RevMan Tutorial - How To Perform Subgroup Analysis
Вставка
- Опубліковано 7 сер 2024
- In this video tutorial, I will show you how to perform subgroup analysis in RevMan as part of a meta-analysis. I will also describe the output of the results.
In the example, I will use a continuous data type. However, most of the settings apply to other data types, such as odds ratios.
RevMan version used: 5.3
THE ONLINE GUIDE
toptipbio.com/subgroup-analys...
VIDEO BREAKDOWN
Step 1: Adding a new subgroup to an outcome (00:45)
Step 2: Moving the studies into different subgroups (01:34)
Step 3: The output (02:00)
MORE HELPFUL HINTS & TIPS
toptipbio.com/
FOLLOW US
Facebook: / toptipbio
Twitter: / toptipbio
LinkedIn: / top-tip-bio - Навчання та стиль
THE ONLINE GUIDE
toptipbio.com/subgroup-analysis-revman/
You don’t understand how much you’ve helped me with my dissertation I love you
Really simple to understand and very informative! Thank u for ur effort!
Thank you!! Excellent video!
Thanks for posting.
Briliant...! Thanks for sharing...
Fantastic explanation
Thank you
Спасибо!
Hello, thank you for the video! I am wondering how many subgroups that you can include under one variable? For example, I would like to see if different teaching method would have an association with different in ES for an intervention. I have divided the teaching methods into 10 different subgroups (active, behavioural, passive, home activities..), does it make sense to perform the subgroup analysis for all of the subgroups and compare the results? Thanks!
Thank you for this clarification but I wonder if I can change the color of the icons of the forest plot and write on it. Also, how can I insert a column to show the population for every study
Hi , thanks for such a great video. Just a question.
Prior to subanalysis my overall effect p=0.8 with 1^2= 93% (p
Hello
Could you please explain the following analysis trial sequential analysis (TSA) in meta-analysis
Hi ..is there a way to add an additional or more studies after performing the analysis?
Love the videos! Off topic question but just wondered if a network meta-analysis could be completed on RevMan?
No
Hello. In order to make a sub-group analysis, do I need at least 10 studies that fit the sub-group criteria? Or that number (10) is the minimum for the broad meta-analysis and if I have those, that I can make a sub-group with just 3 studies, for example? Thanks
Hi Jorge,
There is no specific suitable number of studies to include for subgroup analysis, per say. However, the more you have the better. I have seen subgroups with 2 studies in each. Including more studies will give you more confidence in your results.
Best wishes,
Steven
How to do clustering correlation analysis between subgroup? please
how do i get back the forest plot before subgroup analysis e.g. 3.1
Hi, it looks like you used some studies that had 100% male or female samples and for studies that included males and females, you separated the samples into M & F (example Young et al. 1999 (F) and Young et al 1999 (M)). Is this correct? Also, do you have a reference for the sub-group analysis procedure please?
Hi Claire. That is correct, i created a separate entry for males (M) and females (F). As for a reference, this may be useful for you: pubmed.ncbi.nlm.nih.gov/23479191/
Thanks
Steven
@@StevenBradburn Thank you so much, that's really helpful!
Addionatily, If I want to compare a sub-group analysis comparing male to female, should I use only RTC that compared gender differences? Or can I group studies that have only men or only women and compare them anyway in the sub-group analysis?r
Hi Jorge,
That is a good question. I would presume there would be less variability in the first option since RCTs are designed in a way that makes them more suitable for meta-analysis. This is not always possible, however. So, you could take all of the studies that have a male-only population and put them in one subgroup and then do a separate subgroup for female-only population studies.
I hope that makes sense.
Steven
@@StevenBradburn Thanks, Steven. Yes, it makes sense. One more, kind of on the same spirit... If I was going to make a MA about the effect of pilates on back pain, and some controls have no intervention, others is running, others is educational therapy.... Consider that there are little studies and as such, I want to incorporate all. How should I formulate my research question correctly in order to be able to include them all? Because, I believe, that If I want to determine the effect of Pilates as it is, the ideal would be to use only studies that had control without any intervention, right?
PS: I already read a couple of times that one needs at least 10 studies for sub-group. Not sure if was borenstein....
P
Is sub-group analysis and stratified analysis the same?
As far as I am aware, yes they are
Hi I like your video. I have a question, what if the p-value for both my subgroups are shown (p
Hi Chun,
So in this instance, it sounds like individually, your subgroups pooled effect is significant (P0.05, this means your subgroups are not statistically significant from each other. I presume the pooled effect for each subgroup is quite similar?
Thanks,
Steven
@@StevenBradburn Greetings, thank you for your prompt reply. Both the pooled effects of my subgroups are 0.47 and 0.30 respectively. Are these values considered similar with a difference of 0.17? Sorry for asking so many questions.