We examine public policy toward the use of genetic information by insurers. We study insurance markets with adverse selection due to individual differences in risk type and prevention efficiency, and where individuals can choose to be tested. Prevention gives positive clinical value to genetic information, which countervails the negative informational value of classification risk. We distinguish between the cases where prevention effort is observable and where it is not (moral hazard), consider four different policy regimes and rank them by the incentives to become informed. We also show that these policy regimes can be Pareto ranked, with a duty to disclose being the preferred regime and an information ban the least preferred one.