Irish psychiatric absconders : characteristics and outcome

When a patient goes missing from a psychiatric hospital much time and effort is expended in the subsequent search. Not only may the patient be at risk but so may members of the public. It is important from a management point of view to identify subgroups of patients most likely to abscond. Following studies published in the 1960s (Muller, 1962, Antebi. 1967), in the aftermath of the new open door policy resulting from the 1959 Mental Health Act, there was until recently a paucity of research on patients who abscond from psychiatric hospitals. Despite differences in definitions and methodology, find ings from more recent studies (Tomison, 1989; Falkowski et al, 1990; Farragher et al. 1996) largely concur with earlier work that was carried out. Absconders have been found to be signifi cantly more likely to be compulsorily admitted if they are male (Tomison, 1989) or from ethnic minorities (Falkowski et al, 1990). Farragher et al (1996) in the only Irish study published to date, found absconders significantly more likely to be compulsorily detained and showed a trend for them to be single males with a diagnosis of schizophrenia, personality disorder or alcohol dependence. The aims of this study were to define the socio-demographic and clinical characteristics of a group of patients who absconded from an Irish psychiatric hospital, to assess whether the profile of absconders is similar to British absconders, to identify commonly chosen desti nations, to record the rate of destructive behaviour to self or others, and to examine the outcome of these incidents.

When a patient goes missing from a psychiatric hospital much time and effort is expended in the subsequent search.Not only may the patient be at risk but so may members of the public.It is important from a management point of view to identify subgroups of patients most likely to abscond.Following studies published in the 1960s (Muller, 1962, Antebi. 1967), in the aftermath of the new open door policy resulting from the 1959 Mental Health Act, there was until recently a paucity of research on patients who abscond from psychiatric hospitals.Despite differences in definitions and methodology, find ings from more recent studies (Tomison, 1989;Falkowski et al, 1990;Farragher et al. 1996) largely concur with earlier work that was carried out.Absconders have been found to be signifi cantly more likely to be compulsorily admitted if they are male (Tomison, 1989) or from ethnic minorities (Falkowski et al, 1990).Farragher et al (1996) in the only Irish study published to date, found absconders significantly more likely to be compulsorily detained and showed a trend for them to be single males with a diagnosis of schizophrenia, personality disorder or alcohol dependence.
The aims of this study were to define the socio-demographic and clinical characteristics of a group of patients who absconded from an Irish psychiatric hospital, to assess whether the profile of absconders is similar to British absconders, to identify commonly chosen desti nations, to record the rate of destructive behaviour to self or others, and to examine the outcome of these incidents.

The study
St John of God Hospital provides in-patient psychiatric treatment and accepts referrals from all parts of Ireland.In addition, it serves a population of approximately 180 000 in the local catchment area of south-east County Dublin.The areas of the hospital studied included all hospital wards (two acute locked wards, two open rehabilitation wards, a psychogeriatric unit and an adolescent unit), two day hospitals and the occupational therapy department.
For the purposes of this study an 'absconder' was denned as a patient who was absent from the hospital building and grounds without permission from staff.Those who were missing, but subse quently found within the hospital grounds were excluded, as were those who failed to return from leave, and those discharged against medical advice.
Absconders were identified by reviewing the 24-hourly nursing reports for one calendar year.Data collection for each episode of absconding based on medical and nursing case notes included socio-demographic details, admission status, discharge diagnosis, time of leaving, destination, mode of return and outcome.Diag nosis was recorded using ICD-9 classification (World Health Organization, 1978).Data drawn from the hospital's computerised patient infor mation system were used to compare the study group with the hospital population (n=1237) for that year.The data were analysed using the Statistical Package for the Social Sciences (SPSS/PC+).

Socio-demographic characteristics
A total of 1771 admissions involving 1332 patients were recorded during the time period and 95 patients absconded a total of 156 times.Of the total 156 absconding incidents, females accounted for 55% (n=85) and males 45% (n=71).Forty-nine men absconded a total of 71 times.Twelve men absconded more than once.Fortysix women absconded a total of 85 times.Fourteen women absconded more than once.The average age of the absconders was 32 years for men and 32 years for females versus 35 years for men and 33 years for females in the control group.Seventy per cent of absconders were single compared with 34% of controls (P< 0.001).In addition, absconders were signifi cantly more likely to be compulsory patients (24%) when compared with controls (4%) (P<0.001).

