We study economic decision-making of 284 people with obesity and pre-diabetes who participated in a 6-months randomised controlled trial to control weight and prevent diabetes. To elicit preferences, we use incentive-compatible experimental tasks that participants completed during their medical screening examination. We find that, on average, participants are risk averse, show no evidence of present bias, and have impatience levels comparable to healthy samples described in the international literature. Variations in present bias and impatience are not significantly associated with variations in markers of obesity. But we find a significant negative association between risk tolerance and BMI and other markers of obesity for women. A 1 standard deviation increase in risk tolerance is associated with a 0.2 standard deviation drop in BMI and waist circumference. Impatience moderates the link between risk tolerance and obesity. We replicate the key finding of interaction effects between risk and time preferences using survey data from a nationally representative sample of 6,281 Australians with similar characteristics. Deviating markedly from the literature, we conclude that risk tolerance brings benefits for health outcomes if combined with patience in this understudied but highly policy-relevant population.