Assessment of therapy in patients with rheumatoid arthritis is important but difficult. We examined 4 different methods of analyzing pretreatment data and assessed the difference that each made in detecting a positive effect of intramuscular gold on the patient's overall disability. The methods were (1) calculating the arithmetic mean of prior data points, (2) taking the last data point pretreatment, (3) fitting a straight line to pretreatment points and (4) fitting the pretreatment points with a quadratic equation. After comparison with matched controls (not taking remittive agents) the most significant difference was found by fitting a straight line to pretreatment data. This technique demonstrated about one-third more of intramuscular gold's effectiveness than the usual technique of using the last data point pretreatment. We conclude that statistical power is improved by obtaining and analyzing longitudinal pretreatment data appropriately.