Types of Meta-Analysis
🛑🚧🚧This page has been detected as SPAM and will be removed soon
The term Meta-analysis was coined by Gene Glass (1976)..[1]. Meta-analysis is a statistical procedure that is applied when amalgamating results from several individual studies to create a single, more accurate prediction of an overall effect of treatment or risk factor, or processes or other outcomes[2]. The key objectives of a meta-analysis is to augment the statistical power, to tackle any arguments when there is disagreement between individual studies, to enhance estimates pertaining to effect sizes and to find solutions to new questions that have not been addressed in individual studies contributing to the pooled analysis[3].
Each and every definition of meta-analysis emphasizes that there is an authentic requirement to synthesize the studies following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyes) guidelines. It has been stated that[4]. that reviewers often find it challenging to overcome the temptation to combine studies when there are questions raised on meta-analysis or it is unsuitable. However, meta-analysis could offer an accurate and strong summary estimate following an arduous and systematic integration of evidence that is already available. For an example of meta-analysis, researchers can refer one of the articles published on Medium[5]. Furthermore, meta-analysis can also prove to be beneficial while identifying, outlining and potentially investigating sources for bias, quantifying amongst heterogeneity of the study and recommending certain potential elucidation for revealing authentic heterogeneity from bias[6]. However, such variation cannot be identified in the literature review[7]. Now that we have understood what meta-analysis is all about, let’s take a look at the different types of meta-analysis.
Meta-Analysis Types
Cumulative Meta-Analysis
A cumulative meta-analysis (CMA) is a process of executing a new meta-analysis within every point over the history of the domain of research, thereby could prevent delays in the introduction of effective treatment. In CMA, the results from a new individual study is included to the accumulated database one at a time in a specified order (e.g. according to the date of publication or quality)[8]. Cumulative meta-analysis, as the name suggest accumulates experiments right from the earliest to the latest. In this case, every successive experiment comprises a synthesis of experiments conducted previously. This kind of chronological combination of experiments tends to reveal whether there is any consistency in the outcomes from subsequent experiments and specify the point where no additional experiments are needed as the outcomes favour one particular procedure, treatment or product[9]. CMA identifies two distinct knowledge: sufficiency (“Are additional studies required to establish the existence of phenomenon) and stability (will additional studies change the overall picture of the phenomena?). Thus, CMA provide a framework to describe trends in overall estimates of effects over time or as each study is added to the pool.
Individual Participant Data Meta-Analysis (IPD-MA)
[[Individual participant data|(IPD) meta-analysis or IPD-MA are the gold standard of meta-analysis. In an IPD-MA line-by-line patient data are collected from the relevant studies or directly from the researchers responsible for each study that addresses a particular research question, rather than extracting aggregate data from study publications or from investigators[10]. Using IPD-meta-analyse, both study-level and patient-level factors can be explored and included allowing a more detailed exploration of potential sources of heterogeneity in the study results. Besides, IPD offers several advantages 1) it reduces the risk of publication bias as it allows direct contact with the authors for the missing or incomplete information 2) the availability of original data means, a more detailed information about the study (e.g. allocation concealment and completeness) and thereby allows better analysis of subgroup data. 3) Permits time-to-event analyses for different time points (e.g. time to pregnancy), different outcome measurements, and different scales of measurements.
For example; in a time-to-event analysis, aggregate data meta-analysis tends to compare groups by defining the number of events that takes place over a particular point of time and later calculating the odds-ratio (OR). As opposed, IPD-MA tends to utilize techniques of survival analytics to explain the intervention effect as a hazard ratio (HR). Both the methods generally results in results that are starkly diverse. Individual patient data was utilized by Stewart and Parmar, (1993)[11]. in order to calculate the OR pertaining to survival during every year of follow-up in patients with ovarian cancer. These included patients who did or did not receive platinum-containing chemotherapy. From their findings, it was revealed that the value of OR and its significance from a statistical point of view fluctuated largely on the basis of which ‘time-point’ was utilized during the period of observation, with a view to calculate OR. IPD-MA is particularly beneficial in case of data that have repeat or longitudinal measures[12]. Thus, IPD meta-analyses can improve the quality of both the data and the analyses and so the reliability of the results[13]. Therefore, considered to be a ‘gold-standard’ of systematic review. PRISMA-IPD (2015)[14] provides guidelines for reporting systematic reviews and meta-analyses of IPD.
