When receiving personalized rather than population-based information, agents improve their knowledge about their probability of experiencing adverse events (e.g. health shocks). Being revealed as high or low risk, they may revise their willingness to pay (WTP) for prevention programs. If the WTP changes of the high- and low-risk individuals go in opposite directions, the overall impact on the WTP for prevention depends on whether the relationship between WTP and the initial probability of damage is convex or concave. We address this question in a laboratory experiment. Participants received an endowment and were exposed to a non-financial damage—consisting in electrical shocks—with a certain probability. We elicit subjects’ WTP for self-protection and self-insurance, i.e. actions reducing respectively the probability and the number of shocks, using the Becker-DeGroot-Marschak mechanism. Our results suggest that WTP for self-protection is insensitive to the baseline probability to undergo pain, but reveal that WTP for self-insurance increases at a decreasing rate with this probability. This implies that the diffusion of personalized information should reduce the demand for self-insurance programs.