Mental Ill health and burnout in residential aged care workers

Abstract

Aged care staff are exposed to workplace risk factors that have the potential to considerably impact mental health. This study aimed to explore mental ill health, burnout, and associated occupational factors in a nationwide sample of residential aged care workers in Australia (N = 1085). Cross-sectional online survey data were collected. Rates of depression, anxiety, wellbeing, burnout, and turnover intentions were explored using descriptive statistics. Regression models were used to analyse occupational factors associated with mental ill health, wellbeing, and burnout. One quarter (24%) of participants reported symptoms indicating a probable depressive disorder, and over one third (35%) reported symptoms consistent with an anxiety disorder. Over half (56%) reported burnout at elevated levels. Lower perceived supervisor support and previous assault by a resident/client were associated with significantly higher anxiety, depression, and burnout. These findings suggest there is an urgent need for evidence-based interventions to improve conditions for residential aged care workers, including preventing staff assaults and upskilling managers in supporting the mental health of staff.

Introduction

The aged care workforce provides a vital service to society, but these roles can come with significant psychosocial stressors that impact mental health and wellbeing¹. This workforce faces increasing demands due to the ageing population^2,3, yet they receive low levels of financial reward⁴. They are frequently exposed to abuse and assaults^5,6 and trauma in the form of client/resident death, along with a range of other risk factors for poor mental health including shift work, low job control, and inadequate resourcing^7,8. Within the residential aged care sector, personal care workers constitute the majority of staff in direct care roles (around 70%), with smaller proportions of nurses (23%) and allied health workers (7%)⁹, Relative to nursing professionals, personal care workers are known to be exposed to higher levels of psychosocial risks such as workplace violence and aggression⁶, along with lower wages and less decision-making latitude¹⁰. In recent years, these issues have been exacerbated by the impacts of COVID-19, which caused disproportionate levels of trauma and isolation to those in the aged care sector, and associated moral injury^11,12.

Despite these risks, few population-level surveys have been conducted to estimate the true extent of mental ill health among aged care workers, and those that do exist have largely focused on the impact of COVID-19 and the mental health of workers during that period of major strain^12–15. While this was a time of considerable pressure for the aged care industry, the level of strain has not remained static¹² and it is important to continue to collect mental health data on aged care workers. Furthermore, despite the known workplace hazards, the relationship of risk and protective factors to mental health conditions among these workers has received little attention. It therefore appears salient to investigate whether these workers (and the differing occupational groups within the sector) experience differential rates of mental ill health and burnout. There is also evidence that supervisor support may play a role in mitigating psychosocial risks that are associated with burnout and turnover intention in aged care workers^16–18.

In this study we aimed to explore the rates of depressive and anxiety symptoms, burnout, and overall wellbeing in an Australia-wide sample of residential aged care workers. We also aimed to investigate the associations of these outcomes with a range of occupational factors, specifically: 1) job category; 2) level of perceived supervisor support; 3) workplace environment and individual worker factors; and 4) job turnover intention.

Methods

Study design

A cross-sectional online survey design was used. The study (including data analysis plan) was approved by the UNSW Human Research Ethics Advisory Panel (iRECS4699) and conducted in accordance with the National Health and Medical Research Council’s (NHMRC) National Statement on Ethical Conduct in Human Research (2007).

Participants

Participants were aged 18 or above, were currently employed in a residential aged care facility in Australia (working in direct care or in a management or administration role) and were required to have at least a basic level of English literacy (Level 2/Year 7 or above).

Procedures

Recruitment took place over 15 weeks from February to June 2024 (more than 12 months after the end of the Australian Government’s COVID-19 emergency response). The survey was promoted via paid and unpaid social media advertisements and mailing lists. The promotional material provided a link to the Black Dog Institute website, which included a brief description of the study purpose, eligibility criteria, type of survey questions, estimated time involved, prize draw incentive, confidentiality and anonymity of participant data. Potential participants were then directed to the online survey hosted on the Qualtrics platform. All participants were required to read an online participant information statement and provide informed consent before commencing the survey.

Participants who self-declared that they did not meet inclusion criteria or elected not to proceed in the survey were provided with information on telephone counselling and crisis support services in case of distress. These resources were also listed throughout the survey and were presented as a pop-up notice for participants whose survey responses indicated they were at risk (i.e., any level of suicidality based on PHQ-9 questionnaire) along with a recommendation to contact a health professional for further mental health support.

Upon completion of the survey, participants could choose to list their email address to be contacted for participation in related research. They could also elect to be entered into a prize draw to win a $200 gift voucher.

Outcome measures

The survey collected data on: participant demographics (age, gender, ethnicity, language spoken at home, education, residency status, and state/territory of residency); occupational characteristics (occupation, work location, employment status, size of organization, tenure in aged care, number of staff supervising); experiences of and openness to mental health training; mental health help-seeking (from sources including doctor/GP, counsellor or psychologist, psychiatrist, employee assistance program, online treatment program, and telephone/text support lines); sickness absence; and adverse experiences in aged care work (abuse, assault, patient safety incidents).

