The use of mental health services, psychological distress, loneliness, suicide, ambulance attendances and COVID-19

The AIHW has been compiling detailed data every week on the use of mental health services and the various crisis help lines since April 2020 as part of the National Suicide and Self-harm Monitoring Project.  These data are shared within government to inform the mental health response to the COVID-19 pandemic.

There have been notable rises in the use of crisis lines and mental health services since the onset of the COVID-19 pandemic. For example, calls to Lifeline in the 4 weeks from 10 August to 6 September 2020 were up 15.3% compared with the same time in 2019. Contacts to Beyond Blue (total of call, web chat and email) were up 38.6% over the same period while contacts (total of call, web chat, email and circles sessions) to the Kids Helpline were up 24.5%.

It is not clear to what extent this rise in contacts is driven by rises in psychological distress rather than a higher proportion of people seeking assistance for other reasons such as loneliness and concern about contracting COVID-19. Given that a range of survey data indicate that there have been rises in the level of psychological distress in 2020, increased contact with the crisis lines is almost certainly indicative of a rise in the need for assistance as a result of the pandemic.

The crisis line data also shows variation by state and territory with greater rises in the use of crisis lines evident in Victoria (and to a lesser extent New South Wales) when comparing the 4 weeks from 10 August to September 2020 to the same time in 2019, than in other jurisdictions combined.

There have also been rises in the use of Medicare subsidised mental health services. For example, the total number of mental health related MBS services provided in the 4 weeks to 6 September 2020 was 12.5% higher than in the equivalent period in 2019. The equivalent increase in Victoria over this period (20%) was considerably stronger than in other jurisdictions.

There has been strong take-up of MBS mental health services provided through telehealth—over a third of all MBS mental health services in the 4 weeks to 6 September 2020 were provided through telehealth.

There is a correlation between high levels of psychological distress and common mental health disorders. As a result, instruments such as K10 and K6 can be used to track probable changes in the incidence of these disorders. This is important, as there is an association between mental health issues and deaths by suicide. Data from the Queensland Suicide Register for 2014–16 based on police and coroners reports, suggest that ‘mental health conditions were prominent in those who died by suicide, with 51.5% reportedly having a diagnosed mental health condition.’ (Leske et al. 2020). Just over 65% of people who died by suicide in 2018 had mental and behavioural disorders recorded as an associated cause of death (ABS 2019).

There are several ways to gain insights into the level of psychological distress in the community, and monitor trends over particular time periods.

One way is to look at trends in the use of mental health services. The AIHW has been compiling data each week on the use of mental health services and crisis lines during the pandemic. While this approach is useful, it is not a direct measure of the level of psychological distress as it does not capture those who do not seek out or do not have access to mental health services and crisis lines.

Another way to analyse trends in the level of psychological distress since the onset of the pandemic is to use sample surveys. This approach has been challenging since the onset of COVID-19 due to the fact that face-to-face surveys are very difficult to undertake at this time and pose a potential health and safety risk to interviewers and interviewees. This has led to a number of online surveys being conducted but many of these surveys are not based on probability sampling. In some cases, samples are drawn by inviting all members of the public above a certain age to respond, with unknown response rates. Other samples are drawn from panels where individuals opt-in online. While this sort of approach can provide some useful information, especially regarding associations between factors that may affect outcomes for respondents, results may not be representative of the Australian population. A major report on online panels for the American Association for Public Opinion Research (AAPOR 2010) noted that:

Given the need for representative data, the AIHW collaborated with the Centre for Social Research and Methods at the Australian National University to include questions on loneliness and the level of psychological distress using the Life in Australia^TM Panel, managed by the Social Research Centre. Importantly, this panel exclusively uses random probability-based sampling methods and covers both online and offline populations (that is, people who do and do not have access to the internet). In addition, as a panel it is possible to obtain longitudinal data including from the same respondents prior to the spread of COVID-19 which provides richer information than a series of cross-sectional snapshots, especially with regards to changes through time. Data on psychological distress were collected in April, May and August with further data collections planned for November, 2020.

