Should the recently reported increase in Australian suicide rates alarm us?

Any suicide is one too many. Psychological autopsy studies suggest that many suicides are potentially preventable. Increases in suicide rates commonly spur demands for governments to promote actions and policies aimed at prevention of suicide.

In September 2016, the Australian Bureau of Statistics (ABS) released ‘preliminary’ statistics concerning suicides in 2015. There were 3027 suicides (2292 males, 735 females), the rate per 100,000 having increased significantly since 2013 (95% confidence interval [CI] = 0.99, 2.24; p < 0.001). The male rate increased 16%, from 16.7 in 2013 to 19.4 in 2015; the female rate increased 10%, from 5.6 to 6.1. Between 1998 and 2013, suicide rates of young and elderly men and women had decreased, the former dramatically (Snowdon, 2016). Why the reversal? Should we be alarmed?

How good is the data collection process?

The ABS has recommended caution when interpreting suicide data, especially regarding cases where the cause of death remains undetermined. The ratio of suicides to ‘undetermined’ cases varies between countries and over time. The suicide rate in England and Wales is reportedly much lower than Australia’s, partly attributable to its higher rate of ‘open verdicts’. Furthermore, 50% of medicine poisoning deaths given accident verdicts were rated by researchers as suicides (Gunnell et al., 2013). Since 2006, the ABS has allowed revision of suicide figures during the first 2 years after their first release so that in cases where coronial decisions are delayed, the records can be updated. In 2006–2010 the ratio of suicides to ‘undetermined’ cases was 5.3:1. The ‘undetermined’ rate was 1.6 per 100,000, while in New Zealand (NZ) it was 0.6. The ABS noted a reduction in ‘undetermined’ cases in 2012–2013; in 2014, the ratio was 10:1.

The 2015 figures have been released by the ABS 6 months ahead of expectation. NZ’s Ministry of Health has been more cautious, stating in July 2016 that 2014 data were still too incomplete for official figures to be released.

Why have rates increased?

The reported increases in male and female suicide warrant close examination. In order not to rely too heavily on a single year’s figures, the 2014 and 2015 rates and those of previous 5-year periods have been averaged (Tables 1 and 2). Age patterns changed dramatically between the 1960s and 1980s.

Table 1.

Australian male suicide rates (per 100,000).

	1964–1968	1979–1983	1984–1988	1989–1993	1994–1998	1999–2003	2004–2008	2009–2013	2014–2015
15–19	6.4	10.7	15.8	18.8	17.9	13.8	10.3	11.4	11.9
20–24	15.2	25.9	30.9	34.9	38.5	28.4	22.4	17.5	21.5
25–29	19.8	24.8	28.5	33.0	38.3	34.4	24.4	19.3	21.4
30–34	25.0	21.3	26.1	28.8	34.9	34.6	27.8	22.0	25.7
35–39	26.8	21.1	21.8	27.8	32.1	31.8	26.3	24.6	26.0
40–44	31.6	23.6	26.8	23.4	29.0	29.0	26.5	26.0	30.2
45–49	30.7	24.4	23.0	26.0	24.6	26.5	24.3	26.3	28.3
50–54	35.6	22.7	25.1	24.1	25.6	21.8	21.6	22.8	29.8
55–59	36.6	24.3	25.6	24.8	23.6	18.6	17.3	20.2	23.5
60–64	35.2	24.3	23.5	22.8	23.2	17.9	17.0	18.5	19.7
65–69	40.0	25.4	23.7	24.2	22.8	19.1	17.0	15.4	17.6
70–74	37.9	26.2	30.3	28.4	22.7	20.6	18.4	16.3	17.4
75–79	41.4	30.0	34.3	29.3	28.0	23.9	18.8	21.3	20.3
80–84	36.3	35.4	36.1	36.9	34.7	24.3	23.0	24.6	23.8
85+	33.6	38.9	43.2	43.5	37.7	36.2	32.2	33.8	38.4
All ages	22.1	18.6	20.6	20.7	22.1	19.6	16.6	16.8	19.0

Note: Suicide rates of two male cohorts, born 1960-1968 (bold typeface) and 1940-1948, are underlined to draw attention to cohort differences that persist across adulthood.

