Short training on systematic review and meta-analysis.pptx

Systematic review and Meta-analysis

1. Short refreshment training on meta-analysis Using R Hadush Negash : [email protected]

2. Outlines Introduction to citation managers How to import references to EndNote Evidence based health care?????????? What is meta-analysis Introduction software's for metan Meta-analysis for proportion Meta-analysis for determinant factors Meta-analysis for continuous variables Meta-regression Meta-analysis for diagnostic test accuracy

3. Citation managers EndNote Searching Importing Search full text PDF Avoiding duplicates Transferring and creating groups Citing Exercise ???????? Mendeley

4. Evidence based healthcare Evidence based medicine Evidence based health care

5. Developing review questions Effectiveness review Research question : Is breast feeding replacement reduces HIV transmission from mother to child among HEI (HIV exposed infants)? P (population) = HIV exposed infants I (intervention) = Breast feeding replacement C (comparator group) = On breast feeding O (outcome) = HIV transmission

6. 2. QUALITATIVE REVIEW Research question : Length of stay among recovered SAM patients in TFU (therapeutic feeding unit)? P (population) = Recovered SAM patients I (intervention) = Length of stay Co (condition) = TFU

7. 3. Cost effectiveness Research question : Is using electronic medical record system effective than using paper based service for patient management in health facility? P (population) = Clients I (intervention) = Electronic medical records C (comparator) = paper based service O (outcome) = Cost reduction C (context) = Health facility

8. 4. Prevalence or incidence reviews Topic: What is the prevalence of cervical cancer among HIV patients in Tigrai? Co = Having CC Co = Tigrai Pop = HIV positive

9. 5. Diagnostic test accuracy review Research question : What is the diagnostic accuracy of RDT for cholera comparing to culture for detection V.cholarae of among patients with AWD ? Possible justification for the question : P (population) = Patients with AWD I (Index test) = RDT R (reference test) = culture D (Diagnostic) = V.cholarae

10. 6. Etiology and risk review Review question : Are tobacco passive smokers at risk of developing lung cancer ? Justification: P (population) = passive smokers E (Exposure) = Tobacco Smoke O (Outcome) = Lung cancer Concept map and logit grid for an alternative searching Main topic Keywords MeSH terms

11. Importing references to EndNote Booleans MeSH

12. Study selection Screening of studies for relevance Selection bias Title, abstract, full text article reading Inclusion and exclusion criteria for articles PRISMA flow diagram PRISMA checklist ..\Presentation\PRISMA 2009 checklist.docx Critical appraisal and quality assessment for screened articles Score ..\Presentation\JBI Include Vs exclude Data extraction: At least two authors independently PRISMA ( Preferred Reporting Items for Systematic Reviews and Meta-Analyses) ENTREQ ( Enhancing transparency in reporting the synthesis of qualitative research) JBI Reviewer’s Manual and JBISRIR Author guidelines Cochrane Handbook and MECIR ( Methodological Expectations of Cochrane Intervention Reviews)

13. Data entry and analysis Data entry Word pad Excel SPSS Stata SAS R Statistical analysis Stata R

14. A. Meta-analysis for proportions (prevalence, incidence) Review title: Epidemiology and effect of sex differences on Podoconiosis infection in Ethiopia: A systematic review and Meta-analysis; 2019 Protocol : Epidemiology and effect of sex difference on Podoconiosis infection in Ethiopia; A protocol for systematic review and Meta-analysis; 2019 Review type: CoCoPop (Epidemiology of Podoconiosis, Ethiopia, people with symptoms of Podoconiosis) PEO (Population, sex [Females Vs males], Podoconiosis) Data: ..\Presentation\H.csv commands: meta, metaprop

15. Opening and working R/ R Studio

16. Changing working directory and import/read data

17. Run the meta-analysis: metaprop ……

18. Forest ( hh )……Enter

19. Resolving heterogeneity: 1. subgroup analysis

20. Sensitivity analysis and publication bias Metainf :….. For influential analysis Metabias :…. For publication bias/ Egger’s test

21. 2. Meta-analysis for determinants/effects (OR/RR) Command: metagen / metabin Data: ..\Presentation\HB.csv

22. Heterogeneity test, forest plot

23. Influential analysis, publication bias end event rate Influential analysis: metainf Publication bias: Funnel plot, Egger's test, Trim-and-fil test

24. Trim-and-fill method and funnel plot

25. Event rate Command: labbe.metabin ( nn,col.random =“red”)

26. Subgroup analysis nn < - metabin ( Ef,Tf,Em,Tm,studlab = Author,data = n,sm ="OR", byvar = setting)

27. subgroup cont’d…

28. 3. Meta-analysis of continuous variables Command: metacont Data: ..\Presentation\PCNL.csv Mean difference (with in the same units) Vs standardized mean difference (with different measuring units) Then extract data based on: ne, me, sde , nc , mc, sdc R: aa<- metacont ( mde , sde , ne, mdc, sdc , nc studlab =Author/study/trial, data = PNCL/ hh , sm = “MD/SMD)

29. Importing and analyzing

30. Standardization

31. Sensitivity analysis and publication bias??? > metabias (b) Linear regression test of funnel plot asymmetry data: b t = NaN , df = 20, p-value = NA alternative hypothesis: asymmetry in funnel plot sample estimates: bias se.bias slope NaN NaN NaN

32. 4. Meta-regression Command: metareg Metareg ( dataname , v1+v2) Rma.uni () Plots in metaregression Plot(h)

33. 5. Meta-analysis for diagnostic accuracy Command: madad () > H<- madad ( AuditC ) Forest (H) DOR: TP÷FP/FN÷TN LR+ = TP ÷ D- / TP ÷ D- LR- = TN ÷ D+ / TN ÷ D-

34. THANK YOU!!!! Greetings to all your attentions!!!!