Inequalities in Access to Mental Health Treatment by Australian Youths During the COVID-19 Pandemic

We requested quarterly counts of mental health MBS service sessions held between January 2011 and December 2020 from Services Australia, the Australian Government Department of Health (data were received on May 17, 2021, and descriptions of MBS services are presented in a table in the online supplement to this article). Services Australia provided data that were deidentified and aggregated by Statistical Area Level 3 (SA3) (boundaries approximate local government areas) for all young people ages 12–25 years, stratified by sex (males vs. females) and age (adolescents, ages 12–17, vs. young adults, ages 18–25). Counts <10 were suppressed to protect privacy. Because of various factors, including developmental differences and increasing autonomy (health care and financial), data for minors (ages 12–17) and for those ages ≥18 were examined separately.

The estimated residential population (ERP) for 2011–2020 for the matching age and sex subgroups by SA3 was obtained from the Australian Bureau of Statistics. Index of Relative Socio-economic Advantage and Disadvantage (IRSAD) scores for individual SA3s were estimated as the weighted average of included scores for SA2s (representing smaller geographical units, compared with SA3s, and approximating local towns) in 2016 (16). IRSAD is an area-based socioeconomic status measure that summarizes census information about the economic and social conditions of local residents, including measures of both relative advantage and disadvantage (higher scores indicate a relative lack of disadvantage and greater general advantage) (16). SA3-level IRSAD scores ranged from 748.6 to 1,166.6 (first quartile to third quantile: 943.3–1,044.8) (see figure in the online supplement).

Ethics approval for this study was provided by the University of Melbourne Human Research Ethics Committee and Human Ethics and Advisory Group (2056431.1).

All analyses were conducted with R, version 4.0.2 (2020-06-22). A summary of these analyses is provided below (for detailed descriptions, see the online supplement).

For all young people (across all SA3s), the median increase in the rate of MBS mental health service use between 2019 and 2020 was 14.0% (Table 1). However, we noted variation in median service use rates over this period by sex, age, and location. The highest increase was observed among young females (median increase=20.4%), particularly in the last two quarters of the year 2020 (see figure in the online supplement). Rates among young males remained largely unchanged (median increase=2.8%) in 2020, compared with 2019 (Table 1). Service rates in 2020 varied substantially across SA3s, with lower rates observed in regional areas, compared with metropolitan areas, largely consistent with trends before the pandemic (see figures in the online supplement). Increases in service use were higher in Victoria than in other states for young people ages 18–25 (median increase=21% in Victoria and 14% in other states) and females (median increase=26% in Victoria and 19% in other states) (see table in the online supplement).

Interrupted time-series analyses revealed a positive linear association between MBS service use during COVID-19 (RR) and higher socioeconomic status (as measured by IRSAD scores) (Figure 1). This trend was consistent across age and sex subgroups, but the trend was more pronounced for females in areas of higher socioeconomic status. A reduction in MBS service use was common among males in areas with higher socioeconomic disadvantage. Among females in most SA3s, the increase in MBS service use in 2020 was higher (or the reduction was smaller), compared with the trends among males (Figure 2A). However, this sex difference was relatively comparable among areas with different socioeconomic statuses (Figure 2C). Areas with a higher socioeconomic disadvantage tended to have slightly higher increases (or smaller reductions) in service use among those ages 18–25, compared with those ages 12–17 (Figure 2D)—a pattern that became less apparent in socioeconomically advantaged areas.

Higher increases in MBS service use were observed in Victorian SA3s (except for young people ages 12–17), compared with other states and territories. Nationally, metropolitan SA3 areas also had higher increases, compared with regional SA3 areas (see figures in the online supplement).

On average, a 6.2% (95% CI=5.3%–7.2%) additional increase was noted in use of MBS services in 2020 on top of existing trends (Figure 3). The increase was higher for females (11.8%, 95% CI=10.7%–12.9%) and those ages 18–25 (7.2%, 95% CI=6.2%–8.3%), whereas a decreasing trend was observed for males (3.5% reduction, 95% CI=2.5%–4.5%).

