Characteristics of admissions considered inappropriate by junior psychiatrists

Aimsand method Theaim of the studywasto identify medical and socio-demographic characteristics of admissions considered inappropriate by psychiatrists. Appropriateness of admission was assessed by questionnaire. Medical and socio-demographic characteristics of each admission were obtained by case note review. Results Twenty-five per cent of admissionswere considered inappropriate. Thesepatients showed high levels of adverse socio-demographic characteristics and commonly were diagnosed as personality disordered or as substance misusers.Junior doctors commonly admitted patients despite considering admission inappropriate. Clinical implications Regularaudit of the admission, process should be encouraged.

Admission to a psychiatric hospital is a complex process mediated by a number of behavioural, social and medical variables. Goldberg & Huxley (1980) proposed a model demonstrating how the patient, general practitioner and psychiatrist act as a series of filters in the presentation and subsequent treatment of mental illness.
Concerning the final filter by which the patient is admitted to a psychiatric hospital, adverse socio-demographic characteristics of the patient such as unemployment, divorce and medical characteristics including a past history of admis sion, parasuicide or psychosis may positively predict admission (Stein & Test, 1980;Hoult. 1986;Kent & Yellowlees, 1994). Attitudes and experience of the admitting doctor also influence the admission process (Meyerson et al, 1979;Flannigan et al, 1994).
Studies of the appropriateness of admission have mostly concerned general hospitals with psychiatric admission receiving little attention. However, Beck et al (1997) surveyed 481 psy chiatric admissions over six months in Notting ham and identified alternatives to hospitalisation for 29% of admissions, while Strumwasser et al (1991) analysed 539 admissions in Michigan, and suggested that 40% of bed days utilised by general psychiatric patients and 60% utilised by patients with a diagnosis of substance misuse were occupied inappropriately.
It is important to identify and avoid inap propriate admission. Much has been written from the patient's perspective regarding stigma tisation (Scambler, 1991), mÃ©dicalisation of problems and diminished coping strategies (Hoult, 1986) and ultimately institutionalisation (Goffman, 1961). Less attention has been paid to the cost of inappropriate admission. Hicks (1994) emphasises the importance of avoiding overuse of health services (patients being offered inappropriate health care) and underuse of services (patients not being offered appropriate health care). Both are related in that if beds are filled inappropriately, then those truly in need of admission miss out. particularly in a climate of bed storages. Both Flannigan et al (1994) and Beck et al (1997) suggest that some psychiatric in-patients could be treated in less intensively staffed community alternatives.
In the most recent comparison of costs of day versus inpatient treatment, Creed et al (1997) showed day hospital treatment to be a viable cheaper proposition for a number of potential admis sions, many of them people with schizophrenia.
The alms of this study are: first, to quantify the number of admissions considered inappropriate by the admitting psychiatrist; second, to estab lish the socio-demographic and medical char acteristics of these admissions and to compare them against those of patients whose admission was considered appropriate in a case-control study: and finally, to compare the length of stay of admissions considered inappropriate and appropriate.

The study
The study was performed in York, a psychiatric hospital of 80 acute beds covered by seven consultants and 13 trainees. All patients ad mitted over a four-month period (April-July 1996) were included, the only exclusions being those whose admission the author was involved in.
On admission the admitting psychiatrist re corded his grade and the time of the admission on a questionnaire. He allocated the patient to an ICD-10 (World Health Organization, 1982) diag nostic category and stated whether suicidal ideation was a factor in the decision to admit. He recorded whether he considered the admis sion to be appropriate. If inappropriate, he chose an alternative management strategy from a list of options, all of which were operational. The psychiatrists were blind to the aims of the study and incomplete questionnaires were not com pleted at a later date to avoid bias.
The author identified medical and socio-demographic characteristics of each admission by case note review. The length of each admission was recorded. The characteristics of admissions rated inappropriate (cases) were compared with those of admissions rated appropriate (controls). Two controls were included per case, the controls being the admission rated appropriate that preceded and followed each case. When con secutive admissions were rated as inappropriate, or when inappropriate admissions were sepa rated by a single appropriate admission, two controls preceding the first and two following the second were taken instead.

