Increase in Self-Injury as a Method of Self-Harm in Ghent, Belgium: 1987-2013

Background Self-harm is a major health care problem and changes in its prevalence and characteristics can have important implications for suicide prevention. The objective was to describe trends in the epidemiology of self-harm based on emergency department (A&E departments) visits over a 26-year period in Ghent, Belgium. Methods We analyzed data on all self-harm presentations from the three large general hospitals in Ghent between 1987 and 2013. We investigated trends in prevalence (events by year per 100.000), methods and alcohol use. Results Rates of self-harm steadily decreased during the 26-year study period. In general female rates of self-harm were higher than male rates. The mean patient age was 35 years. The most commonly used method of self-harm was self-poisoning by means of an overdose of medication (80.8%), followed by cutting (10.2%) and hanging (4.2%). Psychotropics (including antidepressants, benzodiazepines, barbiturates and other tranquilizers) were the most frequently used drugs (74.5%). A proportional increase in the use of self-injurious methods in self-harm was highly significant, more specifically in the use of hanging, jumping from heights and the use of other violent methods such as the use of firearms, jumping before a moving object or other traffic related injury. Conclusion This epidemiological study showed an increase in the use of high-lethality methods in self-harm which has important implications for suicide prevention. As restrictions in the availability of these methods are difficult or impossible to achieve, prevention programmes will have to emphasize the role of thorough psychosocial assessment and adequate follow-up care of self-harm patients.

Introduction collected for all individuals (age 15 or older) who presented with an episode of self-harm to the A&E department of the University hospital  of Ghent and two general hospitals in Ghent (since 1996 till 2013) for the 26-year period from 1 January 1987 to 31 December 2013.
Between 1987 and 1994 self-harm was defined according to Hawton and Catalan (1987) [32] and included deliberate self-poisoning and self-injury. The term 'deliberate self-poisoning' was used to describe the deliberate ingestion of more than the prescribed amount of medical substances, or the ingestion of substances never intended for human consumption, irrespective of whether harm was intended. 'Deliberate self-injury' was used to describe any intentional self-inflicted injury, irrespective of the apparent purpose of the act. Since 1994, self-harm was defined as "an act with nonfatal outcome, in which an individual initiates a deliberate, wellconsidered, and unusual behaviour, that without intervention of another will lead to self-harm or destruction, or when an individual deliberately takes a substance in a higher quantity then subscribed or generally suitable doses, with the intention by means of actual or expected physical consequences to initiate desired changes" [33].
Demographic and clinical data on each episode were collected by clinicians and nurses using standardized forms. Through scrutiny of the records of the A&E department and the psychiatric department of the University hospital of Ghent, information was also available on patients who were admitted to the hospital but who were not seen by a psychiatrist of for whom no data sheet was collected due to early discharge, refusal of assessment, or unavailability of staff.

Demographics, clinical characteristics and rates
Data for this study included gender, age, date of self-harm, method and use of alcohol. The data on methods were coded according to ICD-10 codes. Self-injury involved violent methods (X70-X77; X78-X84), while self-poisoning involved less violent methods (X60 to X69). Selfharm episodes involving alcohol alone were not included unless accompanied by other types of self-poisoning or self-injury. Rates were calculated as the number of people (aged 15+) per 100 000 inhabitants in Ghent for each year.

Data-analysis
Statistical analyses were performed using SPSS22. The primary analysis was a descriptive summary for self-harm. The chi square test for trend (linear by linear association) was used to test the level of significance of changes over the period 1987-2013. Groups were compared using chi-square tests. P < 0.05 was considered significant.

Ethics statement
Ethical approval for data collection on self-harm was obtained from the Ethical Committee of the University Hospital Ghent according to the principles expressed in the declaration of Helsinki. The Ethics Committee gave approval to collect patient data without obtaining informed patient consent, this because the data were analyzed anonymously.

