Hippisley-Cox et al (2003) described a survey of the results of various statins on serum cholesterol in general practice. They found that the serum cholesterol for 554 patients on atorvastatin to be 4.99 mmol/l (95% CI 4.90 to 5.09). Suppose a new statin came on the market and we wished to design a trial to see if it was an improvement on this. Assume a difference of 0.5 mmol in serum cholesterol is worthwhile. How many patients would be needed for a two-sided signifi cance of 5% and 80% power?