Model 1 – Prism

The Primary Care Mental Health Service (Prism), which is run by the Cambridgeshire and Peterborough NHS Foundation Trust (2017) vanguard, involves mental health nurses promptly triaging all general practitioner (GP) requests for assistance and signposting to appropriate secondary care services, self-help groups or third-sector agencies. PRISM has shown a saving of at least £650 000 annually in reduced referrals to secondary mental health services in the first 2 years of operation, alongside high user and referrer satisfaction (Elliott Reference Elliott and Allan2018). Although cost savings have not been published, there appear to be equivalent reductions in administrative overheads, such as processing formal referrals, setting up appointments and typing letters. It is estimated by Cumbria, Northumberland, Tyne and Wear NHS Foundation Trust that 50% of the cost of a mental health referral (around £319) is spent on referral processing and other hospital overheads.

Model 2 – The Gold Standards Framework

This framework (Reynolds Reference Reynolds and Croft2010) involves a team of GPs, district nurses and Macmillan nurses routinely reviewing all patients (irrespective of diagnosis) on the palliative care register and intervening when necessary. It has been suggested that community psychiatric nurses (CPNs) looking after patients with severe mental illness (SMI), intellectual disability and dementia could follow a similar regime, by attending Gold Standard meetings and seeing patients as needed, without a formal referral. The City and Hackney Primary Care Psychotherapy Consultation Service (PCPCS) (Parsonage Reference Parsonage, Hard and Rock2019) has experience in collaboration with GPs in managing patients with SMI.

Model 3 – specialist consultancy teams

The third model, a consultancy team of skilled practitioners (occupational therapist, social worker and dementia nurse), was developed for complex dementia care by PMH Consultancy and Education, a private-sector company working in Cheshire (https://pmhcande.com/ola/services/dementia-care-consultancy-1). This team does not carry a case-load, but works alongside any team requesting help (including care homes) to manage behaviours that challenge and to avoid hospital admission. Consultative teams could focus on challenging behaviour and other complexities posed by people with dementia, autism, intellectual disability and personality disorders, with medication advice sought using an internet-based app such as Skype.

What biases are likely in liaison work and how to mitigate them?

Liaison psychiatry involves rapid diagnostic formulations, risk mitigation and treatment planning; but this can result in thinking errors (O'Sullivan Reference O'Sullivan and Schofield2018). Confirmation bias (often associated with overconfidence) is the most common problem, when a doctor – often under time pressure – looks for evidence confirming the initial impression, rather than seeking evidence to the contrary. There is a variant of confirmation bias called ‘diagnostic overshadowing’ (Jones Reference Jones, Howard and Thornicroft2008), when doctors (including psychiatrists) encountering patients with mental illnesses, conclude that their physical symptoms are part of their psychiatric syndrome and fail to investigate adequately.

In addition, trying to fit presentations to diagnostic codes and treat using previously learned ‘rules of thumb’ (heuristics) can also produce errors. An example of a heuristic error is assuming that ‘past risky behaviour predicts future behaviour’ without looking at recent changes in context. Affective bias (‘heart ruling the mind’) can influence risk management; for example, in Mental Health Act (MHA) assessments when assessors are excessively risk averse. As regards psychotropic selection, pharmaceutical promotions, unreplicated drug trials and perceived individual successes (or failures) with specific drugs can be potential biases.

Working as a multidisciplinary team (MDT) can also lead to biases such as groupthink (overconfidence in ‘group wisdom’) and escalation of commitment (‘throwing good money after bad’), particularly in the absence of clinical leadership. Sometimes MDT biases are due to lack of clearly stated objectives (‘mission creep’).

Increasing professional maturity should help a doctor to avoid or mitigate biases, and a clinical lead should assist subordinates in making unbiased decisions. However, in a busy clinical setting, self-regulated professionalism and proactive supervision can easily be lost. For a full discussion on bias see the Royal College of Psychiatrists’ CPD Online module (de Silva Reference de Silva2010).

How do we reverse rising admissions under Section 2 of the Mental Health Act in the context of reduced bed availability?

A consequence of increased investment to the primary–secondary care interface is less investment in mental health beds; these continue to reduce in number (McCartney Reference McCartney2017), partly owing to unsafe staffing levels (including staff preferring to work Monday to Friday in the community). Because of fixed commitments, it is rare for the patient's care coordinator to be part of an MHA assessment, and this results in a loss of ‘soft’ knowledge about resilience factors.

