Background

Suicide happens much more often after an episode of non-fatal self-harm than it does in the general population. Self-harm is one of our strongest predictors of completed suicide, ^{Reference Owens, Horrocks and House1} and the older the person who has self-harmed the greater the risk of subsequent suicide. ^{Reference Owens, Wood, Greenwood, Hughes and Dennis2,Reference Hawton, Zahl and Weatherall3} Research that has demonstrated this effect of age has, however, been unconvincing. Self-harm in older people is rarer than in the young, so studies have been based either on small samples of older people or on cohorts that were assembled over many years – rendering the findings susceptible to imprecision and to temporal and cohort effects. Such limitations have been overcome by Murphy and colleagues in a multicentre study of self-harm in people aged 60 and over, reported in this Journal. ^{Reference Murphy, Kapur, Webb, Purandare, Hawton and Bergen4} They recruited to their 8-year cohort study in three English cities more than a thousand individuals who attended hospital because of non-fatal self-harm. The study used well-validated methods for case identification and comprehensive tracing at follow-up. They found that 1.5% died by suicide in the next 12 months, pointing to a suicide rate 67 times that of the general over-60 population and 3 times that of younger adults who self-harm (where the suicide rate was 24 times greater than that in the general population). Men aged 75 and over had the highest suicide rate.

Suicide rates in England have decreased substantially in the past 20 years and rates in older groups have dropped more than those in the overall populations of males and of females. This fall has been attributed to a variety of factors including better welfare and economic conditions, better access to specialist services, and increased recognition and effective treatment of depression. Male and female rates are currently highest among 30- to 50-year-olds (although recent rates show a second peak in the very elderly, aged 80 years or over). ⁵ Why then, despite suicide rates being generally lower among older than younger adults, does suicide following self-harm show a contrary pattern? First, research has repeatedly emphasised that, more often than not, self-harm in older people represents a failed attempt at suicide; other motives for the behaviour are correspondingly less frequent than in younger adults. ^{Reference Dennis, Wakefield, Molloy, Andrews and Friedman6} Second, the recognised determinants of suicide in the elderly population – depression, poor physical health and poor social support – are evident in older patients who attend hospital because of self-harm. ^{Reference Dennis, Wakefield, Molloy, Andrews and Friedman6}

A succession of guidance statements – consensus and evidence-based – have set out advice on hospital care that is in keeping with the high suicidality of older patients who have self-harmed. These include advice from the Royal College of Psychiatrists (College Reports 32, ⁷ 122 ⁸ and 158; ⁹ and Better Services for People who Self-Harm); ¹⁰ and National Institute for Health and Clinical Excellence (NICE) guidelines on short-term and long-term management of self-harm (Clinical Guidelines 16 and 133). ^11,12 These guidelines, spanning nearly 20 years, consistently and strongly advise that all older people who self-harm should be assessed by an old-age psychiatrist, or at least by a mental health specialist trained to assess risks and needs in this age group. In the cohort study described in the current edition of the Journal ^{Reference Murphy, Kapur, Webb, Purandare, Hawton and Bergen4} the study population represents patients in England who ought to have been subject to these recommendations. The authors found a high but incomplete rate of psychosocial assessment (81%) but we do not know whether the assessors were suitably trained and experienced in old-age care. Although 90% of those assessed received referral for aftercare, more than one-third of the aftercare was in the form only of referral to the general practitioner (GP), thereby not resulting in any specialist care. These findings may suggest that general hospital care is below an optimal level but we know from earlier work that it probably represents better assessment and aftercare than is experienced by patients younger than 60 years. ^{Reference Marriott, Horrocks, House and Owens13}

What are the key messages for management of self-harm in older people?

First, the present findings ^{Reference Murphy, Kapur, Webb, Purandare, Hawton and Bergen4} re-emphasise the need for adherence to the existing guidance concerning specialist assessment following an episode of self-harm in older people. Adherence to guidelines should be monitored and audited by healthcare providers, particularly where there is pressure on bed numbers in the general hospital; older people should not be discharged home from the emergency department before an appropriate assessment has been undertaken by a skilled mental health professional.

