This paper summarizes a group of presentations and panel discussions on chronic insomnia at the 2001 NCDEU meeting. The presentations and discussions focused on the twin issues of efficacy and concerns of abuse liability with long-term hypnotic therapy. The panel concluded that insomnia may be an epidemiological marker for a variety of difficulties including accidents, increased health care utilization and subsequent development of major depression. Whether or not treatment of insomnia will prevent these long-term problems has not yet been determined. Since the mid-1980s there has been a rapid rise in the off-label use of antidepressants, particularly trazodone, for treating insomnia. Some participants expressed concern at the lack of data for this practice, particularly the absence of dose-response and tolerance information, and noted that the small amount of efficacy data available is not encouraging. Similarly, there are minimal data to support the use of antihistamines as sleep aids; moreover, their side effect profile and interactions with other drugs may be under appreciated. The limited data available on nightly long-term usage of the newer non-benzodiazepine hypnotics, primarily of six-months' duration, suggest an absence of tolerance, but more data for both nightly and non-nightly administration are needed. Insomniacs tend to show therapy-seeking, rather than drug-seeking behavior, and patients without histories of drug abuse are unlikely to self-escalate dosage of currently available hypnotics. There is fairly good agreement on the characteristics of an ideal hypnotic. All currently available agents, while effective and safe, do not achieve this ideal. The next few years are likely to see the appearance of a variety of agents with new and promising mechanisms of action.