Volume 26, Issue 7 p. 2787-2795
ORIGINAL ARTICLE
Open Access

Geographic variation in sodium-glucose cotransporter 2 inhibitor and glucagon-like peptide-1 receptor agonist use in people with type 2 diabetes in New South Wales, Australia

Juliana de Oliveira Costa PhD

Juliana de Oliveira Costa PhD

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Jialing Lin PhD

Jialing Lin PhD

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Tamara Y. Milder FRACP

Tamara Y. Milder FRACP

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

Department of Diabetes and Endocrinology, St. Vincent's Hospital, Sydney, New South Wales, Australia

Department of Clinical Pharmacology and Toxicology, St. Vincent's Hospital, Sydney, New South Wales, Australia

Clinical Diabetes, Appetite and Metabolism Laboratory, Garvan Institute of Medical Research, Sydney, New South Wales, Australia

School of Clinical Medicine, UNSW Medicine & Health, St Vincent's Healthcare Clinical Campus, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Jerry R. Greenfield FRACP

Jerry R. Greenfield FRACP

Department of Diabetes and Endocrinology, St. Vincent's Hospital, Sydney, New South Wales, Australia

Clinical Diabetes, Appetite and Metabolism Laboratory, Garvan Institute of Medical Research, Sydney, New South Wales, Australia

School of Clinical Medicine, UNSW Medicine & Health, St Vincent's Healthcare Clinical Campus, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Richard O. Day FRACP

Richard O. Day FRACP

Department of Clinical Pharmacology and Toxicology, St. Vincent's Hospital, Sydney, New South Wales, Australia

School of Clinical Medicine, UNSW Medicine & Health, St Vincent's Healthcare Clinical Campus, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Sophie L. Stocker PhD

Sophie L. Stocker PhD

Department of Clinical Pharmacology and Toxicology, St. Vincent's Hospital, Sydney, New South Wales, Australia

School of Pharmacy, Faculty of Medicine and Health, University of Sydney, Sydney, New South Wales, Australia

Search for more papers by this author
Brendon L. Neuen PhD

Brendon L. Neuen PhD

The George Institute for Global Health, University of New South Wales, Sydney, New South Wales, Australia

Department of Renal Medicine, Royal North Shore Hospital, Sydney, New South Wales, Australia

Search for more papers by this author
Alys Havard PhD

Alys Havard PhD

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

National Drug and Alcohol Research Centre, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Sallie-Anne Pearson PhD

Sallie-Anne Pearson PhD

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

Search for more papers by this author
Michael O. Falster PhD

Corresponding Author

Michael O. Falster PhD

Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia

Correspondence

Michael O. Falster, Medicines Intelligence Research Program, School of Population Health, Faculty of Medicine and Health, Room 219, Samuels Building (F25), University of New South Wales, Sydney NSW 2052, Australia.

Email: m.falster@unsw.edu.au

Search for more papers by this author
First published: 15 April 2024

Juliana de Oliveira Costa and Jialing Lin are Joint first authors.

Sallie-Anne Pearson and Michael O. Falster are Joint senior authors.

Abstract

Aim

Sodium-glucose cotransporter 2 inhibitors (SGLT2is) and glucagon-like peptide-1 receptor agonists (GLP-1RAs) improve glycaemic control and cardio-renal outcomes for people with type 2 diabetes (T2D). However, geographic and socio-economic variation in use is not well understood.

Methods

We identified 367 829 New South Wales residents aged ≥40 years who dispensed metformin in 2020 as a proxy for T2D. We estimated the prevalence of use of other glucose-lowering medicines among people with T2D and the prevalence of SGLT2i and GLP-1RA use among people using concomitant T2D therapy (i.e. metformin + another glucose-lowering medicine). We measured the prevalence by small-level geography, stratified by age group, and characterized by remoteness and socio-economic status.

