Aboriginal and Torres Strait Islander people constitute three per cent of the Australian population.1 Aboriginal and Torres Strait Islander peoples’ understanding of health and wellbeing is generally holistic and grounded in the cultural value of collectivism, which inextricably ties the wellbeing of the individual to the wellbeing of community, culture and Country (Country is a holistic concept that includes identity, spirituality, culture, people, language, law and ceremony – not just physical land).2-4 Wellbeing for Aboriginal and Torres Strait Islander people, like other populations, depends on access to resources to satisfy basic needs, work roles and responsibilities, education, and physical and mental health.3 Aboriginal and Torres Strait Islander people face substantial economic, educational, health and social disadvantages as a result of the continued impact of colonisation, ongoing marginalisation and social injustices (e.g. racism).5, 6 These inequalities persist largely due to past and current government policies and political arrangements that disrupt Aboriginal and Torres Strait Islander people's social structures and societies, which ultimately impacts their wellbeing.5-7

The pandemic of the novel coronavirus (SARS-COV-2) and the disease it causes (COVID-19) has had a serious impact on the health and wellbeing of people in almost every country.8 With rapid global spread of the virus, the response from the Australian federal and state/territories governments was relatively swift. After the first Australian COVID-19 case was reported in late January 2020, lockdown orders and restrictions were in place across most of the country by late March 2020.9 While most public health restrictions were lifted by May 2020, some remained in place for most of 2020 (Figure 1: abridged timeline of the COVID-19 Pandemic in Australia).10, 11 Following this period, Australia has experienced numerous outbreaks across three waves of the pandemic, which necessitated Victoria and New South Wales going into prolonged lockdowns in 2021, while at the same time the government financial support program called JobKeeper, which was implemented in 2020 to support individuals affected by lockdown orders, was terminated in March 2021.12 Moreover, 2021 saw the first death of an Aboriginal person from COVID-19, with further COVID-19 deaths expected.13

Abridged timeline of the COVID-19 Pandemic in Australia.10, 11

Lockdown orders and other pandemic-related restrictions can have a significant impact on peoples’ lives and can be particularly devastating for marginalised and under-served populations who may already experience challenges to maintaining their wellbeing.8 Previous pandemics have had a greater toll on Aboriginal and Torres Strait Islander people and communities than other Australians. For example, the type A H1N1 virus (swine flu) pandemic in 2009 resulted in higher relative rates of infection, hospitalisation and death for Aboriginal and Torres Strait Islander people.14 This has been attributed to the higher prevalence of risk factors and comorbidities within this population,15 however, due to the lack of outcome data reported during previous pandemics, these underlying causes are likely more complex, potentially embedded in social and health services structures.14 During the current pandemic, the success of Indigenous-led strategies to prevent spread of COVID-19 among Aboriginal and Torres Strait Islander people and communities in Australia, and Government's support of these, has been recognised.16 Only 0.8% of all Australian COVID-19 cases have been among Aboriginal and Torres Strait Islander people.16 However, as Australia moves into the recovery phase and the COVID-19 vaccination rolls out, the toll on wellbeing needs to be assessed and addressed. This is particularly urgent for Aboriginal and Torres Strait Islander people.16

This secondary data analysis was conducted using information collected from two survey waves (Time 1 – pre-COVID-19 pandemic, and Time 2 – after the first wave and lockdown in Australia and during the first extended lockdown in the state of Victoria) from the What Matters 2 Adults study,17 to quantify change in wellbeing and HRQoL for Aboriginal and Torres Strait Islander adults associated with the pandemic, and identify factors associated with wellbeing and HRQoL over time.

Methods Context

Our team acknowledges the importance of reflexively considering and describing our own backgrounds, perspectives and values that we each bring to the project.18, 19 The first author (AG) is an Aboriginal PhD candidate with experience in qualitative and mixed methods Indigenous health research. The second author (AD) is a non-Indigenous early career researcher with expertise in epidemiology and Indigenous health research. The third author (GG) is a senior Aboriginal researcher with extensive experience in Indigenous health research and psychosocial research. The fourth author (KA) is a non-Indigenous senior qualitative health researcher with experience in Indigenous health research. The fifth author (DL) is a non-Indigenous early career researcher with expertise in quantitative research methods. The last author (KH) is a non-Indigenous senior researcher with expertise in health economic and Indigenous health research.

The current study is a secondary analysis of data collected as part of a larger study, called the What Matters 2 Adults (WM2A) study.17 The larger study aims to develop a wellbeing measure that is grounded in and underpinned by the values and preferences of Aboriginal and Torres Strait Islander adults. In the broader WM2A study, strategies were implemented to ensure Aboriginal and Torres Strait Islander people's voices were privileged, including the establishment of an Indigenous Project Advisory Group and an Indigenous Researchers Group. The initial decision to include the two measures reported in the current analysis involved input from both of these groups. This current study involved input and leadership from senior and early career Aboriginal researchers.

