pharmacoepidemiology and drug safety 2015; 24: 447–455 Published online 31 March 2015 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/pds.3756

REVIEW

Studies using Australia’s Pharmaceutical Benefits Scheme data for pharmacoepidemiological research: a systematic review of the published literature (1987–2013)† Sallie-Anne Pearson1,2*, Nicole Pesa1, Julia M. Langton1, Annabelle Drew3, Margaret Faedo4 and Jane Robertson5 1

Faculty of Pharmacy, University of Sydney, Sydney, Australia School of Public Health, Sydney Medical School, University of Sydney, Sydney, Australia 3 NSW Therapeutic Advisory Group, Sydney, Australia 4 Research Integrity and Ethics Administration, University of Sydney, Sydney, Australia 5 School of Medicine and Public Health, Faculty of Health, University of Newcastle, Newcastle, Australia 2

ABSTRACT Purpose Research using dispensing claims is used increasingly to study post-market medicines use and outcomes. The purpose of this review is to catalogue more than 25 years of published literature using Australia’s Pharmaceutical Benefits Scheme (PBS) dispensing records. Methods We searched MEDLINE, PreMEDLINE and Embase and conducted author searches for studies published from 1987 to 2013. Independent reviewers screened abstracts of 3209 articles and reviewed 264 full-text manuscripts. Included studies used PBS dispensing data to measure patterns and/or outcomes of prescribed medicines use or dispensing claims to derive a proxy for a specific disease cohort or health outcome. Results Of the 228 studies identified, 106 used PBS claims only (56 using claims-level data and 50 using individual-level data) and 63 studies linked individual-level PBS claims to other health data. Most commonly, studies examined trends in drug utilisation (33%), clinician and patient practices (26%), drug use and outcomes (18%) and evaluations of intervention impacts (17%). Sixty-two percent of studies using individual-level data were based on a subset of elderly Australians. Most studies focused on drug classes acting on the nervous system (36%), cardiovascular system (15%) and alimentary tract (11%). Few studies examined prescribed medicines use in children and pregnant women. Conclusions Pharmaceutical Benefits Scheme claims represent a significant resource to examine Australia’s billion-dollar annual investment in prescribed medicines. The body of research is growing and has increased in complexity over time. Australia has great potential to undertake world-class, whole-of-population pharmacoepidemiological studies. Recent investment in data linkage infrastructure will significantly enhance these opportunities. Copyright © 2015 John Wiley & Sons, Ltd. key words—pharmacoepidemiology; observational studies; drug utilization; prescription databases; record-linkage Received 2 November 2014; Revised 16 December 2014; Accepted 19 December 2014

INTRODUCTION Pharmacoepidemiological research has grown significantly in recent decades.1,2 The growth has been driven by recognition of the limitations of pre-market medicines evaluation, the utility of large routinely collected *Correspondence to: Sallie-Anne Pearson, Pharmacy and Bank Building, Faculty of Pharmacy, The University of Sydney, NSW, 2006, Australia. E-mail: [email protected] † An early iteration of this review was presented as an invited plenary at an International Society of Pharmacoepidemiology meeting (Pearson S. Application of Pharmacoepidemiology Research for Australian Policy Making. 4th Asian Conference on Pharmacoepidemiology, 2009; Taiwan).

Copyright © 2015 John Wiley & Sons, Ltd.

linked datasets to explore drug use and outcomes in routine clinical care and improved processes and governance around data access. Many jurisdictions globally are using administrative health data to undertake observational studies to improve quality use of medicines and inform clinical and policy decision-making.3–6 In the last decade, the Australian government has invested between $4bn and $9bn annually in prescribed medicines subsidies under its national formulary, the Pharmaceutical Benefits Scheme (PBS).7,8 Australia is recognised as a pioneer of cost-effectiveness assessment for pre-subsidy decision-making; it was the first country to introduce assessment of value as a prerequisite for

448

s.-a. pearson et al.

new prescribed medicines for public subsidy.9 This approach is overseen by the Pharmaceutical Benefits Advisory Committee and has been described as a system of purchasing health outcomes rather than drugs.10 As a consequence of its universal health care arrangements, Australia is replete with routine health data to explore prescribed medicines use and outcomes. Australian citizens, permanent residents and clients of the Department of Veterans’ Affairs (DVA) access subsidised prescribed medicines under the PBS and the Repatriation Pharmaceutical Benefits Scheme (RPBS), respectively. The Drug-Utilisation Sub-Committee (DUSC) of the Pharmaceutical Benefits Advisory Committee was established in 1989 to monitor medicines use post-subsidy (particularly in the first 2 years of listing) and to address specific issues related to quality use of medicines.11 DUSC maintains a database estimating total community use of PBS and RPBS prescribed medicines and estimates of nonsubsidised prescriptions.12 The DUSC database is used for routine monitoring activities and forms the basis of its annual publication, the Australian Statistics on Medicines.13 The overall objective of this review is to catalogue the pharmacoepidemiological studies using Australian PBS and RPBS claims data published in the last 25 years. Specifically, we review the study populations, methodological approaches (including data sources and linkage of dispensing claims with other routine data collections), therapeutic focus and main outcomes of interest. Despite the wealth of routinely collected medicines data in Australia, there has not been a comprehensive synthesis of pharmacoepidemiological research to date. A synthesis of the existing outputs will assist in identifying clinical and policy needs and data and linkage priorities. METHODS Data of interest Our review focuses on studies using PBS and RPBS dispensing claims that are processed by the Department of Human Services (DHS; previously Medicare Australia and the Health Insurance Commission). The dispensing claims are provided to the Commonwealth Department of Health and the DVA (RPBS only) for monitoring, evaluation and health service planning. External bodies may apply to any one of these agencies to use the data for research purposes. Until recently, DHS only recorded dispensing claims submitted for the payment of a subsidy under the PBS and RPBS. As such, medicines costing less than the patient co-payment threshold were not ascertained in the Copyright © 2015 John Wiley & Sons, Ltd.

collection. In effect, low-cost medicines dispensed to beneficiaries with the highest patient co-payment threshold (referred to as general beneficiaries) have been under-ascertained; this issue does not have an impact on medicines dispensed to beneficiaries with lower co-payment thresholds (PBS concessional beneficiaries or DVA clients).14,15 In April 2012, DHS began recording below co-payment prescriptions.16 Prior to 2012, DUSC recorded estimates of under copayment and private prescriptions ascertained by an ongoing Pharmacy Guild survey of a representative sample of community pharmacies.12 Combined with RPBS dispensing, this data represented total community use of prescribed medicines in Australia. Study identification We conducted a review of published studies using dispensing claims in the Australian setting. We searched MEDLINE, PreMEDLINE and Embase from January 1987 through December 2013. We combined search terms describing medicines use (e.g. prescription drugs, drug therapy and drug utilisation) with data sources of interest (e.g. Department of Veterans’ Affairs and Pharmaceutical Benefits Scheme, Australia) (Appendix A). Given the heterogeneity of studies and the broad range of key terms likely to be attributed to them, we also conducted searches on key researchers in the field of pharmacoepidemiology in Australia and screened the reference lists of all included studies (Figure 1). Studies eligible for review We included English-language research articles and letters to editors including original data that met the following criteria: • The study used PBS or RPBS claims data or the DUSC database; and • The study used dispensing data to do the following:

○ Measure patterns of use (that may be linked to health outcomes). ○ Derive a disease cohort (patient dispensed medicine/s used as a proxy for a specific disease). ○ Derive a specific health outcome (patient dispensed medicine/s used as a proxy for a specific disease state or symptoms). We excluded the following: • Studies focusing exclusively on drug expenditure or modelling. • Studies using dispensing data obtained directly from pharmacies. Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

australian pharmacoepidemiological research

Figure 1.

• • •

449

Identification of studies included in the review

Cohort studies requiring informed consent to link dispensing data to other datasets such as self-report or survey data. Studies using data derived from state-based registries (e.g. authorisations for drugs of dependence from state health departments). Grey literature such as reports published by the Australian Department of Health and Departments of Human Services and Veterans’ Affairs.

