Processor sharing (PS) models nicely capture the bandwidth sharing and statistical multiplexing effect of TCP at the flow-level. However, these 'rough' models do not provide insight into the impact of packet-level parameters (round trip time, buffer size) on performance metrics such as throughput and flow transfer times. In a previous paper an analytical, integrated packet/flow model was developed, exploiting the advantages of PS approach at the flow-level, while incorporating the most significant packet-level effects. In the present paper we propose an enhancement of this model, which is validated through extensive NS simulations. The numerical results show that this enhancement leads to significantly better performance approximations than the originally proposed model.