Design

An online cross-sectional survey was conducted at two time points, as part of a larger study to develop a new preference-based wellbeing measure for Aboriginal and Torres Strait Islander adults.17 Ethics approval for this study was granted by the Human Research and Ethics Committee of the Northern Territory Department of Health and Menzies School of Health Research (NHMRC Reg no. EC00153), The University of Sydney Human Research Ethics Committee (CRICOS Number: 00026A), and the University of Technology Sydney (UTS) Human Research Ethics Committee (TRIM no. RES20/234).

Participants

Participants were recruited via Dynata, an online survey platform with an established nation-wide panel of Aboriginal and Torres Strait Islander people, as well as through investigator networks. Aboriginal and Torres Strait Islander people aged 18 years or older were eligible to participate. Potential participants were emailed study information and prompted to provide online informed consent to commence the online survey. In total, 309 and 354 Aboriginal and Torres Strait Islander people participated in the survey at Time 1 and Time 2, respectively. Of these, 42 participants completed both surveys and were included in this analysis.

Data collection

Once consent was obtained, participants were directed to the online survey. The two surveys were conducted in October–November 2019 (pre-COVID-19 pandemic – Time 1) and August-September 2020 (after the first wave and lockdown in Australia and during the first extended lockdown in the state of Victoria – Time 2). The surveys collected information on wellbeing and HRQoL, as well as socio-demographic, health and economic variables (i.e. age, gender, Indigenous status, main language spoken at home, area, relationship status, household size, highest level education, employment status, financial situation, comorbidities and mental health comorbidities).

Outcome measures

Wellbeing: Data on wellbeing was collected using a visual analogue scale (VAS), where participants were asked to indicate their wellbeing at that moment on a horizontal scale ranging from zero (the worst wellbeing they could imagine) to 100 (the best wellbeing they could imagine) (Supplementary File 1). This is a standard VAS with slight modification to measure wellbeing.

Health-related quality of life: The Assessment of Quality of Life (4 Dimension) measure (AQoL-4D) is a health-related multi-attribute utility instrument that assesses quality of life across four dimensions: independent living (IL-QoL), mental health (MH-QoL), relationships (R-QoL) and senses (S-QoL). These four dimensions each include three items with four response levels for each question, giving 12 questions in total (Supplementary File 2).20 The AQoL-4D provides both dimension scores and an overall HRQoL score that represents the health state utility ranging from −0.04 to 1.0.20 Utility scores represent preferences for health states anchored on a 0.0–1.0, dead to full health scale; negative values represent health states worse than death. In accordance with previous population-based norms, an overall utility score greater than 0.90 were defined as excellent HRQoL.21

Change in wellbeing and AQoL-4D scores: Mean change in scores from Time 1 to Time 2 was calculated (Time 2 score-Time 1 score). Here, a negative change score indicates a decline in wellbeing/HRQoL over time, while a positive change score indicates an improvement.

Explanatory variables

A priori we decided to focus on examining variation in the change in wellbeing and HRQoL over time by gender, age group, residential remoteness, financial situation (see below), presence/absence of comorbidities and presence/absence of mental health comorbidity. We also described the cohort in terms of Aboriginal and/or Torres Strait Islander identity, household size, relationship status, highest education, employment status and main language spoken at home. Due to the small numbers of participants across categories of variables, categories were collapsed into broad groups: age (18–54yrs, 55–80yrs); gender (male, female); highest level of education (year 12 or below, post high school qual/s); employment status (paid employed, not employed/unpaid employment); financial situation (not enough, just enough and more than enough, money until next pay); and total number of comorbidities (nil, 1–5, 6 or more). Mental health comorbidity was defined as the presence of anxiety, depression and/or other mental health conditions (yes, no). Dynata coded participant residential postcode as either metropolitan or rural/regional based on the population size residing within that postcode and distance to services as per methods used by the Australian Bureau of Statistics.22

Analysis

There were no missing values in our dataset at either timepoint. All analyses were conducted in Stata v15.23 Participant characteristics at Time 1 were described using frequency (n) and percentage (%) for categorical variables and median and either inter-quartile range (IQR) or range (min-max) for age and household size, which were non-normally distributed continuous variables. Characteristics at Time 2 were also examined and reported if they differed from Time 1. Wellbeing scores and AQoL-4D utility scores approximated a normal distribution and were described using mean and standard deviation (sd). Two-tailed paired t-tests were used to examine differences in wellbeing and AQoL-4D utility scores between Time 1 and Time 2, overall and stratified by sub-groups. Between-group variation in Time 1 scores and differences in mean change scores over time (Time 1 – Time 2) were assessed using two-sample t-tests and ANOVA. Statistical significance was set at α = 0.05. There is no evidence for what constitutes a clinically meaningful change in wellbeing score. For AQoL-4D, the minimally important difference in utility has been reported as 0.06.21

