Such an assumption should be explicitly been represented in the prediction model. Cutting off the time window of a model at an arbitrary period of 2 months is ambiguous at best when the model is about estimating the cumulative outcome in time.
The point of these models is to estimate peak utilization of health facilities, not to predict outcomes of different remedies after the initial suppression period.
