SUMMARY This paper concerns a study of the outcome achieved following fieldwork responses to referred client problems within particular client categories. The evidence from the study demonstrates that while referred client problems initiate the social work response, administratively assigned client groups were a better determinant of outcome. It is shown that the category of a case is more directly related to outcome than is the problem referred. If social work is to be predominantly based on client categorization then polarization will result, represented by an increasingly professional child care practice and a more basic servicing role for the worker with elderly and handicapped people. This presents dangers for the organization of teams, the allocation of resources, and the training of social workers. Taking account of client problems as part of caseload manage ment should reduce this effect. The day to day fieldwork practice of social work in Britain is an import ant area for research into client problems. The need to reconsider the basic elements of practice is evidenced by the 'production of welfare' (Davies and Knapp, 1988) which emphasizes the need for 'efficiency' in community care and that efficiency is contingent on 'cost-effective' practices; yet if individual social worker client contacts are based on cost-effectiveness rather than benefit to clients, then the work achieved is of a second order. Renshaw (1988), in voicing the need for individual