• Users Online: 39
  • Home
  • Print this page
  • Email this page
Home About us Editorial board Ahead of print Current issue Search Archives Submit article Instructions Subscribe Contacts Login 


 
 Table of Contents  
EDITORIAL
Year : 2017  |  Volume : 6  |  Issue : 1  |  Page : 1-2

Meta-analysis: Minimizing errors


Editor-in-Chief, Journal of Dental and Allied Sciences, Department of Prosthodontics, Sinhgad Dental College and Hospital, Pune, Maharashtra, India

Date of Web Publication2-May-2017

Correspondence Address:
Jayant N Palaskar
Department of Prosthodontics, Sinhgad Dental College and Hospital, S. No. 44/1, Off Sinhgad Road, Vadgaon Budruk, Pune - 411 041, Maharashtra
India
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/jdas.jdas_19_17

Rights and Permissions

How to cite this article:
Palaskar JN. Meta-analysis: Minimizing errors. J Dent Allied Sci 2017;6:1-2

How to cite this URL:
Palaskar JN. Meta-analysis: Minimizing errors. J Dent Allied Sci [serial online] 2017 [cited 2017 Oct 21];6:1-2. Available from: http://www.jdas.in/text.asp?2017/6/1/1/205438










A meta-analysis is a powerful tool to cumulate and summarize the knowledge in a research field through statistical instruments and to identify the overall measure of a treatment's effect by combining several individual results.[1] Although meta-analysis and its outcomes are gold standards in the field of review, researchers must be aware of what precautions should be taken before taking up any meta-analysis study. Any error committed inadvertently by the researchers can lead to misleading conclusions; an investigator should be eternally vigilant against such errors which may undermine the purpose of meta-analysis. For this reason, it is important to write a prospective analysis protocol, which specifies the objectives and methods of the meta-analysis. Having a protocol restricts the risk of biased and selective reporting.


  Inclusion and Exclusion Criteria of Studies to Be Included in Meta-Analysis Top


Researchers should be absolutely clear about their research question specified in the protocol. The research question framed has to be very narrow to avoid heterogeneity of studies to be included. On the other hand, a very broad question will permit researchers to include many studies which may not be directly related to the research question framed for the meta-analysis and indeed, will lead to increase in heterogeneity of the study that may result in nonconclusive, nongeneralizable outcomes. I would like to reinstate that inclusion of homogeneous (similar) studies will help in drawing generalized conclusions.

For pooling articles for the analysis, at least three to four databases must be searched. Most valid databases such as PubMed, MEDLINE, EMBASE, and The Cochrane Central Register of Controlled Trials are recommended. Other databases such as Google Scholar, Scopus, Ebsco, and ResearchGate may also be utilized.

Monitoring quality of included studies

It is imperative to maintain a strict check on the quality of studies included in the analysis. Cochrane guidelines give structured formats to establish the internal and external validity of the studies to be included. Internal validity helps in minimizing bias and external validity will help in extrapolating the outcomes of the analysis to a larger population. The investigator has to exhibit constant vigilance against bias such as the articles' author's standing in the field, reputation of the journal, in which the said article is published, and country where the study was conducted. Validity results may also be influenced by investigators' opinion. There should be no bias at any stage of meta-analysis as it will adversely affect the outcome of the analysis. Different types of bias have been dealt by me in my last editorial.

Fixed- versus random-effects estimate

There are two common models of a meta-analysis to estimate treatment effects of any study included in the analysis: Fixed- and random-effects estimates. If the random-effects estimate is more beneficial, researchers should consider whether it is reasonable to conclude that the treatment was more effective in smaller studies. This is because the weightage given to each included study through the random-effect model is less influenced by the sample size than that given by means of the fixed-effects model.[2] One must note that if there is no evidence of heterogeneity between studies, the fixed- and random-effects estimates will be identical, so there will be an actual difficulty in identifying the small-study effect.[3] Researchers who are concerned about the influence of small-study effects on the results of a meta-analysis in which there is evidence of between-study heterogeneity should compare and address the fixed- and random-effects estimates of the treatment effect.


  Managing Heterogeneity of Selected Studies Top


The degree of heterogeneity of the selected studies is another important limitation, and the random-effects model should be used during the data analysis phase to incorporate in the treatment effect the identifiable or nonvariability between studies.[4] It is fundamental to observe that exploring heterogeneity in a meta-analysis should start at the stage of protocol writing, by identifying a priori which factors are likely to influence the treatment effect. Visual inspection of the meta-analysis plots may show whether the results of a subgroup of studies have the same overall direction of the treatment effect. One should pay attention to meta-analysis, in which results have a discordant treatment effect for groups of studies and no explanation of variance has been done. Sources of variation should be identified, and their impact on effect size should be quantified using statistical tests and methods such as analysis of variance or weighted meta-regression.[5]

Testing effects suggested by data and not planned a priori considerably increases the risk of false-positive results.[6] To minimize this error, it is important to identify the effects to test before data collection and analysis;[7] otherwise, one may adjust the P value according to the number of analysis performed. In general, post hoc analysis should be deemed exploratory and not conclusive.



 
  References Top

1.
Biondi-Zoccai G, Landoni G, Modena MG. A journey into clinical evidence: From case reports to mixed treatment comparisons. HSR Proc Intensive Care Cardiovasc Anesth 2011;3:93-6.  Back to cited text no. 1
    
2.
Greco T, Zangrillo A, Biondi-Zoccai G, Landoni G. Meta-analysis: Pitfalls and hints. Heart Lung Vessel 2013;5:219-25.  Back to cited text no. 2
[PUBMED]    
3.
Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration, 2011. Available from: www.handbook.cochrane.org. [Updated 2011 Mar; Last accessed date on 2017 Apr 14].  Back to cited text no. 3
    
4.
Ng TT, McGory ML, Ko CY, Maggard MA. Meta-analysis in surgery: Methods and limitations. Arch Surg 2006;141:1125-30.  Back to cited text no. 4
[PUBMED]    
5.
Harrison F. Getting started with meta-analysis. Methods Ecol Evol 2011;2:1-10.  Back to cited text no. 5
    
6.
Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine – Reporting of subgroup analyses in clinical trials. N Engl J Med 2007;357:2189-94.  Back to cited text no. 6
[PUBMED]    
7.
Walker E, Hernandez AV, Kattan MW. Meta-analysis: Its strengths and limitations. Cleve Clin J Med 2008;75:431-9.  Back to cited text no. 7
    




 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Inclusion and Ex...
Managing Heterog...
References

 Article Access Statistics
    Viewed305    
    Printed8    
    Emailed0    
    PDF Downloaded64    
    Comments [Add]    

Recommend this journal


[TAG2]
[TAG3]
[TAG4]