Pre COVID–19 snapshot

To understand how COVID-19 may have affected Australians’ levels of psychological distress, it is important to look at data from before the pandemic. This is possible using results from the Australian Bureau of Statistics’ National Health Survey, which is conducted approximately every 3 years. It is particularly important to consider any existing trends prior to the pandemic—for example, if psychological distress was generally increasing among Australians in the years before COVID-19. Tables 1 to 3 show the proportion of males, females and people with high or very high levels of psychological distress as measured by the Kessler 10 Scale from 2004–05 to 2017–18. While the results vary by age, there is no consistent trend over this period. It is worth noting, however, that young women aged 18–24 consistently have higher levels of psychological distress than other groups. Having said that, there have been small increases in the proportion of both males and females with high or very high levels of psychological distress from 2011–12 to 2017–18, albeit with important fluctuations through time. The National Drug Strategy Household Survey (NDSHS) also shows rises from 2010 to 2019 (AIHW 2020).

Psychological distress and loneliness under COVID-19

The ANU poll results that are based on just over 3,000 respondents suggest that the proportion of the population experiencing severe psychological distress as measured by the Kessler 6 Scale rose from 8.4% in February 2017 to 10.6% in April 2020 with a subsequent fall to 9.7% in May 2020 (still significantly higher than in February 2017). However, there is a distinct pattern by age (Biddle et al. 2020b):

However, between April and May 2020, there was a slight improvement in psychological distress for young Australians, though psychological distress is still significantly and substantially above what it was in February 2017.

For older Australians (those aged 55 years and over), there was a slight improvement in the level of psychological distress between February 2017 and April 2020, and a further improvement between April 2020 and May 2020 (Biddle et al. 2020b). Based on a smaller ANU Pulse Survey, there was some evidence of further improvements in the level of psychological distress in late June/early July, however this was prior to increased COVID-19 infection rates and a second lockdown in Victoria.

The ANU also asked also respondents whether ‘In the past week, how often have you felt lonely?’ Analysis summarised in Biddle et al. (2020b) shows that those who experienced loneliness had higher rates of psychological distress than those who did not.

However, there was a distinct pattern by age with 18–24 year olds the only group to show no statistically significant reduction in the level of loneliness from April to May despite the fact that this group had a higher proportion of respondents saying that they felt lonely at least some of the time in April (63.3%) than any other age group.

There was a rise in reported loneliness from 36.1% in May 2020 to 40.5% in August. However, this rise only occurred in Victoria where ‘the proportion of the population who were lonely at least some of the time increased from 35.7% in May 2020 to 44.5% in August’ (Biddle et al. 2020c). In ‘the other seven States and Territories, there was no significant difference between loneliness in May 2020 (37.1%) and August 2020 (38.8%)’.

According to Biddle et al. (2020c):

The ABS has also asked around 1,000 respondents about their level of psychological distress using Kessler 6 through their Rapid Surveys, which are based on a probability sample. However, the ABS sample does not have information from the same individuals prior to the spread of COVID-19. The results suggested higher levels of distress in April 2020 when compared with equivalent results from the 2017–18 National Health Survey (NHS). However, results in June (24–29 June) had improved so that according to the ABS:

One of the advantages of the data collected through ANU poll is the fact that longitudinal data are available. This makes it possible to model the factors that appear to be contributing to rises in the level of psychological distress during the pandemic. In this modelling (Biddle et al. 2020b), the strongest predictor of psychological distress (K6) is where people say that their stress has worsened (this is not that surprising). Increased loneliness is also a strong predictor of K6 scores even when other factors like changes in employment status are controlled for. This suggests that increased loneliness during the pandemic is of concern and that this is not just being driven by job loss.

Job loss itself was a predictor of K6 scores in May. Controlling for other factors, people who were employed in February 2020 but not in May had higher levels of psychological distress. Interestingly, in all the models those who live outside capital cities had lower rates of psychological distress than those who lived in capital cities, controlling for other factors. This probably reflects the fact that the economic impact of shutdowns has been higher in the major cities than it has been in regional or remote areas and that infection rates have also been higher. ABS employment data, for example, shows faster falls in employment in the major cities than in regional or remote areas.

