Rural youth in distress? Youth self-harm presentations to a rural hospital over 10 years

KEYwords: Adolescent, emergency service, mental health, rural, self-injurious behaviour, young adult

Rural youth are increasingly identified as a group at risk for adverse mental health outcomes, abroad and to a lesser extent in New Zealand (NZ).^1–4 Despite higher suicide rates in NZ in youth aged 15–19 years than in any other Organisation for Economic Cooperation and Development (OECD) country, little is known about youth self-harm rates in NZ, and even less is known about rural–urban differences.^5–8

Self-harm is a strong risk factor for completing suicide, so it is important to establish and monitor rates of self-harm.^9–11 Although more common than suicide, most self-harm behaviours occur with no involvement of health services.^8,11 This raises challenges for accurately capturing rates of self-harm without implementing intensive widespread surveillance. Accordingly, the World Health Organization (WHO) has endorsed monitoring emergency department (ED) self-harm presentation rates as an acceptable alternative, because the frequency of events is higher than suicides but easier to calculate than community-based self-harm.⁹

NZ has a significant proportion of youth living in areas that are formally classified as ‘urban’ yet depend on rural health services.^12,13 In addition, the Ministry of Health reports a trend towards increased suicide rates in rural populations, but does not draw any explicit rural–urban comparison in self-harm data.⁶ Further, there is a concerningly high level of suicide among NZ Māori compared with non-Māori, typical of indigenous populations worldwide.⁹ However, non-Māori generally present to EDs for self-harm at higher rates than Māori.⁶

To gain further understanding of suicidality among NZ rural youth, the present study aimed to quantify youth self-harm presentations to one rural hospital ED from 2008 to 2017. The study was particularly interested in comparing the first and second 5-year periods as the effects of a major natural disaster (the Christchurch earthquakes) impacted this rural area just before the study midpoint. Secondary outcomes included self-harming rates within demographic subgroups and other characteristics. Attempts were made to compare local self-harming behaviour and suicide rates with available national rates.

Patients aged 15–24 years who presented to the Ashburton Hospital ED for ‘self-harm’ from January 2008 to December 2017 and who were domiciled in Ashburton District were included in the study. The definition of self-harm used in this study did not differentiate suicidal intent from non-suicidal intent, because that information was not available. This age group and definition allowed direct comparison with national rates. Both short-stay ED visits (up to 6 h) and longer admissions were included. Patients who were rapidly transferred to the nearest tertiary hospital before admission locally were not included.

The study design was a quantitative retrospective audit using electronic hospital records. Hospital admission records were merged with mental health outpatient records to ascertain the proportion of patients having contact with specialist services and whether this was a marker of severity. The data were provided to the researchers in a deidentified electronic format.

A data request was made to the Canterbury District Health Board (CDHB) data service for each eligible individual’s age at presentation, date and time of presentation, home suburb, sex, ethnicity, presence or absence of a general practitioner (GP) associated with their National Health Index and, if available, details of contacts they had with mental health services and the method(s) they used to self-harm.

The WHO International Classification of Disease codes (https://icd.who.int/browse10/2016/en, accessed 24 June 2019) for hospital admission (X60, X61, X62, X63, X64, X65, X66, X68, X69, X72, X75, X76, X77, X79, X80 X83, X84, U73.8) and ED assessment codes (DSH, PSD) inputted by the CDHB were provided, which matched the codes used by the Ministry of Health. Population, domicile and income information were derived from projections from the 2006 and 2013 census figures.

Once calculated, rates were compared across years with 95% confidence intervals (CIs). The 10-year period was further divided into two periods, namely ‘peri-earthquake’ (2008–12) and ‘post-earthquake’ (2013–17) periods. Annualised rates were calculated from averaged annual self-harm presentations over estimated median populations at the period mid-point. For comparison, the researchers also accessed coronial regional suicide data and national suicide and self-harm rates.^6,7

Three variables were postulated to be potentially associated with more severe clinical presentations: re-presentation to hospital with self-harm, use of multiple methods for self-harm and admission to hospital (not just short stay). Rates of positivity in these putative proxies for severity were compared against age group (15–19 vs 20–24 years), ethnicity (Māori vs non-Māori), socioeconomic status (low median income domicile vs high income) and remoteness from services (resident within domiciles corresponding to Ashburton township suburbs vs living in other domiciles within the district).

