The temporal reasoning of the HDP was implemented a couple of years ago and has gone through one evaluation and soon will go through another. The addition of temporal reasoning has eliminated a major class of the errors that were being made before, such as findings with chronicity longer than the proposed cause or ignoring possible causes because some part of the causal chain was no longer true. Examination of the hypotheses produced by the HDP make sense to cardiologists and do not reveal any shortcomings in the temporal reasoning process.
The domain of heart disease provides a good test bed for developing temporal reasoning in a diagnostic context. There are a wide variety of temporal situations that require explicit reasoning to handle properly. Thus, our expectation is that the lessons learned in this domain will be transferable to a number of other domains.