Aggregate Data Meta-Analysis (ADMA)
An aggregate data meta-analysis (ADMAs) utilizes statistical analyses to generate a summary statistics using effect estimates available from multiple (aggregate or grouped) data of individual studies or data that is sought from authors who have not yet published their completed articles to estimate the outcome of interests[15]. This type of review is much more because of low-cost, completed relatively quickly, the data its uses is easy to access and required far less resources to produce. Despite the above advantages, ADMAs are subject to publication bias, have limited control over the data and interpret summary effect within the context of the heterogeneity that arises between various study design, model selection, analysis approaches. For an instance, at the time of taking into account an outcome which is continuous, in general, every study should essentially report a common measure of an outcome variable and its spread (for example; its standard error and sample mean). However, there are certain studies which might in its place report the median in tandem with diverse measures related to spread. In recent times, the task of integrating medians within meta-analysis has been realized through an estimation of the sample mean and its standard error within every study which reported a median with the objective of meta-analyzing the mean[16]
Network Meta-Analysis (NMA)
Network meta-analysis is a statistical method through which it is possible to compare multiple treatments concurrently within a single analysis through a combination of indirect and direct evidence in a network of randomized controlled trials (RCTs) based on a common comparator. Thus, NMA would aid in evaluating the comparative effectiveness of several diverse treatments that are used frequently within clinical practice through the visualization of a larger amount of evidence, estimation relative effectiveness and rank ordering of the intervention[17]. Thus has turned out to be much preferred amongst clinicians.
A network meta-analysis or NMA which is valid tends to satiate the assumption pertaining to transitivity. This refers to the fact that there are no variations amongst the existing comparisons apart from the treatments that are under comparison[18]. Another manner through which this can be observed is that within an RCT which is hypothetical that comprises of all treatments included in a NMA it is possible to randomize participants to any of the treatments. For instance; topical medication is prescribed as monotherapies for initial treatment in cases of glaucoma. Whereas, combination therapies are exclusively utilized amongst patients whose intraocular pressure is inadequately controlled by monotherapy[19]. Nevertheless, in case proper caution is not accorded in performing and interpreting a network meta-analysis, there is scope for bias to set in within inferences that are derived[20]. PRISMA-NMA (2015)[21] provides guidelines for reporting systematic reviews and meta-analyses of NMA.
Prospective Meta-Analysis (PMA)
A prospective meta-analysis (PMA), is a collaborative research design where individual site performs RCTs and pool the data for meta-analysis. In a PMA, the hypotheses, the criteria to select a study and the analyses are outlined even prior to the outcomes of the studies associated to the PMA research question have been identified, which lowers several of the challenges linked with a traditional (retrospective) meta-analysis. There are several advantages that are presented by a PMA; a prospective meta-analysis can be instrumental in lowering bias and waste in research, they are efficient, adaptive and also allow collaboration[22]
PMAs have of-late been elucidated as the systematic review of the next generation. As a matter of fact, it has also been argued by Ioannidis,(2010)[23] that every primary authentic research might be structured, conducted and inferred as a prospective meta-analysis. In prospective meta-analysis, studies are added in a prospective manner which means prior to any individual study outcomes associated to the PMA research question are available. This in turn is said to lower the scope of risk in terms of publication bias as well as selective reporting bias while facilitating better harmonization within the outcomes of the study.
Conclusion
While these are the many types of meta-analyses, each of these types have a different utility and are used in different situations depending upon the nature and kind of the study being conducted. Though each of these types of meta-analyses is used in different situations and each of these has their own pros and cons however, it is not possible to determine whether one type of meta-analysis is more effective and robust than the other. As each of these serve distinct purposes.
References
- ↑ Glass GV. Primary, secondary and meta-analysis of research. Educ Researcher1976;10: p.3-8. Retrieved from https://journals.sagepub.com/doi/abs/10.3102/0013189x005010003?journalCode=edra
- ↑ Ferrer, R.L. ”Graphical methods for detecting bias in meta-analysis”. Family Medicine, 1998, 30(8), P.579-583p.
- ↑ Hunter & Schmidt.” Methods of meta-analysis: Correcting error and bias in research findings”. Thousand Oaks, CA, US: Sage Publications, Inc., 1990, Retrieved from https://psycnet.apa.org/record/1990-97087-000
- ↑ Egger, M., Ebrahim, S., & Smith, G. D. (2002). "Where now for meta-analysis?". International Journal of Epidemiology, 31(1), p.1–5. Retrieved from https://doi.org/10.1093/ije/31.1.1.
- ↑ Pubrica Healthcare. "Sample work of Written Meta-analysis", "Medium", 6 April 2019. Retrieved on 9th January 2020.
- ↑ Tatsioni, A., & Ioannidis, J. P. A. (2008). “Meta-Analysis”. In International Encyclopedia of Public Health (pp. 442–450). Elsevier. Retrieved from https://doi.org/10.1016/B978-012373960-5.00341-5
- ↑ Pubrica Healthcare. "The difference between meta-analysis and literature review". "Medium", 6 April 2019. Retrieved on 9th January 2020.
