Main findings

We investigated ethnic inequalities in involuntary admission under the MHA and prior clinical care received in a cohort of people with a first hospital admission, in a 13-year observation window. Results show that 10 of the 14 minoritised ethnic groups had higher rates of involuntary admission than White British people, with the exceptions being White Irish people and people in the any Mixed race background group. Additionally, we observed that a higher number of SLaM appointments in the 12 months prior to admission (as compared with zero appointments) was related to higher rates of involuntary admission. Home treatment was also related to higher rates of involuntary admission. Care from an early intervention for psychosis team was not consistently related to involuntary admission. Having received psychological therapies and having a care plan was associated with reduced odds of compulsory admission.

There is evidence of some ethnic inequalities in clinical care and the impact of ethnicity being mediated by clinical care. However, although statistically significant indirect effects were observed, the magnitude of the ethnic disparities in involuntary admission was not substantially explained by differences in clinical care prior to admission.

Ethnicity and involuntary admission

The magnitude of the observed ethnic inequalities in involuntary admission is similar to what has been reported in recent meta-analyses.^{Reference Barnett, Mackay, Matthews, Gate, Greenwood and Ariyo8} Differences observed, compared with previous literature, relate only to the Other White and Mixed race ethnic groups. One study observed that controlling for diagnosis, age and clinical team, men of Other White backgrounds did not have higher odds of involuntary admission than White British men.^{Reference Mann, Fisher, Major, Lawrence, Tapfumaneyi and Joyce20} However, that study included only people within the first year of receiving care from early intervention for psychosis services, and the comparision group included people who were not hospitalised. Other studies have reported that people of Mixed race ethnicity have almost twice the odds of involuntary admission,^{Reference Weich, McBride, Twigg, Duncan, Keown and Crepaz-Keay21} and Mixed race women attending earlier intervention services for psychosis also had higher involuntary admission rates.^{Reference Mann, Fisher, Major, Lawrence, Tapfumaneyi and Joyce20,Reference Weich, McBride, Twigg, Duncan, Keown and Crepaz-Keay21} Again, one major difference between these studies and ours is that their comparison groups include people who were not hospitalised, while our cohort is restricted to people in their first hospital admission episode.

The pattern of ethnic inequalities in involuntary admissions, with only 4 of the 14 minority ethnic groups not having a higher risk than White British, and those groups being White Irish, White and Black African, White and Black Caribbean and Other Mixed race, suggests that social and cultural differences could be affecting the experience and outcomes of care.^{Reference Nazroo, Bhui and Rhodes22} There is evidence that people who face multiple forms of disadvantage (such as poverty, poor literacy, unemployment, poor housing and discrimination) may be in a less favourable position than White British people to negotiate their care.^{Reference Lawrence, McCombie, Nikolakopoulos and Morgan23} Previous experiences of discrimination and associated greater mistrust in mental health services may also affect the quality of therapeutic relationships.^{Reference Henderson, Williams, Gabbidon, Farrelly, Schauman and Hatch2,Reference Lawrence, McCombie, Nikolakopoulos and Morgan23,Reference Clement, Williams, Farrelly, Hatch, Schauman and Jeffery24}

Prior clinical care and involuntary admission

Our results suggest that ethnic disparities in involuntary admission are not because of lack of engagement with services. We observed that people with 6 to 11 appointments, as compared with those with no appointments, had higher odds of involuntary admission (although the magnitude of the effects was small). These findings are against our hypotheses and differ from previous evidence in the literature.^{Reference Rodrigues, MacDougall, Zou, Lebenbaum, Kurdyak and Li4} The associations may reflect that illness severity or ongoing clinical crises are undermining the efficacy of care in preventing involuntary admission. A possible explanation may be related to the diversity of teams providing care in the community, with potentially less continuity of care and a weaker therapeutic alliance. A previous study showed that lack of continuity of care is detrimental to clinical outcomes.^{Reference Macdonald, Adamis, Craig and Murray25} Another likely explanation is that the provision of appointments is dependent on perceptions of risk – an unmeasured factor that is highly related to involuntary admission.^{Reference Rodrigues, MacDougall, Zou, Lebenbaum, Kurdyak and Li4,Reference Walker, Barnett, Srinivasan, Abrol and Johnson26–Reference So, Wierdsma, Kasius, Cornelis, Lommerse and Vermeiren28} Thus, patients perceived to be at higher risk are provided with more appointments, albeit that an increased number of appointments is not contributing to reduced odds of involuntary admission.