To identify potentially relevant studies for inclusion in the review, pairs of reviewers also independently extracted data from all included studies. Any disagreements were resolved by discussion. We Copyright © 2015 John Wiley & Sons, Ltd.

extracted key features of each study, including the following:

• •

•

Study characteristics—publication year, journal, study objectives, funding source, objectives and setting. Study period—defined as the longest period of observation. If different datasets were used across different time periods (e.g. PBS from 2000 to 2005 and hospital separations from 2002 to 2007), the broadest dates were deemed to be the study period (i.e. 2000–2007). Publication lag—calculated as year of publication minus last year of study observation. Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

450

• •

s.-a. pearson et al.

Study population—whether studies were carried out within the general population, children, DVA (in particular, noting those with full entitlements) or concessional beneficiaries. Broad study focus—categorised according to six themes:

○ Drug utilisation—classified according to (i)

• •

• •

trends and patterns of dispensing stratified by gender, age and drug or (ii) trends and patterns of dispensing stratified by the above in addition to other variables such as socioeconomic status, geographic location or involving additional data sources (either national or international). ○ Clinician and patient practices—examined individual-level prescriber behaviour such as patterns of prescribing (e.g. concomitant and inappropriate) or patient behaviour around medicines use such as persistence and adherence to medicines. ○ Drug use and outcomes—investigated the relationship between the medicine dispensing and a subsequent event (e.g. death or hospital admission). ○ Intervention impacts—examined the effect of one or more interventions on prescribing or another outcome. Interventions were classified as educational (e.g. prescriber feedback and education), policy (e.g. subsidy changes and restrictions), media (e.g. advertising campaigns) or multi-faceted (a combination of the above domains). ○ Methods—dispensing data used to develop and refine pharmacoepidemiological techniques (e.g. validation of prescribing indicators and evaluation of preference-based instruments). Drug focus—we identified the drugs of interest in each study and assigned the corresponding Anatomical Therapeutic Chemical classifications.17 Data source(s)—primary and any other data sources, including other Commonwealth datasets (e.g. Medicare Benefits Schedule); State datasets (e.g. hospital separations); administrative datasets from other health care jurisdictions (e.g. Nova Scotia Pharmacare Program, Canada); and drug sales (e.g. IMS Health). Analytical approach—individual-level studies track patients and/or providers over time, and claimslevel studies present aggregate or group-level analyses. Primary outcomes—up to three outcomes reported by the authors.

Copyright © 2015 John Wiley & Sons, Ltd.

Reporting Because of the heterogeneous nature of study methodology, we did not assess the quality of individual studies. Instead, we describe key study features and provide a qualitative analysis of common themes across included studies. RESULTS Studies identified A total of 228 studies were included in the review. We screened the titles and abstracts of 3209 articles and reviewed 264 full-text manuscripts; 199 of these were eligible for inclusion. In addition we identified 16 studies through key author searches and 13 through back references of included articles (Figure 1) (see Appendix B for bibliography of included studies and Appendix C for details of individual study features including aims, study population, datasets and main outcomes in the Supporting Information). Overview of included studies Study characteristics: 156 of the 228 included studies (68%) were published since 2007. The publication lag time was 3–5 years for 57% of studies and within 2 years for 33% (Table 1 and Figure 2). Study populations: Most commonly, studies were conducted across the entire PBS-eligible population (128 studies) or DVA clients with full health care entitlements (40 studies); only seven studies were conducted in pregnant women and four in children. Data source(s): 106 (47%) studies used dispensing data exclusively; 56 studies were based on claims-level data and the remaining 50 studies used individual-level data. One Hundred and twenty two studies used additional health datasets, 63 of which linked person-level dispensing claims to other routine collections including hospitalisations, medical service claims, cancer notifications and fact and cause of death data. Study focus (Figure 2): Drug utilisation studies were the most common (33%), followed by studies of clinician or patient practices (26%), drug use and outcomes studies (18%), and evaluations of intervention impacts (17%). Few studies (6%) focused primarily on methodological issues such as validation. Analytical approach. One hundred and fifteen studies used claims-level data, 56% of which were published since 2007. Nearly all studies using claims-level data were conducted in the entire PBS-eligible population, and most studies examined drug utilisation (65% of studies), investigated intervention impacts (19%) or Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

451

australian pharmacoepidemiological research Table 1. Study characteristics All studies (%)

