PURPOSE: This paper aims to examine whether a computerised system for medication reviews can support physicians' decisions and improve the quality of drug treatment in the elderly.

DESIGN/METHODOLOGY/APPROACH: This is a descriptive intervention study. The study included 275 patients living in community settings and nursing homes in Stockholm, Sweden. Patient data were analysed using computer software and scrutinised by a clinical pharmacologist. Pharmaco-therapeutic advice was sent to the physician responsible for each patient. The main outcome measures were initiation and discontinuation of drugs, changes of doses and rates of identified drug-related problems.

FINDINGS: Expert opinions were given by the clinical pharmacologist, for 275 patients, mean age 85 years; 70 per cent female. An average of 3.3 remarks was given concerning unsuitable drugs, unclear indication, dosing when the kidney function was decreased, drug-drug interactions and quality indicators. On average 1.5 drug-related problems (DRP) per patient were attended to by the responsible physician at each unit. The most common action taken was withdrawal of a drug (n = 208). On average the drug use decreased from 10.4 to 9.5 drugs per patient, and several quality indicators were met. The drug costs decreased, and resulted in a more cost-effective drug therapy.

ORIGINALITY/VALUE: The paper develops and tests a method for intervention in the care of elderly patients. The method is based on a computerised expert support system for medication reviews at a distance and on education of the staff. A safer drug therapy with improved quality and cost-effectiveness is thus provided.