Two stages of a gender-specific treatment program for perinatal cocaine-addicted women were introduced into a coed peer-led day treatment program and evaluated for outcomes. Stage I (N = 21) targeted gender-specific needs, but did not expressly promote family reintegration. Stage II (N = 27) augmented the gender-specific program with a multisystems model for family reintegration. As a control for historical effects, nonperinatal clients whose treatment remained the same during the periods corresponding to the two stages, were evaluated for outcomes (N = 66 for Stage I, N = 75 for Stage II). As hypothesized, urine toxicology and retention data were significantly improved for perinatal clients treated in Stage II, as compared with those treated in Stage I; no such improvements were noted for nonperinatal clients. The data support a conclusion that introducing a multisystems framework into a gender-specific program selectively improves clinical outcomes for inner-city perinatal cocaine-addicted women.