Depressive symptoms were measured using the Patient Health Questionnaire-9 (PHQ-9), a reliable and valid 9-item tool that assesses depression symptom severity over the previous two weeks^19,20. Individual items are scored from 0 (‘not at all’) to 3 (‘nearly every day’). The PHQ-9 can be used as a continuous measure with scores ranging from 0 to 27 (higher scores indicate more severe symptoms). Scores ≥15 correspond to at least moderately severe symptoms and are indicative of a probable depressive disorder¹⁹. The PHQ-9 showed good internal consistency in this study (Cronbach’s alpha = 0.88).

Anxiety symptoms were measured using the General Anxiety Disorder-7 (GAD-7), a reliable and valid 7-item tool that assesses generalized anxiety disorder (GAD) symptoms over the previous two weeks^20,21. Individual items are scored from 0 (‘not at all’) to 3 (‘nearly every day’). GAD-7 total scores range from 0 to 21 (higher scores indicate more severe symptoms). Scores ≥10 correspond to at least moderate symptoms and indicate likely cases of GAD²¹. In the present study there was excellent internal consistency across the 7 items of the scale (Cronbach’s alpha = 0.91).

Wellbeing was measured using the Short Warwick-Edinburgh Mental Wellbeing Scale (SWEMWBS), a validated 7-item measure of mental wellbeing²². Individual items are positively worded and scored from 1 (‘none of the time’) to 5 (‘all of the time’) based on frequency over the previous two weeks. Total scores range from 7 to 35 (higher scores indicate better mental wellbeing). The SWEMWBS showed good internal consistency in this study (Cronbach’s alpha = 0.85).

Work-related burnout was measured using the 16-item Oldenburg Burnout Inventory (OLBI²³), which has been shown to have satisfactory reliability and validity among a wide range of occupational groups²⁴. The OLBI consists of two subscales: the ‘disengagement’ subscale comprising eight items, and the ‘exhaustion’ subscale, also comprising eight items, with each item scored from 1 (‘strongly agree’) to 4 (‘strongly disagree’). Items are both positively and negatively worded, with negatively worded items receiving a reverse-coded score. The total score ranges from 16 to 64, and each subscale ranges from 8 to 32 (higher scores indicate increased burnout). Prior research has identified total scores ≥44 as indicative of elevated burnout levels²⁵. In the present study there was good internal consistency for the total scale (Cronbach’s alpha = 0.86) as well as for each subscale (disengagement: Cronbach’s alpha = 0.75; exhaustion: Cronbach’s alpha = 0.79).

Supervisor support at work was measured via 9 items referring to a range of supportive management behaviours (e.g., “My supervisor pays attention to my feelings and problems and notices if I’m not feeling so well”; “My supervisor would be someone I would speak to if I was experiencing workplace stress”; “My supervisor is accessible and approachable to people in the team”). These questions have been used in previously published research²⁶. Items were measured on a scale from 1 (‘never/hardly ever’) to 5 (‘always’) reflecting the employee’s assessment of the frequency/likelihood of each behaviour occurring. Responses were summed to create a total score ranging from 9 to 45 (higher scores indicate more supervisor support). For statistical analysis, total scores were categorized into three levels of supervisor support: low (9–18), moderate (19–27), and high (28–45). In the present study there was a high level of internal consistency across the 9 items (Cronbach’s alpha = 0.96).

Turnover intention was assessed via a binary yes/no response item: “Do you intend to look for another job in the next 12 months?” Participants who responded ‘yes’ received a follow-up question: “How likely is your next job to be in residential aged care?” measured on a 4-point scale from ‘very unlikely’ to ‘very likely’.

Statistical analysis

Data analysis was conducted using IBM SPSS Statistics (version 27.0). Descriptive statistics (mean/SD or frequency/%) were calculated for demographics, occupational variables, and for mental health and burnout measures. One-way ANOVAs were used to assess any differences in mean depression, anxiety, wellbeing and burnout scores associated with occupational role and level of perceived supervisor support (group comparisons were conducted with Tukey’s HSD post-hoc tests to control the Type 1 error rate). Multiple linear regression models were used to analyse occupational factors associated with depression, anxiety, wellbeing, and burnout (adjusted for age and gender). Binary logistic regressions were used to analyse the individual associations of depression, anxiety, wellbeing, and burnout with turnover intention. All significance tests were set at an alpha level of 0.05.

Results

Sample characteristics

Of 1171 individuals who consented to participate in the survey, 1123 completed the demographics questions and were considered for inclusion in the sample. A further 38 participants were excluded as they did not meet eligibility for the study (predominantly due to being retired or not working in an eligible residential aged care role) or because they submitted multiple responses. The final sample comprised 1085 respondents.

Characteristics of the sample can be seen in Table 1. Participants were on average 50 years old and predominantly identified as women (90.7%) and of Caucasian ethnicity (87.9%). The main language spoken at home was English (95.9%) and the vast majority (97.9%) were Australian citizens or permanent residents. The majority (62.2%) had completed a TAFE course/certificate or a diploma. Individuals from all Australian states and territories participated. The geographic distribution of the sample was broadly representative of the national population, with Tasmania overrepresented, and Australian Capital Territory and Northern Territory somewhat underrepresented²⁷. Around half the sample (51.2%) were personal care workers, while 30.3% worked as a registered nurse or enrolled nurse. The majority worked part-time (57.2%) and had been employed in aged care for more than 10 years (53.0%). Around two in five (40.3%) worked in a regional centre location; fewer participants worked in a rural area (31.4%) or major city (28.3%).

Table 1.

Sample characteristics (N = 1085)

	n (%)
Mean age (SD, range)	50.22 (11.56, 18–75)
Gender
Man	72 (6.6)
Woman	984 (90.7)
Non-binary / Gender fluid or prefer to self-identify	18 (1.7)
Prefer not to say	11 (1.0)
Ethnicity*
Aboriginal and/or Torres Strait Islander	37 (3.4)
White or European	955 (87.9)
Asian	45 (4.1)
Hispanic	11 (1.0)
African	6 (0.6)
Pacific Islander	6 (0.6)
Other	43 (4.0)
Prefer not to say	6 (0.6)
Language spoken at home
English	1040 (95.9)
Other	42 (3.9)
Prefer not to say	3 (0.3)
Education
Year 10 certificate or below	115 (10.6)
Year 12 certificate	61 (5.6)
TAFE course / certificate	422 (38.9)
Diploma	253 (23.3)
Undergraduate degree	135 (12.4)
Postgraduate degree	99 (9.1)
Residency status
Australian citizen	896 (82.5)
Permanent resident	167 (15.4)
Temporary visa	16 (1.5)
Other or prefer not to say	6 (0.6)
State / Territory of residence
New South Wales	309 (28.5)
Queensland	220 (20.3)
South Australia	110 (10.1)
Tasmania	65 (6.0)
Victoria	286 (26.4)
Western Australia	88 (8.1)
Australian Capital Territory or Northern Territory	7 (0.6)
Work location
Major city	307 (28.3)
Regional centre	437 (40.3)
Rural area	341 (31.4)
Occupation
Registered nurse	180 (16.6)
Enrolled nurse	149 (13.7)
Personal care worker	556 (51.2)
Allied health (e.g., Occupational therapist, Physiotherapist, Social Worker, Allied health assistant)	51 (4.7)
Other (e.g., Management/administration, Leisure/lifestyle, Nursing assistant, Support worker)	149 (13.7)
Employment status
Full-time	338 (31.2)
Part-time	621 (57.2)
Casual	100 (9.2)
Other	26 (2.4)
Size of organization
1–20 staff	35 (3.2)
21–50 staff	170 (15.7)
51–100 staff	199 (18.3)
>100 staff	215 (19.8)
Large provider (>100 staff) with multiple facilities	466 (42.9)
How long working in aged care
<1 year	36 (3.3)
1–3 years	130 (12.0)
3–5 years	131 (12.1)
5–10 years	213 (19.6)
>10 years	576 (53.0)

^*Note. Participants could select multiple ethnicities.

Approximately half of respondents (51.1%; 527/1032) reported that they supervised at least one other staff member. Only one in five (19.8%; 204/1032) had received training for their own (or their team’s) mental health or wellbeing, although most (83.9%; 866/1032) declared that they would be likely or very likely to engage in this type of training. Just under half (45.3%; 406/896) had sought help or advice for a mental health or emotional issue in the previous 90 days. Additionally, 42.4% (378/891) had needed this kind of help or advice at some time in the past but had not accessed it. One in nine (11.1%; 97/872) reported a period of sickness absence lasting one week or more due to a mental health or emotional issue in the previous 6 months.

A high proportion of respondents had received abuse from a relative of a resident/client (58.7%; 505/861) and two thirds had been assaulted by a resident/client (66.7%; 574/861). Over a third of respondents (37.5%; 323/861) had been involved in a patient safety incident at some point (13.5% in the previous 3 months).

Depression, anxiety, wellbeing and burnout

Table 2 presents descriptive statistics for mental health symptomatology, wellbeing and burnout. Overall, around a quarter of participants (24.4%) scored ≥15 on the PHQ-9, indicating at least moderately severe depressive symptoms, suggestive of a likely depressive disorder¹⁹. Over a third (35.3%) scored ≥10 on the GAD-7, indicating at least moderate anxiety symptoms, suggestive of probable GAD²¹. A similar proportion (36.2%) had low levels of wellbeing ( ≤ 19 on the SWEMWBS), while over half (56.4%) had elevated levels of burnout (total score ≥44). Allied health workers reported higher mean levels of wellbeing and lower depression, anxiety and burnout compared to workers in every other occupational role; however, one-way ANOVAs showed these differences were statistically significant for burnout only (OLBI total score: F (4975) = 5.296, p < 0.001; disengagement: F (4975) = 4.611, p = 0.001; exhaustion: F (4975) = 6.199, p < 0.001).

