The National Service Framework for Mental Health (Department of Health, 1999) recommends the development of acute home treatment services as part of a spectrum of care to reduce reliance on hospital admission. The key components of this model are short-term acute treatment, mostly in patients' own homes, provided by an intensively staffed team available 24 hours a day. Several services of this kind have now been described in the UK (Reference Bracken and CohenBracken & Cohen, 1999; Reference Brimblecombe and O'SullivanBrimblecombe & O'Sullivan, 1999; Reference Harrison, Poynton and MarshallHarrison et al, 1999) with one controlled trial demonstrating reduced bed occupancy and costs, and improved integration with community teams (Reference Minghella, Ford and FreemanMinghella et al, 1998). Despite recommendations for home treatment to be more widely adopted, the existing literature provides little information about the type of patient suitable for this approach and the practicalities of the referral process.

Background

The Home Options Service in Central Manchester was established in March 1997 as a direct development from the psychiatric day hospital. The day hospital was already managing acutely ill patients as an alternative to inpatient care (Creed et al, Reference Creed, Black and Anthony1990, Reference Creed, Mbaya and Lancashire1997), but was limited by 9 a.m. to 5 p.m. opening hours and a focus on treatment occurring at the team base. Additional funding allowed an increase in staffing to provide 24 hour cover on a shift basis and a choice of treatment at the team base or in patients' own homes. The resulting service model is something of a hybrid between day hospital and home treatment, with a roughly equal number of contacts occurring at home and at the base.

Psychiatrists and community staff can refer patients to the service at any time if they are experiencing an acute mental health crisis that would otherwise require in-patient care. The Home Options staff must feel the level of risk to self or others is manageable and the patient must be able to offer a degree of cooperation. Short-term intensive treatment is offered in patient's homes and at the team base, with patients and carers able to contact the service at any time. There are 30 available places with no waiting-list in operation and a median length of admission of 35 days. Further details of service delivery have been described elsewhere (Reference Harrison, Poynton and MarshallHarrison et al, 1999).

Method

During a 6-month period from January to June 1998 all referrals to the service were tracked using a combination of case notes and local information system. Details of date, time and source of referral were collected plus immediate outcome of the referral and the reasons given by staff if the patient was not accepted. Clinical information included age, gender, ICD-10 diagnosis (World Health Organization, 1992) and previous contact with the service. Information was also collected for all in-patient admissions during the same time period and also for patients accepted by Home Options who subsequently needed to be transferred to in-patient care.

Results

Of the 195 patients referred to the service, 101 were accepted and 94 (48%) were refused. The most common reasons for non-acceptance were that the patient was unwilling to cooperate (22, 23%), was not considered sufficiently acutely ill (22, 23%) or was considered too ill (20, 21%), usually because of the degree of risk to self or others. Of the patients considered too ill, all but three were subsequently admitted to in-patient care, whereas only four of the patients not considered acute enough were admitted. In each case clear reasons could be identified why Home Options was not suitable, for example, the patient may have had a previous admission to Home Options that was not productive.

Female patients were more likely to be accepted by Home Options (58% of women accepted compared to 43% of men, P < 0.05) and the mean age of patients accepted was greater (36 compared to 32 years, P < 0.05). Patients were more likely to be accepted if they were referred by a senior doctor (72% of referrals accepted) rather than a junior doctor (32% accepted, P < 0.0001); if they were referred from the community or out-patients (73% accepted) rather than accident and emergency (28% accepted, P < 0.0001); if they were referred in normal working hours (58% accepted) rather than out of hours (36% accepted, P=0.01); and if they were already known to the service (58% accepted compared to 43% of those not already known, P < 0.05) (Table 1).

Table 1.

Outcome of Home Option referral

	Accepted n (%)	Refused n (%)	P value
Female patients	64 (58)	46 (42)	<0.05
Male patients	37 (43)	48 (57)
Mean age (years)1	36	32	<0.05
Referred by senior doctor	56 (72)	22 (28)	<0.0001
Referred by junior doctor	32 (32)	67 (68)
Referred from community or out-patients	66 (73)	24 (27)	<0.0001
Referred from accident & emergency	20 (28)	51 (72)
Referred between hours of 9 a.m. to 5 p.m.	81 (58)	59 (42)	<0.01
Referred out of hours	16 (36)	29 (64)
Previously known to service	68 (58)	50 (42)	<0.05
Not previously known	33 (43)	44 (57)

Diagnosis was aggregated into four groups: schizophrenia and related disorders; severe mood disorders (including bipolar affective disorder and psychotic depression); less severe depression/anxiety disorders; and personality disorder/substance misuse. Patients were more likely to be accepted by the service if they suffered from schizophrenia and related disorders (62% accepted) or severe mood disorders (77% accepted) and less likely to be accepted if they suffered from milder mood disorders or anxiety (38% accepted) and personality disorder or substance misuse (18% accepted) (P < 0.001). Most of the patients with schizophrenia and related disorders and with severe mood disorders who were not accepted by Home Options were admitted to in-patient care instead, whereas for the less severe mood disorders and personality disorder/substance misuse, the most likely outcome was for other follow-up arrangements to be made (Table 2).

Table 2.

