Quantitative Methods 2017

Heterogeneity

• The results loose credibility

• Different effect sizes in different trials • Different patient cohorts? • Different treatment efficacy? • Quality of data - How often is the TSH measured? The assumption behind a meta-analysis is that 1) The interventio i t same

- In Hodgkin patients? - In H&NSCC patients? • An opportunity to learn • Which patients benefit?

2) The effect size is the same in different cohorts

• Which is the better treatment?