Health service use in the year before and after intentional self-harm among ex-serving ADF members

Intentional self-harm is defined as deliberately causing physical harm to oneself but not necessarily with the intention of dying (AIHW 2025a). Intentional self-harm comes in many forms, and affects people from different backgrounds, ages and lifestyles. The reasons for self-harm are different for each person and are often complex (AIHW 2025b).

Most people who self-harm do not go on to end their lives – but previous self-harm (with or without suicidal intent) is a strong risk factor for suicide, with research showing elevated suicide rates among those with a history of self-harm (Clapperton et al 2024; Carroll et al 2014; Favril et al 2022). Therefore, monitoring of intentional self-harm is important to suicide prevention.

In Australia, there were close to 23,600 hospitalisations for intentional self-harm in 2023–24 (AIHW 2025c). Previous AIHW research showed that of all persons admitted to hospital, a higher proportion of ex-serving Australian Defence Force (ADF) members were admitted for intentional self-harm compared to all admitted Australians (AIHW 2024).

Despite the greater proportion of ex-serving ADF members being treated through hospital for self-harm, little is known about how these members interact with the health system around the time of self-harm events. Understanding patterns of health service use before and after being admitted can help identify those at risk and improve support following the self-harm event. These insights can also inform policy, guide service planning, and support the monitoring of trends, emerging concerns and priority populations.

This report examines use of health services among ex-serving ADF members who were admitted to a public hospital for intentional self-harm in the year before and after an index self-harm admission (the first self-harm hospitalisation in the study period). The report describes patterns of health service use in order to improve understanding of health service use and support efforts to potentially identify people at risk of self-harm or suicide.

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 of health services used in the year before and after self-harm hospital admission (self-harm cohort) with other ex-serving members who were admitted for non-injury reasons such as digestive diseases, mental health, and musculoskeletal conditions (comparison cohort) to contextualise health service use in the self-harm cohort
• cluster analysis of health service users across both cohorts to identify subgroups of ex-serving members based on type and intensity of service use
• transition analysis to examine changes in health service user groups from the year before to the year after self-harm admission

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

Ex-serving members admitted for self-harm accessed and used more health services than ex-serving members admitted for non-injury reasons

Among the 1,634 ex-serving ADF members who were admitted for an index self-harm episode (self-harm cohort) over the study period, almost all (98%) accessed at least one in-scope health service in the year before or after their intentional self-harm hospitalisation. These proportions were higher than the proportion (92%) among the 6,597 of ex-serving members admitted to hospital for non-injury related reasons (comparison cohort).

The proportion of the self-harm cohort who accessed at least one in-scope health service in the quarter before the self-harm admission was 92%, well above the 75% of the comparison cohort. The proportion of those in the self-harm cohort accessing a health service was highest in the quarter after admission (94%).

A higher proportion of the self-harm cohort accessed each health service type compared with the comparison cohort. In the year after admission, the greatest differences between the cohorts were presentations to ED (58% compared with 26%), hospital admissions (55% compared with 31%), use of DVA-funded MBS equivalent services (72% compared with 56%) and PBS (which includes RPBS) services (94% compared with 78%).

A higher proportion of the self-harm cohort accessed mental health and non-mental health services compared with the comparison cohort in the year before and after admission. The proportion of the self-harm cohort who accessed mental health services increased from 86% in the year before to 92% in the year after self-harm admission, higher than the comparison cohort (37% in the year before and 39% in the year after).

Ex-serving members in the self-harm cohort used health services at twice the rate of the comparison cohort in the year before and after self-harm admission. Those in the self-harm cohort averaged 90 health services per person in the year before admission and 107 services per person in the year after admission. The greatest relative differences between the self-harm cohort and comparison cohort in the year after were ED presentations (1.9 compared with 0.5), admitted care (2.1 compared with 0.7) and PBS (including RPBS) prescriptions dispensed (38 compared with 17).

Ex-serving members admitted for self-harm had higher odds of accessing health services and higher rates of health services use than ex-serving members admitted for non-injury reasons

This report includes modelling (to estimate odds ratios and rate ratios) to compare the differences in use of health services, between the self-harm cohort and the comparison cohort.

The odds ratios (OR) and rate ratios (RR) do not indicate the effectiveness of health services, such as whether they prevent or contribute to self-harm, but rather compare the use and frequency of use of health services between the two cohorts.

The modelling approach allows for differences in comorbidities (see Glossary), including physical and mental health conditions between the self-harm and comparison cohorts to be considered. This ensures that comparisons in health service use account for underlying health conditions that may influence health service needs.

When adjusted for comorbidities, those in the self-harm cohort had twice the odds of accessing an in-scope health service in the year before (OR=2.00) and 1.7 times the odds in the year after the admission (OR=1.73), compared with those in the comparison cohort. The odds of accessing mental health services were also higher in the self-harm cohort, with an adjusted OR of 6.97 in the year before admission and 12.64 in the year after admission.

The adjusted modelling showed that ex-serving members in the self-harm cohort were more likely to present to ED (OR=2.69), use a PBS (including RPBS) service (OR=1.79) and more likely to use admitted care (OR=1.64) than the comparison cohort in the year after admission.

To assess differences in the frequency of health service use, rate ratios (RR) were estimated. The adjusted modelling showed that ex-serving members in the self-harm cohort had higher rates of in-scope health service use in the year before (RR=1.31) and after the self-harm admission (RR=1.41). The self-harm cohort used mental health services at more than twice the rate of the comparison cohort before admission (RR=2.55) and over three times the rate after admission (RR=3.07).

In the year after admission, the self-harm cohort had higher rates of ED presentations (RR=2.94), admitted care services (RR=1.46) and MBS services (RR=1.43), compared to the comparison cohort.

Many ex-serving members admitted for self-harm were high or increasing health service users

This report also examines subgroups of ex-serving members in both the self-harm and comparison cohorts, based on their volume of health service use in the year before and after admission. These subgroups were identified using a data-driven method (latent class analysis) conducted separately for the year before and the year after admission. Subgroup classification was derived from observed use of six health service types: GP, specialists, allied health, ED presentations, admitted care, and prescription medications. Subgroups were labelled as low, medium, or high users based on the modelled patterns of volume and intensity of health service use and not determined by predefined thresholds.

The majority of ex-serving members were medium users (over 60% in both cohorts in the year before and after). However, there was also a substantial proportion of high health service users in the self-harm cohort (19% in the year before and 29% in the year after), and this proportion was greater than in the comparison cohort (approximately 11% at both time points).

Comparing health service subgroups between the year before and after admission, 26% of the self-harm cohort transitioned from the low or medium subgroup to a higher health service use group in the year after self-harm admission, compared with 16% of the comparison cohort.