Comparisons of demographic variables were undertaken using the Chi-squared test of independence. Comparisons between the two five-year periods assumed the Poisson distribution for the number of events. The time period selected was large enough to detect (at a two-sided significance level of 0.05 with a power of 0.80) differences between ethnicities and sexes. Census information was extrapolated to obtain an estimate of the total population of demographic groups of interest in the Ashburton region. This estimate was then used as the denominator for calculating presentation rates, to account for group-to-group variances such as a generally younger population curve for Māori.

Ethics approval for the research was granted by the Health and Disability Ethics Committee, Northern B Branch (Reference 17/NTB/218). This study was approved by the local health authority, CDHB Research Office and the local Māori authority, Te Komiti Whakarite.

There were 197 presentations to the Ashburton ED coded as due to self-harming in people aged 15–24 years over the 10-year study period. As indicated in Table 1, of the 197 patients presenting, 75.1% were female, 88.3% were non-Māori and 68% were aged <20 years. Re-presentations numbered 62 (31%), occurring in 27 of a total of 162 patients (17%). Most presentations were for self-poisoning, often with multiple drugs or toxins, and, in line with current hospital policy, most were admitted to hospital for an overnight stay. The prominent substance type ingested was prescription drugs (n = 150; 76.1%). Alcohol, biological and industrial substances were also used.

Table 1.  Demographic and descriptive features of 197 presentations of 162 youth with self-harming behaviour, 2008–17, to Ashburton emergency department
GP, general practitioner

Most participants were able to name a GP (81%), although this was significantly lower for Māori (57%, Chi-squared = 9.8, P = 0.002). Sex, rurality and median income were not associated with presence or absence of a GP.

Over the last two years of the study time period there was an upward trend in presentation rates over baseline, but the increase was not clinically significant for either year individually. Most presentations were after working hours and clustered around weekends rather than weekdays, with two-thirds of presentations occurring over the cooler months of the year.

As shown in Table 2, there were significantly higher rates of self-harm presentations in the second, ‘post-earthquake’ period than in the first ‘peri-earthquake’ period, both overall and specifically among Māori, younger (youth aged 15–19 years) and female youth compared with non-Māori, male and older youth. The marked increase in Māori youth self-harm rates was further scrutinised and the annual rates are shown in Fig. 1.

Table 2.  Annualised rates of self-harm (per 100 000 youth) for 2008–12 compared with 2013–17
CI, confidence interval; SES, socioeconomic status

Fig. 1.  Comparison of annual rates (per 100 000 youth) of youth self-harm presentations between Māori and non-Māori. Data are the incidence rates ± 95% CIs.

Residence in a suburb with a lower median income than neighbouring suburbs was a marker for significantly more presentations for self-harm than made by people with higher-income suburb residence. The annualised incidence over 10 years was 765 per 100 000 (95% CI 626–926 per 100 000) compared with 427 per 100 000 (95% CI 345–524 per 100 000). Presentations increased significantly from the first to the second 5-year period. Increased rurality was inversely associated with presentations for self-harm: 290 per 100 000 (95% CI 213–387 per 100 000) in more rural domiciles versus 778 per 100 000 (95% CI 659–913 per 100 000) within the Ashburton township.

The lowest rate of presentations occurred during 2011, the year of the Canterbury Earthquake. From 2012 onwards, youth self-harming rates for non-Māori remained similar to peri-earthquake rates, but Māori rates rose steeply, with an almost ninefold increase during the 2013–17 compared with 2008–12 period, from 116 to 1106 per 100 000, a difference of 990 per 100 000 (95% CI 471–1509 per 100 000; Chi-squared = 14.0, P < 0.001).

Among three proxy markers of severity (namely admission to hospital rather than ED assessment and treatment only, multiple presentations for self-harm and use of multiple methods (usually multiple drugs or toxins)), the only association found with the demographic characteristics studied was an inverse correlation between the older age group (20–24 years) and admission to hospital (22% not admitted vs 6% of 15–19 year olds; 95% CI for difference in rates 6–26%, Chi-squared = 11.4; P < 0.001).

As shown in Table 3, both the rates of self-harm and completed suicide per 100 000 youth over the 10-year period under study are higher in the whole rural Ashburton District than the national average for these rates in 2013, which is the mid-point of the period.

Table 3.  Local self-harm and suicide rates (per 100 000 youth) compared with available national data

This study examined rates of youth self-harming behaviour presenting to Ashburton ED, which serves a geographically rural area of NZ. The key finding was an increased rate of self-harm over the latter part of the study time period.

The observed rise in self-harm presentations in the recent 5-year period was greatest among Māori, which contrasts with national data showing non-Māori presenting with self-harm more than Māori.⁶ Any potential effect of the younger population curve for Māori compared with non-Māori has been accounted for, at least until the census in 2013, as ethnicity data for specific age group populations in the district were available.¹⁴ It is possible that improved ethnicity data collection in recent times may explain part of the jump in rates. However, it seems more likely that young Māori in this district have a real increased risk of presenting with self-harm due to increased rates of distress or changed help-seeking behaviour. In addition to the outstanding burden of health and socioeconomic inequities that have persisted since colonisation, Māori settled in the Ashburton District have typically migrated from other areas of NZ. Being distanced from ancestral whenua (land) and whanau (family) could potentially be adding a further sense of displacement.^15,16

The further findings of increased presentations for self-harm among the younger age group, among females and among youth from lower socioeconomic groups is not unexpected. This study does not suggest that increased rurality or remoteness from health services in the district is an additional risk factor for presenting with self-harm, because youth living in the district town of Ashburton were more likely to present with self-harm and there was a trend towards an increase in self-harm over the two periods among youth living within the town. Although geographical proximity and ease of access to health services may explain the difference in part, it could also be that youth living outside the town boundary have greater connectedness to people and place than youth from the district town, especially in a time of much internal migration. Nevertheless, as with young Māori and young males, a lower rate of ED presentations for self-harm may contrast with higher rates of completed suicide. No increase in severity of presentations among youth from more remote residences was detected in this study to support this possibility.

Other trends observed include an increase in presentations after hours and in late winter months, which fits with global trends.¹⁷ This may have implications for mental health staffing and targeted helpline marketing.

Aligned with global trends, the common method of harm among Ashburton youth presenting to ED was poisoning, generally paracetamol, antidepressants and antipsychotic agents, with or without alcohol.³ Self-harm involving synthetic opioids currently represents only a small proportion of presentations, but very high rates of use in other OECD countries means that opioids warrant close attention.¹⁸ It was surprising that cutting did not make up more than 10% of presentations, given its apparent frequency in the community. One possible explanation is that cutting has become so commonplace in recent years that health-care providers are no longer noting it in records, especially if patients present with a matter (usually self-poisoning) that requires more urgent attention. Discussions with staff and auditing of 1 year of hospital records (SG Withington, unpubl. data) tend to support the hypothesis that under-reporting of cutting is significant. It is difficult to estimate more accurate prevalence of cutting without conducting a population-wide survey. Local ambulance services attending call-outs for cutting report often directing patients to mental health services where ED care is not required, as with school counsellors taking students to practice nurses, highlighting a further group of missing presentations but another possible target for surveillance.

