Background: Family members affected by a close relative's substance misuse constitute a large but neglected, high-risk group. Interventions for that group are now being developed and tested in trials. More use should be made of qualitative methods to study process in such interventions. Design: Semi-structured interviews with family members 12 weeks after assignment to a full or brief 5-Step intervention. Both involved a primary care professional giving a self-help manual, and the full version also involved up to five sessions of a psychosocial intervention. Participants: 143 patients of GPs, health visitors and practice nurses in general practices in. two areas of England. Patients were the family members of close relatives with drinking or other drug problems. Analysis: Framework analysis and grounded theory analysis of post-interview reports. Results: There was strong support for a form of intervention that involved face-to-face discussion with a primary care professional, and those who received the full intervention were appreciative of being able to talk to a professional who had time to listen and who appeared interested, understanding and caring. The self-help manual itself was reported to contain active ingredients for change, and a number of family members described transformations in their ways of coping with the problem whether they received the full or brief form of intervention. A common constellation of changes included increased consciousness of the nature and extent of the relative's drinking or drug use and its family effects, an acknowledgement of the family member's own needs and rights, a strengthening of resolve to assert plans and expectations, and a calming effect with reduction in stress symptoms. Many participants were unable to describe changes, however, and the following principal limitations of the intervention were described: prior familiarity with the material, perception that the intervention did not go far enough, belief that it was incapable of effecting change for the substance misusing relative, and a perception that sufficient professional expertise or sympathy was not always available in primary care. Conclusions: Findings suggest that the 5-Step family intervention has positive potential in the primary care setting, but has limitations and may need strengthening for family members whose problems are of longer standing or who have already been exposed to relevant information.