Once you control for things like relationships worsening and stress rising and employment loss then young people were no longer worse off than older people when it comes to K6 scores. This suggests that rises in the level of psychological distress among young people are being driven by things like increased stress and job loss.

One final point worth noting. In the regression analysis of K6 scores, previous K6 scores in February 2017 had a significant predictive effect on K6 scores for May 2020. This shows that people who are already experiencing high levels of psychological distress can be particularly vulnerable when the situation worsens.

More recent data show that after improving from April to May the level of psychological distress worsened from May to August and remains higher than it was prior to the pandemic. However, the worsening from May to August was concentrated among women with males showing no change from May to August (Biddle et al. 2020c). In addition while the initial rises in the level of psychological distress were higher among young people (particularly those aged 18–24), the only age group to experience a rise in psychological distress from May to August was people aged 75 and over (Biddle et al. 2020c). However, the level of psychological distress remains considerably higher among young people.

To test whether outcomes have worsened in Victoria relative to the rest of the country since the reintroduction of lockdowns, in July Biddle et al. conducted a Difference-in-difference analysis using linked data for May and August (that is, data across these months for the same people). This showed, a significant worsening in Victoria relative to the rest of the country on several outcomes including: psychological distress, loneliness, life satisfaction, satisfaction with direction of country, likely to be infected by COVID-19 and hours worked (Biddle et al. 2020c).

Table 1: Proportion of persons with high/very high psychological distress, by age group and year

	18–24	25–34	35–44	45–54	55–64	65–74	75+	Total
2004–05	15.5	11.8	14.9	13.7	12.2	11.2	10.7	13.0
2007–08	11.8	12.3	11.5	13.2	13.2	9.0	10.8	12.0
2011–12	11.8	10.9	10.9	11.4	10.4	9.3	9.3	10.8
2014–15	15.4	10.9	11.9	12.4	11.0	10.1	9.7	11.7
2017–18	15.2	13.1	11.8	14.3	14.4	10.7	8.8	13.0

Source: ABS 4364.0.55.001 - National Health Survey

Table 2: Proportion of males with high/very high psychological distress, by age group and year

	18–24	25–34	35–44	45–54	55–64	65–74	75+	Total
2004–05	12.4	9.3	11.4	11.0	11.3	9.8	10.8	10.8
2007–08	6.5	10.9	8.8	11.1	11.6	7.4	8.8	9.6
2011–12	10.6	8.2	9.1	8.9	8.8	7.0	7.8	8.8
2014–15	11.1	9.8	10.0	9.2	9.8	9.7	9.0	9.9
2017–18	12.4	11.8	10.3	13.2	12.1	10.0	7.5	11.3

Source: ABS 4364.0.55.001 - National Health Survey

Table 3: Proportion of females with high/very high psychological distress, by age group and year

	18–24	25–34	35–44	45–54	55–64	65–74	75+	Total
2004–05	18.7	14.3	16.6	16.3	13.1	12.5	10.7	15.1
2007–08	17.4	13.9	14.2	15.3	14.7	10.6	12.3	14.3
2011–12	13.0	13.6	12.7	13.8	11.9	11.6	10.5	12.7
2014–15	20.0	11.7	13.7	15.4	12.3	10.3	10.6	13.5
2017–18	18.4	14.7	13.7	15.3	16.9	11.3	9.7	14.5

Source: ABS 4364.0.55.001 - National Health Survey

There has been considerable commentary since the start of the pandemic on its potential to impact on the incidence of deaths by suicide. Much of this commentary has been based on modelling based on previous experience including the relationship between unemployment and deaths by suicide. A ‘living systematic review’ (John et al. 2020) based on evidence until 7 June, has concluded that:

It is true that some key risk factors associated with deaths by suicide have worsened since the onset of COVID-19. For example, there has been considerable job loss and rises in the level of psychological distress. On the other hand, it is possible that a general sense of ‘we are all in this together’ could have a protective impact. From February to May 2020 there were rises in the level of trust in others and in governments in Australia (Biddle et al. 2020). In addition, some care is required as the vast majority of people who experience unemployment or high levels of psychological distress or mental health issues will never experience a suicide attempt. That said, it is very important to monitor trends in risk factors and trends in deaths by suicide in real time.

Another factor that should be considered is the impact of both JobKeeper and the JobSeeker supplement. This is important given the association between the risk of dying by suicide and socioeconomic outcomes. Modelling undertaken by the ANU suggests that not only are levels of poverty and housing stress lower than they otherwise would be a result of these payments they are also lower than they were prior to the spread of COVID-19 (Philips et al. 2020). Households who mainly relied on the JobSeeker payment prior to COVID-19 and the introduction of the JobSeeker supplement saw their poverty rate fall from 67% prior to COVID-19 to 6.8% (Phillips et al. 2020). On a similar note Biddle et al. (2020d) found that real incomes actually rose for those in the bottom decile of the income distribution from February to August 2020. Using data from the Taking the Pulse of the Nation Survey, Botha et al. (2020) have shown that the level of psychological distress among the unemployed has fallen since May. The ANU modelling suggests that the protective impact of JobKeeper and the JobSeeker supplement on housing stress and poverty have been reduced somewhat by the changes to these payments announced in July (Phillips et al. 2020).

In Australia, data on suspected suicides in 2020 have been released for both Victoria and Queensland from their respective suicide registers. We are not aware of similar data being released for other OECD countries apart from New Zealand. In all cases there is no evidence to date of any increase relative to previous years. It is true that these data are not based on final determinations on the part of Coroners, but in the case of the Victorian Suicide Register:

In recent years differences between the numbers in the Queensland Suicide Register and the data released by the ABS are quite small (Leske et al. 2020).

The number of suspected deaths by suicide in Victoria from 1 January to 30 September 2020 (530) was very similar to the equivalent period in both 2019 (534) and 2018 (520). The monthly data show considerable volatility which: ‘demonstrates the importance of not attributing too much significance to the suicide frequency in any one month’ (Coroners Court 2020b).

Consistent with data from previous years, around three-quarters of suspected suicides in 2020 were among males with most deaths by suicide for both males and females occurring among those aged between 25 and 55 (Coroners Court 2020b). In most age groups the number of deaths by suicide is consistent with previous years. The proportion of suspected suicides in metropolitan and regional areas in 2020 is similar to earlier years with approximately two-thirds occurring in metropolitan locations.

Data from the Queensland Suicide Register (Leske et al. 2020) show that the number of suspected suicides from 1 January to 31 July 2020 (454) was similar to that of the same period in 2019 (445) and 2017 (456).

In analysing trends in the number of deaths by suicide it is important to take population into account given strong population growth in Australia—it is important to focus on trends in suicide rates. Monthly age-standardised suspected suicide rates in Queensland in 2020 for both males and females are similar to the previous 5 years; see Figure 1 and Figure 2 replicated with permission from Leske et al. (2020) and including updated data for August 2020.

Figure 1:  Monthly age-standardised suspected suicide rate per 100,000, Queensland males, 1 January 2015–31 August 2020

Figure 2:  Monthly age-standardised suspected suicide rate per 100,000, Queensland females, 1 January 2015–31 August 2020

While data for Queensland does not show rises in suspected suicide rates compared with previous reports the following is worth noting (Leske et al. 2020).

The Chief Coroner of New Zealand released annual provisional data on suicides for the year to 30 June 2020 in 21 August this year. The data show that the provisional suicide rate is at its lowest in 3 years. According to the Coroner:

The Coroner further noted that:

A key goal of the Suicide and Self-harm Monitoring Project is to establish suicide registers in all jurisdictions. Registers currently exist in Victoria, Queensland, Tasmania and Western Australia. New South Wales are aiming to have a register established by October 2020. The AIHW is currently working with South Australia, the Australian Capital Territory and the Northern Territory to help establish suicide registers.

