Longitudinal Patterns of Chronic Hepatitis B Care and Treatment Adherence in Australia: A National Linked Data Cohort Study

ABSTRACT

Current uptake of chronic hepatitis B (CHB) treatment and care in Australia is well below target levels, leading to an ongoing burden of preventable liver disease and cancer. Understanding individual‐level trajectories of care engagement and variation by sociodemographic factors is essential to guide efforts to improve access to and retention in care. This study analysed records of hepatitis B viral load (VL) testing and antiviral treatment provided through Australia's national health insurance system, which were linked to migration, death, and sociodemographic data from various government datasets. Treatment and VL testing by population group were assessed, and uptake was calculated using modelled estimates of the number living with CHB. VL testing frequency and treatment adherence based on scripts dispensed were also assessed. Of the estimated 219,800 people living with CHB in Australia in 2023, 118,334 (53.8%) received either VL testing or treatment during 2014–2023. Estimated uptake of VL testing and treatment uptake were lower among Aboriginal and Torres Strait Islander people and also varied by region of birth. Of the 81,360 who received any VL testing, only 17,352 (21.3%) received it regularly, while 31,715 (39.0%) had evidence of prior monitoring but no VL testing since 2019. The majority of those who initiated treatment continued therapy and received supply sufficient for adherence. This study shows that substantial gaps exist in CHB care in Australia, particularly in continued monitoring among those not on treatment. Disparities in uptake are evident by priority population. Re‐engaging those lost to follow‐up should be prioritised.

1. Introduction

Chronic hepatitis B (CHB) affects over 250 million people globally [1], and more than 200,000 people in Australia are living with CHB [2]. It has substantial personal and social impacts [3] and leads to an increasing burden of liver disease and liver cancer [4, 5]. Effective (usually lifelong) management, including regular viral load (VL) monitoring and treatment where indicated, reduces this risk significantly [6]. Understanding uptake according to sociodemographic factors is also crucial in Australia, given that the population living with CHB is predominantly people born overseas and the disproportionate burden among Aboriginal and Torres Strait Islander people [7].

The current level of diagnosis, treatment, and monitoring for CHB in Australia is inadequate, despite these services being covered by Australia's national health insurance. Australia is not on track to meet its 2030 national strategy targets [8] for diagnosis (90%), care (80%), and treatment (27%), with current levels at 69%, 25% and 13% respectively [2]. Despite increases in care and treatment uptake over the past decade, progress has stalled since 2020, with evidence that the COVID‐19 pandemic negatively impacted service delivery for CHB [2, 7].

Australia is also not on track to meet mortality reduction targets, with attributable liver cancer and cirrhosis deaths increasing by 15% since 2015 to more than 400 per year, compared with a target reduction of 30% by 2030. These issues are reflected globally, and almost all countries will require substantial improvement to reach hepatitis B goals [1].

Population‐level care and treatment uptake for CHB in Australia have been reported annually by geographic region since 2014 [2, 9]. The frequency and patterns of monitoring and adherence to antiviral treatment over time have not previously been reported outside small cohorts [10]. Adherence to antiviral treatment is correlated with suppression of viral load and is a key goal in reducing adverse outcomes. While treatment is not recommended for all people with hepatitis B, regular HBV VL monitoring is indicated to identify opportunities to initiate treatment and reduce the risk of adverse outcomes [6].

However, detailed patterns of care and longitudinal trends at the individual level have not been assessed. Key gaps include understanding the frequency and patterns of monitoring and adherence to antiviral treatment. These are important to understand as lack of adherence can have negative consequences on viral suppression [10], and inadequate monitoring can result in missed opportunities to start antiviral treatment and so prevent liver cancer and cirrhosis. Analysis according to priority population has also previously been limited by the availability of comprehensive data; however, recent national data linkage initiatives provide the opportunity for robust assessment [11].

This analysis characterises ongoing care for CHB, including treatment adherence, to inform clinical and policy priorities that enhance access to and retention in care. We estimate engagement in CHB care in Australia and examine trends in management and treatment, including variation across priority populations.

