Abstract:
Amongst the most significant problems to confront adult mental
health services are, firstly, how to provide sufficient acute beds,
and secondly, how to evaluate different service models.
Because of the multiple relevant variables and the pressures
to provide according to both needs and demands, models of
service delivery, as opposed to specific treatments, appear
rarely to be evaluated by randomised controlled trials. The
more common technique, and very much a second best, is to
use historical controls. An alternative is to use groups matched
on as many of the variables as possible, and to apply critical
inspection of the available data.
