Sir: I would like to respond to the concerns raised by Sandor (Psychiatric Bulletin, December 2001, 25, 486-487) regarding home treatment. He is correct to highlight the lack of a ‘strong evidenced-based rationale’. However, his focus on ‘model fidelity’ is, in my view, misplaced.

It is tempting to fault models of service delivery on this basis, but surely this ignores more important issues? Instead we should focus on the important factors like patients' clinical and social outcomes. Other factors like service retention, adherence and satisfaction levels should also be borne in mind.

To suggest that an identikit model can be used in vastly different settings seems unrealistic. This creates a problem insofar as it acknowledges that model fidelity is an improbable goal. None the less, I would refer Sandor to the editorial by Slade & Priebe (Reference Slade and Priebe2001), ‘the challenge is to make the important measurable’. We could see this as following the lead of naturalistic pharmacological research (i.e. examining real-life scenarios).

Therefore, I would suggest that those assessing the impact of home treatment should acknowledge the deficiencies as outlined by Sandor. But it is imperative that we embrace the challenge to measure what is important.