Second, mental health services and primary care need to be ready to deal with longer-term management of vulnerable older people with depression and a history of self-harm – because of their hugely increased risk of suicide ^{Reference Owens, Wood, Greenwood, Hughes and Dennis2–Reference Murphy, Kapur, Webb, Purandare, Hawton and Bergen4,Reference Hepple and Quinton14} and its relation to persistent depression. ^{Reference Hepple and Quinton14} Optimum management of any depression (including the use of maintenance antidepressant medication when appropriate), close follow-up, collaboration and liaison between agencies, and engagement with key social factors is crucial in the care of older people who have self-harmed. Collaborative care models have proved beneficial in older patients with depression with suicidal ideation. In the USA, the Improving Mood Promoting Access to Collaborative Care Treatment (IMPACT) trial examined collaborative care for older patients with depression, assigning patients a depression care manager who provided support, education and psychological therapy as well as liaison between the primary care physician and a psychiatrist; there were benefits in the long-term management of depression as well as a significant reduction in suicidal ideation. ^{Reference Unützer, Tang, Oishi, Katon, Williams and Hunkeler15} Importantly, depression care managers provided close follow-up, ensured optimal management of patients who had initial partial response or resistant depression, and assisted in the development of relapse prevention plans.

The Prevention of Suicide in Primary Care Elderly: Collaborative Trial (PROSPECT), a USA study of older patients with depression in primary care, examined the effects of a collaborative and stepped-care approach involving antidepressant medication, care management and psychotherapy; the active intervention was associated with a quicker reduction in suicidal ideation. ^{Reference Bruce, Ten Have, Reynolds, Katz, Schulberg and Mulsant16} These clinical trials investigate collaborative and supportive components of care in depression among older patients and encourage cooperation with taking medication. In both cases they also examined the use of psychological therapies, in particular problem-solving and interpersonal psychotherapy.

In the UK monitoring study set out in this Journal, ^{Reference Murphy, Kapur, Webb, Purandare, Hawton and Bergen4} over half of the patients received specialist mental healthcare following the index self-harm episode. In earlier research by others, ^{Reference Hepple and Quinton14} nearly two-thirds of patients were still receiving psychiatric treatment 1 year after the index self-harm episode. The current research base therefore points to the need for assessment of suicidal ideation as a continuous process during an episode of care: during the course of a depressive episode, older patients can change from being ‘passive ideators’ with death wishes to having active suicidal thoughts. ^{Reference Szanto, Reynolds, Frank, Stack, Fasiczka and Miller17} The recent NICE guidance on longer-term management of self-harm (Clinical Guidance 133) emphasises care planning and risk management strategies. ¹²

When striving to reduce the risk of subsequent suicide in individuals who have self-harmed, attending only to the clinical management of late-life depression is unlikely to be sufficient. In the social domain, integration and support are particularly relevant: self-harm in older people is associated with widowhood, living alone, social isolation and low levels of social support. In the recent past we compared two groups of older people with depression in secondary care: those who had and had not ever self-harmed. ^{Reference Dennis, Wakefield, Molloy, Andrews and Friedman18} Severity of depression and other health and social variables were similar in both groups but individuals who had harmed themselves were more likely to have a poorly integrated social network, and they reported lower levels of community support. The innovative work of De Leo and colleagues in Italy with the large scale TeleHelp–TeleCheck intervention illustrates the potential of increased social support combined with a crisis service. ^{Reference De Leo, Dello Buono and Dwyer19} In that study vulnerable older people were referred to the service by their GP or social worker. The intervention involved twice-weekly telephone support from a trained professional and the provision of an alarm connected to a 24 h response network. The service was provided for over 18 000 elderly people over a 10-year period, resulting in a significant reduction in the suicide rate as a consequence of reduced female suicides. Besides those who are socially isolated, vulnerable older people include those with physical ill health and disability – poor physical health is known to be a common feature of older people who self-harm, ^{Reference Dennis, Wakefield, Molloy, Andrews and Friedman6} and disability is independently associated with suicidal ideation. ^{Reference Dennis, Baillon, Brugha, Lindesay, Stewart and Meltzer20}

Suicide in older people, although less frequent than it once was, is a significant cause of death. The accompanying distress caused to family and friends – and to any staff who have been dealing with the person's healthcare – is often severe. Attendances at the general hospital's emergency department because of non-fatal self-harm are perhaps the best single harbinger of that shocking outcome. We are unlikely to gain very much more from research that elucidates further risk indicators for older people following self-harm: they generally have substantial risk and need for care. We must ensure adequate assessment and aftercare, in particular for men, who have proved – as with younger males after self-harm – especially difficult to engage in treatment and care. The role of the specialist in assessment and longer-term management of older people following self-harm, in particular for people with depression, further highlights the importance of mental health services that specifically cater for the needs of older people. With an increasing number of frail older people being cared for in community settings, the development and increased availability of modern communication technology for those who are vulnerable seems one of the most promising routes for research and development.