Results Participant characteristics

Forty-two participants completed the online survey at both timepoints and were included in analyses. Participant characteristics at Time 1 are described in Table 1. At Time 1, participants had a median age of 55.5 years (IQR 41.0–65.0), approximately 60% were male (n=25), half lived in rural and regional locations (n=21) and 37 (88%) self-reported they had at least one comorbidity, with 23 (55%) reporting they had at least one mental health comorbidity. At Time 1, most participants reported they either did not have enough money (n=17, 40.5%) or just enough money (n=15, 35.7%) to get by until the next payday. At Time 2, 13 (31.0%) participants did not have enough money and 16 (38.1%) had just enough money until their next payday. For all other variables the total number of participants in each category was similar across both timepoints (difference n≤1 person).

Table 1. Participant characteristics at time point 1 pre-COVID-19 (n=42).

Characteristic

Baseline N (%)

Months between surveys, median (IQR)

9.5 (9.2–9.7)

Age in years, median (IQR∗)

55.5 (41.0–65.0)

Age group

18–54 years

17 (40.5)

55–80 years

25 (59.5)

Gender

Male

25 (59.5)

Female

17 (40.5)

Indigenous status

Aboriginal

40 (95.2)

Torres Strait Islander

2 (4.8)

Main language spoken at home

English

41 (97.6)

Torres Strait Islander language

1 (2.4)

Area

Metropolitan

21 (50)

Rural/Regional

21 (50)

Relationship status

Partnered

17 (40.5)

Single

24 (57.1)

Other

1 (2.4)

Household size

Range

1 to 7

Highest level education

Grade 12 or below

14 (33.3)

Post high school qual/s

28 (66.7)

Employment status

Paid employment

16 (38.1)

Not working/unpaid employment

26 (61.9)

Financial situation

Not enough

17 (40.5)

Just enough

15 (35.7)

More than enough

10 (23.8)

Total comorbidities

Nil

5 (11.9)

1 - 5

30 (71.4)

6 or more

7 (16.7)

Total mental health comorbidities

Nil

19 (45.2)

Any

23 (54.8)

Note: IQR = interquartile range Wellbeing

Table 2 presents the wellbeing scores (overall and stratified by participant characteristics) at Time 1 and Time 2, as well as the mean change in scores over time. Overall, wellbeing scores did not significantly change from Time 1 to Time 2 (mean change −6.4, 95%CI −14.2–1.4).

Table 2. Mean (SD) wellbeing score for Aboriginal and Torres Strait Islander participants (n=42), at Time 1(T1) and change over time.

Pre-pandemic (T1)

Changes over time (T1-T2)

Characteristic at baseline (T1)

Wellbeing score

Mean (SD)

P-value

(between groups at baseline)b

Wellbeing score

Mean difference (95%CI)

P-value (between group change over time)b

Total sample

64.2 (26.7)

n/a

-6.4 (-14.2 - 1.4)

n/a

Gender

Male (n=25)

66.4 (27.9)

0.52

-4.5 (-13.9 - 5.0)

0.60

Female (n=17)

60.9 (25.3)

-9.1 (-23.7 - 5.5)

Age group

18–54 (n=17)

55.1 (28.0)

0.07

-12.1 (-29.4 - 5.3)

0.22

55–80 (n=25)

70.4 (24.5)

-2.5 (-9.3 - 4.3)

Financial situation

Not enough (n=17)

52.6 (24.9)

0.02

-12.9 (-28.5 - 2.6)

0.40

Just enough (n=15)

65.8 (31.3)

-0.8 (-12.8 - 11.2)

More than enough (n=10)

81.4 (5.9)

-3.5 (-17.0 - 10.0)

Area

Metropolitan (n=21)

67.4 (24.9)

0.44

-6.4 (-16.7 - 3.8)

1.00

Rural/Regional (n=21)

61.0 (28.6)

-6.3 (-19.0 - 6.4)

Total comorbidities

Nil (n=5)

51.4 (38.8)

0.26

1.0 (-36.1 - 38.1)

0.74

1–5 (n=30)

68.4 (24.7)

-6.7 (-15.8 - 2.5)

6 or more (n=7)

55.0 (24.4)

-10.3 (-35.2 - 14.6)

Total mental health comorbidities

Nil (n=19)

69.4 (26.3)

0.26

-2.0 (-13.2 - 9.2)

0.31

Any (n=23)

59.9 (26.9)

-10.0 (-21.4 - 1.5)

Notes: a: Negative change scores indicate a decline in score over time b: Two-tailed two-sample t-test for variables with two categories and ANOVA for variables with three or more categories

At Time 1, participants aged 55–80 years had higher wellbeing scores than participants aged 18–54 years and experienced a lesser reduction in wellbeing over time compared to participants aged 18–54 years, although these differences were not significant. Higher wellbeing scores were associated with increased financial stability. At Time 1, participants reporting having not enough money had significantly lower wellbeing scores than those reporting more than enough money. Those who reported not having enough money at Time 1 also had a greater, but non-significant, reduction in wellbeing over time compared to those with just enough or more than enough money. At Time 1, participants without mental health comorbidity had higher wellbeing than those with mental health comorbidity and those without a mental health comorbidity experienced a lesser decline in wellbeing from Time 1 to Time 2, although these differences were not statistically significant.

