If an already running process instance is waiting for a message from another organization, the corresponding FHIR Task may never arrive. Either because the other organization decides to never send the message or because some technical problem prohibits the Task resource from being posted to the DSF FHIR server. This would result in stale process instances that never finish.
At the same time, you might also expect to receive one out of a number of different message types at once.
In order to solve both problems we can add an Event Based Gateway to the process waiting for a response and then either handle a Task resource with the response and finish the process in a success state or trigger a Timer Intermediate Catching Event after a defined wait period and finish the process in an error state. The following BPMN collaboration diagram shows how the process at the first organization would look like if we wanted to react to multiple different messages or missing messages:
