In the early stages of Australia’s COVID-19 response, governments moved quickly to publish daily case numbers, hospitalisations and deaths. The scale and speed of reporting were unprecedented and, rightly, praised.
But as the pandemic unfolded, a critical limitation became increasingly visible. Despite repeated commitments to equity, information about ethnicity, language, and cultural background was inconsistently collected across jurisdictions. In some datasets, the largest category was simply “unknown”. In others, the data was not collected at all. The Australian Institute of Health and Welfare (AIHW) has previously noted that gaps in cultural and linguistic data limit the system’s ability to identify inequities and respond early, particularly during public health emergencies.
As outbreaks began to disproportionately affect multicultural communities, public health teams often relied on indirect signals, such as postcode-level trends and workplace clusters, because meaningful data on ethnicity and language was not consistently available. These proxies offered clues, but they could not fully capture the lived realities shaping risk. In many cases, it was community leaders and multicultural organisations who were raising concerns well before the system could clearly see them.
By the time more targeted responses were implemented, such as multilingual communications, community-led outreach, and pop-up clinics, many communities had already experienced disproportionate exposure, mistrust and harm. COVID-19 death rates among migrants were twice as high as those of people born in Australia.
No one set out to exclude multicultural communities from COVID-19 data. But the absence of consistent, meaningful data meant inequities were recognised late, rather than prevented early. There were valuable lessons to be learnt here, ones that the multicultural sector has been repeating for a long time: When we are invisible in the data, we become invisible in the system. Inequity becomes inevitable.
HIV surveillance reveals a similar dynamic. Even in one of Australia’s most established and closely monitored areas of public health, multicultural communities remain only partially visible. National HIV surveillance reports compiled through state and territory notification systems show that variables such as country of birth are collected more consistently than ethnicity or language, resulting in a fragmented picture of risk, diagnosis and care pathways. And if gaps persist here, they are almost certainly present across the rest of the health system.
Nation Leading Health Surveillance Uncovers Gaps
Australia has long relied on HIV surveillance data to guide prevention, treatment and community responses. Compared with many other areas of health, HIV data is often seen as relatively mature: nationally coordinated, routinely reported and closely scrutinised by governments, clinicians and communities alike.
HIV has always been shaped by social and structural factors. In addition to sexuality and gender, migration pathways, language barriers, stigma, access to Medicare, trust in health systems, and experiences of racism all influence who is tested, diagnosed and supported.
For this reason, HIV data functions as an early warning system. When gaps appear in HIV surveillance, they often point to broader failures in system design. And yet, recent work examining HIV notifications among people from culturally and linguistically diverse backgrounds shows that even in this comparatively advanced space, multicultural communities remain only partially visible. While country of birth is generally recorded, other meaningful variables such as ethnicity, language spoken, migration pathway, and visa status are not collected in a consistent or standardised way. As a result, people from diverse cultural backgrounds are often grouped into broad categories that obscure important differences in risk, late diagnosis and access to care.
Recent analysis of HIV notifications highlights persistent limitations in how people from multicultural backgrounds are identified in national datasets. Definitions of “CALD” vary. Ethnicity, language spoken, country of birth and migration status are not consistently captured. Where data exists, it is often incomplete or difficult to interpret.
The result is a partial picture. Communities at higher risk of late diagnosis or disengagement from care may not be clearly identified. Trends that could inform targeted prevention or culturally appropriate responses remain obscured.
These issues are not unique to HIV. They reflect structural weaknesses in Australia’s approach to multicultural health data more broadly.
This is not just a technical issue for epidemiologists. It is a signal of deeper, system-wide problems in how Australia collects, understands, and uses data about multicultural communities across the health system.
HIV data in Australia is often held up as a best-practice example of how surveillance can inform public health action. That reputation is largely deserved. But it also makes HIV a critical test case.
If persistent data gaps exist in one of Australia’s most mature, well-resourced and closely scrutinised surveillance systems, it raises serious questions about how multicultural communities are captured across the rest of the health system.
In this sense, HIV data does not just tell us about HIV. It tells us to what extent Australia’s health system sees, understands and responds to diversity.
If we cannot clearly see multicultural communities in HIV data, we will struggle to see them anywhere else. In public health, what we fail to measure, we also fail to fund.
Blunt categories cannot deliver precise health responses
For decades, Australian health data has relied on broad proxy measures such as country of birth or language spoken at home. While useful, these variables no longer reflect the complexity of Australia’s population. The Australian Bureau of Statistics has acknowledged that these variables were never designed to function as a comprehensive measure of ethnicity, yet they are often used that way in health policy and service planning.
These proxy measures also render locally born cohorts of multicultural communities largely invisible. Second- and third-generation Australians who are born in Australia and speak English at home can still experience structural racism, cultural stigma or health risks shaped by migration histories and community context. When data collection relies primarily on country of birth or language, these cohorts are statistically absorbed into the “Australian-born, English-speaking” majority, masking important inequities and trends.
“CALD” has become a catch-all classification that masks enormous diversity. It can group together people with distinct differences in health needs which are uniquely impacted by compounding experiences of marginalisation across culture, migration, social supports, sexuality, gender, and access to services. In practice, this means that data often lacks the nuance needed to design effective interventions – a low-resolution label applied to a highly complex population.
In the HIV context, this can translate into missed opportunities for prevention, delayed diagnosis, or services that fail to resonate with the communities they aim to reach. In other parts of the health system, the consequences are just as serious, affecting cancer screening, chronic disease management, mental health and perinatal care.