Psychiatric diagnosis
The most common discharge diagnoses for the absconding group were schizophrenia (40%) and personality disorder (19%).These diagnoses were less prevalent among controls, 20 and 3% respectively.Diagnoses of schizophrenia and personality disorder were over-represented in the absconding group (x2=98.73,d.f.=5.P<0.001) with diagnoses of affective disorder, depressive neurosis and alcohol dependence under-represented compared to controls.

Time and place of absconding
The most frequent time of day for absconding was between the hours of 13.30 and 21.00, with 70% (11= 109) of patients leaving during this time.
Absconding was most likely to occur in the earlier part of the week, with the weekend having the lowest rate.The most common months for absconding were May, June, August and Sept ember, showing a definite summer peak that was statistically significant compared to the total number of admissions for each month (x2=36.4.d.f.= ll, P<0.001).The majority of patients left from the two open rehabilitation wards.

Outcome of absconding incident
Patients returned in 91% (n=142) of incidents.In 62% (n=97) of cases patients returned of their own volition, were accompanied by family or friends in 19% (n=30).or by the police in 6% (n=9) and by hospital staff in 5% (n=8).Of those who did return 80% came back within 24 hours.The most common destinations were the public house (28%.n=44), own home (20%, n=31) and friend's home (13%, n=20).It was not possible to establish the patient's destination in 41% (n=64) of cases.
There were no fatalities among the abscond ing group during the study year.Nine patients inflicted harm to themselves while absent from hospital: self-mutilation, self-poisoning, injury in a self-inflicted road traffic accident and hypothermia secondary to exposure.Two ab sconders physically assaulted others and five absconders damaged property outside the hospital.
On return four patients (2%) were transferred to a general medical hospital for treatment, medication was changed in 16% (n=25) of cases, patients were transferred to a secure ward in 15% (n=23) and in 7% (n= 11) status was changed from informal to formal.

Comment
Absconders when compared to the remaining hospital population were significantly more likely to be single, formally admitted patients with a diagnosis of schizophrenia or personality disorder.These findings are in agreement with the previous Irish study and show that the characteristics of Irish absconders conform largely with those in mainland Britain.We confirmed the most consistent finding that compulsorily detained patients have the highest absconding rate (Tomison, 1989: Falkowski et al 1990;Farragher et al. 1996) but failed to find a gender effect.This study replicates Muller's (1962) finding of a seasonal variation in ab sconding.Most absconders left at the time of maximum staffing levels showing that high staffing levels fail to deter absconders.
It is noteworthy that in other studies, most abscondences were from disturbed or locked wards but in this study most absconding took place from the rehabilitation wards.
Previous studies have found that 65 to 90% of patients returned.The 91% returning in this study is the highest documented.
When initiating a search for a missing patient, apart from carefully documenting all infor mation, informing relevant staff, checking the hospital and immediate environs and contacting relatives, we found that the public house may be a good place to start looking.A telephone call to the patient's own home or to that of a close friend is also useful.Nine absconders harmed themselves, two members of the public were hurt and property was also damaged.Apart from the more obvious harm that patients may inflict on themselves while absent, absconding may lead to a host of more subtle complications, for example the breakdown of trust with hospital staff, distress to relatives and disrup tion of the treatment process.Falkowski et al (1990) retrospectively described suicide in six patients within 18 months of absconding and this led to the suggestion that absconding during treatment may be a risk factor for subsequent suicide.This question deserves more attention.

Ethnicity: An Agenda for Mental Health Edited by Dinesh Bhugra
This book sets the scene for identifying and meeting the mental health needs of black and minority ethnic groups.Clinicians, researchers, academics, hospital managers, commissioners and voluntary organisation workers come together to discuss the problems in health care delivery and the way of moving the agenda forward.In addition to multi-disciplinary working the key emphasis here is in involving commissioners and voluntary organisations in deciding how best to meet the needs of the communities.