We would have liked to include the nurse-physician collaboration variable or a proxy. However, it does not exist in any database as far as we know and it was not possible to collect these data in this study. There are also varying definitions of MD-RN collaboration. Ms. Carroll is correct about the age of the database but it was what was available at the time of the study. We now have the 1994-1998 CHOP data and plan to repeat the study.

We apologize for the typographic error in Table 3. The nonunion hospital number should read 237. The reader is correct that some key characteristics were not explored in our study. Unfortunately, that is one of the limitations of using secondary data. We believe we included a number of conceptually relevant variables, although, as indicted by the overall R-squared, there is a substantial amount of variation not explained by the models.

Jean Ann Seago, PhD, RN

Michael Ash, PhD