Table 2.

Mean scores and prevalence for depression (PHQ-9), anxiety (GAD-7), wellbeing (SWEMWBS) and burnout (OLBI), overall and by occupational group

Outcome measure	Statistic	Overall	Registered nurse	Enrolled nurse	Personal care worker	Allied health	Other
PHQ-9 (n = 954)	Mean (SD)	10.4 (6.1)	10.4 (5.9)	11.3 (6.1)	10.2 (6.1)	9.0 (4.8)	11.0 (6.3)
	% scoring ≥15†	24.4%	22.8%	28.9%	23.9%	12.8%	28.0%
GAD-7 (n = 917)	Mean (SD)	8.1 (5.4)	8.0 (5.1)	8.5 (5.4)	8.0 (5.5)	7.1 (4.1)	8.8 (5.7)
	% scoring ≥10‡	35.3%	32.3%	35.9%	35.0%	27.7%	42.3%
SWEMWBS (n = 900)	Mean (SD)	21.0 (4.4)	20.9 (4.2)	20.6 (4.1)	21.0 (4.5)	22.4 (3.0)	21.0 (4.8)
	% scoring ≤19§	36.2%	31.9%	37.9%	38.2%	13.0%	41.4%
OLBI (n = 980)	Mean (SD), total score	44.5 (6.7)	44.9 (6.8)	45.2 (6.0)	44.4 (6.7)	40.3* (6.5)	44.9 (6.7)
	% scoring ≥44||	56.4%	57.7%	64.7%	55.2%	29.8%	60.4%
	Mean (SD), disengagement	21.1 (3.7)	21.0 (3.9)	21.8 (3.3)	21.2 (3.7)	19.2* (3.9)	21.1 (3.7)
	Mean (SD), exhaustion	23.3 (3.6)	23.9 (3.4)	23.4 (3.2)	23.2 (3.7)	21.2* (3.4)	23.8 (3.8)

^*p < 0.05 compared to every other occupational group; PHQ-9 Patient Health Questionnaire-9 item, GAD-7 General Anxiety Disorder-7 item, SWEMWBS Short Warwick-Edinburgh Mental Wellbeing Scale, OLBI Oldenburg Burnout Inventory.

^†Cutoff for moderately severe depressive symptoms / probable depressive disorder.

^‡Cutoff for moderate anxiety symptoms / probable GAD.

^§Cutoff for low wellbeing.

^||Cutoff for elevated burnout.

Supervisor support

The level of supervisor support received by workers was evenly spread with 34.6% (295/852) reporting low levels of support, 32.3% (275/852) reporting moderate support, and 33.1% (282/852) reporting high support. Participants with higher levels of perceived supervisor support reported significantly better mental health and wellbeing and lower burnout scores compared to those with less support (see Appendix A. Supplementary Materials).

Occupational factors associated with depression, anxiety and wellbeing

The results of the multiple regression analyses examining associations between occupational factors and depression, anxiety and wellbeing can be seen in Table 3. Separate regression models were conducted for each outcome, adjusted for age and gender.

Table 3.

Occupational factors associated with depression (PHQ-9), anxiety (GAD-7) and wellbeing (SWEMWBS), adjusted for age and gender

