I'm conducting a systematic review and found 5 RCTs measured the depression (outcome) using different scales (Hamilton Depression Rating Scale/ Montgomery-Åsberg Depression Rating Scale/Beck Depression Inventory etc) with the different period (3 months, 4, months, etc.). I was able to conduct a meta-analysis using the same scales and period (2 RCTs), however I'm wondering whether it is sensible to conduct a meta-analysis using different sales and period?
Potentially yes - could be very interesting but I have some concerns or issues that would need to be addressed
- number of studies is very small
- you would need to look at the differences between the scales and whether the thresholds are different and their sensitivity at the bottom end (i.e. high depression scores) which could vary
- you would also need to look at the literature on timescales because there are patterns to the impact of treatment (i.e. dose/response, decreased effectiveness over time etc) and to recurrence of symptoms which I imagine have very complex natural histories.
I would therefore recommend getting advice from three sources - a psychometrist on the scales, a clinical expert on the longitudinal natural hisotry of treatment, recovery and relapse, and a methodologist on the number of studies (critical I suspect).
Yes, it is possible by combining the data using "standardized mean difference" then perform subgroup analysis according to the duration of the follow-up.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis