Prevalence of deliberate self harm and attempted suicide within contemporary Goth youth subculture: longitudinal cohort study

Introduction

Deliberate self harm is relatively common among young people, with rates of 7%-14% in the United Kingdom.1 2 Common acts of self harm include cutting, burning, and punching, usually resulting in relatively minor injury; rarer, more serious, acts include self poisoning.

Self harm is understood to be a maladaptive coping strategy intended to relieve negative emotions such as anger, anxiety, frustration, or guilt. It is usually unrelated to an immediate suicide attempt. Knowledge about risk factors is limited, but previous research has implicated peer modelling and depression.1 2 Self harm is related to later risk of suicide and psychiatric disorder and has a high prevalence among certain subpopulations, notably prisoners and homosexual and bisexual people.1 2 3 The media have linked contemporary Goth youth subculture with self harm,4 but evidence for this is sparse. “Goth” could be described as a subgenre of punk with a dark and sinister aesthetic, with aficionados conspicuous by their range of distinctive clothing and makeup and tastes in music.4 5 We investigated whether identification with Goth is associated with self harm.

Methods

We collected data on participants at age 19 through the west of Scotland 11-16 study, a longitudinal survey of health and lifestyles.6 7 Respondents were recruited during their final year (1994) of primary school (age 11, n = 2586) and resurveyed at ages 13, 15, and 19 (2002-4, n = 1258). As weights to adjust for attrition bias did not alter the results we report unweighted data. For youths aged 15 and 19 we used a computerised version of the diagnostic interview schedule for children (Voice-DISC)6 to collect data on psychiatric diagnosis, including a question on suicide attempts.

At age 19, during the Voice-DISC, participants were asked “have you ever in your whole life, tried to kill yourself or make a suicide attempt?” One section of the survey interview asked “have you ever tried to hurt yourself or harm yourself deliberately,” the methods used, and age at first act of self harm. Self harm was coded as any method and methods involving cutting, scratching, or scoring. In a separate section participants were also asked at what age and how much they identified (present and past) with a variety of youth subcultures, including Goth, on a five point scale. The two most extreme categories were collapsed and comprise the most highly identified group. We focus on current and lifetime peak (defined as highest current or past) identification.

We used logistic regression, with lifetime self harm and lifetime suicide attempt as outcomes, adjusted for sex, social class of head of household (coded, non-manual, manual, or unclassifiable according to the registrar general's schema of occupational social class),7 8 lifetime parental separation or divorce, substance use (smoking, any drug, alcohol), and the highest score on a depression scale administered at ages 11, 13, and 15.9

Results

Table 1 shows the characteristics of the sample according to level of Goth identification. No differences were found by social class, parental separation, rates of smoking, alcohol use, or previous depression, but males were more likely to identify with Goths (Pearson χ² test, 8.582, df = 3, P = 0.035), and rates of drug use were slightly higher among those who were most highly identified (Pearson χ² test, 7.318, df = 3, P = 0.062).

Table 2 shows the results for lifetime rates of self harm (any method); self harm from cutting, scratching or scoring; and attempted suicide (rates for overall sample: 7.1%, 4.1%, and 6.4%). Lifetime self harm (any method) and lifetime suicide attempt were highly correlated (r = 0.59).

Current Goth identification was strongly associated with lifetime prevalence of self harm and attempted suicide, with a prevalence of 53% for self harm (any method); 47% for self harm involving cutting, scratching, or scoring; and 47% for lifetime suicide attempt among the most highly identified, and evidence suggesting a dose-response relation. Predictors of self harm and suicide attempt were being female, having divorced or separated parents, smoking and any drug (not alcohol) use, and prior depression. Adjusting for these factors did not attenuate the Goth identification effect, which remained the single strongest predictor of either self harm or suicide attempt (table 2). Lifetime identification produced similar results (see table A on bmj.com).

Of 25 participants with a high identification (at some point in their lifetime) with the Goth subculture, 12 had harmed themselves; five before identification as Goth, two after, and four at about the same time (one participant had poor recall).

To determine how specific this identification effect was to Goth, as opposed to a general effect attributable to any other subculture, we carried out a series of additional analyses substituting Goth identification with 14 other common youth subcultures (table 3). Model 1 shows the association (odds) between each of the subcultures (dichotomised as “heavy” or “I am one,” compared with “none,” “just,” or “quite a bit”) and lifetime self harm, after adjusting for confounders. Although some other subcultures were also associated with self harm (Punk, odds ratio 4.42, 95% confidence interval 1.28 to 15.33; Mosher, 3.49, 1.08 to 11.27), the association was strongest for Goth (14.16, 4.42 to 45.39). Goth identification remained the only subculture which significantly predicted self harm after adjusting for other subcultures (model 2, table 3). Results were similar for self harm involving cutting, scratching, or scoring and for lifetime suicide attempt (see tables B and C on bmj.com).

Discussion

Identification by youth aged 19 as belonging to the Goth subculture was the best predictor of self harm and suicide attempt. This effect was not attenuated by adjusting for identification with any other youth subculture. Self harm could be a normative component of Goth subculture including emulation of subcultural icons or peers who self harm (modelling mechanisms). Alternatively, it could be explained by selection, with young people with a particular propensity to self harm being attracted to the subculture.

Although our study is based on small numbers, our data suggest that both processes are involved, with selection mechanisms possibly being more likely. Replication in alternative locations is needed to determine if this is widespread or localised, and a persistent or transient phenomenon.