I would to apply the logic of Nested Logit choice models to an EWS for financial crises.

My work consists on finding a set of regressors able to describe the probabilities associated to the events depicted in photo. To infer the model I need a Nested Logit, the problem is that most of the softwares provide nested models for discrete choice problems, which use a utility function, something which seems unadequate for this kind of model.

So my question is: how should I handle a nested logit for discrete choices, in order to use it for my EWS? What are the mathematical passages which allow me to convert the model from a utility-based model to a general one?