 |
 |
Dear Craig, I think plots are useful even if there is not combined estimate. The units depend on which plot you are using: 1. For a standard "forest plot" of RR or OR the raw data is OK because its being transformed in to RR or OR - under the assumption that the RR or OR is constant over time (which is only ever approximately true) 2. For the L'Abbe plot of treated RATE versus control RATE its better to convert to rates to allow a comparison across studies. Regards, Paul Glasziou -----Original Message----- From: Craig Lockwood To: [log in to unmask] Sent: 22/08/03 10:24 Subject: Meta view or meta analysis - urgent help please Dear EB group, I am reading a review where meta analysis was not possible, but many meta graphs of individual studies are presented, and I am wondering if there is a consensus on the value of doing this, given the point of meta-analysis is to present combined study effects. The primary question of interest I want to explore is on presentation of data in a meta graph from a single study using events per patient year versus using the raw data. EG In a study measuring the effectiveness of an intervention at reducing the incidence of infection, the raw data indicates infection rates as follows: 16/116 for the treatment group 36/148 for the control group In the trial, it is also reported as: 0.19 infections per patient year in the treatment group 0.39 infections per patient year in the control group The meta graph has been constructed using the event per patient year data rather than the primary data. I lack a statistical background, and would greatly appreciate advice on which if either is considered the proper convention. Sincerely, Craig Craig Lockwood Deputy Director; Centre for Evidence-based Nursing South Australia (CENSA) Margaret Graham Building Royal Adelaide Hospital North Terrace Adelaide 5000 [log in to unmask] phone: (08) 8303 4880 fax: (08) 8303 4881 _____ Apologies if you have received this transmission in error. Please notify me immediately by return e-mail and delete all copies _____