Claims-level analysis (%)

Individual-level analysis (%)

n = 228

n = 115

n = 113

7 (3.1) 15 (6.6) 22 (9.6) 28 (12.3) 114 (50.0) 42 (18.4)

7 (6.1) 12 (10.4) 14 (12.2) 18 (15.7) 46 (40.0) 18 (15.7)

0 (0.0) 3 (2.7) 8 (7.1) 10 (8.8) 68 (60.2) 24 (21.2)

2 (1.7) 53 (46.1) 54 (47.0) 5 (4.3) 1 (0.9)

2 (1.7) 23 (20.0) 75 (65.2) 13 (11.3) 0 (0.0)

128 (56.1) 30 (13.2) 30 (13.2) 40 (17.5) 7 (3.1) 4 (1.8)

100 (87.0) 15 (13.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (2.6)

28 (24.8) 15 (13.3) 30 (26.5) 40 (35.4) 7 (6.2) 1 (0.9)

75 (32.9) 40 (17.5) 47 (20.6) 13 (5.7) 39 (17.1) 14 (6.1)

75 (65.2) 10 (8.7) 0 (0.0) 0 (0.0) 22 (19.1) 8 (7.0)

0 (0.0) 30 (26.5) 47 (41.6) 13 (11.5) 17 (15.0) 6 (5.3)

Dataset Dispensing claims only Dispensing claims and other health data

106 (46.5) 122 (53.5)

56 (48.7) 59 (51.3)

50 (44.2) 63 (55.8)

Journal Australian International

103 (45.2) 125 (54.8)

61 (53.0) 54 (47.0)

42 (37.2) 71 (62.8)

Publication year 1987–1991 1992–1996 1997–2001 2002–2006 2007–2011 2012–2013

Publication lag (time between last observation year and publication year) Same year 4 (1.8) 1–2 years 76 (33.3) 3–5 years 129 (56.6) >5 years 18 (7.9) Not reported 1 (0.4) Study population All beneficiaries Concessional beneficiaries Veterans (all levels of entitlement) Veterans (full entitlements) Pregnant women Children Study focus Drug utilisation Drug use and outcomes Clinician practices Patient practices Intervention impacts Methods

examined drug use and outcomes (9%), all of which used ecological study designs. One hundred and thirteen studies used individuallevel data, 81% of which were published since 2007, and 62% were conducted in the DVA population. Approximately half of the studies using individual-level data examined clinician or patient practices (53%), 27% investigated drug use and outcomes, and 15% evaluated intervention impacts (Table 1 and Figure 3).

Figure 2. approach

Number of studies according to study focus and analytical

Copyright © 2015 John Wiley & Sons, Ltd.

Medicines focus. The most commonly studied drug classes were those acting on the nervous system (36%), followed by cardiovascular medicines (15%), drugs acting on the alimentary tract (11%), anti-infectives (6%) and antineoplastic and immunomodulating agents (6%). Overall, the five most commonly studied drug classes also accounted for the top five PBS-listed drug classes by volume and cost in 2013. Cardiovascular medicines Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

452

s.-a. pearson et al.

Figure 3. approach

Number of publications (cumulative) according to analytical

accounted for the highest PBS volume and expenditure, yet studies examining this drug class are published at half the rate of those acting on the nervous system. Moreover, more than half of the 40 drug use and outcomes studies focused exclusively on medicines acting on the nervous system. Finally, antineoplastic and immunomodulating agents account for 1% of PBS drugs by volume and 13% by cost; however, only a small number of studies (6%) examined drugs in this class (Table 2).