Results

The prevalence of SGLT2i (29.7%) and GLP-1RA (8.3%) use in people with T2D aged 40-64 increased with geographic remoteness and in areas of greater socio-economic disadvantage, similar to other glucose-lowering medicines. The prevalence of SGLT2i (55.4%) and GLP-1RA (15.4%) among people using concomitant T2D therapy varied across geographic areas, with lower SGLT2i use in more disadvantaged areas and localized areas of high GLP-1RA use (2.5 times the median). Compared with people aged 40-64 years, the prevalence of SGLT2i and GLP-1RA use was lower in older age groups, but with similar patterns of variation across geographic areas.

Conclusions

The prevalence of SGLT2i and GLP-1RA use varied by geography, probably reflecting a combination of system- and prescriber-level factors. Socio-economic variation in GLP-1RA use was overshadowed by localized patterns of prescribing. Continued monitoring of variation can help shape interventions to optimize use among people who would benefit the most.

1 INTRODUCTION

The management of type 2 diabetes (T2D) centres on glycaemic control and the use of specific disease-modifying therapies to prevent or delay disease progression and the development of comorbidities. This is achieved through comprehensive lifestyle modification to address risk factors as well as pharmacotherapy. Metformin is the historically recommended first-line pharmacotherapy1, 2; however, many people require add-on therapy to maintain glycaemic targets.3 Sodium-glucose cotransporter 2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP-1RA) are preferentially recommended as add-on therapy because of their substantial cardio-renal benefits for people with T2D. The SGLT2is reduce cardiovascular death or heart failure-related hospitalization by 23% and kidney disease progression by over 37%,4, 5 while the GLP-1RA reduces major cardiovascular events by up to 14% and kidney disease by 21%.6

Despite their proven efficacy in reducing cardiovascular and kidney disease-related morbidity and mortality, SGLT2is and GLP-1RAs are yet to be widely adopted by people with T2D. International studies have found that, despite modest increases in their use over time, the majority of people who might benefit are not receiving one.7-9 Recent qualitative studies suggest clinicians are hesitating in prescribing these newer therapies for people with T2D because of safety concerns, a limited understanding of cardio-renal benefits, and a lack of confidence in prescribing across different medical specialties.10, 11 Several jurisdictions showed increases in SGLT2i and GLP-1RA use with corresponding decreases in the use of sulphonylureas and dipeptidyl peptidase-4 inhibitors (DPP-4i),12-17 suggesting that changes in prescribing preferences for add-on therapy are occurring, although these variations are not well understood.

International studies have indicated there are substantial disparities in the uptake of SGLT2is and GLP-1RAs. While older people are often at higher cardiovascular risk, they also have lower use, potentially reflecting concerns about frailty and increased rates of adverse effects. US studies have found lower use among people of lower socio-economic status, as well as among non-white people, suggesting potential barriers to access and affordability of these medicines, despite their higher clinical need in these populations.14, 18 Notable disparities also exist at the health service level, highlighting the multifaceted challenges of evidence implementation.19 Previous research has identified pronounced variation in the early uptake of SGLT2is and GLP-1RAs among people living in remote and socio-economically disadvantaged areas of Australia.16 However, health systems differ within broad geographic areas, and little is known about the variation in uptake of SGLT2is and GLP-1RAs across smaller, more granular regions. Understanding such variations is critical to identifying potential barriers to uptake as well as health care planning and implementing targeted prescriber and patient education programmes to improve individual and population outcomes.

We therefore examined the geographic variation in the prevalence of SGLT2i and GLP-1RA use across Australia's most populous state and how this varied across regions with differing socio-economic status and remoteness from services. We estimated the prevalence of specific glucose-lowering medicines among people with T2D and the prevalence of SGLT2i and GLP-1RA use among people using metformin and at least one other concomitant glucose-lowering medicine.

2 METHODS

2.1 Setting and data sources

In Australia, all citizens and permanent residents are Medicare-eligible, entitling them to publicly funded universal health care and subsidized prescription medicines via the Pharmaceutical Benefits Scheme (PBS).

