We investigate adaptation of subjective well-being using a randomised controlled trial. We find that providing medical equipment to a random sample of Ugandan adults with lower limb disabilities has a positive effect on their physical health, using both objective and self-reported measures. Treated patients experience a significant improvement in life satisfaction initially, but the effect is not prolonged. After one year, life satisfaction returns to the pre-treatment level. This evidence of adaptation is supported by observations of changes in reference levels and is robust to alternative estimation methods including instrumental variable estimation and intention-to-treat analysis.