Findings
Two hundred and ten patients were admitted during the study period. Eighteen admissions were excluded due to the author's involvement in the admission process. One hundred and fiftyseven questionnaires were returned on the remaining 192 admissions (75%), with 152/157 (97%) completed by senior house officers or registrars.
Thirty-eight (24%) admissions were considered inappropriate. There was no significant differ ence in the proportion of cases and controls occurring out of hours (63 v. 59%, /2=0.17. P=0.7).
Cases showed higher levels of adverse sociodemographic characteristics compared to con trols (Table 1). High levels of significance were reached for being single, separated, divorced or widowed as opposed to being married or cohabit ing, for living alone, being of no fixed abode and for having no further education, a history of alcohol misuse, current debt or forensic involve ment. High levels of significance were not reached for unemployment/retirement or a history of drug misuse.

Outcome of admission
The mean length of stay for cases was 7.7 days (range 1-27 days, s.d.=6.4) compared with a mean of 35.7 days (range 3-124 days, s.d.=28.8) for controls (P<0.001). Table 3 shows the frequency of suggested alternative management plans for the 38 cases as a percentage of the total number of admis sions on whom questionnaires were completed (157). All four cases of schizophrenia were considered better placed in rehabilitation beds.

Comment
The response rate was satisfactory with ques tionnaires completed on 75% of admissions. One-quarter of admissions were considered inappropriate by the admitting psychiatrist, a finding similar to the Nottingham study by Beck et ai (1997). The proportion afeases and controls occurring out of hours was similar. Cases showed significantly higher levels of adverse socio-demographic characteristics than controls, and more commonly presented with personality disorders or substance misuse, with controls more likely to be diagnosed as suffering from severe mental illness. Cases more commonly had a past history of parasuicide, although they showed no significant excess of parasuicide or suicidal ideation at the time of admission.

Alternative management strategy Frequency
Acute day ward 13 (8.3%) Rehabilitation bed 9 (5.7%) Urgent out-patient appointment 2 (1.3%) Referral to community mental health team 1 (0.6%) Referral to social services 4 (2.6%) Other (e.g. medical referral, police) 9 (5.7%) The mean length of stay of cases was sig nificantly shorter than that of controls (7.7 v. 35.7 days). Although the number of bed days used by cases seems few. this translates as a considerable usage of psychiatric beds. Thirtyeight inappropriately placed patients staying for eight days each utilise 304 bed-days over a 120day period, or In this study these patients occupied 2.5 beds per day in an 80-bed hospital.
The study has limitations. No formal definition of appropriateness was employed and ratings were made by junior rather than senior psychia trists. The questionnaire was not formally tested for reliability or validity. However, no uniform definition of appropriateness is employed in previous studies and no rating scale for appro priateness has been developed for psychiatric use. In addition, one must consider that much of our judgement in psychiatry is subjective. The opinion of the junior psychiatrist is important since often he and the referrer act as gatekeepers to the hospital. A previous unpublished study in York demonstrated that 70% of admissions resulted from liaison between the general practi tioner and junior psychiatrist, with no consul tant involvement (further details available from the author upon request). This contrasts with Beck et afs (1997) finding that consultants or senior registrars made the decision to admit for 75% of admissions. At the time of the study bed occupancy levels in York rarely reached 100% and perhaps with less pressure on beds, con sultants were less likely to be involved in the admission process.
In the study by Creed et al (1997), a large proportion of patients with acute psychotic episodes were considered appropriate for the day ward, whereas in this study admission was considered appropriate for the vast majority of people with schizophrenia. No-one with schizo phrenia was considered better placed on the day ward. Again this may reflect local policy. In York the day ward caters more for patients whose psychotic symptomatology have been stabilised by a period of in-patient treatment or for the neurotic spectrum of disorders.
Short lengths of admission have been consid ered to represent low levels of illness and inappropriate admission (Flannigan et al, 1994). If so, junior psychiatrists appear to be good judges of appropriateness of admission, since most admissions considered inappropriate lasted only a few days. However, despite con sidering admission to be inappropriate and having the option of other services, junior psychiatrists repeatedly admitted patients. Per haps the ease of admission was attractive when alternative resources may be difficult and time consuming to access, or junior psychiatrists may consider particular patient groups threatening or unpredictable. Although statistical analysis did not reach significance, there is a suggestion that suicidal ideation was more common in cases than controls. It is possible that the expression of suicidal ideation acts as a form of talisman in the patient gaining admission. This along with previous suggestions warrants further study.
Perhaps the key point to emphasise is the importance of auditing the admission process. Flannigan et al (1994) suggested that clinicians need to identify those patient groups that benefit from admission with consultants offering super vision to more junior gatekeepers. The findings of this study support this view.