The COVID-19 interruption effect was strongly associated with area socioeconomic status (Table 2). That is, a 100-unit increase in IRSAD score (indicating increased socioeconomic advantage) was associated with an increase of between 4.6% and 7.7% in use of MBS services in 2020, after the analysis controlled for differences between regional and metropolitan areas and between Victoria and other states and territories. The highest association was found for young people ages 12–17 (7.7% per 100-unit increase in IRSAD, 95% CI=5.5%–10.0%). Despite overall lower increases observed in regional areas (see figure in the online supplement), meta-regression indicated higher increases for those ages 18–25 (3.1%, 95% CI=1.0%–5.2%) and females (3.4%, 95% CI=1.3%–5.6%) in regional areas, compared with metropolitan areas, with the same level of socioeconomic status and state category. Higher increases in use of MBS services were observed in Victoria for those ages 18–25 (10.1%, 95% CI=7.8%−12.5%) and females (6.6%, 95% CI=4.3%–8.9%), when the analysis controlled for IRSAD and location.

Results from the sensitivity analysis suggested that excluding the autocorrelation structure (i.e., the degree of similarity between a given time series and a lagged version of itself over successive time intervals) had little impact on results (see figure in the online supplement). A slightly smaller COVID-19 interruption effect was estimated from longer time-series analyses (from 2011 to 2020) (see figure in the online supplement), because increases in MBS service use were slightly higher in the early 2010s—likely because of changes in funding—which resulted in an overestimation of the historical trend and an underestimation of the COVID-19 impact.

Overall use of MBS-funded mental health services among young people increased by 6.2% in 2019–2020, likely reflecting both increasing demands and changes in MBS funding policy. Consistent increasing patterns have been reported for mental health–related prescriptions and calls to helplines in Australia (15). However, our estimated increases are much lower than recently observed population-based increases in psychological distress and prevalence of common mental disorders during the pandemic (3, 6, 11). For example, a recent review suggested that the pandemic has led to an 11% increase in anxiety and depression at the population level in Australia (3), which could be higher for young people (17). The National Study of Mental Health and Wellbeing in 2020–2021 found that about half (46.6%) of young females and one-third (31.2%) of young males ages 16–26 had a mental disorder in the past 12 months, compared with 26% of all young people ages 16–26 in the same survey conducted in 2007 (18).

Although it is possible that increasing needs were met by other types of mental health care (e.g., private mental health services) that were not funded by MBS, many news articles and reports, as well as anecdotal information, have suggested long waitlists for accessing care and high levels of unmet needs (19–22). Some have argued that the Australian Government’s Better Access Pandemic Support does not go far enough to meet existing needs, fails to appreciate disparities in service availability and long waitlists, and imposes prohibitive gap fees (i.e., out-of-pocket costs that are not covered by Medicare) (23).

Globally, mental health has declined as a result of the COVID-19 pandemic (3, 24), particularly among young people (17). Findings on the uptake of mental health services are more mixed (25, 26). A few studies observed reductions in service uptake during lockdowns, which was followed by an increase in demand (27–29). Changes in service use may be related to both prepandemic factors shaping help-seeking behaviors (e.g., illness severity, mental health literacy, stigma, and cost) (30, 31) and pandemic-related factors (e.g., fear of infection, reduction in service availability, and uneven burdens imposed on people of different socioeconomic status) (26). However, to the best of our knowledge, no study has evaluated how population-level changes in service utilization were affected by factors such as socioeconomic status and lockdown experiences.

The sex differences in mental health service use observed in this study (11.8% increase for females and 3.5% decrease for males) suggest a widening access gap for males. Rates of MBS service use among males were slightly more than half the rates among females in this population of young people, which further fell during the pandemic. Although prevalence rates of anxiety and depression were higher among females than among males both before and during the pandemic (6, 32), studies also have reported increased psychological distress among males during the pandemic (32, 33). Sex differences in symptom presentation and help seeking might contribute to a reduction in service use by males. Known barriers to mental health help seeking among young men include health literacy, cultural norms that endorse masculinity, stigma, and service acceptability (34).