Health-related quality of life - AQoL-4D

Table 3 presents mean HRQoL utility scores (overall and stratified by participant characteristics) at Time 1 and Time 2, and changes in utility scores over time. Overall, the mean difference in HRQoL utility score between Time 1 and Time 2 was −0.06. While not being statistically significant, this difference was consistent with previously reported minimally important differences. When scores from the separate AQoL-4D domains (Relationships; Independent Living; Senses; and Mental Health) were examined, all domain scores decreased from Time 1 to Time 2. While there was a minimally important difference in Relationship scores over time (mean change Time1-Time2: −0.06), this was not statistically significant (p=0.09), nor was observed for other dimensions (Mean change: −0.04, −0.01, and −0.01, respectively, all p<0.05). At Time 1, HRQoL utility score was associated with financial situation, with utility scores increasing with increasing financial stability, although the only statistically significant difference was between those reporting not enough money and those reporting more than enough money. Those who reported having more than enough money at Time 1 had a greater reduction in HRQoL from Time 1 to Time 2, compared to those who had just enough until next pay day and those who did not have enough until next pay day. At Time 1, there were statistically significant differences in mean HRQoL utility score between those who self-reported having six or more comorbidities, and those who had one to five comorbidities (and those who had none). Wellbeing scores declined over time for those with comorbidity, with the degree of the decline increasing with increasing comorbidity, although these differences were not significant. Similarly, at Time 1 there were statistically significant differences in HRQoL utility score for those with and without mental health comorbidity; HRQoL was higher for those reporting no mental health comorbidity compared to those reporting having a mental health comorbidity. However, both experienced a similar decline over time, which was clinically meaningful, but not statistically significant.

Table 3. Mean (SD) AQoL-4D scores for Aboriginal and Torres Strait Islander participants (n=42), at Time 1(T1) and change over time.

Pre-pandemic (T1)

Changes over time (T1-T2)

Characteristic at baseline (T1)

AQoL-4D utility score

Mean(SD)

P-value

(between groups at baseline)b

AQoL-4D utility score

Mean difference (95%CI)

P-value

(between group change over time)b

Total sample

0.51 (0.28)

n/a

-0.06 (-0.12 - 0.01)

n/a

Gender

Male (n=25)

0.53 (0.29)

0.70

-0.06 (-0.13 - 0.01)

1.00

Female (n=17)

0.49 (0.28)

-0.06 (-0.19 - 0.07)

Age group

18–54 (n=17)

0.46 (0.26)

0.35

-0.04 (-0.17 - 0.08)

0.72

55–80 (n=25)

0.55 (0.30)

-0.07 (-0.14 - 0.01)

Financial situation

Not enough (n=17)

0.38 (0.26)

0.01

-0.01 (-0.13 - 0.10)

0.43

Just enough (n=15)

0.54 (0.28)

-0.06 (-0.14 - 0.01)

More than enough (n=10)

0.71 (0.23)

-0.12 (-0.31 - 0.06)

Area

Metropolitan (n=21)

0.55 (0.28)

0.37

-0.06 (-0.18 - 0.05)

0.83

Rural/Regional (n=21)

0.47 (0.29)

-0.05 (-0.12 - 0.02)

Total comorbidities

Nil (n=5)

0.60 (0.31)

0.005

-0.04 (-0.50 - 0.42)

0.71

1–5 (n=30)

0.57 (0.27)

-0.05 (-0.12 - 0.03)

6 or more (n=7)

0.21 (0.11)

-0.12 (-0.24 - 0.01)

Total mental health comorbidities

Nil (n=19)

0.65 (0.26)

0.004

-0.06 (-0.16 - 0.04)

1.00

Any (n=23)

0.40 (0.26)

-0.06 (-0.15 - 0.04)

Notes: a: Negative change scores indicate a decline in score over time b: Two-tailed two-sample t-test for variables with two categories and ANOVA for variables with three or more categories

Figure 2 shows that the proportion of participants reporting the lowest HRQoL scores increased from Time 1 to Time 2 (23.8% to 38.1%, respectively), while the proportion reporting in the highest HRQoL scores decreased (23.8% to 16.7%, respectively).