Multi-dimensional Treatment Foster Care (MTFC), recently renamed Treatment Foster Care Oregon for Adolescents (TFCO-A) is an internationally recognised intervention for troubled young people in public care. This paper seeks to explain conflicting results with MTFC by testing the hypotheses that it benefits antisocial young people more than others and does so through its effects on their behaviour. Hard-to-manage young people in English foster or residential homes were assessed at entry to a randomised and case-controlled trial of MTFC (n = 88) and usual care (TAU) (n = 83). Primary outcome was the Children’s Global Assessment Scale (CGAS) at 12 months analysed according to high (n = 112) or low (n = 59) baseline level of antisocial behaviour on the Health of the Nation Outcome Scales for Children and Adolescents. After adjusting for covariates, there was no overall treatment effect on CGAS. However, the High Antisocial Group receiving MTFC gained more on the CGAS than the Low group (mean improvement 9.36 points vs. 5.33 points). This difference remained significant (p < 0.05) after adjusting for propensity and covariates and was statistically explained by the reduced antisocial behaviour ratings in MTFC. These analyses support the use of MTFC for youth in public care but only for those with higher levels of antisocial behaviour. Further work is needed on whether such benefits persist, and on possible negative effects of this treatment for those with low antisocial behaviour.