We used data from the Medicine Intelligence Data Platform for people aged ≥18 years who resided in New South Wales (NSW), the most populous state of Australia with a population of approximately 8 million people. The platform includes PBS dispensing claims with detailed geography information linked to other administrative datasets, such as Medicare enrolment, Medicare services provided in the community, hospitalizations and mortality data. The Australian Institute of Health and Welfare and the NSW Centre for Health Record Linkage performed the data linkage.

We used PBS data to identify dispensings for the medicines of interest, Medicare enrolment data to ascertain participants' year and month of birth, sex, residential postcode, Medicare service, hospitalization and mortality data to exclude people with possible linkage errors.

2.2 Medicines of interest

We identified the dispensing of PBS-listed glucose-lowering medicines, including SGLT2is, GLP-1RAs, metformin, sulphonylureas, DPP-4is and insulin. We derived dispensing dates and PBS item numbers from the PBS data and mapped PBS item numbers to World Health Organization Anatomical Therapeutic Chemical (ATC) codes (Table S1). During the study period, SGLT2is and GLP-1RAs were listed on the PBS as second- or third-line therapy in combination with specific glucose-lowering medicines among people with glycosylated haemoglobin ≥7% (Table S2).

2.3 Study population

We included NSW residents alive and aged ≥40 years in 2020, with at least one metformin dispensing in 2020. We excluded people with data inconsistencies indicating potential linkage errors (e.g. hospitalization records or service-related claims occurring after the date of death) and people without information on residential postcodes.

We included people with at least one metformin dispensing in 2020 as a proxy for T2D, given that metformin is the recommended first-line T2D therapy in Australia.2 We restricted this group to people aged ≥40 years, as metformin is also used in younger women to manage gestational diabetes and polycystic ovary syndrome, and our data did not capture the indication of therapy. Hereafter, we refer to people aged ≥40 years who dispensed metformin as people with T2D.

We also identified people with evidence of concomitant T2D therapy (SGLT2i, GLP-1RA, DPP-4i, sulphonylurea or insulin) with metformin as a proxy for second-line T2D therapy because of a lack of glycaemic control.2 At the time of this analysis, neither SGLT2i nor GLP-1RA were subsidized as first-line T2D therapy in the PBS.2 We defined concomitant T2D therapy as the use of metformin and another glucose-lowering medicine in 2020. To do that, we first estimated the length of metformin treatment episodes based on the date of dispensing plus the 75th centile of time between metformin dispensings for the population.20 We then assessed if the date of dispensing of another glucose-lowering medicine overlapped with metformin episodes, determining this to be concomitant use.

2.4 Geographic areas

We used Statistical Area Level 2 (SA2), the standard for measuring small-level geography in Australia as our unit of analysis,21 representing communities interacting together socially and economically. We mapped individuals’ postcodes of residence to SA2s.21 For people with multiple records in 2020, we assigned the most recent postcode. We also excluded records where the postcode indicated the address was a post office box. Some residential postcodes have multiple corresponding SA2s, so we estimated a weighted population for each SA2 based on the proportion of people mapped from each postcode.

We classified each SA2 by geographic remoteness using the Accessibility/Remoteness Index for Australia (ARIA).22 ARIA is a measure of relative geographic access to services, categorized into major cities, inner regional, outer regional and remote/very remote areas. We obtained information on socio-economic status using the Socio-Economic Indexes for Areas (SEIFA) Index of Relative Socio-economic Disadvantage (IRSD).23 SEIFA IRSD is a general measure of relative socio-economic disadvantage at the area level, and it is divided into quintiles, where quintile 1 (Q1) indicates the most disadvantaged areas and quintile 5 (Q5) indicates the least disadvantaged areas.

