In Reply: We thank Dr. Kumar and his colleagues for their interesting comments on our crowding study. One of the potential problems with research on the consequences of ward crowding is that it has to depend on natural fluctuations in patient numbers; an experimental approach with patients would be unethical. Also, in clinical practice, variations in patient numbers are usually kept small; that is, most beds are occupied. For these reasons, it is hard to get a clear picture of the strength of the association between ward crowding and aggression.
The study by Lanza and associates found no association between ward density and violence, but the variance in patient numbers was rather low, with ranges of 39 to 44 patients on ward A, 36 to 41 on ward B, and 34 to 42 on ward D. In Hardie's study, admission rates fluctuated from ten to 17 patients and from 12 to 16. Again, no connection between the number of patients and the frequencies of individual aggression was found. But all patients in that study had private—and lockable—bedrooms, which may have counteracted the effects of ward crowding.
In our study, weekly data were aggregated, which increased fluctuations in occupancy rates and aggression frequencies. Calculated that way, the correlation between crowding and the number of incidents per patient was modest (r=.21). This correlational approach clearly has disadvantages; for example, the choice for weekly clusters is arbitrary. Yet the outcome variable is not dichotomous, as Dr. Kumar and his colleages suggest, nor was it in the previous studies.
As to the extension of ward space, our study relied on a ward rebuilding that was already scheduled. Although patients could and did use the courtyard any time they wanted, the modest extension of ward space did not produce a decline in aggressive incidents. As far as the antecedents of inpatient aggression are concerned, indexes of social space and privacy, such as the daily hours patients can spend alone, the number of undesired interactions, and the number of private bedrooms, might be more relevant than just physical space.