Recent psychiatric epidemiological studies using standardized interviews in the community have yielded high rates of non-psychotic disorders. The implications for service provision in terms of treatment and planning remain unclear. No methodology exists to link the individual needs for care and services to problems associated with disorders. The Needs for Care Assessment Schedule (NFCAS) is a relatively new procedure for assessing the needs of long-term mentally ill patients, mostly psychotic and attending psychiatric services. We report here a trial application of a modified version of the NFCAS on a sample of 39 non-psychotic patients, most of whom were attending psychiatric outpatient services. The results show that the modified procedure requires further refinement to achieve acceptability and reliability. Some improvements are suggested for refining items and for the collation of others. The difficulties encountered underline the key issues in developing such technology: specifying the threshold for recognizing the problems, detailing the interventions considered appropriate, defining the model of care and specifying the composition of the research team.