2.5 Outcomes and statistical analyses

For our first aim, we calculated the prevalence of SGLT2i, GLP-1RA, sulphonylureas, DPP-4i and insulin use among people with T2D, expressing the results as percentages. We considered a person to be using these medicines as prevalent use if these medicines had been dispensed at any time in 2020. People receiving multiple medicines within 2020 contributed to all relevant groups. We then presented the prevalence of specific glucose-lowering medicine use in people with T2D by remoteness and socio-economic status.

For our second aim, we focused specifically on the use of SGLT2i and GLP-1RA among people with an indication for second-line therapy. We calculated the prevalence of SGLT2i and GLP-1RA use among people with concomitant T2D therapy and presented the results by remoteness and socio-economic status. We further examined the prevalence of SGLT2i and GLP-1RA use by small area geography, producing geographic maps of the prevalence of their use.24 We also visualized the distribution of prevalence within each small-level geographic area by the socio-economic status of the area within different remoteness categories (major cities, inner regional, outer regional, remote/very remote). We then summarized the prevalence within each of these strata and tested for differences across socio-economic status categories using χ2 tests.

Age is an important risk factor for T2D severity, and different factors are associated with glucose-lowering medicine use at different ages (e.g. frailty, use of multiple other medicines). We therefore stratified all results according to three age groups: people aged 40-64, 65-74 and ≥75 years old.

All data analyses were completed using R version 4.3.1 (R Core Team 2017).

3 RESULTS

We identified 367 829 NSW residents with T2D, aged ≥40 years and alive in 2020. Of these, 162 904 (44.3%) were aged 40-64 years, 113 775 were aged 65-74 years (30.9%) and 91 150 (24.8%) were aged ≥75 years. The majority of people with T2D lived in major cities or inner regional areas, with a slightly higher proportion living in areas of higher socio-economic disadvantage (Table 1).

TABLE 1. Characteristics of the study population by age group, socio-economic status and geographic remoteness of their area of residence in New South Wales, Australia, 2020.
Characteristic People aged 40-64 years People aged 65-74 years People aged ≥75 years
With T2Da With concomitant T2D therapyb With T2Da With concomitant T2D therapyb With T2Da With concomitant T2D therapyb
N (%) N (%) N (%) N (%) N (%) N (%)
Number of people (%) 162 904 (100) 87 268 (100) 113 775 (100) 67 323 (100) 91 150 (100) 50 785 (100)
Geographic remotenessc
Major cities 125 675 (77.1) 65 906 (75.5) 82 658 (72.7) 48 507 (72.1) 66 741 (73.2) 37 173 (73.2)
Inner regional 27 722 (17.0) 15 584 (17.9) 23 704 (20.8) 14 189 (21.1) 18 732 (20.6) 10 393 (20.5)
Outer regional 8651 (5.3) 5234 (6.0) 6916 (6.1) 4306 (6.4) 5341 (5.9) 3041 (6.0)
Remote/very remote 856 (0.5) 544 (0.6) 497 (0.4) 321 (0.5) 336 (0.4) 178 (0.4)
Socio-economic statusd
Q5 (least disadvantaged) 23 527 (14.4) 10 269 (11.8) 17 008 (14.9) 8922 (13.3) 15 354 (16.8) 7856 (15.5)
Q4 22 955 (14.1) 11 519 (13.2) 14 949 (13.1) 8581 (12.7) 12 402 (13.6) 6715 (13.2)
Q3 31 669 (19.4) 16 649 (19.1) 21 892 (19.2) 12 882 (19.1) 18 389 (20.2) 10 276 (20.2)
Q2 40 514 (24.9) 22 574 (25.9) 30 634 (26.9) 18 545 (27.5) 23 261 (25.5) 13 143 (25.9)
Q1 (most disadvantaged) 44 239 (27.2) 26 257 (30.1) 29 292 (25.7) 18 393 (27.3) 21 744 (23.9) 12 795 (25.2)
  • a Defined as people treated with metformin.
  • b Defined as people treated with either sodium-glucose cotransporter 2 inhibitors, glucagon-like peptide-1 receptor agonists, sulphonylureas, dipeptidyl peptidase-4 inhibitors or insulin, while using metformin.
  • c Accessibility/Remoteness Index for Australia.
  • d Index of Relative Socio-economic Disadvantage quintiles.
  • Abbreviation: T2D, type 2 diabetes.

