Testing a behavior/model

Dear senior researchers,
I am using a secondary data to test the Extended parallel process model(EPPM). Out of the 4 constucts of the model, no item has measured 1 construct (response efficacy). What options I have now to test the model on a behavior? I do have multiple items which have measured the other contructs of the model. Is there a way I can test this model to explain health behavior?
Thanks all in advance!