2. Methods

2.1. Study Design, Data Sources and Study Population

A retrospective population‐based cohort was constructed using linkage of records derived from routinely collected data. The cohort was established using records of all those who received hepatitis B VL testing or treatment through national health insurance during 2014–2023, which were then linked to four additional Australian Government datasets capturing sociodemographic factors and vital status. These data were accessed via the Australian Bureau of Statistics Person‐Level Integrated Data Asset (PLIDA) [11].

2.2. Health Services

Australia's universal government‐subsidised medical care and testing (Medicare Benefits Schedule, MBS [12]) and prescription medication (Pharmaceutical Benefits Scheme, PBS [13]) systems provided data for hepatitis B VL testing and hepatitis B treatment, respectively (see Table S1 for item codes used). MBS and PBS data record each instance of care provided, and specify the indication (e.g., distinguishing use of tenofovir for CHB from HIV). The cohort used for analysis included all people who had any record of hepatitis B VL testing or hepatitis B antiviral treatment via these systems during 1‐Jan‐2014–31‐Dec‐2023.

2.3. Sociodemographic Factors and Vital Status

Sociodemographic data and vital status were generated using three linked datasets: The Australian Census of Population and Housing (Census), government payments data (Data Over Multiple Individual Occurrences dataset, or DOMINO), and the National Death Index (NDI). These datasets are used by the ABS to generate a combined demographic information dataset using a rules‐based approach that consolidates information from multiple source datasets. Region of birth was categorised using the ABS Standard Australian Classification of Countries [14]. Aboriginal and Torres Strait Islander individuals are those ever coded as such in any linked dataset.

Data regarding vital status were also sourced from the NDI. Migration data are used by the ABS to generate a ‘physically present’ status for each person based on travel history, allowing identification of those no longer present in Australia. These were used to exclude those who had died or who no longer lived in Australia, so that the cohort reflected only those living in Australia at the end of 2023 (Figure 1).

FIGURE 1.

Diagram of datasets used in analysis. DOMINO, Data Over Multiple Individual Occurrences dataset.

2.4. Denominator Data

The number of people receiving CHB treatment and care was divided by the number estimated to be living with CHB in Australia in 2023. This denominator was calculated using a deterministic compartmental mathematical model of hepatitis B virus infection in the Australian population, which incorporates existing mathematical models, surveillance notifications, epidemiological research, clinical studies, and demographic and mortality data [15]. Further information regarding the model can be found in the associated literature [2, 16].

Denominator data specific to Aboriginal and Torres Strait Islander people and those born overseas is also generated using this model. Variation in prevalence by population group relies predominantly on age‐specific antenatal seroprevalence data [17, 18, 19] applied to Census population estimates. Additional sources such as population‐level clinical data [20, 21] and notifications data sourced from Australia's National Notifiable Diseases Surveillance System [22] are also incorporated into estimates for Aboriginal and Torres Strait Islander people to account for geographic variation in prevalence.

2.5. Care Outcomes

Our key outcomes of interest were a set of clinical care provision metrics. People were assigned into one of three overarching categories: (1) those who had ever received hepatitis B antiviral treatment, based on PBS records; (2) those who did not receive hepatitis B treatment but had ever received VL testing, based on MBS records; and (3) those with CHB who had neither hepatitis B VL testing nor hepatitis B treatment during the period, derived by subtracting groups 1 and 2 from the estimated total number with CHB (see Denominator data).

Those who received only interferon treatment were excluded from analysis of care patterns due to the time‐limited nature of treatment and small cohort size. For each outcome group, the number and proportion of the total estimated to be living with CHB were calculated. For a detailed description of the items used to define these categories, see the Table S1.

For those who received antiviral treatment, additional categories were derived based on treatment duration and adherence, and the number who received VL testing while also receiving antiviral treatment was derived. For those who received VL testing only, categories were derived based on monitoring frequency. These subcategories are described below, and summarised in Figure 2.

FIGURE 2.

