The association between hospital admission and drug-related problems was evaluated in 285 consecutive admissions to two medical wards in a Swedish university hospital. Standardised definitions and criteria for causality were used. A drug-related problem was judged to have been the main reason for admission of 36 patients, and a strongly contributory reason for 9. These 45 patients comprised 16% of all patients, and 19% of those receiving medication prior to admission. For 19 patients the problem was considered to be failure to achieve the desired therapeutic effect. 11 of these 19 took less medication than prescribed, and an inadequate dose had been presented for the other 8 patients. In 26 patients there was an excessive or otherwise adverse effect. In 10 it was an intentional or accidental poisoning, and 16 had an adverse drug reaction. Non-compliance with the prescribed regimen caused almost half of the drug-related admissions: 11 took too little and 10 took too much of the prescribed drugs. The majority of the other problems could probably have been prevented by better application of pharmacokinetic principles to the prescribing.