There is mounting evidence that males experiencing affective disorders often report a constellation of externalizing signs and symptoms, beyond those included in current diagnostic criteria (35, 36). Indeed, the recently published DSM-5-TR in part acknowledges this evidence, stating that men with major depressive disorder (compared with women) report greater frequencies and intensities of alcohol or other drug misuse, risk taking, and poor impulse control (37). The magnification of externalizing symptoms may also affect the ability of parents to identify and support the needs for care among boys and young men.

MBS-funded services may not adequately address the clinical and engagement needs of young men experiencing or presenting with externalizing symptoms. Population-level data show that compared with young females, young males are less likely to engage with MBS-funded mental health care provided by headspace services, and when they do engage, they are more likely to discontinue care early (38). The degree to which help-seeking and symptom-presentation factors affect the mental health of young men both during and after the COVID-19 pandemic needs further scrutiny.

We observed a concerning inequitable socioeconomic distribution of the growth in mental health service use. For all young people, the areas with average IRSAD scores (995.5) increased by 3.8% during 2020; this change was expected to be a 14.9% increase in areas with the highest IRSAD (1,166.6) and a 10.3% reduction in areas with the lowest IRSAD (748.6) (calculated with the RRs for the intercept and the IRSAD for all young people in Table 2). Socioeconomic status was found to have a higher impact on service use among young people ages 12–17, compared with other age and sex subgroups. Although regional areas had smaller increases in service use than metropolitan areas, findings from multivariate meta-regression models suggested that this differential increase was likely driven by differences in socioeconomic status, particular for those ages 12–17, compared with other age and sex subgroups. The pandemic had a greater impact on low-income families (e.g., higher infection rates, more social isolation, greater income reduction, and decreasing access to resources) (39, 40). These additional challenges may have limited the capacity of these parents to assist young people seeking needed care (e.g., financial difficulties in paying out-of-pocket fees) (40).

We note that marked disparities in mental health problems and service access based on socioeconomic status existed in Australia before the pandemic (41, 42). The pandemic has likely exacerbated existing inequities (43), heightening barriers related to service availability and accessibility (44). The estimated steep socioeconomic gradient in this study also raises questions as to whether “more-of-the-same” policy responses based on MBS subsidies can reduce existing unequal service access distributions during a crisis or are likely to entrench such inequities (23). Out-of-pocket expenses for accessing psychiatrists and psychologists might be a greater barrier for young people from socioeconomically disadvantaged backgrounds (23). The current concentration of service providers in more affluent areas might also contribute to these issues (45).

The higher increase in service uptake in Victoria is consistent with reports of greater mental health impacts in this state because of strict restrictions and lengthy lockdowns (12). Higher service uptake might be related to the higher rate of MBS-subsidized telehealth use in Victoria (15), which may have made access to care easier for young people.

Key strengths of this study included the use of advanced statistical techniques and the most currently available MBS data. A range of sensitivity analyses ensured the findings’ robustness. Limitations of this study are related to the availability and nature of the MBS data. For example, aggregated SA3-level service counts were obtained to protect privacy, limiting the ability to further understand differences at the individual level (e.g., examination of more narrow age bands and whether services were provided to the same individuals) and service level (e.g., type of service provider and mode of service delivery, such as telehealth). Future efforts are needed to improve timely access to individual-level MBS data to better monitor and evaluate trends in service uptake. The quarterly time units did not reflect specific dates of events, such as the timing of lockdowns or the introduction of changes to the MBS. At the time of this evaluation, data were available from 2020 only, which limited our capacity to estimate the continued impacts of the pandemic and policy changes. Further, our data indicated only changes in service use; they did not reflect the severity of disorders or psychological distress and the needs and demands on services. For example, the data did not include those who failed to get help because of lack of service availability and long waitlists or those who accessed mental health services through other funding sources (e.g., private health insurance and schools). Future studies using individual-level and linked MBS data (e.g., examining emergency department presentations) are needed to better understand gaps in service provision and demands.

The ERPs used in this study did not reflect likely changes in population structure (primarily in areas with a higher international student population), which may have caused bias in rates of service use. However, this limitation was unlikely to have affected the generalizability of the study findings, because we used pooled effects.