This paper proposes a data envelopment method to separate avoidable and unavoidable mortality risks. As unavoidable mortality is the result of nature, only avoidable mortality is of relevance in measuring wellbeing and inequality. The new method is applied to a dataset consisting of life tables for 191 countries in the year 2000 to obtain a reference distribution of unavoidable mortality risks. The reference distribution is used to improve on the standard age-at-death measure to obtain an age-at-avoidable-death measure. Comparing with the original measure, age-at-avoidable-death provides a very different picture of wellbeing, and more so when it comes to inequality measures.