Health service use in the year before death by suicide among ex-serving ADF members

Suicide is a concern in the Australian community due to its widespread societal impact and preventable nature (National Suicide Prevention Office, 2025). This has led to an increased focus on identifying effective intervention strategies.

The National Mental Health and Suicide Prevention Agreement identifies Australian Defence Force (ADF) members and veterans as a priority population group, as ex-serving ADF members have higher rates of suicide than those in the general population, particularly those who separated from ADF involuntarily for medical reasons (AIHW, 2024).

Studies on health service use patterns in the year before death by suicide highlight potential opportunities for intervention. Research in general populations show that the majority of individuals who died by suicide have had contact with health services at least once in the last year (89%), month (61%) or even days (34%) before death, with primary care (the general practitioner) being the most common point of contact in the year before death by suicide (average 80%) (AIHW 2022; Ahmedani et al 2019; DelPozo-Banos et al. 2024; John et al. 2020; Laanani et al. 2020; Stene-Larsen et al. 2019). These health service interactions represent potential opportunities for risk assessment, mental health support, and crisis prevention across health service settings.

Despite the high rates of health service use in the general population before death by suicide, little is known about how ex-serving ADF members who died by suicide access and interact with these services. Knowledge of how ex-serving ADF members interact with health services can be used to inform development of policy and service planning and enable the identification of trends, emerging areas of concern and priority groups to support prevention initiatives.

This report examines ex-serving ADF members who died by suicide and their use of health services in the last year of life. However, it is important to recognise that service contact does not necessarily mean suicide risk was present, recognised, or addressed during those interactions. Rather, it reflects a complex interplay between individual health needs, the care received, and broader social and psychological factors, and should not be interpreted as a missed opportunity on its own or evidence of inadequate care.

This report examines health service use across some primary care (Medicare Benefit Schedule or MBS services), public hospital services (admitted care and emergency department or ED care), Pharmaceutical Benefits Schedule (PBS) and Repatriation Pharmaceutical Benefits Scheme (RPBS) and Department of Veterans’ Affairs or DVA-funded public hospital and primary care, referred to in this report as ‘in-scope services’. Information on the specific types of health services that were included are in the Technical notes.

The research included three components:

• comparisons between health services used in the last year of life for ex-serving members who died by suicide with services used over the same period by other ex-serving members
• cluster analysis to identify subgroups of ex-serving members by health service use in the last year of life before death by suicide
• trend analysis to examine patterns of health service use during the last year of life before death by suicide.

This report is part of a series of publications that analyse the use of health services by ex-serving members to 30 June 2020. The publications add to a picture of health service use and the factors associated with variations in health service use.

A similar proportion of ex-serving members who died by suicide accessed at least one health service compared with other ex-serving members

Among the 522 ex-serving ADF members who died by suicide over the study period, almost 9 in 10 accessed at least one in-scope health service in the year before death (86%), similar to other ex-serving members (87%). In the month before death, 61% of ex-serving members who died by suicide accessed a health service, compared with 45% of other ex-serving members; in the week before death, the figures were 37% and 20%, respectively.

When examining specific types of health service, a similar proportion of ex-serving members who died by suicide accessed primary care services in the year before death (81%) compared with other ex-serving members (82%), but a higher proportion did so in the month (41% compared with 29%) and week (18% compared with 11%) prior to death. Ex-serving members who died by suicide also had higher access rates of PBS services, and hospital services (admitted and ED care) across the year, month and week before death.

In addition to healthcare settings, this report examined the type of care accessed. A higher proportion of those who died by suicide accessed mental health services in the year, month and week before death, compared with other ex-serving members.

Ex-serving members who died by suicide used more health services than other ex-serving members

While the proportion of ex-serving members who died by suicide who accessed any health service was similar to other ex-serving members in the year, they used a much higher volume of health services. Ex-serving members who died by suicide used an average of 55 services in the year (76% higher than the 31 services used by other ex-serving members). The 55 services included 33 primary care services (MBS or DVA-funded), 19 prescriptions dispensed and 2 hospital services.

Ex-serving members who died by suicide were more likely to have separated involuntarily due to medical reasons or reasons not in service interest, have separated from the ADF recently, be DVA clients, have multiple health conditions, have had low continuity of GP care (not seeing the same GP each time) and low regularity of GP care (not having regular GP appointments) in comparison with other ex-serving members.

