Combining follow-up and change data is valid in meta-analyses of continuous outcomes: a meta-epidemiological study · Study