Furthermore, despite being told that the assessment is ‘under the Mental Health Act’ it is rare for patients to appreciate that anything they voluntarily disclose could be used as evidence for detention. All three assessors – the two doctors and the approved mental health practitioner (AMPH) – are under obligation to minimise risk, which potentially gives rise to affective bias and the ignoring of the patient's human rights (for example, privacy, right to family life). Indeed, MHA assessments are being legally challenged under the Human Rights Act.

Pragmatically, the best alternatives are to provide an advocate to maintain fidelity to the Human Rights Act; or to invite the relevant home treatment team to attend the assessment, with the sole purpose of providing alternatives to admission, for example utilising supervised ‘safe flats’: an initiative developed by South London and Maudsley NHS Foundation Trust in association with a local housing provider, Thames Reach. Furthermore, before the two doctors attend, the AMPH could carry out an initial review (not under the Act) with the assistance of the home treatment team, to ensure that all alternatives to hospital admission are considered with the patient and carers.

Lean working (de Silva Reference de Silva2018a) as practised in Middlesbrough (Tees, Esk and Wear Valleys NHS Foundation Trust), using daily ward reviews, early pre-discharge meetings and links with community providers, can assist early hospital discharge. Furthermore, in cases of prolonged admission (for example, of people with personality difficulties and risky behaviour), a visiting senior clinician (a ‘trusted advisor’) could be of value. This is currently in use in Newcastle (Cumbria, Northumberland, Tyne and Wear NHS Foundation Trust) and can empower ward teams to discharge patients.

As regards patients with chronic psychoses or dementia, advance care planning (ACP) could help decision makers (such as ambulance crews) to avoid inappropriate admissions. ACP use is patchy around the country (Lund Reference Lund, Richardson and May2010). Some GPs provide admission avoidance documents, but these are not consistent with the algorithms used by ambulance crews. Furthermore, carers and families influence admissions in crisis, suggesting the need to seek agreement with the family and carers when preparing admission avoidance documents.

Is zero suicide among patients under mental healthcare realistic?

This aspiration is viewed as controversial by both clinicians and patients, as the implicit assumption is that suicide is equated to failure of psychiatric treatment and follow-up. Perhaps mortality due to suicide is better viewed as part of reduced life expectancy in people with mental illness generally (see the next section).

Although evidence over the past 10 years suggests that the number of suicides by patients under secondary mental healthcare has gradually reduced (National Confidential Inquiry into Suicide and Safety in Mental Health 2018), suicides among younger women and among patients soon after hospital discharge have not (Haglund Reference Haglund, Lysell and Larsson2019). It is not known whether the rise in young women killing themselves is a result of pre-existing trauma: the inquiry has not collected these figures routinely. As regards post-discharge suicides, the high-risk period for suicide following a hospital admission is now recognised as the 3 days immediately after discharge. It is preferable for this follow-up visit to be carried out by a person already known to the patient; ideally a ward staff member.

Furthermore, information sharing between emergency services and mental health is patchy. With rapid transfer of patients between crisis teams, in-patient wards and CMHTs, patients and carers can have difficulty establishing a trusting relationship with any of the mental health staff they encounter during a single episode of care. Perhaps a patient-held chip and PIN card might be of value, similar to maternity cards from the past.

Standardised risk assessments have been largely ineffective in predicting completed suicide (Large Reference Large, Kaneson and Myles2016): they are mainly used as ‘defensible documentation’. Despite the main remediable cause in suicide being the method selected, specific inquiry via carers is not always carried out owing to anxieties about patient confidentiality and loss of trust. The tool most accepted by patients and carers is the safety plan (Cole-King Reference Cole-King, Green and Gask2013), describing a succinct action list for the person if they are harbouring suicidal ruminations: who to ring, where to go and what to say when they get there.

On staff education, most NHS trusts have mandatory suicide awareness updates, although this should be an annual undertaking, with discussion of local trends in completed suicide. There is scope for closed-circuit television (CCTV) observation of local suicide ‘hot spots’, although there are resource implications. CCTV could be also used in ward settings (Warr Reference Warr, Page and Crossen-White2005) instead of ‘close observation’, which is typically seen by patients as non-therapeutic, with exceptions for patients with persecutory delusions.

Should severe mental illness, intellectual disability and autism spectrum disorder be seen as life-limiting conditions, similar to dementia?

Actual figures suggest that people with these conditions have a reduction in life expectancy of around 15 years, possibly through poor diet, medication side-effects, loneliness, lack of purposeful activity and suicide. There is no evidence that psychiatric treatments (pharmaceutical or otherwise) improve this situation. Furthermore, people who have experienced trauma early in life show shorter telomere length (Tyrka Reference Tyrka, Price and Kao2010), consistent with a shorter lifespan. Chronic stress, as experienced by most psychiatric patients independent of diagnosis, also speeds up the biological clock (Solanas Reference Solanas, Peixoto and Perdignero2017).

