In late 1990, it became standard practice at the palliative care unit of the Edmonton General Hospital to regularly administer the Mini-Mental State Questionnaire (MMSQ) and to undertake opioid rotation and hydration upon detection of cognitive failure. We retrospectively reviewed the charts of 117 and 162 patients admitted in 1988-1989 and 1991-1992, respectively, to assess the impact of these maneuvers on the prevalence of agitated impaired mental status (IMS). All patients underwent regular cognitive assessment in 1991-1992 versus none in 1988-1989. Seventy-three percent of patients received hydration in the second period versus 32% in the first (P < 0.01). The frequency of opioid rotation was also greater in the second period (41% versus 21%, P < 0.001). The incidence of agitated IMS decreased from 26% in 1988-1989 to 10% in 1991-1992 (P < 0.001). This was reflected by a lower mean dose of the major drug used to treat this condition, haloperidol (3.6 +/- 2.4 versus 5.6 +/- 3.8 mg/day, P < 0.01), and less frequent use of other neuroleptics and benzodiazepines (0.12 versus 0.38 prescriptions per patient, P < 0.01) in the second period. Our data suggest that routine cognitive monitoring, opioid rotation, and hydration may reduce the incidence of agitated IMS in terminal cancer patients.