Finally, the present study found that local rates of both youth self-harm presentations and suicide averaged across 10 years were higher than national rates at the mid-point of the time period under study. Further qualitative investigations of the mechanisms behind this difference are warranted. Potential factors may include differential access to mental health services and school counselling services, the effects of rapid social change, a potential waning of resilience after the earthquake, and specific effects of rurality, access to self-harm means and remoteness of the district.

This study was limited by the amount of data available from electronic hospital records. Because only specified codes relating to self-harm were used, other methods of intentional self-harm (e.g. single-occupant motor vehicle accidents or self-withholding of essential medication) were not captured, but are likely to be a small proportion. When the number of admissions for self-harm was checked manually using the hospital admission book in one year (2016) for audit purposes, an additional 10% of cases was identified, notably cases that were rapidly transferred out of Ashburton Hospital because of the severity of the self-harm episode. These cases were instead registered electronically at the closest tertiary hospital in Christchurch, 1 h away by road. These severe cases could not be included in the analysis, but suggest the problem is greater and more severe than observed in this study. A subsequent review of electronic data from Christchurch ED confirmed an additional 22 presentations of Ashburton domiciled youth self-harm cases.

A potential explanation for the recent increase in self-harm is a change in primary care services. In December 2016, the evening time limit for primary care cover by GPs in Ashburton was reduced from 11 pm to 8 pm. Although this may have contributed to the increase in cases presenting to Ashburton Hospital, a subanalysis of rates of youth presentations during the evening between 8 pm and 11 pm showed these did not increase in 2017, suggesting self-harming youth presenting after hours to general practice would be referred to the hospital in any case.

There is no central collection of sexuality and gender diversity information, although these are associated with differential vulnerability to self-harm.^8,11 Occupational data were also not available. Data was sourced only from the CDHB because there are no inpatient non-government mental health providers in Ashburton. No information was available on duration of stay in the district, which may be relevant and warrants further investigation. The lack of identification of a GP may partially correlate with transient residence in the district and, in this study, was more prevalent among Māori than non-Māori. This study was also limited by the small amount of up-to-date national data available, with the most recent self-harm data from 2013. National self-harm data before 2013 was of low comparability due to the non-collection of short-stay ED visits for self-harm. The most recent census data used was also from 2013, so population denominator estimates are somewhat old in a rapidly growing and changing district.

Importantly, this study was quantitative only, so did not contain the perspectives of youth receiving care and, being centre based, could not explore issues of differences in access to help versus severity of distress as conflicting explanations of differential presentation rates. The term ‘self-harm’ is used in this study to incorporate both non-suicidal self-injury and self-injury with suicidal intent, because International Classifications of Diseases (ICD) codes do not allow for differentiation. However, both are strongly associated with completed suicide, so combined analysis was appropriate for the purpose of this study. Internationally, self-poisoning has been reported as more often involving suicidal intent, whereas cutting is mostly used in non-suicidal self-injury and relief from distress.¹¹ The low rates of cutting presenting at this ED could imply that a large number of the presentations were suicide attempts and not non-suicidal self-injury. Finally, this study was limited by low absolute numbers. In future, the combining of data from other rural hospitals would be beneficial.

This study is one of very few in NZ specifically examining self-harm among rural youth. It identifies a potentially growing and significant self-harm problem in the Ashburton rural community. Young Māori are over-represented in recent years, alongside females, younger youth aged 15–19 years, youth from poorer suburbs and those living in the more built up area of a rural district. This study found rates of self-harm (and suicidality) over 10 years that appear to be higher than in the general population, which raises questions about what is occurring in other rural communities. Further studies, including qualitative research and community-level studies, are warranted to further explore the issues raised. Previous self-harm with or without suicidal intent is the strongest risk factor for completing suicide.⁹ By advancing the understanding of these trends and driving development of effective community interventions, a decrease in youth self-harm would contribute to the ultimate goal of reducing youth suicide.^9,11

Isobel Ferguson received financial support for the study from the Advance Ashburton Community Foundation.

The authors declare no conflicts of interest.