A retrospective study of charts according to size: implications for management and prevention?

services provided by Oxford and Wessex Regions. British Medico/Journal. 296, 1779-1784. DAY, K. (1988) A hospital-based treatment programme for male mentally handicapped offenders. British Journal of Psychiatry. 153,635-644. INTER-REGIONALTECHNICALCOMMITTEE(1982) Psychiatric Care in Eight Regional Areas. Statistics from Eight Psychiatric Case Registers in Great Britain 1976-1981. University of Southampton. MARKS, I. & THORNICROFT,G. (1990) Editorial. Private inpatient psychiatric care. British Medical Journal, 300, 892. MARSHALL,M. (1989) Collected and neglected: are Oxford hostels for the homeless filling up with disabled psychi atric patients? British Medical Journal, 299, 706-709. MENTAL HEALTH ACT COMMISSION(1989) Third Biennial Report 1987-1989. London: HMSO. O'GRADY, (1990) The complimentary roles of regional and

A retrospective study of charts according to size: implications for management and prevention? BRIANO'SHEA,Consultant Psychiatrist, Newcastle Hospital, Greystones, Co. Wicklow, Ireland; and JANEFALVEY, Senior Registrar, Eastgate Unit, Middlewood Hospital, Sheffield In 1975 Sims published his celebrated paper on 'Factors Predictive of Outcome in Neurosis' (Sims, 1975). A big chart was associated with a poor prognosis. We were interested in determining the significance of having a big file. A prospective study, while superior in terms of information gathering, would take many years to complete and would not neces sarily reflect what happens in routine clinical practice. Therefore we decided to carry out a retrospective study. We predicted that large file patients would be characterised by extreme social disadvantage.

The study
We chose the In-patient Medical Records Depart ment of a large public urban district psychiatric hospital. The study was cross-sectional and retro spective. All charts were constructed in a similar manner and using the same materials. Chart size was defined using one pair of steel calipers throughout. The width of the chart at mid-spine was chosen as the accepted size of the file.
Thick charts were more than 1.5" in size, the patient's first recorded address belonged to the catch ment area of the hospital, and the chart was currently in the department. Thin charts were less than 0.5"in size, were adjacent to a thick chart, belonged to the hospital's catchment area, and were currently in the department.

Findings
The inclusion criteria were met by 56 thick charts (T group) and 64 thin charts (t). Of the T group, 19.6% were males, 80.4% being female (t = 37.5% and 62.5% respectively). Admissions averaged 25.7 (range 5-28) for females and 28.3 (7-53) for males in the T group (t = 4.3 and 7 respectively). The mean number of recorded offspring was 2.06 v. 1.91 for T and t patients respectively. One of the T group had children in care (a woman with four children in care), and also one (a male with three children in care) of the t group. The mean age at presentation for T females and males was 26.7 and 24.5 years respectively, compared to 31.2 and 34.1 years in the t group.
Of the T group, 48% were employed at the time of first admission, falling to 21% at the final admission; t group patients fell from 41% to 36% employed. For the sake of clarity, employed status included house wives and retired persons. The small drop in employ ment in the t group may be due in part to less marital separation and higher marriage rates, but there is also a real drop.
Known significant medical diagnoses (SMDs) were defined as recorded conditions which were seri ously disabling or were potentially so, e.g. systemic hypertension requiring treatment or diabetes mellitus. On this basis, 34 (61%) of the T group and 20 (31%) of the t group had recorded SMDs.
Of the T group, 20%, 16%,and 18% were reared in care, experienced early paternal death, and suffered early maternal death respectively (t = 0.0%, 8.0%, and 0.0%). Ten (18%) of the Tcharts and one (1.5%) of the t charts contained solicitors' letters, whereas all of the T charts and 31% of the t charts contained other letters (including social worker letters), not including routine admission and discharge letters.
The five most frequent diagnoses in the T group at first admission (absolute numbers) were personality disorder (16), schizophrenia (11), affective disorder (9), alcohol abuse (8), and mental handicap (5). The ranking order for the t group was schizophrenia (21), affective disorder (15), personality disorder (9), alcohol abuse (9), and reactive depression (6). The numbers of T patients with a diagnosis of either personality disorder or schizophrenia increased over time (to 20 and 16respectively), whereas diagnosis in the t group remained stable.
The numbers of patients having only one admis sion were: zero in the T group, 9 (14%) in the t group. Thirty-four per cent (19/56) and 16% (10/64) of the patients had different first and last primary diagnoses in the T and t groups respectively. Four secondary diagnoses were sufficiently represented to necessitate recording here (numbers in T/t groups in brackets): personality disorder (7/0), mental handicap (9/1), alcohol abuse (5/3), and other substance abuse (5/3).
Thin chart patients spend 6-10% as long in hospital as do thick chart patients.* 'More detailed information is available on request to Dr O'Shea.

