Discussion

This study examined the characteristics of people aged over 60 who presented following self-harm and suicidal ideation. It found that self-harm was significantly associated with admission to a psychiatric facility, but not after controlling for gender, substance use and need for medical admission, and these covariates remained significantly associated with subsequent psychiatric admission in this model.

We found that poisoning or overdose is the most common form of self-harm in older people. Troya et al reported a similar pattern in their 2019 systematic review (68.1%), and similarly, lacerations were the second most common method of self-harm. ^{Reference Troya, Babatunde, Polidano, Bartlam, McCloskey and Dikomitis18}

The clinical programme has reported overall psychiatric admission rates of 16.8% of the population as a whole undergoing assessment in the emergency department following a presentation with self-harm or suicidal ideation. The admission rates reported in older people are significantly higher for both those presenting with self-harm and suicidal ideation, with the highest rates in older people presenting with self-harm (26.8%). This database does not collect diagnoses, but given that psychiatric admission is a rare outcome reserved for those with more severe mental illness, which is decided only after a detailed mental health assessment and discussion with a consultant psychiatrist, it may be considered a proxy measure for severe mental illness. ^{Reference Walker, Mackay, Barnett, Sheridan Rains, Leverton and Dalton-Locke23,Reference Degli Esposti, Ziauddeen, Bowes, Reeves, Chekroud and Humphreys24} While depression is the illness more commonly associated with suicidal ideations and behaviours, self-harm and suicidal ideations can also be seen in a range of conditions from psychosis to adjustment disorders. ^{Reference Yates, Lång, Cederlöf, Boland, Taylor and Cannon25,Reference Casey, Jabbar, O’Leary and Doherty26} These results suggest that self-harm is not more likely than suicidal ideation to result in psychiatric admission once other variables are controlled for, and that both presentations (self-harm and suicidal ideation) are associated with a need for admission to a similar extent.

Our findings suggest that there are significant differences between older patients who present to the emergency department with self-harm, compared with those presenting with suicidal ideation. Regarding the relationship between self-harm in older people and gender, we found that there were higher proportions of older women presenting with self-harm (58.4%) compared with men (41.5%), and higher proportions of men presenting with suicidal ideation (58.3%) compared with women. This is consistent with the international literature as summarised in the systematic review by Troya et al, which found that among older people, more women than men presented with self-harm. ^{Reference Troya, Babatunde, Polidano, Bartlam, McCloskey and Dikomitis18} This is at variance with the higher suicide rate in older men, and it suggests that while more women engage in self-harm, there may be a higher burden of illness associated with self-harm in older men.

The relatively low proportion of older patients who presented with self-harm or suicidal ideation who were previously known to mental health services is surprising. However, it is consistent with Szücs et al’s finding that an older age of onset of mental illness is associated with suicidal behaviour in older people. ^{Reference Szücs, Perry-Falconi, O’Brien, Keilp, Bridge and Maier27} It suggests that this population may present to hospital in a suicidal crisis, rather than to their GPs at an earlier stage, or perhaps the crisis evolves more rapidly. It is likewise consistent with the findings of the study by Ho et al, which similarly reported that 37.5% of older people had a history of mental illness, in their study of older people who died by suicide in Singapore. ^{Reference Ho, Ho, Tai, Ng and Chia28} They reported that despite the mental illness rate of 37.5%, only 17% had an antidepressant detected on toxicology, which suggests that perhaps adherence to medication was suboptimal. ^{Reference Ho, Ho, Tai, Ng and Chia28} The majority of older people stayed for assessment, and this was higher in those presenting with self-harm, perhaps due to the urgency of medical treatment.

Alternatively, the lower rates of referrals via GPs suggest that access to GPs has become more limited for this population. It raises the question of whether this may have been due in part to COVID restrictions and reduced in-person visits to GPs, so that the emergency department was perceived as the only option. ^{Reference Troya, Corcoran, Arensman and Kavalidou21,Reference Maguire, Kavalidou, Bannan, Doherty and Jeffers29} There is an international move towards the digital provision of care. ^{Reference Schifferdecker, Yang, Mackwood, Rodriguez, Shortell and Akré30,Reference Mackwood, Fisher, Schmidt, Yang, O’Malley and Rodriguez31} It has been identified in the literature that the indiscriminate use of digital solutions may inadvertently exclude older patients. ³²

Similarly, medical admission can be seen as a proxy and objective measure for the medical lethality of self-harm. This is consistent with the existing literature which suggests that an older age of onset of symptoms is associated with higher lethality of self-harm. ^{Reference Szücs, Perry-Falconi, O’Brien, Keilp, Bridge and Maier27} If we consider suicidality as a spectrum, it might be seen to extend from suicidal thoughts, through acts with death by suicide as the most severe end of this clinical spectrum. Those patients who require medical admission for self-harm are likely to represent the more severe end of the spectrum than those who can be medically discharged from the emergency department. ^{Reference Mohan, Tembo, McNicholas and Doherty33} Existing research suggests that in addition to having high suicide rates, older people are more likely to die following self-harm. This is thought to be related to the use of lethal means, a greater degree of planning (less impulsivity) and physical frailty ^{Reference Szanto, Gildengers, Mulsant, Brown, Alexopoulos and Reynolds34} (as represented by the need for medical admission as a proxy measure).

Substance use, in this study, was associated with self-harm in older people, and there were similar proportions of people using alcohol among those who presented with self-harm and suicidal ideation. There were significant differences, however, in the use of other substances, with higher proportions of those with self-harm using either drugs and alcohol, or just drugs. Ho et al reported that alcohol was a feature in older people dying by suicide in 23.1% of cases, which is slightly lower than our cohort. ^{Reference Ho, Ho, Tai, Ng and Chia28} That said, the majority of older patients presenting with self-harm and suicidal ideation did not use substances at all, with a higher proportion of those presenting with self-harm using substances, with or without alcohol.

Finally, older people attending the emergency department with suicidal ideation had longer waiting times in the emergency department than those attending with self-harm. This finding is somewhat counterintuitive: we expected that those presenting with self-harm, who likely require medical/surgical input for injuries or poisoning, would be longer in the emergency department, awaiting bloods and a range of clinical opinions. This is important from a service perspective, as the emergency department is a high-stimulus environment which can present challenges for older persons, or indeed anyone experiencing acutely distressing symptoms.