Table 2.

Australian female suicide rates (per 100,000).

	1964–1968	1979–1983	1984–1988	1989–1993	1994–1998	1999–2003	2004–2008	2009–2013	2014–2015
15–19	2.9	2.4	3.8	3.7	4.6	4.4	4.4	5.3	6.5
20–24	8.2	6.0	6.3	6.2	7.1	5.2	5.0	6.4	6.9
25–29	11.2	7.4	6.4	6.8	7.5	7.7	5.4	5.6	6.7
30–34	13.2	5.9	6.0	7.1	7.0	8.1	6.1	6.4	7.1
35–39	15.4	8.5	7.3	6.5	8.7	7.9	6.8	7.5	8.9
40–44	17.1	8.8	6.8	6.9	7.8	8.0	6.9	7.9	8.3
45–49	22.9	10.7	10.0	7.2	6.8	7.4	6.5	7.1	9.2
50–54	20.6	11.6	8.8	7.2	8.2	6.7	7.0	8.2	9.0
55–59	20.2	10.6	8.7	8.0	6.9	6.2	5.7	7.0	8.1
60–64	21.2	9.7	8.1	7.2	6.3	5.4	5.3	5.6	5.7
65–69	20.0	8.7	7.8	7.4	5.6	5.0	4.7	4.1	5.3
70–74	15.7	9.3	7.7	7.6	6.1	4.9	5.4	4.9	6.5
75–79	11.3	8.7	8.6	8.1	6.7	4.4	5.7	5.0	4.4
80–84	7.8	8.2	8.1	8.1	7.0	5.8	4.0	6.2	6.3
85+	8.5	3.2	6.9	6.9	4.8	4.8	3.9	5.7	6.1
All ages	11.9	6.3	5.8	5.2	5.4	5.1	4.7	5.3	6.1

Period effects are the impacts caused by environmental circumstances or changes. For example, socio-economic crises, increased unemployment, and availability of addictive drugs and health and supportive services may affect one or more groups within a population. Analysis of Australian suicide data has revealed age, period and cohort effects (Snowdon and Hunt, 2002). The latest data have been examined to assess whether recent increases in rates could be attributable to any of these effects. Tables 1 and 2 show that

1. The suicide rates of almost all male and female age-groups increased between 2009–2013 and 2014–2015. Male suicide rates at age 20–39 in 2014–2015 were higher than in 2009–2013, but lower than in 1989–2008. Male rates per 100,000 at age 40–54 in 2014–2015 were higher (by 4.5 per 100,000) than the corresponding averaged rates in 1979–2013. Rates at age 55 or more were higher (by 1.75) than in 1999–2013, but substantially lower than in 1979–1998. The female rate at age 15–19 in 2014–2015 (6.5) was higher than previously at that age and at 35–59 was somewhat higher (by 1.4) than the average it had been since 1984–1988. Female rates at age 60 years or more were lower in 1999–2013 than before, but rose by 0.7 in 2014–2015. In 1964–1968, suicide rates of males aged 40–79 and females aged 20–79 were higher than throughout 1979–2015.

2. The peak suicide rate among non-elderly males was at age 20–29 in 1984–1998, 25–34 in 1999–2003, 30–44 in 2004–2008, 40–49 in 2009–2013, and 40–54 in 2014–2015. A cohort effect was evident: those with a high rate (30.9) at age 20–24 in 1984–1988 (born1960–1968) still had a high rate at age 45–51 (2009–2015) (bold lettering in Table 1). Those with suicide rates of 21.1 and 26.8 at age 35–44 in 1979–1988 still had relatively low scores at age 65–71 in 2009–2015. No definite cohort effect was seen among Australian women (Table 2). A comparable cohort effect was noted in the United States, where (e.g.) suicide rates of men born in 1985–1989 were twice as high as those born in 1930–1934 (Phillips, 2014).

