Companies often make this mistake. They extrapolate the aggregate event curve from the aggregate event curve seen during active enrollment. The active enrollment masks the tail of the curve - so as soon as enrollment stops the aggregate event curve slows down.