Categories of hepatitis B care and treatment used for analysis, Australia, 2014–2023 VL, viral load. *Adherent treatment represents those who received 80% or more of the required pill supply during the total period they received treatment.

2.6. Antiviral Treatment Adherence and Patterns

Analysis of treatment adherence was limited to those who received two or more years of treatment in total during 2014–2023, as those treated for shorter periods may only have had one script dispensed, limiting adherence calculations. This also excludes short‐term therapy indications (e.g., pregnancy).

Categories of treatment adherence were derived from script dispensing records, using a pill supply threshold of 80% [10]. People were considered “adherent” if they received 80% or more of the required pill supply during the total period they received treatment, and “non‐adherent” if they received less than 80%. Pill supply calculation accounted for the combined hepatitis B therapy, where people were prescribed multiple drugs concurrently.

People receiving treatment for two years or more were categorised based on whether they initiated treatment during the period 2014–2023, or whether they were already on treatment at the beginning of the period of observation available.

Among those who had ≥ 2 years of treatment and then ceased treatment, we extracted the number and proportion of people who received either a hepatitis B VL test or a hepatitis serology test at any time after the date of their last script. When this occurred, it was assumed the treatment cessation was a clinically guided treatment decision as opposed to loss to follow‐up.

For those treated for less than two years, we derived the number who were initiated on treatment after 2021 (thus could not receive more than two years of therapy) and the number who were treated for less than two years but then ceased therapy.

2.7. Monitoring Frequency

For those who had not received any treatment but received VL testing, the frequency of monitoring was assessed by deriving the number of people who received a VL test during each of the three time periods during follow‐up (2014–2016, 2017–2019, 2020–2023). Those who had a VL test during all three time periods were classified as having ‘regular VL testing’. Those who did not were classified as having had ‘recent VL testing’ if they had a test during 2020–2023, and ‘prior VL testing’ if they did not (Figure 2). This ‘prior VL testing’ category was considered indicative of clinical loss to follow‐up.

2.8. Statistical Analyses

The number of people in each care outcome category outlined above was extracted, and uptake was calculated by dividing the number in each category by the total estimated number of people living with CHB (see Denominator data). This was calculated by 20‐year age group (based on age in 2023), Aboriginal and Torres Strait Islander status, and region of birth (see Sociodemographic factors and vital status) for antiviral treatment and VL testing, where numbers were sufficient for assessment. Pearson chi‐squared analysis was performed to assess differences in uptake between categories. Data management and analysis were conducted within a secure environment using Stata, R, and Microsoft Excel.

2.9. Approvals

This research was conducted in accordance with the Declaration of Helsinki. Human Research Ethics Committee approval was obtained from the Royal Melbourne Hospital (reference HREC/77929/MH‐2022), including a waiver of individual consent for the use of individual‐level data with identifiers removed.

3. Results

3.1. Overall Treatment and Monitoring Uptake

Between 1‐Jan‐2014 and 31‐Dec‐2023, 129,040 people received hepatitis B VL testing and/or treatment. After exclusion of those who had died (7,832 people) or who had emigrated (2,874 people), there were 118,334 people in the final analysis cohort (Figure 1).

This represented 53.8% of those estimated to be living with CHB in Australia in 2023 (Table 1). Of those, 36,846 people received hepatitis B antiviral treatment, 16.8% of the total estimated to be living with CHB in 2023. An additional 128 people received interferon therapy and no antiviral therapy during the period (< 0.1% of those living with CHB).

TABLE 1.

Number of people who received hepatitis b antiviral treatment and viral load testing, and estimated uptake of all people with CHB, according to demographic characteristics in australia during 2014–2023.

	Estimated number of people living with CHB	Had HBV antiviral treatment	Proportion of total with CHB (%)	Had HBV VL testing but no treatment	Proportion of total with CHB (%)
TOTAL	219,800	36,846	16.8%	81,360	37.0%
Age group
< 20	14,024	87	0.6%	632	4.5%
20–39	86,745	4754	5.5%	15,943	18.4%
40–59	77,471	16,983	21.9%	38,738	50.0%
60–79	34,842	13,684	39.3%	24,230	69.5%
80+	6720	1338	19.9%	1,816	27.0%
Aboriginal and Torres Strait Islander status
Aboriginal and/or Torres Strait Islander person	14,785	1092	7.4%	4,208	28.5%
Non‐Indigenous	205,015	35,754	17.4%	77,152	37.6%
Born in Australia or overseas
Australia	60,224	4,136	6.9%	16,719	27.8%
Overseas	159,574	31,588	19.8%	61,751	38.7%
Overseas region of birth
Americas	2923	91	3.1%	378	12.9%
North Africa and the Middle East	7582	1152	15.2%	3,189	42.1%
North‐East Asia	51,840	11,601	22.4%	20,318	39.2%
North‐West Europe	4514	333	7.4%	1296	28.7%
Oceania (excluding Australia)	11,980	1500	12.5%	3729	31.1%
South‐East Asia	49,995	13,767	27.5%	23,695	47.4%
Southern And Central Asia	8234	1021	12.4%	2974	36.1%
Southern And Eastern Europe	11,848	1354	11.4%	3371	28.5%
Sub‐Saharan Africa	10,658	769	7.2%	2801	26.3%

Note: Totals will not add due to the occurrence of missing demographic data.

Abbreviations: CHB, chronic hepatitis B; VL, viral load.

Of those who received hepatitis B treatment, 31,688 (85.7%) also received VL testing. There were 81,360 people who received no hepatitis B treatment but did receive VL testing during 2014–2023, representing 37.0% of people living with CHB in 2023 (Table 1).

The remaining 101,466 people (46.1% of the number estimated to be living with CHB) had neither hepatitis B treatment nor VL testing during 2014–2023.

3.2. Uptake by Sociodemographic Factors and Priority Population

Hepatitis B antiviral treatment and VL testing uptake were estimated to be highest among those aged 60–79 years (Table 1, p < 0.001). Treatment uptake was estimated to be higher overall among those born overseas (19.8%) than those born in Australia (6.9%, p < 0.001), as was VL testing (38.7% compared to 27.8%, p < 0.001). Among those born overseas, uptake of both VL testing and treatment was estimated to be highest among those born in South‐East Asia (Table 1; p < 0.001 for difference with those born in Australia).

Treatment uptake was estimated to be lower among Aboriginal and Torres Strait Islander people than non‐Indigenous people overall (7.4% compared to 17.4%, p < 0.001); however, it was similar to uptake in all Australian‐born people (6.9% uptake, p = 0.03). VL testing uptake was also lower among Aboriginal and Torres Strait Islander people than non‐Indigenous people (28.5% compared to 37.6%, p < 0.001).

3.3. Patterns of Treatment and Adherence

Of the 36,846 who received hepatitis B antiviral treatment, 26,013 (70.6% of those treated) received it for two or more years and were included in the analysis of treatment patterns and adherence. The majority (20,908 people, 80.4% of those treated for two or more years) were adherent to therapy using the defined 80% threshold for pills dispensed over time, while 5105 people (19.6% of those treated for two or more years) were dispensed at a frequency below the 80% threshold (non‐adherent; see Table 2).

TABLE 2.

Patterns of Hepatitis B viral load testing and antiviral treatment, 2014–2023.

Category of care provision	Number of people	Proportion of total estimated number of people living with CHB (N = 219,800)
Had HBV antiviral treatment	36,846	16.8%
Had ≥ 2 years of antiviral treatment	26,013	Proportion of people who received ≥ 2 years of antiviral treatment (N = 26,013)
Adherent treatment	20,908	80.4%
Consistent treatment during 2014–2023	9,020	34.7%
Started treatment after 2014	9518	36.6%
Ceased treatment prior to 2023	2370	9.1%
Non‐adherent treatment	5105	19.6%
Had < 2 years of antiviral treatment	10,833	Proportion of people who received < 2 years antiviral treatment (N = 10,833)
Started treatment prior to 2021, but ceased	5311	49.0%
Started treatment after 2021	5522	51.0%
Had HBV VL testing but no treatment	81,360	37.0%
		Proportion of people who had a VL test (N = 81,360)
Regular VL testing	17,352	21.3%
Recent VL testing only	32,293	39.7%
No recent testing	31,715	39.0%

Note: Adherent treatment was defined as sufficient scripts dispensed to above 80% pill supply threshold.

