We aimed to provide robust estimates, accounting for censoring and ascertainment biases.