Article type
Abstract
Background: Stroke is the second leading cause of death and the third leading cause of disability worldwide, with acute ischemic stroke (AIS) accounting for 87% of all strokes. Randomized controlled trials (RCTs) are the gold standard for evaluating treatment efficacy. However, the reports of some RCTs of AIS are inconsistent with actual results, posing a risk of selective outcome reporting bias (ORB), which could distort the results of systematic reviews (SRs).
Objectives: The objective of this study is to investigate the occurrence of ORB in SRs of AIS, and use CORB method to correct the results of SRs for clinical practitioners and policy makers to comprehensively and correctly judge the therapeutic effect of AIS.
Methods: Cochrane Database of Systematic Reviews、Medline、Embase were used to identify SRs of RCT on AIS on June 12, 2023. Screening and data extraction were carried out back-to-back by four independent researchers. The Outcome Reporting Bias In Trials (ORBIT) method was employed to assess the likelihood of ORB for dichotomous outcomes in RCTs, and the correcting for outcome reporting bias (CORB) method was employed to correct the results of systematic reviews. Stata software was employed to data analysis.
Results: Among the 2,142 articles retrieved, 33 SRs were included in this study. These SRs included 97 dichotomous outcomes, with reported outcomes being more likely to have statistical significance than unreported outcomes (RR=2.01; 95% CI, 2.51-1.68). These 33 SRs contained 287 unique RCTs, nearly half of which (138/287, 48.08%) were at risk of selective reporting, and with 27.54% (38/138) were assessed as high risk. The CORB results indicated that among the 31 outcomes with ORB, 25.81% of the OR values were flipped, while among the 23 outcomes without flip, the OR values of 13 outcomes were out of the acceptable range, and the P-values of 51.61% outcomes were out of the acceptable range.
Conclusions: The results of this study suggest that ORB may have a severe impact on the field of AIS. It is recommended that registries, journals, peer reviewers, and institutions involved in RCT research prioritize the prevention of ORB.
Objectives: The objective of this study is to investigate the occurrence of ORB in SRs of AIS, and use CORB method to correct the results of SRs for clinical practitioners and policy makers to comprehensively and correctly judge the therapeutic effect of AIS.
Methods: Cochrane Database of Systematic Reviews、Medline、Embase were used to identify SRs of RCT on AIS on June 12, 2023. Screening and data extraction were carried out back-to-back by four independent researchers. The Outcome Reporting Bias In Trials (ORBIT) method was employed to assess the likelihood of ORB for dichotomous outcomes in RCTs, and the correcting for outcome reporting bias (CORB) method was employed to correct the results of systematic reviews. Stata software was employed to data analysis.
Results: Among the 2,142 articles retrieved, 33 SRs were included in this study. These SRs included 97 dichotomous outcomes, with reported outcomes being more likely to have statistical significance than unreported outcomes (RR=2.01; 95% CI, 2.51-1.68). These 33 SRs contained 287 unique RCTs, nearly half of which (138/287, 48.08%) were at risk of selective reporting, and with 27.54% (38/138) were assessed as high risk. The CORB results indicated that among the 31 outcomes with ORB, 25.81% of the OR values were flipped, while among the 23 outcomes without flip, the OR values of 13 outcomes were out of the acceptable range, and the P-values of 51.61% outcomes were out of the acceptable range.
Conclusions: The results of this study suggest that ORB may have a severe impact on the field of AIS. It is recommended that registries, journals, peer reviewers, and institutions involved in RCT research prioritize the prevention of ORB.