People with a history of specialist homelessness services support who have died

People who experience homelessness have a higher prevalence and severity of physical and mental health conditions than other Australians (see Health of people experiencing homelessness; Gordon et al. 2025). Any experience of homelessness – from couch surfing to rough sleeping – increases the risk of an early death (see People receiving specialist homelessness services support in the last year of life; Seastres et al. 2020; Zordan et al. 2023). In Australia, there is an estimated life expectancy gap of more than three decades between people who have experienced homelessness and people who have not (Tuson et al. 2024).

Understanding the characteristics, service needs and causes of death among people with a history of specialist homelessness services (SHS) support who have died can help to inform improvements to service provision and outcomes for clients.

Analyses presented in this article are based on data from the NACS linked dataset. Linkage rates and SHS data coverage issues in the early years of the collection mean that the results are likely to be an underestimate of people with a history of SHS support who have died.

Characteristics of people with a history of SHS support who have died

Throughout 2012–13 to 2022–23, around 43,200 people with a history of SHS support died: around 25,600 males (59%) and 17,700 females (41%) (Figure 2). Almost half (48%) of deaths occurred within 2 years of receiving SHS support, including:

• 3,100 (7.1%) people who died while support was ongoing
• 10,900 (25%) people who died up to 12 months after ceasing support
• 6,700 (15%) people who died between 1 and 2 years after support ended.

The remaining 22,600 deaths (52%) occurred more than 2 years since support ceased.

Throughout 2012–13 to 2022–23, the median age at death was 55 years (55 for males, 56 for females) (Supplementary table 2.1). People who died closer to the time they last received SHS support tended to die at younger ages than people who died after more time had passed (Figure 2). The median age at death was:

• 49 years for people who died while support was ongoing
• 60 years for people who died more than 2 years since support ceased.

The difference in median age at death was larger among females: 46 years when SHS support was ongoing and 60 years when support ended more than 2 years prior to death, compared with males (50 and 59, respectively) (Supplementary table 2.1).

Potential years of life lost

Increasing data coverage of the SHSC contributes to observed changes over time in the scale of PYLL (see Definition and data quality issues). Earlier years are likely to be an underestimate, and caution should be used when interpreting time series data.

In 2022–23, people with a history of SHS support who died contributed around 15% (121,400) of PYLL nationally – reflecting the low median age at death (55 years) among the cohort (Supplementary tables 2.2 and 2.3).

Potential years of life lost by sex

In 2022–23, as a proportion of PYLL among females nationally, females with a history of SHS support contributed around 17%. Comparatively, males with a history of SHS support contributed around 14% of PYLL among males nationally (Supplementary table 2.3).

Each year from 2012–13 to 2022–23, males accounted for a majority (57–63%) of PYLL among people with a history of SHS support (Supplementary table 2.3). The difference between males and females narrowed over the 5 years to 2022–23:

• In 2017–18, males contributed 62% of PYLL, females 38%
• In 2022–23, males contributed 57% of PYLL, females 43%.

Potential years of life lost by time since support

In 2022–23, people who died while SHS support was ongoing or up to 12 months after support ended (28% or 34,100 years combined) contributed more than a quarter of PYLL among people with a history of SHS support. This equates to almost 100 years (94 years) of potential life lost each day among people who had received SHS support in their last year of life (Supplementary table 2.3). For more information about this group, see People receiving specialist homelessness services support in the last year of life.

In 2022–23, among people with a history of SHS support who died, the majority (72%) of PYLL were among those who died more than 1 year since last receiving SHS support, including (Supplementary table 2.3):

• 15,300 (13%) PYLL among people who died between 1 and 2 years after support
• 72,000 (59%) PYLL among people who died more than 2 years after support.

Experience of homelessness

The homelessness status and last known housing situation of people with a history of SHS support who died are explored below. For more information about how these variables are derived, refer to Data presentation and derivations.