A key part of the National Suicide and Self-harm Monitoring Project is the compilation and coding of data from ambulance attendances by Turning Point. This coding exercise is quite large as not only do Turning Point have to code data for 2020 they also have to code data for previous years to allow for analysis of trends.

The AIHW asked Turning Point if they could prioritise the coding of data for Victoria. Importantly, Turning Point are able to separate suicide attempts from self-injury and suicide ideation. Figure 3 shows monthly data on the number of ambulance attendances related to suicide attempts from January to June for 2020 and the equivalent periods in both 2019 and 2018. As the chart shows there is no clear difference from 2019 to 2020. The total number of ambulance attendances related to suicide attempts in Victoria from January to June 2020 was 2% lower than for the equivalent period in 2019. The number of ambulance attendances relating to suicidality (thinking about suicide) was also fairly similar across the 2 years (it was 4% higher in 2020).

Interestingly there was a pick up in the total number of mental health attendances (up 13% in the first half of 2020 compared with the same time in 2019). This is consistent with the overall greater use of mental health services in 2020 that is evident in other data. This highlights the fact that greater use of, and need for, mental health services does not necessarily equate to trends in the number of suicide attempts. The vast bulk of people who use mental health services will never have a suicide attempt but timely access to mental health services could reduce the number of deaths by suicide.

The number of self-injury ambulance attendances are up (17.6%) compared with 2019. This highlights the fact that self-injury and suicide attempts are not the same thing.

Figure 3: Monthly (Jan–June) ambulance attendances for suicide attempts, Victoria

AAPOR (American Association for Public Opinion Research) 2010. Report on Online Panels.

ABS (Australian Bureau of Statistics) 2019. Causes of Death 2018, ABS Catalogue number 3303.0.

ABS 2020. 4940.0 - Household Impacts of COVID-19 Survey, 24–29 June 2020.

AIHW (Australian Institute of Health and Welfare) 2020. National Drug Strategy Household Survey 2019. Drug Statistics series no. 32. PHE 270. Canberra AIHW.

Biddle N, Edwards B, Gray M, and Sollis K 2020a. Initial impacts of COVID-19 on mental health in Australia. COVID–19 Briefing Paper, ANU Centre for Social Research and Methods, Australian National University, Canberra.

Biddle N, Edwards B, Gray M, and Sollis K 2020b. Mental health and relationships during the COVID-19 pandemic. COVID–19 Briefing Paper, ANU Centre for Social Research and Methods, Australian National University, Canberra.

Biddle N, Edwards B, Gray M, and Sollis K 2020c. Tracking outcomes during the COVID-19 pandemic (August 2020) – Divergence within Australia. COVID–19 Briefing Paper, ANU Centre for Social Research and Methods, Australian National University, Canberra.

Biddle N, Edwards B, Gray M, and Sollis K 2020d. Hardship, distress, and resilience: The initial impacts of COVID-19 in Australia ANU Centre for Social Research and Methods.

Botha, F Butterworth, and Wilkins, R (2020). Mental distress in Australia over the course of the first wave of the COVID-19 Pandemic. Melbourne Institute Research Insights 24/20.

Chief Coroner of New Zealand, Media Release 21 August 2020. 

Coroners Court of Victoria 2020a. Monthly Suicide Data Report, Report 1 -27 August 2020.

Coroners Court of Victoria 2020b. Monthly Suicide Data Report, Report 2 – 5 October 2020.

John A, Okolie C, Eyles E et al. The impact of the COVID-19 pandemic on self-harm and suicidal behaviour: a living systematic review [version 1; peer review: awaiting peer review] F1000Research 2020, 9:1097.

Leske S, Adam G, Schrader I, Catakovic A, Weir B, & Crompton D (2020). Suicide in Queensland: Annual Report 2020. Brisbane, Queensland, Australia: Australian Institute for Suicide Research and Prevention, School of Applied Psychology, Griffith University.

Phillips B, Gray M, and Biddle N 2020. JobKeeper and JobSeeker impacts on poverty and housing stress under current and alternative economic and policy scenarios, ANU Centre for Social Research and Methods.