Research on preventing neurodegenerative conditions (such as Alzheimer's disease, Parkinson's disease and stroke) has moved away from specific disease-modifying treatments to extending healthy lifespan by triggering cell ‘cleansing’ called autophagy (Cuervo Reference Cuervo, Bergammi and Brunk2005). Specifically, patients could have their lifespan prolonged by reduced meal sizes (Martin Reference Martin, Golden and Egan2007), intermittent fasts (Wei Reference Wei, Brandhorst and Shelehichi2017), supplements such as resveratrol and nicotinamide adenine dinucleotide (NAD) boosters (Wang Reference Wang, Hu and Yang2006) and exercise (Reimers Reference Reimers, Knapp and Reimers2012). These interventions will probably reduce morbidity rather than closing the mortality gap.

Clearly, viewing SMI, intellectual disability and autism as palliative conditions is controversial; however, this view fits the available evidence and dovetails with the Gold Standards Framework method of oversight (model 2 above) – which secondary mental health services cannot provide owing to their move to episodic care, typically a specialist assessment and two follow-up contacts. Primary care SMI management has been piloted in Whitby (the Whitby SMI monitoring project; de Silva Reference de Silva2018b), with a CPN rotating between 6 GP practices (population 26 000), working with practice nurses responsible for cardiovascular screening to cover all patients on antipsychotic medication (including clozapine). This approach had widespread staff and patient approval, leading to all patients on long-term depot antipsychotics being discharged from the CMHT, as well as pick up of previously undiagnosed hypertension, dyslipidaemia and glucose intolerance.

What key areas of emerging medical knowledge should community psychiatrists be aware of?

How do we encourage patients and carers to talk freely about difficulties?

Approaches to facilitate patient and carer input to assessments have included open dialogue (Freeman Reference Freeman, Tribe and Stott2018) and trauma-informed mental healthcare (Sweeney Reference Sweeney, Clement and Filson2016). Open dialogue involves helping the patient and carers to articulate their understanding of the presenting problem(s), to agree a joint narrative and to suggest solutions. Evidence from Finland with first-episode psychosis presentations suggests high patient approval coupled with earlier hospital discharges, reduced readmissions, reduced psychotropic prescriptions and increased employment (Bergström Reference Bergström, Seikkula and Alakare2018). A trial of open dialogue is underway in England involving selected CMHTs working entirely using this approach (ODESSI trial with Principal Investigator Steven Pilling UCL (2017-2022)).

What is the most cost-effective liaison psychiatry service in primary care for improving access and patient flow?

The simplest solution would be to transplant the established general hospital psychiatric liaison team (PLT) model to primary care. This has been recently piloted in Nottingham (O'Shea Reference O'Shea2019), with an average annual saving of over £500 per patient on discharge. PLT staff triage all referrals within 1 day, independent of age and circumstances, including substance misuse, intellectual disability and dementia.

Instead of formal written referrals, an email including an ‘intermediate summary’ from GP records (consisting of a problem list, medication list and the last three consultations describing the presenting problems) would be sufficient to commence an assessment, saving referral processing costs. Typically, information is summarised in the ‘situation, background, assessment and response’ (SBAR) format recommended by NHS Improvement, upgraded for mental illness by adding in a ‘risk’ section (de Silva Reference de Silva2013). This summary is used thereafter as the core document when care is transferred to other mental health teams. Consultant supervision is available immediately, including the option of discussing short-term treatment using an app such as Skype.

Alternatively, a liaison service in primary care could use the Prism model for rapid triage, coupled with the Gold Standards Framework to monitor people with chronic mental health conditions, including psychosis, dementia and intellectual disability. Furthermore, consultancy teams can back up primary care PLTs on ‘behaviours that challenge’ and on optimising psychotropic medication prescription.

Clearly 24 h coverage in primary care is not required as GP surgeries generally operate Monday to Friday ‘office hours’. However, there could be close working with the hospital liaison teams that cover accident and emergency departments where patients attend out of hours. A joint set of electronic records, including primary care, mental health and acute hospital (as operated in Sheffield at South Yorkshire and Bassetlaw Integrated Care System), would be ideal in facilitating seamless care.

Hospital liaison psychiatry services have demonstrated significant savings through rapid assessment and diversion (Bell Reference Bell2018). This is consistent with the above-mentioned financial savings demonstrated by Prism. Rapid access will also limit unnecessary referrals of ‘functional’ patients to other specialties.