Head To Head Maudsley Debate

Does the emphasis on risk in psychiatry serve the interests of patients or the public? Yes

BMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f902 (Published 12 February 2013) Cite this as: BMJ 2013;346:f902
  1. John F Morgan, consultant psychiatrist
  1. 1Leeds and York Partnership NHS Foundation Trust, United Kingdom
  1. Correspondence to: Yorkshire Centre for Eating Disorders, Seacroft Hospital, Leeds LS14 6UH, UK jmorgan{at}sgul.ac.uk

Identifying patients who are likely to harm themselves or others has become central to psychiatry. John Morgan argues that though the methods are flawed, identifying risk is essential, but Matthew Large (doi:10.1136/bmj.f857) thinks we should focus on the wider harms that patients may experience

In a biblical failing to understand sensitivity and specificity, King Herod applied demographic factors to kill all male children near Bethlehem, wishing to avoid losing his throne to a newborn king. The Massacre of the Innocents provides a potent example of inadequate predictive value of risk assessment categorisation, with sex and age failing to detect a rare outcome. Likewise, emphasis on risk in psychiatry has gone wrong, and psychiatrists baulk at playing Herod.

But medicine is risky business. Assessment and communication of risk permeates orthopaedic examination of a knee and psychiatric examination of mental state. Both orthopaedic surgeon and psychiatrist systematically elicit signs and symptoms, with due knowledge of pathology, making judgments about diagnosis, treatment, and prognosis based on awareness of risks. Our social function as doctors requires that we accept emphasis on risk in psychiatry. Psychiatric risk assessment processes are flawed, misdirected, and innumerate, but risk remains a fundamental component of psychiatry, as in all medicine.

Imperfect system

However, worst practice in psychiatric risk differs from, say, surgical risk assessment in several regards. Risk assessment has become horribly separated from clinical examination. Risks of common outcomes, such as social exclusion, are subordinated to less common outcomes. Systems are seen as entirely failing and individuals scapegoated when mathematically inevitable failures to detect less common outcomes emerge. As a result, psychiatrists become agents of social control rather than physicians of the mind, with consequent poor morale. Risk assessment categorisations carry inadequate predictive value. We require a re-emphasis of risk, with patient safety at the forefront.

Psychiatric risk assessment is an inexact actuarial science operating in a political arena. The recent history of psychiatric risk assessment is one of scepticism and criticism, to which my own voice has been added in three coauthored policy documents. Giving up the Culture of Blame noted that many psychiatric approaches to risk assessment lacked an evidence and were not validated for the populations to whom they were applied, generating a false sense of security.1 The Kennedy report called for risk assessment to be better reintegrated with clinical assessment,2 and the Alderdice report voiced fears that patient wellbeing was jeopardised by defensive practice, including preoccupation with “tick box” risk assessment and relative neglect of clinical skills.3

Risk is unavoidable

I agree substantially with the published opinions of my opponents in this debate.4 However, while I continue to criticise current risk management practice with a reformer’s zeal, we cannot and should not abandon public responsibilities. A dichotomised debate is unhelpful, and our professional responsibility is to provide nuance. A perfect risk management system would have modest effects on rates of homicide, and undue emphasis on homicide has skewed debate, as well as resources, away from areas where greatest good can be done to the greatest numbers of people. However, the limitations of risk assessment do not mean that emphasis on risk is misplaced. Psychiatrists retain a duty towards patient safety, and this includes a contribution to the protection of the public, including our patients.

Both surgeon and psychiatrist can hazard judgments of risk for common outcomes, embedded within clinical examination, but balance between sensitivity and specificity becomes more challenging as outcomes become less probable. For uncommon outcomes, such as homicide, the mathematics of risk dictates that the balance will be elusive. Likewise, the history of psychiatry swings between poles of liberalism and authoritarianism—and is frequently damned for both. Liberal culture damns psychiatry for wrongful incarceration while authoritarian culture damns psychiatry for failing to act. Herod’s measures of sex and age held high sensitivity but inadequate specificity, with unintended consequences.

Innumerate risk assessment in psychiatry also carries substantial unintended consequences. Dialogue around psychiatric risk requires awareness of its mathematics. Public and political debate is needed about the unintended consequences of focusing on unidimensional, rare outcomes. How many incorrect risk categorisations are acceptable in a liberal democracy? With finite resources, how much funding should be diverted away from psychiatric safeguarding and towards the prevention of societal violence? Anorexia nervosa has the highest standardised mortality ratio of any psychiatric disorder 5 but slips below the radar of commissioning in favour of new constructs such as “dangerous and severe personality disorder.”

Most worrying, the process of “risk assessment” becomes a tick box process separated from clinical examination. In the largest survey of its kind,3 the Royal College of Psychiatry’s members reported undue reliance on non-evidence based risk assessment tools, and these were often divorced from mental state examination. Just as the orthopaedic surgeon systematically conducts a clinical examination to evaluate orthopaedic risk, so must the psychiatrist apply knowledge of phenomenology to assess mental state. Thereby we evaluate many and varied types of psychiatric risk that defy simplistic categorisation. Risk assessment tools are not without value as aide mémoires, just as checklists have value in surgery. But psychiatry subverts its skills and acumen by separating risk from clinical examination.

My manifesto is this. We need to return emphasis on risk in psychiatry to clinical examination, by systematic evaluations of signs and symptoms. We need to reposition risk assessment forms as nothing more than helpful aide mémoires. We need to ensure that common psychiatric risks are not overshadowed by rare risks. We need to reject innumerate risk assessment processes with inadequate predictive value. We need nuanced debate and engagement with the public in a political arena. But we cannot remove risk from psychiatry.


Cite this as: BMJ 2013;346:f902


  • Competing interests: I have read and understood the BMJ Group policy on declaration of interests and declare I am chair of the Royal College of Psychiatrists patient safety steering group and carry out research on risk assessment in eating disorders.

  • “Risk in psychiatry” is the subject of the latest Maudsley Debate to be held on Tuesday 19 February at 6 pm at the Wolfson Lecture Theatre, Institute of Psychiatry Main Building, De Crespigny Park, London SE5 8AF.

  • Provenance and peer review: Commissioned; not externally peer reviewed.