving 54,101, participants were included. The quality of the included studies was found to be either low or moderate quality. Random model single-arm meta-analysis showed that both the CIRT and PBT have favorable efficacy and safety, with similar 5-year overall survival (OS) (94 vs 92%), the incidence of grade 2 or greater acute genitourinary (AGU) toxicity (5 vs 13%), late genitourinary (LGU) toxicity (4 vs 5%), acute gastrointestinal (AGI) toxicity (1 vs 1%), and late gastrointestinal (LGI) toxicity (2 vs 4%). However, compared with CIRT and PBT, photon radiotherapy was associated with lower 5-year OS (72-73%) and a higher incidence of grade 2 or greater AGU (28-29%), LGU (13-14%), AGI (14-19%), and LGI toxicity (8-10%). The meta-analysis showed the 3-, 4-, and 5-year local control rate (LCR) of CIRT for prostate cancer was 98, 97, and 99%; the 3-, 4-, 5-, and 8-year biochemical relapse-free rate (BRF) was 92, 91, 89, and 79%. GRADE assessment results indicated that the certainty of the evidence was very low. Meta-regression results did not show a significant relationship based on the variables studied (P < 0.05). Conclusions: Currently available evidence demonstrated that the efficacy and safety of CIRT and PBT for prostate cancer were similar, and they may significantly improve the OS, LCR, and reduce the incidence of GU and GI toxicity compared with photon radiotherapy. However, the quantity and quality of the available evidence are insufficient. More high-quality controlled studies are needed in the future.