Lattice-based cryptosystems typically employ either direct modulation or its combination with error-correcting codes (ECC) for error correction mechanisms. In this paper, we propose a scalable nested lattice coding framework based on Barnes-Wall lattices that unifies modulation and ECC, enabling efficient and reliable error correction. Our design can serve as a drop-in replacement for direct modulation in lattice-based schemes, significantly improving overall robustness and communication efficiency. By carefully integrating low-dimensional lattice codes into high-dimensional module-lattice-based constructions, we incorporate this technique into MLWE-based key encapsulation mechanism and present an enhanced version of \(\mathsf {ML\text {-}KEM}\) , the NIST-standardized lattice-based KEM. We conduct a detailed analysis of the joint distribution of noise polynomial coefficients, from which we derive a methodology for evaluating the decryption failure rate (DFR) without any independence assumptions on the noise coefficients. Our scheme, \(\mathsf {BW\text {-}KEM}\) , offers stronger security guarantees, more compact ciphertexts, and significantly reduced DFRs across all security levels compared to \(\mathsf {ML\text {-}KEM}\) . Extensive benchmarking demonstrates that \(\mathsf {BW\text {-}KEM}\) outperforms \(\mathsf {ML\text {-}KEM}\) in key generation and encapsulation, while incurring only marginal overhead in decapsulation. For instance, at the highest security level, \(\mathsf {BW\text {-}KEM}\) achieves a 14-bit and 13-bit improvement in classical and quantum security estimates, respectively, or equivalently, a 16.33% reduction in ciphertext size, while maintaining significantly lower DFRs. These advantages are further complemented by an 8% and 12% acceleration in key generation and encapsulation, respectively, with only a 1% performance loss in decapsulation. Our results suggest that \(\mathsf {BW\text {-}KEM}\) is a promising candidate for practical and quantum-resistant key encapsulation mechanism.