DISCUSSION Our systematic review has assembled Australia’s pharmacoepidemiological research output, representing

a significant resource for the international research community. In this period spanning more than a quarter of a century (1987–2013), 228 studies were published utilising Australia’s dispensing databases. This collection of research has generated evidence relating to quality use of medicines across a number of domains. There has been a heavy focus on descriptive drug utilisation studies based on aggregated claims data. Because of greater access to individual-level dispensing data in the last decade, there has been a growth in studies addressing clinician and patient behaviours. Access to dispensing data linked to other routine data collections such as medical services, hospitalisations and death records has created opportunities to conduct drug safety studies and those evaluating the impact of specific interventions. However, this review also highlights significant gaps and challenges related to the conduct of pharmacoepidemiological research in Australia. The Australian research effort and the populations under study have been driven by the available data. Many of the studies accessed data held by DUSC and the DVA. The studies based on DUSC data provide important information on overall trends in total drug use in the entire Australian population.18–22 However, they shed no light on individual-level medicines use or the outcomes of this use; it is these studies that will provide the depth and detail of evidence for our community to determine the return on our PBS investment. Without the goodwill and proactive approach to research of the DVA, our understanding of drug use and outcomes in Australia would be significantly diminished. The DVA

Table 2. Number of studies by pharmacological group compared with PBS volume and PBS expenditure for year ending June 2013. Study could be classified under more than one pharmacological group Claims-level studies

Individual-level studies

Anatomical therapeutic classification first level grouping

n

n

n

A B C G H J L M N R

10 — 13 6 — 11 8 7 43 3 7 15

13 — 22 0 — 3 5 2 37 4 3 26

23 — 35 6 — 14 13 9 80 7 10 41

Alimentary tract and metabolism Blood and blood forming organs Cardiovascular system Genito-urinary system and sex hormones Systemic hormonal preparations Anti-infectives for systemic use Antineoplastic and immunomodulating agents Musculo-skeletal system Nervous system Respiratory system Other ATC groups* All ATC groupings

PBS volume 2013#

PBS cost 2013#

%

%

%

10 — 15 3 — 6 6 4 35 3 4 18

14 4 34 1 2 7 1 4 21 6 6

14 4 26 1 1 4 13 3 19 7 8

All studies

PBS, Pharmaceutical Benefits Scheme; ATC, Anatomical Therapeutic Chemical. Six studies were removed from the analysis. These studies used individual-level drug data to define their cohort but did not derive drug-related outcomes. *Other ATC groups: dermatologicals, sensory organs, and various. # Data derived from the PBS: Expenditure and prescriptions 12 months to 30 June 2013. Canberra; 2013. http://www.pbs.gov.au/info/statistics/expenditure-and-prescriptions-30-06-2013. The figures include prescriptions on the general Section 85 schedule only; prescriptions dispensed under Section 100 (highly specialised drugs) are not included in these estimates.

Copyright © 2015 John Wiley & Sons, Ltd.

Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

australian pharmacoepidemiological research

oversees the whole-of-health care for its clients. As the sole payer of these health services, the DVA maintains custodianship over a wide variety of routine data collections including dispensing and medical service claims and hospitalisations. Not surprisingly, the vast majority of studies using individual-level data with linkages to other health datasets are based on this subset of elderly Australians.23–27 However, the DVA client population is small (approximately 220 000 veterans and their dependents with health care entitlements across Australia in 2014) and therefore of insufficient size to power a broad range of comparative effectiveness and safety evaluations.28 While we applaud the DVA for their continued research support, there has been a strong push to extend linkage between PBS claims and other collections to undertake pharmacoepidemiological research in the wider Australian population.29 However, this requires engagement of data custodians across Commonwealth and state boundaries, which is yet to become commonplace. The push to extend data linkage initiatives is of particular importance for pharmacoepidemiological research given the diminishing DVA treatment population who had an average age of 75 years in 2014,28 rendering it unfeasible to rely on this sole linked data source in the future. Moreover, most of the studies in this review have been necessarily restricted to populations where complete ascertainment of medicines is possible. Therefore, the studies have been restricted to Australians with concessional benefits or DVA clients with full health care entitlements. Finally, the significant lag between study end dates and year of publication is likely to influence the clinical and policy impact of the research output. Our review also highlights significant blind spots with respect to quality use of medicines and our country’s return on the PBS investment. Not surprisingly, the Australian research output has focussed on medicines prescribed primarily in general practice such as statins, antidepressants and diabetes medications for diseases that represent a significant proportion of disease burden in Australia.30 However, the most recent PBS statistics highlight the growth, particularly in relation to PBS spend, on specialist-initiated medicines such as antineoplastics (cancer drugs). We have very limited research in this important and growing area of therapeutics31–34 given cancer represents a significant disease burden in the Australian population.30 Moreover, populations such as the elderly, children, pregnant women and those with multiple morbidities are studied rarely in pre-market evaluation.1 The post-market setting provides the only opportunity to understand the benefits and risks of drug use in these populations. Compared to the elderly Copyright © 2015 John Wiley & Sons, Ltd.

