The present invention concerns methods for efficiently supporting Voice over Internet Protocol (VoIP) on the Forward Packet Data Channel (F-PDCH) in CDMA 2000 1xEV-DV systems. Active speech in VoIP is encoded using, for example enhanced variable rate codec (EVRC), which produces 171, 80 and 16 bits per 20 ms of speech for Rate 1, Rate ½ and Rate ⅛, respectively. However, about 60% of the time a user is inactive during a speech session, so an inordinate amount of system bandwidth is comprised of rate ⅛ VoIP packets. In one embodiment of the present invention the apparatus of the present invention identifies the Rate ⅛ voice frame packets and discards them. In another embodiment of the present invention, the apparatus of the present invention identifies the Rate ⅛ voice frame packets and selects some of them for further transmission. In both embodiments the efficiency of channel utilization is increased since the amount of channel band width used to communicate relatively little information, e.g., gaps of silence, is decreased.