Approximately half (n = 205 376, 55.8%) of people with T2D had concomitant use of metformin with another glucose-lowering medicine. Compared with people with T2D, a slightly higher proportion of people who received concomitant T2D therapy were living in regional and remote areas, as well as in areas of higher socio-economic disadvantage. This pattern was broadly consistent between age groups (Table 1).

We present all subsequent results for people aged 40-64, unless otherwise specified, given that this age group represented the largest population of our study, had greater potential for benefits of long-term cardiovascular prevention, and had a higher prevalence of glucose-lowering medicine use.25

3.1 Prevalence of glucose-lowering medicine use in people with type 2 diabetes

The prevalence of SGLT2i and GLP-1RA use among people with T2D was 30.1% and 8.6%, respectively. The prevalence of these medicines tended to be higher in more disadvantaged areas and in regional and remote areas. This pattern was similar for DPP-4i, sulphonylureas and insulin (Figure 1).

Details are in the caption following the image
Prevalence of glucose-lowering medicines among people aged 40-64 years with type 2 diabetes, by remoteness and socio-economic status. People with type 2 diabetes defined as people treated with metformin. ARIA, Accessibility/Remoteness Index for Australia; DPP-4i, dipeptidyl peptidase-4 inhibitors; GLP-1RA, glucagon-like peptide-1 receptor agonist; IRSD, Index of Relative Socio-economic Disadvantage; Q, quintile; SGLT2i, sodium-glucose cotransporter 2 inhibitor.

3.2 Prevalence of sodium-glucose cotransporter 2 inhibitor and glucagon-like peptide-1 receptor agonist use in people with concomitant type 2 diabetes therapy

Among people with concomitant T2D therapy, 55.4% were using SGLT2i. This varied across geographic areas, with a median of 56.1% (interquartile range 51.4%-59.4%) (Figure 2A). We found an overall prevalence of GLP-1RA use of 15.4%, with a median of 15.0% (interquartile range 12.1%-19.0%) across geographic areas. We also observed areas in the north-east of NSW where the prevalence of GLP-1RA use was atypically high. This included the majority of SA2s with a prevalence between 30% and 39% and all three SA2s with a prevalence of ≥40%, up to 2.5 times higher than the median (Figure 2B).

Details are in the caption following the image
Prevalence of (A) sodium-glucose cotransporter 2 inhibitor or (B) glucagon-like peptide-1 receptor agonist use in people aged 40-64 years with concomitant type 2 diabetes therapy, by small-level geography in New South Wales. People with concomitant type 2 diabetes therapy defined as those treated with either sodium-glucose cotransporter 2 inhibitors, glucagon-like peptide-1 receptor agonists, sulphonylureas, dipeptidyl peptidase-4 inhibitors or insulin, while using metformin. Small-level geography measured using Statistical Area Level 2. Numbers less than 6 were not shown because of data confidentiality.

When stratifying by remoteness and socio-economic status, we observed a similar overall prevalence of SGLT2i use between categories of geographic remoteness (between 55.2% and 55.8%; Table S2). Within each remoteness category, we found the prevalence varied across socio-economic strata (Figure 3, Table S2), tending to be lower in more disadvantaged areas. This was most pronounced in inner regional areas (p = .054), where use in the most disadvantaged socio-economic stratum was 4.4% points lower than the highest socio-economic stratum, followed by outer regional areas (p = .007; 3.1% difference) and major cities (p < .01; 2.5% difference).

