Hello Patrick,
please accept my apologies for the late reply. Your question sent me running around in circles a bit, which is partially because I'm not a statistician (we are here to support using EPPI-Reviewer, we are not specialists in Meta-Analysis) and partially because I could see different ways in which I could read this question:
> Is there a way to calculate the missing data in order to use eppi for the meta-analysis?
What I think pointed me in the correct direction was a direct look at your data, to try and figure out what you actually mean. Reason why I was confused is that if a study reports the mean effect and SD for intervention and control arms, then how to extract the numerical outcome data should be quite obvious...
By looking at your specific data, however, I think I've understood your problem. Specifically, I've looked at "Bang (2013)" (item Id: 46668653) and at the "10-m walk test" therein. Found the "Table 2" in the full text, and suddenly your question made sense:
Table 2 contains 4 Means and 4 Standard Deviations (for: intervention and control arms, before and after the intervention), when in fact you would need two (representing the differences between the before and after measures for each arm), so your question would translate (in my own words/mind): "how do I calculate the latter values, given the first?".
Answer is I don't know, and I suspect that it isn't possible. Which then confused me even more...
Luckily I then noticed that in fact Bang and colleagues actually do provide the data we need, but they report it in the main text, not the summary table:
The mean change of 10-m walk test in the action observational training group was 0.35 (SD 0.13) seconds and control group was 0.16 (SD 0.08) seconds, which was significantly different between groups (P = 0.001)
Which is what we're after and can be recorded directly in an outcome of type "Continuous: Ns, means and SDs". The same applies to the other measures reported in this particular study, so perhaps my overall understanding of both problem and solution isn't that bad, after all.
Does the above help?
Re your second question:
> How is it inteded to enter the timepoints of the values in eppi? Can you make inputs for baseline and post or just the duration between the measurements?
In this case, I don't think you'll need timepoints. They would be useful if the study in question had also repeated the measures again, a week after the intervention, for the purpose of verifying whether the effect of the intervention disappears after some time. In such cases, EPPI-Reviewer provides the Timepoints structure, that allows to record when consecutive measures were taken. You would first enter the "timepoints" described in the full text via the "Arms and timepoints" tab, then when recording the specific values for a given measure, you would pick the appropriate timepoint from the values that will be now available in the "timepoint" drop-down menu. In turn, this will become useful data at the Meta Analysis stage, as it would allow to include in a given MA only studies that reported measures taken at comparable/equivalent time points. For example, you would not place in a same MA two measures of survival, coming from two different studies of the same medication, if one was taken 6 months after the intervention while the other was taken after 18 months. I hope this makes sense to you and that it helps!
Best wishes,
Sergio