Outcome of referral to Home Options by diagnosis

	Accepted n=97	Refused but admitted to hospital n=39	Refused and not admitted to hospital n=46
Schizophrenia and related disorders (n=76)	47 (62%)	16 (21%)	13 (17%)
Severe mood disorders1 (n=30)	23 (77%)	6 (20%)	1 (3%)
Less severe mood disorders/anxiety (n=65)	25 (38%)	15 (23%)	25 (38%)
Personality disorder/substance misuse (n=11)	2 (18%)	2 (18%)	7 (64%)

Using logistic regression analysis with forward stepwise regression, the strongest predictors of acceptance by Home Options were diagnostic group (P=0.0005), location of referral (P=0.0007) and the identity of the referrer (P=0.0048).

Twenty (20%) of the patients accepted by Home Options later had to be transferred to in-patient care. This included eight out of 23 (35%) patients with severe mood disorder, nine out of 47 patients with schizophrenia (19%) and three out of 27 patients with other diagnoses (11%) (χ² P=0.05). The other variables that predicted original acceptance by the service (time and source of referral, previously known to the service age and gender) did not appear to influence whether patients later required in-patient care.

Two hundred and three patients were admitted to in-patient care during the same 6-month period, of whom 43 (21%) had been assessed and refused by Home Options immediately prior to admission. Forty-six per cent of patients admitted to in-patient care were detained under the Mental Health Act. Using the same diagnostic groupings, patients admitted to in-patient care had very similar diagnoses to those admitted to Home Options (Table 3).

Table 3.

Diagnosis for patients admitted to Home Options and in-patient care

	Successfully treated by Home Options1 n=78 (%)	Admitted to in-patient care n=184 (%)
Schizophrenia and related disorders	38 (49)	101 (55)
Severe mood disorders	15 (19)	33 (18)
Less severe mood disorders/anxiety	23 (29)	39 (21)
Personality disorder/substance misuse	2 (3)	11 (6)

Discussion

Home treatment is being advocated as an important component of comprehensive mental health services (Department of Health, 1999) but little has been written about the practicalities of running a home treatment service. In keeping with other reports (Reference Brimblecombe and O'SullivanBrimblecome & O'Sullivan, 1999), our data suggest that a high proportion of patients referred will not be accepted and account needs to be taken of the amount of staff time spent assessing patients who are not considered suitable. As yet there has been little national debate about whether general practitioners should be able to refer new patients directly to a service of this kind, but we would require significant extra staffing to deal with the increased referrals such a change would generate.

As decisions about acceptance to a service of this kind shift away from doctors, non-medical staff involved in assessing patients need training in risk assessment and carefully considered policies and procedures for decision-making. The outcomes demonstrated for patients considered either too ill or not ill enough for our service suggest staff are making correct decisions about referrals and that many of the referrals received were not appropriate. This was particularly likely to be the case for patients referred from casualty by junior doctors. Emergency psychiatric services in the UK remain highly dependent on both casualty and junior doctors (Reference Johnson and ThornicroftJohnson & Thornicroft, 1991) and it is therefore not surprising that in the absence of other back-up services and with in-patient beds hard to find, junior doctors appear to have a lower threshold for referral.

Diagnosis appears to exert a strong influence on outcome for patients referred to home treatment. In general, patients with severe disorders were more likely to be accepted by Home Options and also more likely to require in-patient care, either because the referral was refused or the patient was accepted and later required transfer to in-patient care. This suggests appropriate targeting, with over 70% of the patients accepted by Home Options suffering from schizophrenia and related disorders or severe mood disorders, an almost identical figure to the in-patient unit. Home treatment studies in Bradford (Reference Bracken and CohenBracken & Cohen, 1999) and Birmingham (Reference Minghella, Ford and FreemanMinghella et al, 1998) have reported similar figures, though the service in more rural Hertfordshire treated a much lower proportion of patients from these categories (Reference Brimblecombe and O'SullivanBrimblecombe & O'Sullivan, 1999).

Patients with severe mood disorders may present particular difficulties in home treatment. Although this group of patients was most likely to be accepted, it was also the most likely to later require transfer to in-patient care. Other studies have also reported difficulty treating patients with bipolar affective disorder, particularly those who have mania, in home treatment (Reference Bracken and CohenBracken & Cohen, 1999; Reference Brimblecombe and O'SullivanBrimblecome & O'Sullivan, 1999) or a day hospital setting (Reference Creed, Black and AnthonyCreed et al, 1990), suggesting particular attention should be paid to the initial assessment and treatment plan for such patients.

Patients with personality disorder or who misused substances were the least likely to be accepted by Home Options and the least likely to require in-patient care either at the time of the original referral or as a transfer from Home Options. Similarly, patients with personality disorder were least likely to be accepted by the home treatment service in Hertfordshire (Reference Brimblecombe and O'SullivanBrimbecome & O'Sullivan, 1999). Patients with personality disorder can be difficult to manage in a home treatment service as the 24 hour accessibility of staff may lead to an abandonment of adult coping mechanisms and an increase in maladaptive behaviours. With time we have successfully managed some patients with personality disorders presenting in crisis by adopting firm boundaries at the start of the admission, for example a fixed duration of treatment and degree of contact.

In the light of these findings we intend to provide more information and training for junior doctors about the role of Home Options, to emphasise that all potential referrals should be discussed first with a senior doctor and to bid for additional funding to increase the range of services available for patients presenting in crisis out of hours, who would not usually warrant in-patient care or admission to Home Options.