Factors Associated With Multiple Readmissions to an Urban Public Psychiatric Hospital

Abstract

Discussion Although the incidence of patient assaults and property destruction declined significantly in the four months after the ban, no definite causal relationship linking this de dine tO the removal of caffeinated beverages from sale at the hospital can be established. Several external factors could account for our results. First, because the ban took effect on April 1, the onset of spring and in creased outdoor activities could ex plain the change in behaviors. How ever, examination of the same measures for four months after April 1ofthe previous yearshowed a slight increase in these behaviors, mitigat ing against a seasonal influence. Second, a shift in the index popu lation could have influenced our data. However, the only significant change in hospital census or diagnos tic mix recorded during the study period was a transfer out ofall devel opmentally disabled patients. Data from this group were therefore not included in the analysis. Third, although expectations may have changed behaviors and neither staffnor patients could be blinded to the change in caffeine sales policy, no one other than study team members was aware that behavioral data would be analyzed. In addition, it is also possible that the behavioral changes were part of an observed gradual long-term trend toward less use of restraint and seclusion at the hospi tal. However, no otherperiod within the 2 5 months showed a change with this degree of significance. Finally, no changes in the hospital formulary, prescribing practices, staffallocation, or staff-patient ratios occurred during the period studied. Taking into account the reported effects ofcaffeine on psychotic mdi viduals, and at least one other report of reduced use of behavioral re straints after caffeine withdrawal (6), it is tempting to associate the ob served decline in assaults and prop erty destruction with the change in caffeine availability. Restoring sales of caffeinated coffee and soft drinks to the hospital and observing behav ioral change might settle the ques tion, and we would suggest such a design for future studies. So far, members ofour treatment staff have rejected this suggestion, pointing out that any possible resulting in crease in assaults by patients would constitute a danger to other patients, as well as to themselves. The behavioral changes identified in this study are being monitored for stability. We expect that reduced availability of caffeine will continue to be associated with a reduced in cidence of aggressive behavior. As saultive and destructive behaviors are serious problems at psychiatric hospitals; removing caffeine from sale could be a simple and safe step