Details are in the caption following the image
Prevalence of (A) sodium-glucose cotransporter 2 inhibitor or (B) glucagon-like peptide-1 receptor agonist use in people aged 40-64 years with concomitant type 2 diabetes therapy, across small-level geographic areas, stratified by remoteness category and socio-economic status of the area. People with concomitant type 2 diabetes therapy defined as those treated with sodium-glucose cotransporter 2 inhibitor, glucagon-like peptide-1 receptor agonists, sulphonylureas, dipeptidyl peptidase-4 inhibitors or insulin while using metformin. Each dot represented one small-level geography area (Statistical Area Level 2) and those with numbers less than 6 were not shown because of data confidentiality. IRSD, Index of Relative Socio-economic Disadvantage; Q, quintile.

The prevalence of GLP-1RA use differed across categories of geographic remoteness, from 14.0% in major cities to 23.3% in outer regional areas (Table S2). Within each remoteness category, we found the prevalence significantly differed across categories of socio-economic status (particularly for inner and outer regional areas), but we did not see clear trends in prevalence across categories of socio-economic disadvantage (Figure 3, Table S3).

3.3 Variation in prevalence among other age groups

The prevalence of both SGLT2i and GLP-1RA use was lower in people with T2D aged 65-74 years (29.0% and 7.2%, respectively) and lowest in people aged ≥75 years (17.2% and 3.4%, respectively), compared with people aged 40-64 years (Figures S1 and S2).

Similarly, when examining prevalence among people with concomitant T2D therapy, we found a lower prevalence of both SGLT2is and GLP-1RAs in people aged 65-74 years (48.4% and 11.9%, respectively) and the lowest in those aged ≥75 years (30.3% and 5.9%, respectively), compared with people aged 40-64 years. Despite this, we observed similar patterns of geographic variation in prevalence of SGLT2i and GLP-1RA use across all age groups: there was a localized area of higher prevalence of GLP-1RA use in north-east NSW, with no such pattern for SGLT2i (Figures S3 and S4).

When stratifying by remoteness and socio-economic status in people with concomitant T2D therapy (Figures S5 and S6, Tables S4 and S5), in people aged 65-74 years, we observed a similar prevalence of SGLT2i use across remoteness categories, with decreasing prevalence in areas with more socio-economic disadvantage (similar to that for people aged 40-64 years). This was less pronounced among people aged ≥75 years. For GLP-1RA, we found similar patterns in people aged 65-74 and ≥75 years, as with people aged 40-64 years. That is, higher prevalence within inner and outer regional areas than major cities, but no clear pattern by socio-economic status.

4 DISCUSSION

The pharmacological management of T2D has observed a paradigm shift with the emergence of SGLT2i and GLP-1RA as treatments for glycaemic control and improved cardio-renal outcomes. In our population-based study of people with T2D in Australia's most populous state, we found about half of people receiving concomitant T2D medicines were using an SGLT2i, and about one in seven people a GLP-1RA. This prevalent use varied across geographic areas, particularly with somewhat lower use of SGLT2i in areas of greater socio-economic disadvantage, as well as a local region with atypically high GLP-1RA use. Together, these findings suggest strategies at both the system (e.g. prescribing restrictions, cost) and prescriber (e.g. education, integrated care) levels may be required to increase use of these medicines. Importantly, they also show that variation in medicine use by socio-economic status cannot be interpreted without considering the influence of localized policies and prescribing practices. As the emerging clinical trial evidence and recommendations from major clinical guidelines continue to shift to a broader indication beyond glycaemic control,26 continued monitoring of geographic variations may ensure use is optimized among people who would benefit the most.2, 27, 28

Similar to other Australian studies,16, 29 we found prevalent use of both SGLT2i and GLP-1RA in people with T2D was higher in regional and remote areas relative to major cities and areas of greater socio-economic disadvantage relative to more advantaged areas. These results possibly reflect a higher need for therapy in these areas, given that the prevalence of T2D is 1.3-1.9 times higher in remote areas and socio-economically disadvantaged groups than in the general population.30 This is supported by the fact that we observed similar patterns of prevalence by remoteness and socio-economic status across all glucose-lowering medicines (as well as other alternatives to add-on therapy) examined.