453

(represented primarily by DVA clients), we know little about medicines use in children35–38 and pregnant women39–44 and found no studies examining the outcomes of medicines use in Australian children. The studies of drug use and outcomes in pregnant women highlight Australia’s capacity for pharmacoepidemiological research using linked routinely collected data in populations other than the elderly.39–45 Australia is in a powerful position to conduct wholeof-population pharmacoepidemiological research due to its universal drug coverage for 23 million citizens and the more than 200 million prescriptions generated each year.8 The significant investment in data linkage infrastructure through the Population Health Research Network has supported Australian states and territories to build a network enabling health data collected around the nation for health-related research purposes;28 the recent establishment of Data Integrating Authorities will support the integration of Commonwealth data with state-based collections.46 Finally, the collection of under co-payment dispensing claims from 2012 will, for the first time, facilitate research in all PBS-eligible populations (not just concession card-holding Australians).16 All of these developments will be pivotal for enhancing pharmacoepidemiology research in Australia. To date, our research has been relatively fragmented, focusing on specific sub-populations. These restrictions are also common to jurisdictions such as the USA and Canada where data are only available on selected patient populations such as beneficiaries of Medicare and Medicaid programs, veterans, enrollees in health maintenance organisations, or ‘seniors’ in some of the Canadian provinces. However, these countries, along with Scandinavia, lead the pharmacoepidemiological research enterprise globally. Notably, Scandinavia has a combined population equivalent to that of Australia and universal health care arrangements. A recent review of the pharmacoepidemiological literature based on the Nordic databases, representing Denmark, Sweden, Norway and Iceland, demonstrated an output in 6 years that was double the Australian output in more than a quarter of a century.6 These established traditions of data access demonstrate Australia’s future potential. Our review has some limitations. We may not have included all relevant studies, highlighted by the number of studies identified via back references and key author searches. We also had a defined focus whereby we did not include cohort studies or those requiring specific patient consent for data linkage using PBS and RPBS data or publically available reports based on regular monitoring activities of the Australian Departments of Health, Human Services and Veterans’ Affairs. In addition, we Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

454

s.-a. pearson et al.

did not assess the methodological quality of studies, but this would be an important addition to elucidate the inferences made from the research studies. Finally, we developed an arbitrary classification to group studies by their main focus. Given the heterogeneity of studies, some could clearly fall into more than one category. In conclusion, Australia has a global reputation as a leader in pharmaceutical policy because of its pioneering work in assessing medicines prior to listing for public subsidy. Our research effort assessing the use and impact of prescribed medicines post-subsidy has been restricted mainly to descriptive studies of use across the whole-population or more detailed analyses in specific sub-populations. As a result, we know little about the outcomes of our annual multi-billion dollar spend on prescribed medicines. Our country’s potential is significant in this realm. We have the capacity to join the global leadership in pharmacoepidemiology as our data infrastructure and access processes mature and we build a capacity for the timely analysis, interpretation and translation of the outcomes of this research to support clinical and policy decision-making. CONFLICTS OF INTEREST The authors report no actual, potential or perceived conflict of interest with regard to the submission of this manuscript. KEY POINTS Pharmaceutical Benefits Scheme and Repatriation Benefits Scheme (PBS and RPBS) data represent significant resources to examine Australia’s multibillion dollar annual investment in prescribed medicines. • 228 pharmacoepidemiological studies based on PBS and RPBS dispensing records were published between 1987 and 2013, and the body of research has increased exponentially in the past decade. • Drug utilisation studies, based mostly on claimslevel data, were the most common (33%); drug use and outcomes studies, based mostly on linked person-level data, were less common (18%); studies were often limited to elderly patient subpopulations; few studies examined prescribed medicines use in children and pregnant women. • Most studies focused on drug classes acting on the nervous system (36%), cardiovascular system (15%) and alimentary tract (11%). • Recent investment in data linkage infrastructure will facilitate a growth in drug use and outcomes studies in the future.

