In doing a meta-analysis using articles done with time-to-event analysis, I faced the problem that some authors reported the effect size (incidence density) with person-year observations, while others reported person-month observations. Some also reported the incidence density for only the exposed group (single arm). Others reported the rate ratio (the ratio of IR in the exposed group to IR in the unexposed group).
Hello Nebiyu,
If you have a computed rate per person-month, and wish to convert that to a rate per person-year, you would multiply numerator (aggregate instances per month) and denominator (months) by 12.
Have a look at this link for an illustration of conversions: https://sph.unc.edu/wp-content/uploads/sites/112/2015/07/nciph_ERIC4.pdf
As for consolidating studies that report only one arm vs. those which report both, your choices are: (a) incorporate study type as a moderator/explanatory variable; or (b) accept only one study type for data points.
Good luck with your work.
I would like thank you very much dear David Morse for giving me your genuine answer and clarification to my question.
If you are interested only on the observation time, divide it by 12. But if you are interested in the IR, multiply it by 12.
- Badges
- Science method