Any history of homelessness

Throughout 2012–13 to 2022–23, over half (56%) of people with a history of SHS support who died had a recorded history of having experienced homelessness, including around (Supplementary table 2.2):

• 11,600 people (27%) with a recorded history of rough sleeping 
• 12,500 people (29%) with a recorded history of other homelessness.

Any experience of homelessness as recorded by the SHS agency, particularly rough sleeping, was associated with a younger age at death. The difference in median age at death between people who had experienced rough sleeping (50 years) and people who had not experienced any form of homelessness (61 years) while in contact with SHS agencies was around 11 years (Supplementary table 2.2).

Among people who died with a history of rough sleeping, 3 in 4 (75%) were male. Males (34% of males who had died) were twice as likely as females (17%) to have died with a history of rough sleeping recorded by the SHS agency (Figure 3). Although females are generally less likely to experience rough sleeping than males (AIHW 2024b; Bullen 2019), the health effects of rough sleeping may be more impactful on females (Box et al. 2022). Females (48 years) with a history of rough sleeping had a younger median age at death than males (51 years) (Supplementary table 2.2).

Last known housing situation

Throughout 2012–13 to 2022–23, the most common last known housing situation among people with a history of SHS support who have died were (Supplementary table 2.4):

• private or other housing (30%; 11,700 deaths)
• public or community housing (27%; 10,700 deaths)
• short-term temporary accommodation (17%; 6,500 deaths).

In part, this reflects the most common housing situations among all SHS clients (see Specialist homelessness services annual report 2023–24).

Throughout 2012–13 to 2022–23:

• among people who died while SHS support was ongoing, 1 in 4 (25%) were living in short-term temporary accommodation (Figure 4)
• among people who died more than 2 years since SHS support, 1 in 3 (34%) were last known to be in private or other housing (Figure 4).
• Regardless of time since SHS support, females were more likely than males to have died with a last known housing situation of private or other housing (37% compared with 25%). By contrast, males were more likely to have died with a last known housing situation of short-term temporary accommodation (19% compared with 12%) (Supplementary table 2.4).

Underlying cause of death

For information about how causes of death are classified, see Injury in Australia: Accidental poisoning and Technical notes.

Throughout 2012–13 to 2022–23, accidental poisoning, coronary heart disease and suicide were the most common underlying causes of death among both males and females with a history of SHS support, together accounting for 12,500 deaths (29%). The ranking of these leading underlying causes of death varied by sex and time since last support (Figure 5).

Underlying cause of death by sex

Throughout 2012–13 to 2022–23, females with a history of SHS support were less likely than males to have died from accidental poisoning (9.0% of females compared with 12% of males) and coronary heart disease (6.9% of females compared with 11% of males) (Supplementary table 2.5). Females were more likely to have died from suicide while support was ongoing (16% of deaths among ongoing female clients) compared with males (13%) (Figure 5).

Underlying cause of death by time since last SHS support

As more time passed between last SHS support and death, people with a history of SHS support were more likely to have died from a wider range of chronic health-related causes. Throughout 2012–13 to 2022–23, death from coronary heart disease became more likely, and death by accidental poisoning or suicide became less likely as the time since SHS support increased (Figure 5). Together, accidental poisoning and suicide accounted for (Supplementary table 2.5):

• around 1 in 3 (33% or 1,000 deaths) people who died while support was ongoing
• over 1 in 4 (28% or 3,000 deaths) people who died up to 12 months after ceasing support
• around 1 in 5 (20% or 1,300 deaths) people who died between 1 and 2 years since support ended
• around 1 in 7 (14% or 3,200 deaths) people who died more than 2 years since support ceased.

Underlying cause of death by last known housing situation

Throughout 2012–13 to 2022–23, around 1 in 3 (32%) people with a history of SHS support who died by coronary heart disease had a last known housing situation of homeless. Comparatively, around half of people who died by accidental poisoning (50%) or suicide (46%) had a last known housing situation of homeless (Supplementary table 2.6). 