Variable	Group	PHQ-9		GAD-7		SWEMWBS
		B (95% CI)	P	B (95% CI)	P	B (95% CI)	P
Occupation	Allied health worker	†	†	†	†	†	†
	Registered nurse	0.92 (-1.20 to 3.04)	0.396	0.68 (-1.17 to 2.53)	0.472	-0.93 (-2.41 to 0.56)	0.222
	Enrolled nurse	1.45 (-0.74 to 3.65)	0.194	0.91 (-1.00 to 2.83)	0.349	-0.56 (-2.10 to 0.91)	0.471
	Personal care worker	0.37 (-1.60 to 2.34)	0.711	0.61 (-1.12 to 2.33)	0.490	-0.13 (-1.50 to 1.25)	0.858
	Other occupation	1.97 (-0.15 to 4.08)	0.069	2.02 (0.17 to 3.87)	0.033*	-1.18 (-2.66 to 0.30)	0.118
Level of supervisor support	High	†	†	†	†	†	†
	Low	3.86 (2.88 to 4.84)	0.000***	3.00 (2.15 to 3.86)	0.000***	-3.45 (-4.14 to -2.77)	0.000***
	Moderate	2.07 (1.08 to 3.06)	0.000***	1.53 (0.67 to 2.40)	0.001**	-1.96 (-2.65 to -1.27)	0.000***
How long working in aged care	<1 year	†	†	†	†	†	†
	1-3 years	-0.34 (-2.88 to 2.21)	0.796	-0.67 (-2.90 to 1.55)	0.553	-1.08 (-2.87 to 0.70)	0.233
	3-5 years	-0.34 (-2.85 to 2.17)	0.791	-0.98 (-3.18 to 1.21)	0.379	-0.97 (-2.72 to 0.79)	0.281
	5-10 years	0.56 (-1.88 to 3.00)	0.654	0.16 (-1.97 to 2.30)	0.881	-2.01 (-3.72 to -0.30)	0.021*
	>10 years	0.15 (-2.24 to 2.54)	0.901	0.07 (-2.02 to 2.16)	0.950	-1.60 (-3.27 to 0.08)	0.061
Work location	Rural area	†	†	†	†	†	†
	Major city	-0.09 (-1.17 to 0.98)	0.863	0.35 (-0.59 to 1.28)	0.468	0.72 (-0.03 to 1.47)	0.061
	Regional centre	0.09 (-0.90 to 1.07)	0.865	0.30 (-0.56 to 1.16)	0.488	0.44 (-0.25 to 1.12)	0.214
Organization size	1-20 staff	†	†	†	†	†	†
	21-50 staff	2.19 (-0.40 to 4.77)	0.097	2.45 (0.19 to 4.70)	0.034*	-0.98 (-2.79 to 0.82)	0.286
	51-100 staff	1.75 (-0.82 to 4.32)	0.181	1.49 (-0.75 to 3.73)	0.193	-0.36 (-2.16 to 1.44)	0.697
	>100 staff	1.21 (-1.36 to 3.77)	0.355	2.08 (-0.16 to 4.32)	0.069	-0.64 (-2.44 to 1.16)	0.484
	Large provider (>100 staff) with multiple facilities	1.92 (-0.56 to 4.41)	0.129	1.71 (-0.46 to 3.88)	0.123	-0.50 (-2.24 to 1.24)	0.572
Supervise staff‡		0.20 (-0.75 to 1.14)	0.681	0.65 (-0.17 to 1.48)	0.121	0.14 (-0.52 to 0.80)	0.672
Abused by relative of a client‡		0.23 (-0.67 to 1.12)	0.620	0.45 (-0.33 to 1.23)	0.262	0.22 (-0.40 to 0.85)	0.485
Assaulted by a client‡		1.11 (0.19 to 2.03)	0.019*	0.91 (0.10 to 1.71)	0.028*	-0.59 (-1.24 to 0.05)	0.072
Involved in patient safety incident‡		0.20 (-0.63 to 1.03)	0.638	0.28 (-0.44 to 1.00)	0.445	-0.03 (-0.61 to 0.55)	0.929

^* p < 0.05; ^** p < 0.01; ^*** p < 0.001; PHQ-9 Patient Health Questionnaire-9 item, GAD-7 General Anxiety Disorder-7 item, SWEMWBS Short Warwick-Edinburgh Mental Wellbeing Scale, CI confidence interval.

^† Reference group. ‡Binary variable. Reference group is “No”.

Overall, the model for depression was statistically significant and explained 9.8% of the variance in depressive symptoms (PHQ-9 score: F(22,829) = 4.977, p < 0.001, R² = 0.098). Only the employee’s level of perceived supervisor support and being assaulted by a resident/client were significantly associated with depression. Both low and moderate levels of support were associated with significantly higher depressive symptoms compared to the group with high supervisor support (p < 0.001 for both). Low perceived support was also associated with significantly higher depressive symptoms relative to those experiencing moderate levels of support (p < 0.001). Previous assault by a resident/client was associated with significantly higher depressive symptoms compared to those who had not been assaulted (p = 0.019).

The model for anxiety was also statistically significant and explained 9.6% of the variance in anxiety symptoms (GAD-7 score: F(22,829) = 5.913, p < 0.001, R² = 0.096). This model showed a similar pattern to the model for depression, i.e., lower levels of perceived supervisor support and previous assault by a resident/client were associated with significantly higher anxiety (see Table 3). In addition, employment in ‘other’ occupations (compared to working in allied health; p = 0.033) and in organizations with 21 to 50 staff (compared to working in smaller organizations; p = 0.034) were associated with significantly higher anxiety.

The model for wellbeing was statistically significant and explained 13.3% of the variance in SWEMWBS scores (F(22,829) = 6.740, p < 0.001, R² = 0.133). Similarly to the models for depressive and anxiety symptoms, lower levels of supervisor support were associated with significantly lower wellbeing (see Table 3). Working in aged care for 5 to 10 years (relative to <1 year) was also associated with significantly lower wellbeing (p = 0.021).

Occupational factors associated with burnout

The results of the multiple regression analyses examining associations between occupational factors and burnout can be seen in Table 4. Separate regression models were conducted for overall burnout and each burnout subscale, adjusted for age and gender.