Demographics
During the 26-year study period, i.e. from 1 January 1987 to 31 December 2013, the monitoring system reported 8 692 episodes of self-harm aged 15+ years in the three A&E departments in Ghent. Age and gender were known for 8 377 persons (3.7%, n = 315 were missing) ( Table 1). The majority of self-harm patients were female (56.0%). The mean age for both genders was 35.5 years. More than the half of the patients (n = 4 676, 55.8%) were 35 years or younger. The largest number of females was in the 20-24 age group (n = 733, 15.6%), and the largest number of males was in the 25-29 age group (n = 591, 16.0%).
Decreases in person based self-harm rates for Ghent were seen in both genders during the first three years of the study period. In the period from 1987 to 1989 the rate of males decreased significantly (χ 2 (1) = 6.18, p < 0.05) respectively from 398 (95% CI 360-440) to 331 (95% CI 300-370) and the rate of females decreased significantly (χ 2  years of the study period the rates decreased slightly and non-significantly from 249 (95% CI 220-280) in 2009 to 208 (95% CI 180-240) in 2013 among males (i.e. a decrease of 16.5%) and from 267 (95% CI 240-300) in 2009 till 257 (95% CI 230-290) in 2013 among females (i.e. a decrease of 3.7%). Overall, female rates of self-harm were significantly higher than male rates. In 1989, 1990, 1992 and 1997 the male rate was not significantly higher than the female rate, while in 1993 the male rate was significantly higher than the female rate (i.e. 414 and 257 (χ 2 (1) = 36.86, p < 0.001).
Between 1993 and 1995 the use of self-poisoning decreased, while the combined use of selfpoisoning and self-injury increased, particularly involving the combination of cutting or hanging and medication.

Involvement of alcohol
The use of alcohol before or during self-harm was more common among males (32.6%; 95% CI 31.2%-34.1%) than among females (23.4%; 95% CI 22.3%-24.6%; (χ 2 (1) = 91.16, p < 0.001). During the study period, the use of alcohol before or during self-harm increased significantly among males (χ 2 for trend: 7.53, p < 0.01). There was no significant change over the years in the use of alcohol among females. Over the 26 years, the use of alcohol was most common in the 35-54 years age groups.

