This paper extends the ex ante moral hazard model to allow healthy lifestyles to reduce the probability of illness in future periods, so that current preventive behaviour may be affected by anticipated changes in future insurance coverage. In the United States, Medicare is offered to almost all the population at the age of 65. We use nine waves of the US Health and Retirement Study to compare lifestyles before and after 65 of those insured and not insured pre 65. The double-robust approach, which combines propensity score and regression, is used to compare trends in lifestyle (physical activity, smoking, drinking) of the two groups before and after receiving Medicare, using both difference-in-differences and difference-in-differences-in-differences. There is no clear effect of the receipt of Medicare or its anticipation on alcohol consumption nor smoking behaviour, but the previously uninsured do reduce physical activity just before receiving Medicare.