Audit in psychiatry : " failed discharges "

Medical audit has been defined as the systematic, critical analysis of the quality of medical care, includ ing the procedures used for diagnosis and treatment, the use of resources, and the resulting outcome and quality of life for the patients (Department of Health, 1989). The medical profession has been under press ure to extend and improve audit procedures in recent years (McKee el al, 1989),but there have been doubts about the most satisfactory methods, particularly in psychiatry (Garden & Oyebode, 1989). There are numerous methodological problems in measuring the outcome of psychosocial care (Shaw, 1989; Royal College of Psychiatrists, 1989).Indicators of outcome which have been used in medicine include incidence of adverse events. Reintervention rates do offer some measure of outcome, and have been used widely in other medical specialities. In psychiatry, readmission to hospital within a relatively short period of discharge could be viewed as an adverse event requiring reintervention. Despite the fact that the majority of psychiatric admissions in the UK are readmissions, there has been a paucity of interest in "failed discharges" as a topic suitable for

Medical audit has been defined as the systematic, critical analysis of the quality of medical care, includ ing the procedures used for diagnosis and treatment, the use of resources, and the resulting outcome and quality of life for the patients (Department of Health, 1989).The medical profession has been under press ure to extend and improve audit procedures in recent years (McKee el al, 1989),but there have been doubts about the most satisfactory methods, particularly in psychiatry (Garden & Oyebode, 1989).There are numerous methodological problems in measuring the outcome of psychosocial care (Shaw, 1989; Royal College of Psychiatrists, 1989).Indicators of outcome which have been used in medicine include incidence of adverse events.Reintervention rates do offer some measure of outcome, and have been used widely in other medical specialities.
In psychiatry, readmission to hospital within a relatively short period of discharge could be viewed as an adverse event requiring reintervention.Despite the fact that the majority of psychiatric admissions in the UK are readmissions, there has been a paucity of interest in "failed discharges" as a topic suitable for medical audit.

The study
St Luke's Hospital is a traditional psychiatric hospital serving the district of South Cleveland, with a popu lation of approximately 300,000; there are 90 beds for acute adult psychiatry divided between fiveadmission wards.A case note survey was performed for all acute in-patients on one date in March 1990(n = 71) and for all subsequent admissions over two months (n = 44).For the total 115 patients the following information was noted: age; sex; marital status; employment status; accommodation arrangements; status with respect to 1983 Mental Health Act; main psychiatric diagnosis (ICD-9); other medical diagnoses; the number of previous psychiatric admissions; the date of last discharge; any recorded comments about the last psychiatric discharge (for example, against medical advice); any recorded comments about abnormality of premorbid personality.
"Failed discharges" were defined as patients who had been readmitted to hospital within three months of the last discharge date.For this subgroup case note examination was supplemented by an interview with the patient and/or ward staff to obtain further infor mation about the reason(s) for rapid readmission.

Findings
The 115 acute admissions comprised 28 (24%) first admissions and 87 (76%) readmissions, of which 24 (21% of total) were readmissions within three months or "failed discharges"."Failed discharges" showed a point prevalence of 27% (19/71) of the total acute in-patient population and an incidence of 11% (5/44) of total acute admissions over two months, suggesting that, once readmitted, "failed discharges" are likely to have longer in-patient stays than other patients.
The "failed discharges" (n = 24) could not be differentiated from the remainder of the total (first admissions plus readmissions after three months, (n = 91) or the readmissions after three months (n = 63) in respect of age, sex, marital status, employment status, accommodation arrangements, status with respect to 1983 Mental Health Act, main psychiatric diagnosis, or other medical diagnoses.However the "failed discharges" had had significantly more pre vious admissions than the remainder of the total (t = 3.606, df=113, P < 0.001) or the readmissions after three months (t = 1.981, df = 85, P< 0.05).They were also more likely to have been ascribed with abnormal personality traits (x: = 8.352, df=l, P < 0.01 and x: = 5.712, df = !,/><0.02 respectively).
The reasons for readmission of the "failed dis charges" are shown in Table I.The majority of patients re-presented with the same symptoms (83%) rather than a new illness.Failure of compliance with treatment or follow up arrangements was common (58%), as was failure of community support (54%).The single most common precipitant to rapid readmission was discord with key others (42%), followed closely by failure to take medications as prescribed (33%).Previous discharge against medi cal advice was a factor in four patients.Possible premature discharge due to pressure on beds was relevant to two patients.Unnecessary readmission through a doctor deputising service was a possible factor in two patients.

Comment
In psychiatry the percentage of readmissions com pared to first admissions has been increasing in recent decades, mainly because of changes in service pro vision.Readmission has been viewed as a sign of failure by both patient and psychiatrist, but it does not necessarily represent mismanagement and patients can benefit from readmissions.
There is no consensus on what time period defines a "failed discharge" which makes it difficult to compare rates; authors have adopted definitions of one month, three months, six months and even 12 months.According to the Camberwell Register, 9-14% of all admissions during 1965-70 were readmissions within one month of discharge (Marks, 1977).
Descriptions of "revolving door" patients have not always been consistent.Reported associations include younger age, male sex, unmarried status, marital or family discord, socially disruptive behaviour, and diagnoses of schizophrenia, manicdepressive psychosis, alcohol and drug dependency, and personality disorder.However, the best predictor of a failed discharge is probably the number of previous admissions.Audit of "failed discharges" would be useful if high risk groups of patients could be recognised as a result.This study confirmed that patients with abnormal premorbid personalities and more pre vious admissions were at greater risk of rapid 27 readmission.However no significant associations with demographic factors or psychiatric diagnoses were demonstrated.
Could audit of "failed discharges" lead to improve ments in patient care?It has been suggested that readmission should only be viewed as a failure of management if the readmission is for the same prob lem (Pablo et al, 1986).As the majority (83%) of "failed discharges" in this study re-presented with the same symptoms, possible ways of preventing such recurrences should be examined.The need for a water tight follow up system is highlighted by this study.Patients at higher risk of a "failed discharge" will require particularly vigilant monitoring, especially with regard to compliance with medication.Out patient follow-up could be supplemented by com munity psychiatric nurse involvement in such cases.Other authors have drawn attention to the fact that family support may help avoid readmission.The results of this study suggest that greater involvement of relatives in discharge plans would be a worthwhile preventive measure.

Conclusion
Rapid readmission to hospital ("failed discharge") is a topic worthy of medical audit.In this study "failed discharge" was associated with multiple previous admissions and personality abnormality, but not other demographic or diagnostic factors.Improve ment in the care of such patients might be effected by greater involvement of relatives and community workers in discharge planning and reinforcement of follow up arrangements.