Systematic reviews and meta-analysis of time-to-event outcomes can be analysed on the hazard ratio (HR) scale but are very often dichotomised and analysed as binary using effect measures such as odds ratios (OR). This thesis investigates the impact of using these different scales by re-analysing meta-analyses from the Cochrane Database of Systematic Reviews (CDSR), using individual participant data (IPD) and a comprehensive simulation study. For the CDSR and IPD, the pooled HR estimates were closer to 1 than the OR estimates in most meta-analyses. Important differences in between-study heterogeneity between the HR and OR analyses were observed. These caused discrepant conclusions between the OR and HR scales in some meta-analyses. Situations under which the clog-log link outperformed the logit link and vice versa were apparent, indicating that the correct method choice does matter. Differences between scales occurred mainly when event probability was high and could occur via differences in between-study heterogeneity or via increased within-study standard error in OR relative to HR analyses. In many simulation scenarios, analysing time-to-event data as binary using the logit link did not substantially affect bias and coverage apart from those where large percentage random censoring and long follow-up time was present. The method though lacks precision particularly for small meta-analyses. Analysing the data as binary using the clog-log link consistently produced more bias, low coverage and low power. If a HR estimate cannot be obtained per trial to perform a meta-analysis of time-to-event data, a meta-analysis using the OR scale (using the logit link) could be conducted but with awareness that this would provide less precise estimates in the analysis. Investigators should avoid performing meta-analyses on the OR scale in the presence of high event probability, large percentage random censoring and therefore longer follow-up times assuming of large event rates of the trials included.