For clinical trials in which the outcome variable is trichotomous (regression, no change, improvement), a measure of net improvement in health status is proposed. Test statistics are derived for the null hypothesis of zero net improvement in a single patient group and for the null hypothesis of no difference in net improvement between two groups. The methods are applied to a clinical trial that evaluated the effect of argon laser photocoagulation in patients with macular edema in branch vein occlusion.