Hi there,

Is it possible to calculate the covariance of two paired, continuous measurements in a before-after clinical trial design without the raw data?

For example, I am reviewing some clinical trials that measure drug-metabolising enzyme activity at baseline then again in the same participants following a change to their diet. Unfortunately I usually only get the mean, SD or SE of the baseline measurement, the post-intervention measurement or the difference in these two means and the SD or SE of the difference.

Is it possible to work out the covariance of these two measurements (activity at baseline, activity post-intervention) with just the mean, SD or SE of the respective measurements or alternatively with the difference in means and its SD or SE?

Thank you in advance for anyone who answers this as you'll rescue my literature review and by proxy my PhD thesis!