In cooperative networks, despite its simplicity, the decode-and-forward protocol suffers from the problem of error propagation when the source-relay channel is noisy. By forwarding erroneous decoded messages, the relay significantly degrades the diversity order of the system. One trivial solution was to restrict the use of decode-and-forward protocol to reliable sourcerelay transmissions. However, the forwarded messages from the relay can be seen as data correlated with that transmitted by the source. This property can be exploited in order to allievate the error propagation phenomenon. The problem of joint decoding of correlated sources is then directly transposed to the relay network. In this paper, we propose to exploit the potential correlation existing between the source and the relay messages in order to achieve enhanced end-to-end performance. Performance comparison with classical decoding of the decode-and-forward protocol is conducted especially under degraded source-relay channel conditions. We derive an analytical upper bound on the probability of error at the destination, which is verified by simulation results.