Introduction

A number of outreach services providing comprehensive specialist psychiatric assessment and care have been evaluated (e.g. Fenton et al., 1979; Stein & Test, 1980; Dean & Gadd, 1990). All achieved significant reductions in inpatient care and none of these ‘alternatives to hospitalisation’ studies found the hospital based service superior (Braun et al., 1981). The improved clinical outcome and consumer satisfaction with a possible cost saving demonstrated in Madison (Weisbrod et al., 1980; Stein & Test, 1980) has stimulated a number of replication studies of their ‘Training in Community Living’ approach (Hoult et al., 1983; Muijen et al., 1992).

Despite their consistency, these findings are slow to translate into policy. In the UK, reservations about their generalisability to local conditions persist. Firstly, British sector services must meet virtually all the mental health needs of their defined populations and some patient groups might not respond to experimental approach or perhaps suffer relative therapeutic neglect (Stefansson & Culberg, 1986). Secondly, study staffing levels are much higher than is customary here. Stein and Test documented a rapid loss of their gains after special funding was withdrawn (as did Langley et al. (1969) and Davis et al. (1972)). Thirdly, neither Stein and Test's nor Hoult's control services offered continuity of inpatient and outpatient care or any significant contribution from primary care services. Lastly, how much is due to novelty and the investigator's zeal? Were the experimental services based on work practices and levels of motivation that are compatible with durable service provision?