Table 4.

Occupational factors associated with of burnout (OLBI total and disengagement/exhaustion subscales), adjusted for age and gender

Variable	Group	OLBI Total		OLBI Disengagement		OLBI Exhaustion
		B (95% CI)	P	B (95% CI)	P	B (95% CI)	P
Occupation	Allied health worker	†	†	†	†	†	†
	Registered nurse	2.04 (-0.17 to 4.25)	0.070	0.65 (-0.59 to 1.88)	0.304	1.39 (0.17 to 2.62)	0.026*
	Enrolled nurse	1.45 (-0.83 to 3.74)	0.212	0.94 (-0.34 to 2.22)	0.149	0.52 (-0.75 to 1.78)	0.425
	Personal care worker	1.98 (-0.07 to 4.03)	0.059	0.96 (-0.19 to 2.10)	0.102	1.02 (-0.12 to 2.16)	0.079
	Other occupation	3.41 (1.20 to 5.61)	0.002**	1.38 (0.15 to 2.61)	0.028*	2.03 (0.80 to 3.25)	0.001**
Level of supervisor support	High	†	†	†	†	†	†
	Low	5.21 (4.19 to 6.23)	0.000***	3.01 (2.44 to 3.58)	0.000***	2.20 (1.64 to 2.77)	0.000***
	Moderate	2.96 (1.93 to 3.99)	0.000***	1.68 (1.10 to 2.26)	0.000***	1.28 (0.70 to 1.85)	0.000***
How long working in aged care	<1 year	†	†	†	†	†	†
	1-3 years	3.03 (0.38 to 5.69)	0.025*	2.01 (0.53 to 3.50)	0.008**	1.02 (-0.45 to 2.49)	0.175
	3-5 years	3.47 (0.86 to 6.09)	0.009**	2.26 (0.80 to 3.72)	0.002**	1.21 (-0.24 to 2.66)	0.103
	5-10 years	4.16 (1.61 to 6.70)	0.001**	2.55 (1.13 to 3.97)	0.000***	1.61 (0.19 to 3.02)	0.026*
	>10 years	4.08 (1.59 to 6.57)	0.001**	2.57 (1.18 to 3.96)	0.000***	1.51 (0.13 to 2.89)	0.032*
Work location	Major city	†	†	†	†	†	†
	Regional centre	0.15 (-0.85 to 1.15)	0.773	0.05 (-0.51 to 0.61)	0.863	0.10 (-0.46 to 0.65)	0.730
	Rural area	0.18 (-0.94 to 1.29)	0.753	0.25 (-0.38 to 0.87)	0.436	-0.07 (-0.69 to 0.55)	0.828
Organization size	1-20 staff	†	†	†	†	†	†
	21-50 staff	1.00 (-1.69 to 3.69)	0.466	0.31 (-1.19 to 1.81)	0.687	0.69 (-0.80 to 2.18)	0.364
	51-100 staff	-0.20 (-2.87 to 2.48)	0.886	-0.17 (-1.66 to 1.33)	0.824	-0.03 (-1.51 to 1.46)	0.972
	>100 staff	0.18 (-2.50 to 2.85)	0.897	-0.02 (-1.52 to 1.47)	0.977	0.20 (-1.29 to 1.68)	0.794
	Large provider ( > 100 staff) with multiple facilities	1.05 (-1.53 to 3.64)	0.425	0.26 (-1.18 to 1.71)	0.723	0.79 (-0.65 to 2.23)	0.280
Supervise staff‡		1.13 (0.15 to 2.11)	0.024*	0.51 (-0.04 to 1.06)	0.068	0.62 (0.07 to 1.16)	0.027*
Abused by relative of a client‡		0.94 (0.01 to 1.87)	0.047*	0.44 (-0.08 to 0.96)	0.098	0.50 (-0.01 to 1.02)	0.056
Assaulted by a client‡		1.53 (0.57 to 2.49)	0.002**	0.81 (0.27 to 1.35)	0.003**	0.72 (0.19 to 1.25)	0.008**
Involved in patient safety incident‡		1.11 (0.25 to 1.98)	0.011*	0.64 (0.16 to 1.12)	0.009**	0.47 (-0.01 to 0.95)	0.053

^* p < 0.05; ^** p < 0.01; ^*** p < 0.001; OLBI Oldenburg Burnout Inventory, CI confidence interval.

^† Reference group.

^‡ Binary variable. Reference group is “No”.

The model for burnout as a unidimensional construct was statistically significant and explained 19.2% of the variance in burnout (OLBI total score: F(22,829) = 9.985, p < 0.001, R² = 0.192). Working in aged care for one or more years (relative to <1 year), supervising at least one other staff member, abuse by a relative of a resident/client, assault by a resident/client, involvement in a patient safety incident, and employment in ‘other’ occupations (relative to allied health) all were associated with significantly higher burnout (see Table 4). Low and moderate levels of perceived supervisor support were associated with higher burnout in comparison to high support (p < 0.001 for both). Those with low perceived support were also significantly more burnt out than those experiencing moderate levels of support (p < 0.001).

