The first year experience of an innovative experiment in undergraduate medical education is described. The study investigated the educational effectiveness of prolonged clinical attachments for medical undergraduates in community hospital-based general practice. It has also assessed the ability of students to take some responsibility for their own learning in a clinically challenging environment. A retrospective evaluation of the experience obtained during the 3 month attachments for a self-selected group of fourth year Dundee medical school undergraduates was made. These undergraduates were placed in 10 mainly rural Scottish general practices with attached community hospitals providing a wide spectrum of inpatient and outpatient medical and surgical care. Students were assessed on the satisfactory completion of a portfolio of learning experiences and a practical clinical skills list. They were also required to submit a clinical project based on some aspect of their work during the attachment. The initial results showed a high degree of student and tutor satisfaction with the attachments. The assessment of all 10 of the students' educational achievements in their attachment were regarded as satisfactory and two were assessed as outstanding. Tutor assessment confirmed the validity of the initiative. Prolonged attachments in community hospital-based general practice for medical undergraduates have proved educationally valid and popular with both students and tutors. The development and dissemination of this model on a wider scale has resource issues which require to be addressed.

Plasma phenytoin levels were measured in 60 patients under steady-state conditions for a period of six weeks. During the trial, the preparation of phenytoin was changed from Phenytoin BP (Regent) to Epanutin Infatabs. A significant increase in plasma phenytoin levels following the change of tablet was matched by a decrease in the number of seizures.

A survey of 351 accidents occurring in a two-year period in a psychiatric hospital showed that 77 per cent of the accidents involved female patients and 48 per cent involved patients with organic psychoses; 236 accidents were falls and 280 occurred in the ward setting. In 277 instances adequate controls were available. Seventy-five per cent of the accident patients had received a psychotropic drug on the day of the accident as opposed to 61 per cent of the controls. The possibility that the side-effects of psychotropic drugs may have contributed to some of these accidents is discussed.

218 deaths occurring in a psychiatric hospital during a 3-year period were reviewed and their drug treatment was compared with that of a control group. 75% of the patients suffered from an organic psychosis. Patients received a mean of 2.4 different drugs on the day of death and only 23.6% of the prescriptions were for psychotropic drugs. 20 patients died suddenly and although more of these individuals had received tricyclic antidepressants and phenothiazines than their controls, these differences were not statistically significant.

Previous studies have shown varying rates of discrepancy in drug administration in psychiatric hospitals, and the present survey attempted to elucidate this problem further, using two methods of investigation simultaneously. Two wards in a psychiatric hospital and one ward in a mental subnormality hospital were studied. Urine specimens were tested for psychotropic drugs, and ward drug rounds were observed to detect discrepancies in administration. Of the 266 instances of drug administraion tested for in urine specimens, prescribed drugs were not detected in 7-9 per cent of cases and non-prescribed drugs were found in another 29 instances. Observation was made of 3,689 instances of drug administration, and in 1-7 per cent of these the drug was not given or taken as prescribed, whereas no instances of ingestion of non-prescribed drugs were observed. The reasons for the discrepancies and the different results obtained by the two methods are discussed.