Throughout 2012–13 to 2022–23, the last known housing situation of people who died by accidental poisoning and suicide were relatively similar (around half – 50% and 46%, respectively – were last known to be experiencing homelessness). However, people who died by suicide were more likely to have a last known housing situation of private or other housing (29% compared with 19%), while those who died by accidental poisoning were more likely to be in public or community housing (21% compared with 17%) (Supplementary table 2.6). 

Natural causes of deaths

Throughout 2012–13 to 2022–23, SHS clients (between 55% and 76%) were less likely to have died from natural causes than non-SHS clients (92%). However, as the time since last SHS support increased, SHS clients who died were more likely to have died from natural causes (Supplementary table 2.7). Compared with non-SHS clients who died of natural causes (92%):

• Over half (55%) of people who died while SHS support was ongoing died from natural causes
• Over 3 in 5 (61%) people who died 12 months after ceasing SHS support died from natural causes
• Around 7 in 10 (70%) people who died between 1 and 2 years after ceasing SHS support died from natural causes
• Over 3 in 4 (76%) people who died more than 2 years since SHS support ceased died from natural causes.

Among clients experiencing rough sleeping and receiving ongoing support, under half (46%) died of natural causes – a lower proportion than clients who died more than 2 years after support ended with a last known housing situation of rough sleeping (70%) (Supplementary table 2.7).

Psychosocial factors involved in deaths

People with a history of SHS support who died by accidental poisoning

Throughout 2017 to 2023, an average of 2 psychosocial factors were recorded per death by accidental poisoning among people with a history of SHS support (Supplementary table 2.8). Issues surrounding support systems – often involving the death or disappearance of a family member – were relatively common for both males and females who died by accidental poisoning, involved in around:

• 1 in 3 (34%) deaths among females
• 1 in 4 (24%) deaths among males.

Throughout 2017 to 2023, policing and justice factors – such as recent release from prison or problems related to other legal circumstances – were the most common psychosocial factor among males who died by accidental poisoning (43%) and less likely among females (24%). Issues relating to intimate partners – such as relationship distress or breakdown – were more likely among females who died by accidental poisoning (22%) than males (12%) (Supplementary table 2.8).

People with a history of SHS support who died by suicide

Throughout 2017 to 2023, people with a history of SHS support who died by suicide had an average of 2.6–3 psychosocial factors recorded per death. Among people who died by suicide, suicide ideation was the most common factor (recorded in around 2 in 5 deaths for both females and males) (Supplementary table 2.8).

For people with a history of SHS support who died by suicide between 2017 and 2023 (Supplementary table 2.8):

• Issues with support system factors, a personal history of self-harm and issues surrounding intimate partners were the next most common recorded psychosocial factors for both males and females, albeit with different rankings.
• A personal history of self-harm was more common among females (51%) than males (39%).
• Issues surrounding policing and justice were more common among males (33%) than females (15%).

Potentially avoidable deaths

Potentially avoidable deaths in people who died between the age of 1–74 with a history of SHS support are explored below. Increasing data coverage of the SHSC contributes to observed changes over time in the number and proportion of potentially avoidable deaths (see Definition and data quality issues). Earlier years are likely to be an underestimate, and caution should be used when interpreting time series data.

In 2022–23, over half of deaths (56%; around 3,100 deaths) among people with a history of SHS support were potentially avoidable, equating to more than 9 potentially avoidable deaths per day (Supplementary table 2.9).

Each year from 2012–13 to 2022–23, deaths among SHS clients were more likely to have been potentially avoidable (56–64% of all deaths) than deaths among non-SHS clients (43–47%) – particularly among females (53–61% compared with 35–39%) (Figure 6).

In 2022–23, deaths among people who died while SHS support was ongoing (62%) or up to 12 months after ceasing support (62%) were more likely to have been potentially avoidable, compared with people who died 1–2 years after (56%) or more than 2 years after (53%) support (Supplementary table 2.9).