The models for burnout subscales were both statistically significant, explaining 19.0% and 14.4% of the variance in disengagement and exhaustion, respectively (OLBI disengagement: F(22,829) = 9.502, p < 0.001, R² = 0.190; OLBI exhaustion: F(22,829) = 7.434, p < 0.001, R² = 0.144). Statistically significant variables for subscales were largely the same as for total burnout, with some notable exceptions (see Table 4). For disengagement, supervising staff and abuse by a relative of a resident/client were not significant (p = 0.068 and p = 0.098 respectively). For exhaustion, neither abuse by a relative of a resident/client nor involvement in a patient safety incident were significant (p = 0.056 and p = 0.053 respectively), while registered nurses reported significantly higher exhaustion (relative to working in allied health; p = 0.026). Although tenure in aged care was associated with significantly higher total burnout and disengagement at one year (relative to those new to the area), this association with exhaustion was only seen after five years.

Turnover intention

A large proportion of respondents (42.5%; 371/872) intended to look for another job in the next 12 months. Of these, two thirds (65.5%; 243/371) were unlikely or very unlikely to continue working in residential aged care. Binary logistic regression analyses showed that depression (χ²(1) = 50.31, p < 0.001), anxiety (χ²(1) = 34.68, p < 0.001), wellbeing (χ²(1) = 35.30, p < 0.001), and burnout (χ²(1) = 123.28, p < 0.001) were all significantly associated with turnover intention, explaining 7.5%, 5.2%, 5.3%, and 17.7% (Nagelkerke R²) of the variance in turnover intentions respectively. For each additional unit increase on depression, anxiety, and burnout total score, respondents were respectively 8.6%, 7.9%, and 13.7% more likely to state they intended to change jobs. Conversely, each unit increase in total wellbeing score was associated with 9.3% lower likelihood of intending to change jobs.

Discussion

This study highlights the concerning rates of poor mental health (depressive and anxiety symptoms and burnout) among residential aged care workers in Australia. Importantly, the impact of specific occupational factors on mental health outcomes is also explored. These findings demonstrate the major mental health impacts of staff experiencing assault and, importantly, the considerable potential protective effect of manager support. Being able to retain experienced healthcare staff remains a major challenge for aged care systems internationally. Given this, our finding of strong associations between various mental health measures and turnover intentions highlights the importance of addressing mental ill health within the industry.

In our sample, 24% reported at least moderately severe symptoms of depression. This figure is concerning and comparable with pandemic rates in health and aged care workers reported by McGuinness and colleagues (16.4–22.6%¹²), despite our study using a higher symptom threshold. A third of the sample reported at least moderate anxiety on the GAD-7, considerably higher than rates reported by McGuinness and colleagues (8.8–16%¹²). Rates of burnout were not dissimilar to those reported in previous studies¹², with more than half the sample reporting clinically relevant burnout scores. Consistently, allied health workers reported the best mental health and wellbeing outcomes. A recent scoping review in the area highlights the relevance of environmental stressors in burnout among aged care workers²⁸. Our findings lend support to the importance of what the authors referred to as stressors related to responsibility (namely supervisory role) and reaction (in this case, assaults and patient safety incidents) but also emphasize the role of interpersonal factors (particularly manager support). Perhaps unsurprisingly, burnout disproportionately affected those who had worked in the industry longer.

In our study these mental health issues were underscored by the fact that one in nine respondents reported taking a period of at least one-week sick leave for mental health reasons in the previous 6 months, along with notable rates of recent help-seeking (almost half had sought professional help or advice for a mental health or emotional issue in the prior 90 days). Mental illness is a major contributor to work incapacity globally²⁹ and has been found to account for 7.9% of sickness absences in the UK³⁰. The rates reported in the present study are alarming as they indicate not only the frequency of this form of absenteeism but also that the related impairment is significant enough to cause extended periods of leave. Turnover intentions were similarly concerning, with two in five workers planning to leave their role in the next year and two thirds of these looking to move out of residential aged care work. The known worker crisis facing this sector is not new. Evidence suggests that rates of intended job turnover increased after the pandemic¹², while recent reports have forecast an annual shortfall of around 35,000 direct aged care workers in Australia³¹. Sustained rates of attrition would undoubtably have crippling effects on aged care, the wider health care system, and society more broadly. Our findings suggest that turnover intentions were linked to all mental health outcomes evaluated and consequently, that addressing mental health and wellbeing issues (and associated psychosocial hazards) may be critical to retention within this workforce. Wholistic, organization-wide approaches that consider both prevention and response should be the focus of intervention³².

