Bias and power in group-based epidemiologic studies of low-back pain exposure and outcome - Effects of study size and exposure measurement efforts