Discussion
Using a unique long-term dataset of self-harm presentations to the three general hospitals in Ghent, this study investigated trends in the rates and characteristics of self-harm between 1987 and 2013. Overall rates of self-harm appeared to peak in 1995 and then declined in both genders. There were obvious changes in the methods used for self-harm during the study period with a significant proportional increase in the use of self-injury as the method of self-harm. More specifically, there was a strong increase in the use of hanging, jumping from heights and the use of other violent methods including the use of firearms, jumping before a moving object or other traffic-related injury. Cutting was most common between 20-29 years and people aged 60 and older used commonly drowning as method. Overall, the use of self-injurious methods increased from 11 to 19%, or approximately from one in ten episodes to one in five episodes of self-harm.
Preceding a discussion of implications of the present findings for suicide prevention potential methodological limitations need to be addressed. We were able to use data on a large sample of self-harm patients, which were collected by means of a consistent monitoring system during a prolonged period of time in the three major hospitals in the area. The reliability of the findings is supported by the comparability of demographic characteristics to those from other epidemiological studies of self-harm [15,16,[18][19][20][21]29]. However, the data most probably refer to a selected group of self-harm patients. The sample did not include individuals who selfharmed but did not present to hospital. There are patients who self-harmed but did not need medical attention or are treated by a general practitioner without referral to an A&E department. This indicates a possible under reporting of the real number of people who self-harm in the study area.
The marked decline in rates of self-harm since 1995 coincides with a change in the definition of self-harm in the hospitals since 1994 as described in the method section. However, as the decline in rates of self-poisoning was much more outspoken than the change in rates of self-injury it is unlikely that the decrease in rates is attributable to the change of definition. It is possible that local prevention activities have contributed to the decrease in self-harm rates. For example, in 1995, we demonstrated a near-significant effect of home visits by a community nurse in case of non-compliance with referral to outpatient aftercare among self-harm patients after discharge from the A&E departments involved in the current study, on repetition of selfharm [34]. Since 1997, this activity has been integrated in a suicide prevention programme in which community mental health services target high-risk groups such as self-harm patients. The prevention programme includes the availability of help for self-harm patients within 48 hours after a self-harm episode, educational campaigns targeting health care professionals (e.g. general practitioners, nurses and police officers) by organizing seminars and workshops about suicide prevention and assistance in developing suicide prevention strategies in schools and companies. The Flemish government has implemented two subsequent suicide-prevention programmes during the study period. The first programme ran from 2006 to 2010 [35]. The second Flemish suicide prevention programme was launched in 2012, aiming at a 20.0% decrease in suicide rates in 2020 when compared to 2000 [36]. Suicide rates in Flanders have stabilized between 2009 and 2012, thus not showing the substantial increases, which became apparent in surrounding countries such as The Netherlands during this period, and which are thought to be related to the economic turmoil since 2008 [37]. The extent to which the suicide prevention programme has countered the negative impact of the economic problems and thus contributed to the stabilisation of suicide rates and the decrease in self-harm rates in times of economic problems remains to be elucidated. The continued monitoring of self-harm will demonstrate whether the second prevention programme is associated with a further decrease in rates of self-harm. Other targets of the Flemish suicide prevention programme such as improved media reporting of suicide may also have contributed to the decline. There may have been other reasons for the declining rates of self-harm, such as changes in help-seeking behaviour including the increased use of counseling services (chat and forums), telephone support [38] and internet websites for help and support [39]. The increased proportional use of self-injury in self-harm may be due to changes in patterns of hospital presentations of self-harm over time. For example, the proportional increase in selfinjury may be due to a decrease in hospital presentations of self-poisoning. Such a change in hospital presentations cannot be ruled out, but might have been reflected by a stronger decrease in rates in females than in males, as females more commonly self-harm by means of self-poisoning than males. As this was not the case, it appears that there indeed is a relative increase in self-injurious methods of self-harm during the study period. In addition, a similar trend in the use of self-injury has been found in other studies [15,31]. Given the up to 5-fold increased risk of subsequent suicide following self-injury when compared to a non-violent method such as self-poisoning [40,41] and the increased suicidal intent associated with the use of highly lethal methods of self-injury [25], this finding may be of particular relevance for suicide prevention and e.g. the A&E department staff when making decisions about follow-up care [25].
The current findings have important implications for suicide prevention. First, with regard to methods used in self-harm the data show that the medication used in self-poisoning mainly consists of psychotropic drugs. This is in contrast with findings in other countries such as the UK in which shows paracetamol and paracetamol-containing compounds are the most commonly used drugs in self-poisoning [29]. Such differences most probably reflect an effect of availability of drugs whether or not via prescriptions. This issue therefore needs to be addressed in educational campaigns and training programmes targeting physicians. Such training programmes may lead to marked effects in reducing suicide rates [42,43]. A reduction in the accessibility of medication may indeed contribute to suicide prevention [44]. In Ireland, the withdrawal of a prescription-only analgesic compound was associated with a marked reduction in the rate of intentional drug overdose presentations to hospital involving the drug [45]. Restricting the availability of medication has led to marked reductions in suicidal behaviour in England, Wales, Scotland and Scandinavia [46][47][48]. Secondly, the relative increase in selfinjury, due to increases in the use of hanging, jumping from heights and the use of other violent methods such as the use of firearms, jumping before a moving object or other traffic related injury, poses an important challenge to prevention. These methods are generally available and a restriction in the access to these methods appears to be difficult or even impossible to achieve [49].
The findings may also have implications for the A&E departments in the general hospitals [31]. Psychosocial assessment following self-harm, is a necessary starting point for preventive interventions. One of the strategies in the Flemish suicide prevention programme [36], aims at increasing the use of a standardized instrument to assess self-harm patients in general hospitals. As this may lead to reduced rates of subsequent repetition of self-harm [50,51]. In case when it is impossible to conduct a full psychosocial assessment, the A&E department staff should be aware that the use of highly lethal methods of self-injury such as the use of firearms and hanging is associated with high suicidal intent and a high risk of suicide so that intensive follow-up care and risk reduction measures are indicated for these individuals [9,25,26].
In conclusion, using data from A&E departments in major general hospitals this long-term monitoring study of the occurrence and characteristics of self-harm shows decreasing rates of self-harm between 1987 and 2013, however with a significant increase in the use of self-injurious methods. Given the fact that self-poisoning continues to be the most common method of self-harm, prevention strategies targeting physicians and their prescription practices continue to be important components of suicide prevention strategies. The proportional strong increase in the use of self-injurious methods in self-harm poses however new challenges to public health approaches in suicide prevention. In any case, the current findings underline the necessity of adequate management of self-harm patients in A&E departments of general hospitals as a very important component of suicide prevention programmes.