Harold & DeGarmo correctly refer to points regarding sample size and power that we already made in the discussion section of our paper. Despite this, we did point to the strengths of the study in the representativeness of the cohort within a real-world implementation setting, the fact that the study was conducted independently of treatment originators and UK implementation team, careful attention to triangulation and masked rating of primary outcome data (something often not undertaken in this kind of context), and the low attrition rates to endpoint. We stated that the convergence of findings from our mixed-method design and the confidence intervals of the outcome estimations gave some confidence to inferences from the results.

Harold & DeGarmo also question whether there was indeed a difference in the standard control condition (usual care) for participants in the US and UK studies. There are certainly likely to be differences in the nature and uses of group care between the two countries, given the differences in their child-welfare and juvenile-justice systems. However, the point we were making is that, in the USA, the MTFC programme for adolescents has been principally found to be successful when targeted at young offenders, in studies that have used a variety of measures of recorded reoffending to assess its effectiveness. ^{Reference Chamberlain1-Reference Leve, Chamberlain and Reid3} This emphasis on the effectiveness of MTFC-A with young offenders is also clear from the programme developers’ own website (www.mtfc.com). By contrast, the participants in our study were young people with complex emotional and behavioural difficulties, 93% of whom were in care because of abuse or neglect and less than a third of whom had a recent criminal conviction. The differences between the populations served by MTFC-A are clearly evident in an article comparing outcomes for high-risk adolescent girls written by the programme developers in the USA and their English colleagues ^{Reference Rhoades, Chamberlain, Roberts and Leve4} and may perhaps partly explain why the results of the English evaluation were less positive than those in the USA.