Of the 118 patients, 33.9% were found to be attending Narcotics Anonymous meetings, 44.1% were employed, 52.5% had not been arrested since leaving treatment, and 65.2% claimed not to be using drugs at the time of the interviews. Although 16% were known to be in jail and may be counted as failures of the program, the claims of 20.3% to have abstained from use of all illicit drugs since leaving Impact [Los Angeles, California, USA] was substantiated by independent verifiers. The major interest in studies of this type stems from any potential they have in suggesting ways to improve success in treating drug addicts. Such improvement could come from concentration on variables strongly associated with high success rates. These variables in the case of the Impact program describe personal characteristics of addicts. Other variables strongly associated with success at Impact were number of days of residential treatment and frequency of attendance at Narcotics Anonymous meetings after leaving the halfway house. The most likely way to improve the success rate at Impact is to reduce the high number of individuals dropping out, perhaps by improved counseling and treatment services during the 1st phase of treatment or by actual program changes. Another suggestion for improving the success rate may be to make attendance at Narcotics Anonymous functions more attractive or determine why they are not as attractive to some addicts as to others who attend meetings after leaving residential treatment. The small percentage of individuals now completing the program probably warrants the risks associated with program changes if they are limited to the first phase of the program.