Drs. Linden and Wilson note that physicians' awareness of our study may have influenced the care provided to study participants. However, any such effect was substantially mitigated by the fact that our study was conducted by an "outside" group of Stanford-based investigators without the incentive or means to provide feedback to Kaiser Permanente physicians or administrators about the quality of care provided. Given the nature of the excluded patients, an analysis of changes in medication prescription rates in this group would be difficult to interpret. As Drs. Linden and Wilson suggest, the evidence-based "baseline" care provided by Kaiser Permanente physicians for low-risk patients with heart failure might not have been amenable to improvement. However, our group has conducted a series of multicenter randomized trials in Kaiser Permanente hospitals in which care management produced superior outcomes (1-4). These trials focused on behaviors (exercise, smoking, diet, drug adherence) that were under the direct control of patients. As we noted in our paper, rehospitalization is only partly influenced by patients' adherence to medication regimens for heart failure. Indeed, most rehospitalizations in our study were not for heart failure but for coronary artery disease and other medical conditions. This underscores the need to address the multiple comorbid conditions associated with chronic diseases (5). We agree with Drs. Linden and Wilson that our study was not designed to evaluate the generalizability of our findings to settings other than health maintenance organizations (HMOs). Regarding the value of disease management in low-risk patients treated elsewhere, our conclusions speak for themselves: "Although nurse care management did not statistically significantly reduce the rate of rehospitalization compared with the Kaiser Permanente HMO–treated group, its potential value in non-HMO settings should not be diminished."

Drs. Linden and Wilson raise an interesting methodologic problem with evaluating nurse care management: Such studies rely on physicians' willingness to collaborate closely with the nurse care managers. It is our impression that physicians practicing in HMO settings are more willing to do this than those practicing in other settings. A future challenge is how to organize and conduct rigorous studies of care management for various chronic conditions outside of environments such as HMOs, where physicians place a high value on collaboration.