Heterogeneity in a meta-analysis: randomized controlled trials versus observational studies. Author's reply.