Case vignettes
The following cases were picked out at random from the T and t groups with the exception of case no. T5 which was chosen because of the striking aspects of management.
(a) Thick Tl. Male (28 years of age), personality disorder and tem poral lobe epilepsy, 17 admissions, aged 23 years at first admission, single, no offspring, unemployed, numerous solicitors' letters and communications from a special hospi tal, many letters from general hospitals and from Inspector of Mental Hospitals, repeated attempts to enucleate eye (associated with epileptic aura), went through rehabili tation programme, now in supervised hostel, repeatedly seeks readmission to hospital.
T2. Female (30),. hysterical personality disorder, 14 admissions, aged 16 years at first admission, single, no offspring, unemployed, large quantity of general hospital letters, repeated self-injury by cuts and swallowing of objects, possible history of incest, born in index hospital.
T3. Female (35), schizophrenic, 10 admissions, aged 23 years at first admission, single, one child, originally a typist but now unemployed, many letters from community psy chiatric nurses and social workers, father died when patient 4 years old, orphanage 8-16 years of age, went through resocialisation project, now in hostel.
T4. Female (70), depression with obsessional features, 20 admissions, 61 years old at first admission, single, no offspring, retired teacher, ward of court letters, arthritic, previously in private psychiatric hospital.
T5. Female (57), personality disorder with depression and alcohol abuse, 38 admissions, aged 16 at first admission, single, no offspring, worked in laundry, now retired, letters from general and psychiatric hospitals, mother died when patient aged 3 years, reared in convent orphanage from age 3-13 years, leucotomy at age 16years (she stole 2 shillings a day from a factory and was promiscuous!), numerous admissions to other psychiatric hospitals.
t3. Female (27), acute schizophrenia/became chronic, two admissions, aged 26 at first admission, common law marriage, two children, unemployed, letters from social workers re children, whereabouts unknown.

Comment
Because Sims (1975) had included personality dis order in his criteria for a diagnosis of 'neurosis' (Sims, 1985),and because one theory of the aetiology of personality disorder is a social learning one (O'Shea & Falvey, 1988), we predicted that patients with the maximum utilisation of institutional services would have experience the greatest social disadvantage.
Thick-chart patients are characterised by being female, being admitted more often, spending more time in hospital, presenting earlier in life, being separated or single, losing employment during the course of their psychiatric career, having an excess of medical illness, being reared in care, and having lost a parent, especially the mother, early in life. They are more likely than thin-chart patients to have had con tact with the law and to have had significant inter action with other agencies, including social workers. Their psychiatric diagnoses are relatively unstable over time and the label of personality disorder is over-represented in this group. They may also be less intellectually bright and more likely to abuse drugs, including alcohol.
No consistent attempt was made to place persons with the above high risk characteristics into a special programme ab initio or to collect the necessary data in a systematic way.
The large-chart patients, apart from any humani tarian consideration, place a considerable burden on health service resources and represent a major chal lenge to any move away from chronic dependence on asylum care.
The excess of females in our study, while greater in the thick-chart group, is present in both groups, and may reflect social influences or a real excess of psychi atric illness in women, perhaps with an hormonal basis.
We strongly suggest that the criteria identified in this study be used for identifying patients at risk of O 'Shea and Falvey chronic psychosocial breakdown. A pilot study of a high intensity medico-social intervention, incorpor ating procedures such as social skills training and attention to self-image (Mylet et al, 1979) in psycho therapy, using only brief hospital or high-support hostel stay during psychosocial crises, would, we believe, pay dividends in the long run. Diagnoses in this survey were applied by doctors at all levels of experience, in varied settings, and using largely unspecified criteria. The Royal College of Psy chiatrists suggest that even juniors try to make an early diagnosis. We suggest that senior psychiatrists only should make diagnoses which carry a potential social stigma, and which tend to stick. Perhaps juniors could be encouraged to use differential diagnoses or formulations, or to emphasise the tentativeness of their diagnoses. However, the requirements of statistical agencies militate against delayed labelling.