Abbreviations: CHB, chronic hepatitis B; VL, viral load.

A total of 9020 people (34.7% of those on treatment for two or more years) were on treatment consistently during 2014–2023, and a further 9518 people (36.6%) commenced therapy after 2014 but were on treatment through 2023 (Table 2, Figure 3). Another 2370 people (9.1%) received continuous treatment while receiving therapy, but ceased it prior to 2023. Of those treated for less than two years (10,833), half (5522, 51.0%) were those who initiated treatment after 2021; therefore, they had < 2 years of follow‐up available (Table 2).

FIGURE 3.

Categories of hepatitis B care for (A) those who received hepatitis B treatment [N = 36,846] or (B) received hepatitis B viral load testing while not receiving treatment [N = 81,360], 2014–2023. Adherent treatment defined as sufficient scripts dispensed to above 80% pill supply threshold. VL, viral load.

For those who had two or more years of adherent treatment but ceased prior to 2023 (2370), the majority (1529, 64.5%) had a hepatitis serology test or hepatitis B VL test after cessation, compared to only a minority of those who had two or more years of non‐adherent treatment before ceasing (2343 of 5105, 45.9%).

3.4. Patterns of VL Testing

Of the 81,360 people who had a VL test but no treatment, only 17,352 (21.3% of those who had a VL test) had regular testing during all three time periods examined (Figure 3). A total of 31,715 people (39.0% of those with a VL test) had a VL test during 2014–2016 and/or 2017–2019, but no test during 2020–2023, indicating cessation of previous monitoring.

4. Discussion

This study provides the first comprehensive longitudinal assessment of engagement in ongoing care for CHB in Australia, including estimation of variation between key priority populations and measurement of treatment continuation. It provides evidence that intermittent VL monitoring is common, as is cessation of previous monitoring (suggesting loss to guideline‐based follow‐up), identifying opportunities for intervention. While treatment adherence overall is moderately high at a population level, it is concerning that a substantial minority of people are ceasing treatment without further evidence of ongoing care.

We also estimated that there is substantially lower uptake of treatment and VL testing among Aboriginal and Torres Strait Islander people, indicating a need for improved access to care. This concords with other sub‐national cohorts, including NSW [23, 24] and Queensland [21], which found treatment and VL testing uptake to be lower among Aboriginal and Torres Strait Islander people and higher in those born overseas. This variation is also supported by existing evidence, which identifies wide variation in treatment and care uptake according to geographic region, given the distribution of priority populations in different areas [7, 25], and highlights the need for a focus on equity in access to care for CHB.

There are established gaps in chronic disease care for Aboriginal and Torres Strait Islander populations, which have been associated with a lack of access to culturally safe services, competing health priorities, gaps in systematic follow‐up, and, for hepatitis B, shame and stigma related to the condition [26, 27]. This contrasts with the high uptake observed in some Aboriginal and Torres Strait Islander populations where comprehensive programmatic approaches have been undertaken, which emphasise the importance of implementing locally tailored programs supported by effective engagement and education [24, 28, 29].

Despite the higher proportional uptake among those born overseas, there remains a need to improve access to care. Given their relative population size, individuals born in the Asia‐Pacific region represent the largest estimated group living with CHB without evidence of care engagement. Interventions which transition CHB management to primary care in conjunction with community engagement and workforce development have been shown to be effective in these populations [30], and their expansion will be necessary to close the remaining significant gaps.