Our study demonstrates the strong relationship between mental health and manager support, which was consistently associated with all mental health outcomes. While a causal link between manager support and mental health of aged care workers cannot be drawn from these cross-sectional findings (e.g., it is possible that those with poorer mental health may have more negative perceptions of their supervisor), the critical influence of managers upon symptoms of common mental disorder has been highlighted in other frontline workers²⁶. Manager mental health training has been found to improve a range of outcomes including supportive behaviour in managers³³. It has also been shown to be associated with improved staff recruitment, retention, customer service, business performance, and reduced absenteeism³⁴. The findings of the current study demonstrate the unmet need for this form of training within the aged care sector, with only one in five respondents having received some form of mental health training. Not only could this help to address the issues of mental distress and burnout, but it may help reduce staff turnover. Encouragingly, the current sample indicated overwhelming interest in receiving this form of training.

Respondents reported alarming rates of abuse from a relative of a resident/client (59%) and assault by patients (67%). Having experienced a workplace assault was associated with significantly elevated symptoms of depression and anxiety, and higher ratings of disengagement, exhaustion, and overall burnout. The issue of violence in aged care facilities has attracted recent media attention³⁵, and the findings of our study suggest little has changed over time, with similar rates reported in 2016⁶. Not only does violence lead to physical and psychological harm, the vulnerability and helplessness that victims can experience with repeated exposure may exacerbate negative health outcomes³⁶. Relatedly, compensation data³⁷ indicates that workplace violence makes up around 9.2 per cent of claims. There is thus a critical need to implement interventions to reduce occupational violence and its impacts within this workforce. Shea and colleagues⁶ suggest workplaces that prioritise employee safety and provide greater supervisor support are likely to have lower rates of violence. Furthermore, implementation of training programs that teach staff how to identify and avoid triggers or to de-escalate aggressive behaviour should be considered as part of standard training^38–40. There is evidence to indicate that in aged care, workers’ trust in managers may be a critical factor in predicting the effectiveness of anti-violence human resource management practices⁴¹. The present findings provide further support for the role of managers in supporting and actively facilitating the wellbeing of staff.

Despite the important implications from these findings, there are some limitations to this study that should be considered. The cross-sectional nature of the data does not allow us to draw any causal inferences. For example, the experience of workplace violence cannot be interpreted as a direct cause of poorer mental health in this workforce; an alternative explanation is that organizations with more frequent assaults from residents and worse mental health outcomes among staff might have other factors that contribute to staff mental ill health, or that workers experiencing poorer mental health may be more likely to report being assaulted. This study is also limited by restrictions around English literacy which may have excluded those from non-English speaking backgrounds who are also more likely to hold temporary visas. Indeed, the demographics suggest these groups were underrepresented in the sample, with most (96%) speaking English at home and less than 2% reporting temporary visa status. Temporary visa holders make up a large proportion of the aged care workforce in Australia⁴² and face additional occupational and environmental stressors contributing to poor mental health such as limited social support, racial discrimination, workplace exploitation, poorer job security, and visa insecurity^43–45. As such, the present study may underestimate mental health concerns of the aged care workforce by not adequately capturing this high-risk subgroup. A lack of engagement from these groups may have been a result of the online recruitment method, reluctance from different ethnic groups to disclose this information, and/or concerns about the survey affecting their employment and visa status. Despite the above potential for underestimating mental health concerns, the self-selection of respondents into the study may have also led to oversampling among those with mental health concerns, potentially leading to an over-estimate of mental ill health symptoms.

This study provides critical insight into the mental health challenges affecting the residential aged care workforce. There is a clear need to address not only the high rates of significant mental ill health symptoms and burnout, but the underlying hazards implicated in these patterns of distress. Perhaps most notable are the experiences of assault and abuse which are not only experienced at unacceptably high rates but appear to also play a critical role in the mental health issues facing this group. Aside from preventing staff assaults, supervisor support appears central and may be a promising target of team-level interventions to improve the psychological health of this workforce. Further research may also benefit from specifically targeting migrant aged care workers with temporary visa status. Urgent action is required to address these issues and avoid staffing crises, which would not only affect service provision and client/patient welfare but further affect staff wellbeing.

Author contributions

M.D. and D.A.J.C. prepared the initial manuscript draft. C.F. and D.A.J.C. conducted data curation and preparation of the dataset. D.A.J.C. analysed the data and prepared the manuscript tables and figure. M.D., A.G., A.S.G., J.O., and S.B.H. provided project guidance and supervision. All authors critically reviewed and edited the manuscript and provided approval for publication.

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to the UNSW Human Ethics conditions of approval for this study, which require the data custodian (M.D.) to obtain evidence of human research ethics approval before data can be shared with other researchers. However, data are available from the corresponding author (M.D.) on reasonable request and with evidence of human research ethics approval.

Competing interests

Authors M.D., D.A.J.C., A.G., C.F., M.C., and S.B.H. are employed by the Black Dog Institute, which provides a variety of workplace mental health services to various populations. All other authors declare no financial or non-financial competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s44184-026-00200-x.