- ↑ Egger, M., Davey Smith, G. & Sterne, J. A. (2001). Uses and abuses of meta-analysis. Clinical Medicine, 1(6), p. 478-484 Retrieved from https://oce.ovid.com/article/00129491-200111000-00015
- ↑ Mullen, B., Muellerleile, P., & Bryant, B. (2001). “Cumulative Meta-Analysis: A Consideration of Indicators of Sufficiency and Stability”. Personality and Social Psychology Bulletin, 27(11), p.1450–1462. Retrieved from https://doi.org/10.1177/01461672012711006
- ↑ Stewart, L. A., & Tierney, J. F. (2002). To IPD or not to IPD?: Advantages and Disadvantages of Systematic Reviews Using Individual Patient Data. Evaluation & the Health Professions, 25(1), 76–97. https://doi.org/10.1177/0163278702025001006
- ↑ Stewart, L. A., & Parmar, M. K. (1993). "Meta-analysis of the literature or of individual patient data: is there a difference?" Lancet, 341(8842), p.418–422. Retrieved from https://doi.org/10.1016/0140-6736(93)93004-k
- ↑ van Walraven, C. (2010). "Individual patient meta-analysis—rewards and challenges". Journal of Clinical Epidemiology, 63(3), p. 235–237. Retrieved from https://doi.org/10.1016/j.jclinepi.2009.04.001
- ↑ Stewart, L.A., Tierney, J.F. and Clarke, M. (2008). Reviews of Individual Patient Data. In Cochrane Handbook for Systematic Reviews of Interventions (eds J.P. Higgins and S. Green). doi:10.1002/9780470712184.ch18
- ↑ Stewart LA, Clarke M, Rovers M, Riley RD, Simmonds M, Stewart G, Tierney JF; PRISMA-IPD Development Group. Preferred Reporting Items for Systematic Review and Meta-Analyses of individual participant data: the PRISMA-IPD Statement. JAMA. 2015;313(16):1657-1665
- ↑ Tudur, C., Williamson, P. R., Khan, S., & Best, L. Y. (2001). “The value of the aggregate data approach in meta‐analysis with time‐to‐event outcomes”. Journal of the Royal Statistical Society, Series A (Statistics in Society), 164(2), 357–370. Retrieved from https://doi.org/10.1111/1467-985X.00207
- ↑ McGrath, S., Zhao, X., Qin, Z. Z., Steele, R., & Benedetti, A. (2017). “One-sample aggregate data meta-analysis of medians”. Methodology. Retrieved from retrived from http://arxiv.org/abs/1709.03016
- ↑ Debray, T. P., Schuit, E., Efthimiou, O., Reitsma, J. B., Ioannidis, J. P., Salanti, G., & Moons, K. G. (2018). An overview of methods for network meta-analysis using individual participant data: when do benefits arise? Statistical Methods in Medical Research, 27(5), 1351–1364. https://doi.org/10.1177/0962280216660741
- ↑ Salanti, G. (2012). “Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: many names, many benefits, many concerns for the next generation evidence synthesis tool”. Research Synthesis Methods, 3(2), p.80–97. Retrieved from https://doi.org/10.1002/jrsm.1037
- ↑ Li, T., Lindsley, K., Rouse, B., Hong, H., Shi, Q., Friedman, D. S., … Dickersin, K. (2016). “Comparative Effectiveness of First-Line Medications for Primary Open-Angle Glaucoma: A Systematic Review and Network Meta-analysis”. Ophthalmology, 123(1), 129–140. Retrieved from https://doi.org/10.1016/j.ophtha.2015.09.005
- ↑ Rouse, B., Chaimani, A., & Li, T. (2017). “Network meta-analysis: an introduction for clinicians”. Internal and Emergency Medicine, 12(1), p.103–111. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/27913917
- ↑ Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, Ioannidis JP, Straus S, Thorlund K, Jansen JP, Mulrow C, Catalá-López F, Gøtzsche PC, Dickersin K, Boutron I, Altman DG, Moher D (2015). The PRISMA Extension Statement for Reporting of Systematic Reviews Incorporating Network Meta-analyses of Health Care Interventions: Checklist and Explanations. Ann Intern Med. 2015;162(11):777-784
- ↑ Seidler, A. L., Hunter, K. E., Cheyne, S., Ghersi, D., Berlin, J. A., & Askie, L. (2019). “A guide to prospective meta-analysis”. BMJ, l5342. Retrieved from https://doi.org/10.1136/bmj.l5342
- ↑ Ioannidis, J. P. A. (2010). “Meta-research: The art of getting it wrong”. Research Synthesis Methods, 1(3–4), p.169–184. Retrieved from https://doi.org/10.1002/jrsm.19
This article "Types of Meta-Analysis" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Types of Meta-Analysis. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.