Ex-serving members who died by suicide had lower odds of accessing health services after adjusting for differences in health status

AIHW conducted modelling using odds ratios (OR) to compare the odds of having accessed a health service at least once in the year before death between ex-serving members who died by suicide and other ex-serving members. These OR reflect whether or not an individual accessed a particular health service, but do not indicate how frequently health services were used, whether care was effective, or whether health service use (or lack thereof) was associated with suicide. However, this information can be used to identify the health services that may offer opportunities for earlier intervention. In particular, the health services with higher odds ratios may represent points of contact where individuals at increased risk of suicide were more likely to engage with the health system.

ORs include adjustment for differences in underlying health conditions (comorbidities see Glossary). For example, using information on individuals from prescription medication and hospital data provides a better comparison between ex-serving members with similar health needs. In unadjusted analysis, ex-serving members who died by suicide had lower odds of accessing any in-scope health service in the year before death compared to other ex-serving members (OR=0.92).

However, after adjusting for comorbidities, those who died by suicide had 65% lower odds of accessing health services in the year before death (adjusted OR=0.35). This suggests that there is a large group of ex-serving members who died by suicide who did not access a health service when compared with those who did not die by suicide that had similar health needs.

Ex-serving members who died by suicide had over four times the odds of accessing mental health services in the last year before death compared to other ex-serving members (adjusted OR=4.04).

Ex-serving members who died by suicide had higher odds of accessing health services in the week before death than other ex-serving members

As the time period examined was closer to the date of death, ex-serving members who died by suicide had relatively higher odds of accessing a health service compared to other ex-serving members. In the six months prior to death by suicide, adjusted OR showed that ex-serving members who died by suicide had 26% lower odds of accessing any in-scope health service (adjusted OR=0.74). In the 3 months and 1 month prior to death, the odds of health service access were not significantly different between those who died by suicide and those who were alive (adjusted OR=0.96 and 1.02 respectively).

However, in the last week before death, those who died by suicide had 39% higher odds of accessing any in-scope health service (adjusted OR=1.39). This increase was driven largely by ED presentations, which were associated with nearly 11 times the odds of access among those who died by suicide (adjusted OR=10.84) and admitted care, which had 145% higher odds of access (OR=2.45) in the week before death.

While the pattern of increased health service use near death was observed for each individual health service type (primary care, PBS medications and hospital services), a different pattern was observed for mental health services. Rather than peaking in the last weeks, mental health service use was highest around 9 months before death. In the year before death, ex-serving members who died by suicide had over four times the odds of accessing mental health services compared to other ex-serving members (adjusted OR=4.04), with 147% higher odds in the last week before death (adjusted OR=2.47).

Many ex-serving members who died by suicide were minimal or low health service users

AIHW identified subgroups of ex-serving members who died by suicide based on their volume of health service use in the year before death. The majority of those who died by suicide were minimal or low users (65%) of health services, with over a quarter being moderate or high users (29%) and the remainder being very high users (6%) in their last year of life. These groups were identified using a data-driven method (latent class analysis), and the labels reflect relative levels of use across subgroups rather than fixed cut-offs. This highlights that while average service use was high, it was driven by a smaller group with frequent service use, and most ex-serving members had relatively limited engagement with health services.

Moderate and high health service users among ex-serving members who died by suicide were more likely to be female and have comorbidities (see Glossary) than minimal users. Very high users among ex-serving members who died by suicide were more likely to be older, have served for 20 or more years, have separated involuntarily from the ADF for medical reasons, be DVA clients and DVA Gold cardholders and have comorbidities compared with minimal users.

Some ex-serving members had increasing use of health services closer to the date of death by suicide

This report examines how patterns of health service use evolved over time during the year before death among ex-serving members who died by suicide. Understanding these trends provides insights into critical time periods for intervention and how service use may change or intensify before death by suicide.

Trajectory modelling of combined hospital care (ED and admitted care) identified two user groups, with 18 per cent of ex-serving members classified as high and increasing users of hospital services. One year prior to death by suicide, the high and increasing hospital use group on average used 0.4 hospital services per person per month, increasing to 0.8 services per person in the month immediately preceding death by suicide. The ex-serving members in the high and increasing hospital use group were more likely to have been female, have lower socioeconomic status, have separated involuntarily from the ADF due to medical reasons, be DVA Gold cardholders or have comorbidities than those in the lower use group.

High-use groups were identified for both mental health and non-mental hospital care. For both mental health and non-mental health hospital use, a spike in use was observed around 7 to 10 months prior to death. For mental health hospital care, this peak (average 1.1 services were 9 months prior) was higher than use in the final month prior to death (0.6 services).