Overall, we estimated that approximately 45% of all people living with CHB in Australia had received neither VL testing nor treatment in the past decade. This is a significant barrier to Australia reaching its strategic goals for increasing hepatitis B treatment and reducing adverse outcomes, as well as delivering a potential future functional cure to those affected. Given that approximately 30% of people living with CHB in Australia are undiagnosed [15], most of those not receiving care are likely to be unaware of their condition. We also identified that a substantial number are only engaged in intermittent care or drop out of care. Further research is needed to identify drivers of this and potential clinical factors affecting these trajectories. Although this study cannot fully account for individualised decision‐making in the provision of care, the large number of individuals estimated to have received no care during the past decade suggests that there are substantial gaps in evidence‐based care among those eligible for it.

The finding that those aged 60–79 years had the highest estimated care uptake is consistent with estimates of treatment eligibility, which increase with age [2]. However, the substantial gaps in VL testing in those aged 40–59 years highlight the need to engage those approaching ages associated with increased risk of adverse outcomes into regular monitoring to facilitate timely guideline‐based treatment.

Additional analysis of this cohort is planned to assess the influence of socioeconomic status, migration history, English language proficiency, and comorbidities on care engagement. Further work is also required to assess elements of hepatitis B care not assessed here, such as hepatocellular carcinoma surveillance and treatment uptake for specific indications, such as pregnancy.

As infant immunisation continues to be delivered, particularly in high‐prevalence countries, the number of people living with CHB in younger age groups is predicted to decline, including in Australia [2]. As the cohort of people living with CHB continues to age, it will be increasingly important to address the ongoing challenges of comorbidities in the population affected, to support the overall and liver‐related health of those affected.

The major strength of this study is the very high population coverage provided, which overcomes the substantial selection bias inherent in analysing clinical cohorts of people already engaged in care for their hepatitis B. Additional strengths include the ability to account for population dynamics such as emigration and mortality; the very high completeness of sociodemographic data derived from multiple linked datasets; and the relatively long time period available, allowing for assessment of longitudinal trends. The use of these linked data helps to address the issue of loss‐to‐follow‐up, which is a considerable limitation in many cohort studies and here allows for distinction between those who have died or emigrated from those living in Australia but not engaged in care.

The denominator estimates used to generate uptake measures for this study are regularly updated and validated against other sources, including population‐level antenatal seroprevalence studies [15, 16, 31, 32, 33, 34]. The concordance between these outputs and previous findings (which do not rely on modelling) regarding treatment and care coverage by population group [23] also provides reassurance. However, the derivation of these key denominators using modelling introduces unavoidable uncertainty, and uptake could be underestimated, or population disparities mischaracterised, due to this approach.

Other study limitations relate to gaps in the available data sources. Although the coverage of the Australian population provided by the MBS and PBS is high, those residing in Australia on temporary visas who are not eligible for Medicare coverage will not be captured [35]. Further, MBS and PBS data do not include test results, preventing identification of reasons for treatment cessation or outcomes that result from lack of treatment adherence (such as inadequate viral suppression). These data gaps also limit the interpretation of variation in treatment uptake by population group, as these reasons may vary due to demographic or clinical factors. Finally, the lack of data regarding hepatitis B diagnosis is a gap in the data, and incorporation of notifiable diseases data into the PLIDA environment should be a priority [36] to address this component of the cascade and to further validate modelled estimates of prevalence.

This analysis emphasises the challenges in reaching global targets for guideline‐based care for hepatitis B [37]. We have shown that there is considerable need for increased access to diagnosis, ongoing engagement in regular monitoring, and support for treatment adherence and continuation to reduce the increasing burden of adverse outcomes attributable to unmanaged hepatitis B in Australia and globally.

Funding

This work was supported by the Australian Government Department of Health, Disability and Ageing (4‐E0CQDHO).

Ethics Statement

This research was conducted in accordance with the Declaration of Helsinki. Human Research Ethics Committee approval was obtained from the Royal Melbourne Hospital (reference HREC/77929/MH‐2022), including a waiver of individual consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: List of Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Scheme (PBS) item codes used.

Data Availability Statement

The data that support the findings of this study were accessed with permission from the Australian Bureau of Statistics, via a secure repository with restricted access. Data are not publicly available and are not able to be provided by the authors due to restrictions applied by the data custodians to protect confidentiality.