Encouragingly, the prevalence of SGLT2i and GLP-1RA use was higher among people on concomitant T2D therapy (i.e. those receiving at least one second-line therapy), with more than half (55%) of people aged 40-64 years receiving an SGLT2i and one in seven (15%) receiving a GLP-1RA. However, we also observed lower SGLT2i use in more disadvantaged areas. While the extent of these differences was not as great as disparities observed in international studies, such as by race or insurance coverage in the United States,14, 18 they do highlight that socio-economic barriers in the use may remain even in a jurisdiction with universal health care. A previous nationwide Australian study similarly showed people living in disadvantaged areas were more likely to receive older therapies such as sulphonylureas and DDP-4i as second-line therapy/add-on therapy, although the extent of these differences decreased over time.16 Our study further showed that these socio-economic differences occur within both major cities and regional areas and across age groups.

The cost of therapy may contribute to disparities by socio-economic status observed in other jurisdictions.18, 31 Recent Australian research did not specifically identify the cost of therapy as a barrier to prescribing10; the monthly cost of therapy for GLP-1RA or SGLT2i for people receiving government benefits (i.e. pensioners and low-income earners) was AUD $6.60 in 202032 and AUD $41 for other Australians (general beneficiaries).33 From January 2023, PBS copayments were reduced by 30% for general beneficiaries,32 meaning their monthly cost is closer to that of other add-on therapies. Broadening the PBS prescribing restrictions for SGLT2i and GLP-1RA to align more closely with the most recent clinical evidence and international guideline recommendations around cardiovascular prevention will also increase uptake. It will be important to monitor the impact of this co-payment policy change and other potential listing changes on access to these medicines.

Ongoing geographic variation analyses may also support targeted local strategies to increase cardiometabolic medicine use. In most jurisdictions, glucose-lowering agents are predominantly prescribed by primary care clinicians12, 34; the pharmacological management of T2D has been a priority educational need for general practitioners across Australian regions,10, 35, 36 including integrating care between hospital specialists and primary care teams.37 Indeed, we observed a higher prevalence of GLP-1RA use in a local health district situated in north-east NSW and its nearby areas. This region has capacity-enhancing activities for primary care clinicians by endocrinologists,38 although our study is not designed to evaluate the impact of this activity. While the prevalence of risk factors, such as obesity and heart failure, is higher in this region than in urban Sydney, it is not notably higher than in other surrounding regions in outer and regional NSW.39, 40 Ongoing monitoring of geographic variation could be used for monitoring uptake of cardiometabolic medicines, to support targeted local strategies to increase use, and to ensure successful health policies can be reviewed for implementation in other local health areas.

Together, our study shows the myriad of factors underpinning socio-economic variation in health care. We found the overall prevalence was probably driven by clinical complexity and the need for add-on therapy. We found socio-economic gradients in SGLT2i use existed within both urban and regional areas, probably influenced by cost and restrictions on prescribing. Importantly, we found that socio-economic differences in GLP-1RA use were clearly overshadowed by local patterns of prescribing. This latter finding would be all but invisible without exploring variation by small-level geography, and it suggests we cannot interpret socio-economic variation in medicine use without considering the influence of localized prescribing practices. Further research unpacking the relative contributions of people's health, socio-demographic factors, access to services and treatment patterns of service providers is needed to fully understand and take action on the inequities in the use of these important medicines.

