A critical stage in drought risk assessment is the measurement of drought hazard, the probability of occurrence of a potentially damaging event. The standard approach to assess drought hazard is based on the standardized precipitation index (SPI) and a drought intensity classification established according to a fixed set of SPI values. We show that this method does not allow for the assessment of region-specific hazards, and we propose an alternative method based on the extreme value theory. We model precipitation using an extreme value mixture model, with a normal distribution for the bulk, and a generalized Pareto distribution for the upper and lower tails. The model estimation allows us to identify the threshold value below which precipitation can be qualified as extreme. The quantile function is used to measure the intensity of each category of droughts and calculate the drought hazard index (DHI). By construction, the DHI value varies according to the specific characteristics of the left tail of the precipitation distribution. To test the relevance of our approach, we estimate the DHI over a gridded set of rainfall data covering West Africa, a large and climatically heterogeneous region. The results show that our mixture model fits the data better than the model used for SPI calculation. In particular, our model performs better to identify extreme precipitation in the left tail of the distribution. The DHI map highlights clusters of high drought hazard located in the central part of the region under study.