The regression effect as a neglected source of bias in nonrandomized intervention trials and systematic reviews of observational studies.

The regression effect as a neglected source of bias in nonrandomized intervention trials and systematic reviews of observational studies.