There is a marked disconnect between research and practice here. Clearly some organizations have found ways to improve follow-up rates. However, only a few quality improvement projects are reported. They rarely describe how the aftercare levels were achieved, and rigor is suboptimal, limiting validity, generalizability, and opportunities for best practice dissemination. In addition, much research on the topic remains descriptive (analyses of administrative data); such studies are also very limited for illuminating the mechanisms that produce better follow-up. Although these analyses will continue to be helpful for monitoring, substantial improvement of practice will depend on rigorous studies of innovative interventions.