Editors note: The above letter was referred to the authors of the original paper, and their reply follows. In reply: Dr. Landi and colleagues have raised certain issues related to the study we reported in September that might explain the results we obtained. The first is that ‘… case managers' home visits alone are insufficient….' We agree with this premise, which is why we did not test an intervention of home visits alone but, rather, one including the coordination of care within a multidisciplinary context and with extensive family involvement. The second concern voiced is ‘… the lack of active family physician participation…' Family physicians were, however, intimately involved in the intervention. The case managers were members of existing interdisciplinary teams in their respective community health centers. These teams consisted of community-based family physicians, nurses, psychogeriatricians or psychologists, geriatricians, social workers, occupational therapists, and dieticians. As such, family doctors were part of all case conferences and resource allocation decisions. Family doctors in the community that were identified in patients' records were automatically notified by the hospital at the discharge of the patient. If the family doctor was one affiliated with the hospital's Family Practice Center, he or she would have been notified upon admission of the patient and would also have been involved in the care of the patient during his or her hospital stay. Overall, the family doctors of most of the study patients were involved with the case managers in discussing and planning care (as were social workers and physical and occupational therapists). Dr. Landi and his colleagues go on to suggest that the reason their randomized trial showed beneficial effects was the ‘close collaboration between case managers, the community geriatric evaluation unit, and primary care physicians.' On the surface, the intervention they tested does not seem to be very different from the one we tested. We do, however, suspect that the nurse case managers hired for our study were not fully accepted as integral members of the healthcare team, probably because of the temporary nature of conducting research studies. They may, therefore, have had insufficient credibility and authority to facilitate the organization of services optimally on behalf of the individuals in their care. Nurse case managers were also not in a position to change the healthcare structures within which they worked. Another reason put forward for the success of the study of Dr. Landi and colleagues' was the intensive training provided to the case managers. Again, our two studies do not appear dissimilar. The training and on-going support provided to the nurse case managers in our study was extensive: 24 hours of initial training followed by weekly 2-hour case conferences with 24-hour per day on-call availability of a clinical supervisor. This support built on an existing geriatric knowledge base that was already strong. An additional important fact to remember when interpreting our study results is that a new method for coordinating the care of frail older people was being implemented province-wide at the same time this study was being conducted.