Home treatment is associated with increased risk of involuntary admission. This finding is unexpected, as home treatment is a way to reduce admissions^{Reference Durcan and Harris1} and a previous review of randomised controlled trials suggests that receiving care from a crisis resolution team is associated with reduced involuntary admissions.^{Reference Bone, McCloud, Scott, Machin, Markham and Persaud13} However, there is also evidence of poor experience of home treatment, with some people reporting negative experiences with the teams and the need for more funding for these teams.^{Reference Durcan and Harris1}

Psychological therapies were associated with reduced rates of involuntary admission. A recent review of the use of CBT to prevent involuntary hospital admission revealed no significant protective effects on preventing compulsory admission.^{Reference Bone, McCloud, Scott, Machin, Markham and Persaud13} However, the studies included in the review were randomised controlled trials with people with psychosis only. It may be that patients with more complex needs, or acute symptomatic presentations, who are the ones more likely to be involuntary admitted are not getting access to psychological therapies or are refusing them. Concomitantly, a potentially positive effect of psychological therapies can be associated with greater opportunity to develop self-management, a strategy associated with reduced risk of involuntary admission.^{Reference Bone, McCloud, Scott, Machin, Markham and Persaud13} It may also be that CBT and talking therapies enhance patients’ trust in the care system and improve therapeutic alliance, leading to voluntary seeking of in-patient care when there is clinical deterioration. Furthermore, our study shows that Black African people were less likely to receive psychological therapies. This is in line with findings from a nationally representative study with people with schizophrenia,^{Reference Das-Munshi, Bhugra and Crawford11} and with evidence that, among people with common mental disorders, people from minoritised ethnic groups were less likely to receive treatment, including counselling.^{Reference McManus, Bebbington, Jenkins and Brugha29} Another study reports that, after referral for IAPT, several minority ethnic groups were less likely to be assessed.^{Reference Harwood, Rhead, Chui, Bakolis, Connor and Gazard12} Among those who did not receive an assessment, Black African people had the highest percentages for declining such assessment.^{Reference Harwood, Rhead, Chui, Bakolis, Connor and Gazard12}

Having a care plan was protective against involuntary admission. The effect may be related to a patient having the opportunity to discuss the management of their health with their clinicians and thus feel more in control of their clinical care. To our knowledge, this is the first study to show a relationship between having a care plan and involuntary admission. Care plans assessed in this study were not crisis plans, or advanced statements, but there is evidence that formulating an advanced statement is protective of involuntary admission.^{Reference Molyneaux, Turner, Candy, Landau, Johnson and Lloyd-Evans15,Reference De Jong, Kamperman, Oorschot, Priebe, Bramer and Van De Sande30} It may be that some of the underlying mechanisms that confer the protection provided by crisis plans are present during the formulation of common care plans, namely shared decision-making and a good therapeutic alliance. We observed that Other White, Black African and Asian Chinese people were less likely to have a care plan formulated. These findings are partially in line with a previous study with a nationally representative sample of people with schizophrenia; it was observed that Asian/Asian British (excluding Chinese) had reduced odds of having a care plan.^{Reference Das-Munshi, Bhugra and Crawford11}

Limitations and strengths

One limitation regards the non-inclusion of other predictors of involuntary admission, namely social support, marital status and household, and the source of referral to secondary care.^{Reference Walker, Mackay, Barnett, Sheridan Rains, Leverton and Dalton-Locke7,Reference Halvorsrud, Nazroo, Otis, Brown Hajdukova and Bhui9} About 1% of people that met the inclusion criteria were not included in the cohort as they were part of the national opt-out, and data for about 8% of patients using IAPT services are not linked to CRIS, because of non-disclosure of their NHS number. This may have led to an underrepresentation of people benefiting from psychological therapies.

Another important limitation is the crude assessment of the severity of illness based only on diagnoses. The level of severity of illness could preclude receiving some types of care (such as home treatment, psychological therapies). Thus, we are not able to rule out potential confounding by indication when assessing clinical care prior to admission. Having access to more intensive care (such as a higher number of appointments or home treatment) could be the result of greater severity of illness. Thus, it is likely that people who are given this type of care are already at higher risk for involuntary admission.

Our study strengths include using real-world data to investigate types of care and their relationship with involuntary admission. The study sample is representative of the SLaM's catchment population because of free access to NHS medical care in the UK. Also, the SLaM catchment population is very diverse (the majority of its residents have a minority ethnic background), which allowed investigation of ethnic inequalities without the need to merge distinct ethnic groups into larger categories. Our findings are likely to be generalisable to other larger urban areas in the UK.

Implications

This study shows that the majority of minoritised ethnic groups (10 of 14) have higher rates of involuntary admission, and these higher rates are not explained by differences in clinical care prior to admission. Future research is needed to uncover the reasons for the observed higher odds of involuntary admission rates among those who have had more SLaM out-patient appointments in the 12 months prior to admission and those who received home treatment. Access to psychological therapies and the formulation of care plans could be promoted as a way to potentially reduce involuntary admission.