Our study has some limitations. First, as people with prediabetes and polycystic ovary syndrome may also be treated with metformin, there is potential misclassification associated with using metformin dispensing data to identify people with T2D. Similarly, we may underestimate the prevalence of T2D where pharmacotherapy is not used or metformin is contraindicated, such as in people with impaired renal function. Second, PBS data do not capture the indication of therapy, so we assumed SGLT2is and GLP-1RAs were prescribed for T2D. This is a reasonable assumption given at the time of our study, SGLT2is were not PBS-listed for other conditions, and we only included people who were dispensed metformin. Third, PBS data do not capture private dispensings and those within public hospitals, possibly resulting in an underestimated prevalence of glucose-lowering medicines.41 Fourth, our study was cross-sectional in design, meaning we did not assess a range of clinical characteristics (e.g. glycated haemoglobin, body mass index, specific comorbidities) relating to prescribing appropriateness nor the pathway of use of SGLT2i and GLP-1RA as potential add-on therapies. However, the pattern of utilization we observed among people with concomitant T2D therapy differs from the known distribution of clinical risk factors, which are typically more common in remote areas and areas of greater socio-economic disadvantage.30, 40 Finally, our analyses were at the geographic level, and such an area-level assessment of socio-economic disadvantage may not capture the types of barriers experienced at the individual level.

We found that about half of the people receiving concomitant T2D medicines were using an SGLT2i, and about one in seven people a GLP-1RA. This use varied across geographic areas, probably reflecting a combination of both system- and prescriber-level factors. Broader socio-economic inequities in use, as seen internationally, point towards potential issues in affordability and treatment of people with more complex risk factors and disease. Many system- and prescriber-level factors, conversely, are more amenable to local policy intervention, such as targeted clinical capacity-enhancing programmes, supply chains for medicines with probable shortages, as well as models of integrated care (e.g. between general practitioners, cardiologists and endocrinologists). As strong clinical evidence for the use of SGLT2i and GLP-1RA as preventive cardio-renal medicines continues to shape international guidelines and therapeutic indications, ongoing monitoring of geographic variation may help ensure use is optimized among people who would most benefit from these medicines.

AUTHOR CONTRIBUTIONS

JDOC, JL and MOF wrote the first draft of the manuscript, and all authors edited, reviewed, and approved the final version of the manuscript. All authors were involved in the conception, design and conduct of the study. JL conducted the analysis under supervision from JDOC and MOF. All authors were involved in the interpretation of the results. JDOC, JL and MOF is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis.

ACKNOWLEDGMENTS

This research is supported by the University of New South Wales, Cardiac, Vascular and Metabolic Medicine (CVMM) Theme Collaborative, the National Health and Medical Research Council (NHMRC) Ideas Grants (grant numbers 2002889, 1183273) and National Health and Medical Research Council (NHMRC) Medicines Intelligence Centre of Research Excellence (grant number 1196900). BLN is supported by an National Health and Medical Research Council (NHMRC) Emerging Leader Investigator Grant (grant number 2026621) and a Ramaciotti Foundation (grant number 2023HIG69). MOF is supported by a National Heart Foundation of Australia Future Leader Fellowship (grant number 105609). We acknowledge Melisa Litchfield for her assistance with project data and governance. This research was completed using the Medicines Intelligence Data Platform. Data were provided by the Australian Government Department of Health and Aged Care and NSW Ministry of Health. Record linkage was conducted by the Australian Institute of Health and Welfare (AIHW) and Centre for Health Record Linkage (CHeReL). Secure data access was provided through the Sax Institute's Secure Unified Research Environment (SURE). Open access publishing facilitated by University of New South Wales, as part of the Wiley - University of New South Wales agreement via the Council of Australian University Librarians.

    CONFLICT OF INTEREST STATEMENT

    SAP is a member of the Drug Utilization Sub Committee of the Pharmaceutical Benefits Advisory Committee. The views expressed in this article do not represent those of the Committee. BLN has received fees for travel support, advisory boards, scientific presentations and steering committee roles from AstraZeneca, Alexion, Bayer, Boehringer and Ingelheim, Cambridge Healthcare Research, Cornerstone Medical Education, Janssen, the Limbic and Medscape, with all honoraria paid to The George Institute for Global Health. The other authors have no conflicts of interest to declare.

    PEER REVIEW

    The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.15597.

    DATA AVAILABILITY STATEMENT

    The data were provided by the Australian and NSW Governments. Access to these data by other individuals or authorities are not permitted without the express permission of the approving human research ethics committees and data custodians.