Hello there,

I am a Ph.D student in forensic psychology. I am currently about to conduct a systematic review and meta-analysis. For the meta-analysis, many eligible studies reported Hazard Ratios (HRs), meaning that recidivism was continuously measured (i.e., time to (recidivism) event analyses) and that coefficients indicate the time to recidivism, not just binary recidivism with a fixed follow-up. Some of these were from univariate analyses, others were from multivariate analyses. I am aware that one cannot meta-analyze effect sizes derived from multivariate analyses, but what about the HRs for the univariate analyses? See below for my specific questions:

1) Can HRs be aggregated with other HRs? (provided they were from univariate analyses)?

2) Can HRs be converted to Odds Ratios (ORs) and aggregated with other ORs (or other types of coefficients) stemming from analysis wherein the follow-up time was fixed and time to event analyses were not used?

3) If HRs CAN be aggregated with, for example, ORs, what is the correct formula? I have yet to find one

Thank you very much in advance to anyone who could help answer my questions and/or point me to the right source

Best regards,

Cathrine Pettersen