3. There is clear evidence of an age effect: suicide rates of males were progressively higher in each age-group older than 65 from 1979 onwards. In 1964–1968, the rate was above 35 per 100,000 among those over 50 years. Among women, rates have tended to be higher in middle age and lower in late old age, although more obviously so in 1964–1968.

4. Period effects have been apparent, interacting with cohort effects. Suicide rates of males aged 20–39 years in 1979–1983 increased across the 5-year periods up to 1994–1998. Rates for those aged 50–69 were not affected in those years, but rates of almost all age-groups fell from then until 2009–2013. It is plausible that a period effect led to the increased male rates in 2014–2015, variably influenced by cohort effects. Suicide rates at age 20–29 of cohorts born in 1975–1993 were much lower than those of male cohorts born 20 years earlier. This provides good news: we can expect that, providing period effects can be controlled, the suicide rate of males now aged 15–39 will, as they get older, be much lower than that of currently middle-aged males.

Ethnicity

The ABS reported that in 2015 the suicide rates per 100,000 in the Aboriginal and Torres Strait Islander population (constituting 2.5% of the Australian population) was 25.5 (male: 39.2, female: 12.9); the non-indigenous rate was 12.5 (male: 19.0, female: 6.2). The suicide rates in 2011–2015 of persons from indigenous backgrounds were almost double those of non-indigenous persons (binomial proportion test p < 0.001), but with disproportionately high rate ratio differences among young people (Table 3). This high and increasing suicide rate of younger indigenous persons partly accounts for the latest increase in suicide rate of younger Australians. A decade ago, Hunter and Milroy (2006) pointed to a dramatic increase in suicides of young indigenous Australians. Now the increase in the indigenous female suicide rate is equally striking; for example, the suicide rate of indigenous women aged 35–44 is four times higher than that of same-aged non-indigenous women.

Table 3.

Indigenous versus non-indigenous suicides 2011–2015 in New South Wales, Queensland, South Australia, Western Australia and the Northern Territory.

Age	Number of suicides	Suicide rate per 100,000	Number of suicides	Suicide rate per 100,000	Ratio indigenous to non-indigenous
	Male indigenous		Male non-indigenous
1–14	13	2.6	23	0.3	7.8
15–24	173	53.7	772	14.4	3.7
25–34	137	63.6	1226	21.0	3.0
35–44	96	55.0	1420	25.7	2.1
45+	70	24.0	3456	16.7	1.4
All ages	489	31.8	6897	17.4	1.8
	Female indigenous		Female non-indigenous
1–14	13	2.7	24	0.4	7.3
15–24	77	25.4	299	5.9	4.3
25–34	48	22.1	337	5.9	3.7
35–44	40	21.3	269	4.9	4.4
45+	22	6.8	1327	6.0	1.1
All ages	200	11.90	2256	4.2	2.8
	All indigenous male and female		All male, female non-indigenous
All ages	689	22.3	9153	11.5	1.9

The Australian Bureau of Statistics (ABS) provided combined 2011–2015 data concerning the most common causes of death in different age-groups. Suicide was among the 10 most common in all age-groups below age 45 years, but not in those aged 45 years or more and not in the female population as a whole. In the above table, the rates for those aged over 45, and for the total female populations, have been calculated using estimated population numbers (total and aged 45+ years) based on the ABS 2011 census figures.

Conclusion

Suicide Prevention Australia has called for Australia to ‘match its prevention efforts and investment to the magnitude of the public health problem we face’. It recommends prioritisation of culturally appropriate indigenous suicide prevention strategies. The reported increase in Australian suicide rates since 2013 is of concern. The main need at present is to discover what combination of factors has contributed to that increase. Partly it is attributable to the rising suicide rate of indigenous children and younger adults. What else? Socio-economic concerns? The suicide rate of non-indigenous Australians is much lower than 50 years ago; the increase since 2013 need not alarm us, except in regard to younger indigenous persons. Special attention is needed for cohorts now in middle age. As younger cohorts reach middle age, a reduced rate can be predicted – as long as period effects do not persist. Enthusiastic bipartisan support for a National Suicide Prevention Strategy is needed.