Aviation accident statistics provide essential information about the development of safety of air transport. They provide insight into the progress made by air transport industry and may indicate possible safety bottlenecks. Based on these insights, the aviation industry can set their strategy and priorities right. Important accident statistics are accident rate, trend and uncertainties based on recorded accident data. One established approach in jointly estimating rate and trend from accident statistics is to perform a regression analysis. However, the uncertainty is not estimated. Another established approach is not to bother about the trend, and just estimate rate and uncertainty for each year separately. This leads to rapid variations in year to year estimates. This paper overcomes the limitation of the established approaches by developing a Bayesian estimation approach for the joint estimation of accident rate, trend and uncertainty. Using worldwide statistical data on aviation accident and flights, it is demonstrated that the novel developed approach works well. Subsequently the novel approach is used to show that there is statistically significant evidence that since 2003 ground related accident rate tends to overtake lead from air related accident rate.