The challenge is not simply about collecting more data. It is about purposefully collecting the right data, in ways that are culturally safe, with interpretation and applications that are meaningful and useful to community.
What invisibility shows us
Data is not neutral, it underpins policy. What we do not measure, we cannot provide focus. What we are not focused on, we fail to change. When multicultural communities are poorly captured in data, health needs are underestimated, funding allocations fail to reflect patterns of need, and accountability is weakened. AIHW reporting has repeatedly shown that where population characteristics are missing or inconsistently recorded, disparities tend to be recognised later and addressed less systematically.
Over time, this reinforces existing inequities. Communities that are already marginalised become further invisible within the system, and disparities persist despite good intentions.
In HIV, where prevention relies heavily on trust, early engagement, and culturally appropriate messaging, these gaps are particularly damaging. But the same dynamics play out across the health system.
From problem identification to system reform
The value of recent HIV data work is that it clearly articulates the problem. The next step is translating those insights into system reform. From a national multicultural health perspective, several priorities emerge.
First, Australia needs greater consistency in how ethnicity, language and related variables are collected across health datasets. This does not require reinventing the wheel. Clear standards already exist, but uptake is uneven and accountability is limited.
Second, data collection must move beyond compliance and towards purpose. Communities are more likely to share information when they understand why the data is being collected and how it will be used to improve care. This requires transparency, feedback loops and genuine engagement.
Third, data alone is not enough. Without culturally safe services, community-led program design, an and an appropriately trained workforce, better data will not translate into better outcomes. Data reform must be embedded within broader efforts to address structural barriers in the health system.
Finally, there is a need for national leadership. Fragmented approaches across jurisdictions and programs make it difficult to build a coherent picture of multicultural health needs. A coordinated multicultural health framework would allow insights from HIV to inform reforms across preventive health, primary care, and population health more broadly.
The HIV classroom
HIV has often led broader health reform, from community engagement models to peer-led prevention and rights-based approaches to care. Multicultural health data should be no different.
The lessons are clear: Early investment in meaningful data pays dividends in prevention and care; community involvement improves both data quality and service uptake; structural barriers, not culture, drive inequities; and, systems must adapt to diversity, not the other way around.
By treating HIV data challenges as isolated technical problems, we miss the opportunity to address the root causes. By recognising them as system signals, we can use them to drive broader change.
What better practice looks like
Other countries have demonstrated that capturing multicultural data well is both feasible and valuable.
While the United Kingdom and New Zealand are not perfect analogues for Australia, and their systems are not without flaws, they show that routinely collecting and using ethnicity data at scale is possible, even in politically and socially complex contexts.
In the United Kingdom, ethnicity is routinely collected across primary care, hospitals and public health surveillance using nationally consistent self-identified categories. This information is linked across systems and analysed alongside outcomes, enabling disparities in areas such as HIV, cancer, cardiovascular disease and COVID-19 to be identified early. The UK Government regularly publishes these insights through public health dashboards and equity reports, creating a clear line of sight between data, policy decisions and accountability.
New Zealand takes a similarly deliberate approach. Ethnicity is treated as a core equity variable across health datasets, not an optional add-on. Maori and Pacific peoples’ outcomes are explicitly monitored, and this data is used to inform funding decisions, service design and performance monitoring. Importantly, ethnicity data in New Zealand is understood as a tool for addressing structural inequity, rather than as a descriptive demographic characteristic.
In both contexts, the value of multicultural data lies not simply in collection, but in use. Data is gathered consistently, interpreted carefully, and applied to improve services and outcomes. When diversity is built into the core architecture of health data, inequities become visible earlier and responses can be more targeted and effective.
The contrast with Australia is not about ambition or intent. It is about system design. Without consistent, meaningful data on ethnicity, language, and cultural background, Australia’s health system will continue to rely on indirect signals to identify inequity, often recognising problems only after communities have already been harmed.
A role for national collaboration
Improving multicultural health data is not the responsibility of any single organisation or program. It requires collaboration across governments, health services, researchers and communities.
National peak bodies have a role to play in bridging these conversations, ensuring that insights from specific areas, like HIV, inform whole-of-system reform. This includes advocating for consistent data standards, supporting culturally safe research practices, and amplifying community voices in policy development.
The work underway in HIV surveillance provides a strong foundation. The challenge now is to ensure those lessons are not siloed but used to strengthen Australia’s health system as a whole.
Looking ahead
Australia’s population is becoming more diverse, not less. Health systems designed around homogenous assumptions will continue to fall short unless deliberate action is taken.
During Victoria’s second COVID-19 wave in 2020, public health data initially suggested that outbreaks were broadly distributed across the population. Case numbers were rising, but there was limited clarity about which communities were most affected.
Behind the scenes, however, multicultural and settlement organisations were reporting a very different picture. They were seeing clusters among newly arrived migrants, refugee communities and temporary visa holders, particularly in essential workplaces. Language barriers, insecure work and limited access to health information were clearly shaping exposure and response.
The problem was not the absence of impact but the absence of data.
HIV data has shown us where the cracks in the system are. The question is whether we are prepared to fix them.
Getting multicultural health data right is not about better statistics for their own sake. It is about visibility, accountability and equity. And ultimately, it is about ensuring that all communities can access care that is timely, effective and respectful.