To the Editor: D.P. Devanand, M.D., and colleagues (1) reported successful control with haloperidol of severe behavioral disturbances in outpatients with Alzheimer’s disease. They rightly claimed that "the efficacy of nonpharmacologic interventions…remains to be established" (p. 1513). We wish to support the efficacy of nonpharmacologic interventions on behavioral disturbances with personal data taken from a controlled European study of demented individuals who were admitted to nursing homes.

Special care units for behaviorally disturbed patients with dementia are largely, although not exclusively, based on nonpharmacologic interventions such as appropriate staff attitudes and specific environmental features and have been proposed as an effective model of care. In 1995, the Regione Lombardia of northern Italy funded the establishment of a number of special care units in long-term care facilities. A pilot study show that care in special care units could effectively reduce behavioral disturbances with no increased resort to psychotropic drug load or physical restraints (2). A controlled study has recently been completed that confirms these preliminary results (3).

Patients who were capable of walking and had moderate to severe dementia (Mini-Mental State examination mean score=7, SD=5) and severe behavioral disturbances on the modified Neuropsychiatric Inventory (maximum score=108) were enrolled in 18 special care units (patients, N=39) and 25 traditional nursing homes (control subjects, N=41) (3). The patients were assessed at baseline (10 days after admission) and after 6 months. Although at baseline the patients had more severe behavioral disturbances (modified Neuropsychiatric Inventory mean score=37, SD=18, and mean score=28, SD=12, for patients and control subjects, respectively) (p=0.02), at the follow-up examination, these disturbances had significantly improved in both groups (decrease of 38% and 41%, respectively) (p<0.0001); there was no increase in the percentage of patients taking neuroleptic or other psychotropic drugs. Extrapyramidal signs were not assessed, but proxies of adverse effects of psychotropic medications (cognitive performance and falls) remained unchanged. It is noteworthy that in the patient group, the reduction of behavioral disturbances was achieved with a lower use of physical restraints: 10% of patients and 32% of control subjects (p=0.02) had to be restrained with a chest vest or a belt at follow-up.

Although implementing and testing the efficacy of environmental interventions in outpatients such as those in the study by Dr. Devanand and colleagues is doubtless more difficult than with institutionalized patients, we believe that environmental interventions should be the first-line option for inpatients, whereas they need further investigation in outpatients.