Indirect comparison refers to a comparison of different healthcare interventions using data from separate studies, and is often used because of a lack of, or insufficient evidence from head-to-head comparative trials. We aimed to summarize the definition, fundamental theory, type, relevant statistical contents, and to clarify some question on how to use indirect comparison, in order to attract more researchers' attention and promote methodological development of indirect comparison.
Due to the lack of head to head direct comparison evidence, applying indirect comparison (ITC) as well as network meta-analysis to compare multiple interventions becomes a new popular and powerful statistical technique. However, its theoretical system still needs improvement. In this article, we briefly introduce and summarize its progress concerning basic concepts, method assumptions, influencing factors of effectiveness, and software for analysis, so as to help researchers better understand the method and promote its application in evidence production.
ObjectiveTo investigate the status of research and development methods of Cochrane overviews. MethodsThe Cochrane Library and PubMed were searched up to March 2014 to identify Cochrane overviews. According to the inclusion and exclusion criteria, two reviewers independently screened literature, extracted data, and assessed and analyzed search strategy, quality assessment method, data analysis, and study results. ResultsA total of 18 Cochrane overviews were included. Among them, 4 (22.2%) overviews included formal statistical indirect comparison; 8 (44.4%) included only results from direct comparison; 6 (33.4%) only systematically analyzed current studies without data pooling; 12 (66.7%) only searched The Cochrane Library, while 6 (33.3%) expanded search to other databases; 14 (77.8%) applied the AMSTAR tool to assess methodological quality of included literature; 12 (66.7%) applied the GRADE system to assess the quality of evidence; and 9 (50%) yielded new outcomes. ConclusionCurrently, the development and reporting standards of Cochrane overviews are still immature. Investigators should choose proper methods based on research objectives when developing Cochrane overviews.
ITC (Indirect Treatment Comparison) software and indirect procedure of Stata software are especially used for indirect comparison nowadays, both of which possess the characteristics of friendly concise interface and support for menu operation. ITC software needs the application of other software to yield effect estimation and its confidence interval of direct comparison firstly; while Stata-indirect procedure can complete direct comparison internally and also operate using commands, which simplifies complicated process of indirect comparison. However, both of them only perform "single-pathway" of data transferring and pooling, which is a common deficiency. From the results, their results are of high-degree similarity.
It is a challenge for clinicians and diagnostic systematic reviewers to determine the best test in clinical diagnosis and screening. Meanwhile, it also becomes the new chance and challenge for diagnostic test meta-analysis. Network meta-analysis has been commonly used in intervention systematic reviews, which can compare the effect size of all available interventions and to choose the best intervention. Network meta-analysis of diagnostic test can be defined as comparing all available diagnostic technologies in the same conditions based on the common reference tests. In order to provide the guide for diagnostic systematic reviewers, we aims to introduce four methods of conducting diagnostic test accuracy network meta-analysis, and to explore two ranking methods of network meta-analysis of diagnostic test accuracy.
ObjectivesTo assess the accuracy of different types and magnetic field intensity of cardiac magnetic resonance for coronary artery disease.MethodsPubMed, The Cochrane Library, EMbase, WanFang Data, CNKI and CBM databases were searched to collect the studies on different types and magnetic field intensity of cardiac magnetic resonance for coronary artery disease from inception to May 15th, 2017. Two reviewers independently screened literature, extracted data and assessed the risk of bias of included studies. Then, data were synthesized by using MetaDisc 1.4, RevMan 5.3 and Stata 12.0 softwares. The pooled sensitivity (Sen), pooled specificity (Spe), pooled positive likelihood ratio (+LR), pooled negative likelihood ratio (–LR), pooled diagnostic odds ratio (DOR) and the area under curve (AUC) of the summary receiver-operating characteristic curve (SROC) were used to assess the diagnostic value of different types and magnetic field intensity of cardiac magnetic resonance.ResultsTwenty diagnostic studies were included, which involved 1 357 patients. The results of meta-analysis showed that (1) based on patient: compared with the gold standard, the pooled Sen, Spe, +LR, –LR, DOR and the AUC of SROC, pre-test probability, post-test probability were (0.87, 95%CI 0.82 to 0.90), (0.88, 95%CI 0.82 to 0.92), (7.33, 95%CI 4.74 to 11.32), (0.15, 95%CI 0.11 to 0.20), (49.53, 95%CI 27.46 to 89.36), (0.93, 95%CI 0.91 to 0.95), 20.00% and 65.00%, respectively. (2) Based on blood vessels: the pooled Sen, Spe, +LR, –LR, DOR and the AUC of SROC, pre-test probability, post-test probability were (0.81, 95%CI 0.76 to 0.85), (0.87, 95%CI 0.81 to 0.91), (6.37, 95%CI 4.37 to 9.30), (0.22, 95%CI 0.17 to 0.27), (29.58, 95%CI 18.53 to 47.22), (0.89, 95%CI 0.86 to 0.92), 20.00% and 61.00%, respectively. (3) Subgroup analysis showed that there was no difference in AUROC of different types of cardiac magnetic resonance, but significant difference was found in AUROC of 1.5T and 3.0T magnetic field intensity.ConclusionsCurrent evidence shows that, compared with gold standard, cardiac magnetic resonance can be regarded as an effective and feasible method for preoperative staging of breast cancer.
Randomized controlled trials (RCTs) are currently the gold standard for the treatment effect comparisons; however, it is sometimes not feasible to conduct an RCT due to ethical and economic reasons. In the absence of evidence for head-to-head RCT direct comparison, the indirect comparison technique is an effective and resource-saving alternative. Matching-adjusted indirect comparison (MAIC) is an attractive method in the field of population-adjusted indirect comparisons between two trials. It can adjust for between-trial imbalances in the distribution of observed covariates by weighting the available individual patient data of the studied intervention and then match the aggregated data of the controlled intervention. Subsequently, the treatment effect comparison can be evaluated through the post-matched population. Although MAIC is gaining increasing attention in clinical research, especially in the evaluation of new drugs, efforts are still largely required for knowledge dissemination in China. In this paper, we briefly introduced the concepts, research value and examples, and pros and cons of MAIC.
Objective To provide methodological guidance for the application of matching-adjusted indirect comparison (MAIC). Methods The methodology literature on MAIC was examined to clarify key steps and methodological points, and MAIC application literature in the non-small cell lung cancer field published after January 2016 was systematically reviewed to compare and analyze the current status and problems of MAIC. Results MAIC consisted of five key steps: data source and sample selection, matching variable screening, individual weight calculation, matching validity evaluation, and relative efficacy calculation. The systematic review revealed that studies primarily employed literature reviews to screen data sources, used statistical analysis and other scientific methods to screen matching variables, employed software for individual weight calculation, evaluated matching validity by reporting effective sample size (ESS), calculated relative efficacy using Cox, logistic, and other models, conducted sensitivity analyses to evaluate the uncertainty caused by different data sources and matching variable combinations, and the studies demonstrated good overall reporting standardization but significant differences in particular aspects. Concerning the connection between MAIC and pharmacoeconomic research, studies included mainly used target drugs as the reference group of survival data extrapolation, and proportional hazards (PH) assumptions were considered when utilizing hazard ratios (HR) in extrapolation. Conclusion There are some deficiencies in the method application and reporting standards of MAIC research, such as lack of explanation of data source selection criteria and matching variable screening criteria, insufficient reporting of weight distribution, and inadequate consideration of PH assumptions. It is recommended that future MAIC research systematically screen data sources and report covariate distribution evaluation, covariate status evaluation, weight distribution, uncertainty measurement, etc. Additionally, considering PH assumptions after calculating HR is suggested.