•

Copyright © 2015 John Wiley & Sons, Ltd.

ACKNOWLEDGEMENTS This study was funded, in part, by the Centre for Research Excellence in Medicines and Ageing (ID: 1060407); SP is a Cancer Institute New South Wales Career Development Fellow (ID: 12/CDF/2–25) and an Australian Health Policy Research Fellow. We thank Mr Leigh Mellish for his support with literature searches and preparation of this manuscript. AUTHOR CONTRIBUTIONS SP conceived the study. All authors extracted data and were responsible for data interpretation. SP and NP drafted the manuscript. JL, AD, MF and JR provided editorial comments and approved the final version of the manuscript.

REFERENCES 1. Avorn J. In defense of pharmacoepidemiology-embracing the yin and yang of drug research. N Engl J Med 2007; 357: 2219–2221. doi:10.1056/ NEJMp0706892. 2. Schneeweiss S, Avorn J. A review of uses of health care utilization databases for epidemiologic research on therapeutics. J Clin Epidemiol 2005; 58: 323–337. doi:10.1016/j.jclinepi.2004.10.012. 3. Forrow S, Campion DM, Herrinton LJ, Nair VP, Robb MA, Wilson M, Platt R. The organizational structure and governing principles of the Food and Drug Administration’s Mini-Sentinel pilot program. Pharmacoepidemiol Drug Saf 2012; 21(Suppl 1): 12–17. doi:10.1002/pds.2242. 4. Platt R, Carnahan RM, Brown JS, Chrischilles E, Curtis LH, Hennessy S, Nelson JC, Racoosin JA, Robb M, Schneeweiss S, et al. The U.S. Food and Drug Administration’s Mini-Sentinel program: status and direction. Pharmacoepidemiol Drug Saf 2012; 21(Suppl 1): 1–8. doi:10.1002/pds.2343. 5. Psaty BM, Breckenridge AM. Mini-Sentinel and regulatory science–big data rendered fit and functional. N Engl J Med 2014; 370: 2165–2167. doi:10.1056/ NEJMp1401664. 6. Wettermark B, Zoega H, Furu K, Korhonen M, Hallas J, Norgaard M, Almarsdottir A, Andersen M, Andersson Sundell K, Bergman U, et al. The Nordic prescription databases as a resource for pharmacoepidemiological research–a literature review. Pharmacoepidemiol Drug Saf 2013; 22: 691–699. doi:10.1002/pds.3457. 7. Pharmaceutical Benefits Scheme: expenditure and prescriptions twelve months to 30 June 2003. Canberra, 2003. Available at: http://www.pbs.gov.au/statistics/ 2002–2003-files/expenditure-and-prescriptions-2002-03.pdf [1 July 2014]. 8. Pharmaceutical Benefits Scheme: expenditure and prescriptions twelve months to 30 June 2013. Canberra; 2013. Available at: http://www.pbs.gov.au/info/statistics/expenditure-and-prescriptions-30–06–2013 [1 July 2014]. 9. Birkett DJ, Mitchell AS, McManus P. A cost-effectiveness approach to drug subsidy and pricing in Australia. Health affairs (Project Hope) 2001; 20: 104–114. 10. Lopert R, Elshaug AG. Australia’s ‘fourth hurdle’ drug review comparing costs and benefits holds lessons for the United States. Health affairs (Project Hope) 2013; 32: 778–787. DOI: 10.1377/hlthaff.2012.1058. 11. Pharmaceutical Benefits Scheme: Drug Utilisation Sub Committee (DUSC). Canberra, 2014. Available at: http://www.pbs.gov.au/info/industry/listing/participants/drug-utilisation-subcommittee [1 July 2014]. 12. Edmonds DJ, Dumbrell DM, Primrose JG, McManus P, Birkett DJ, Demirian V. Development of an Australian drug utilisation database: a report from the Drug Utilization Subcommittee of the Pharmaceutical Benefits Advisory Committee. Pharmacoeconomics 1993; 3: 427–432. 13. The Pharmaceutical Benefits Scheme: Australian Statistics on Medicines 2011. Canberra; 2011. Available at: http://www.pbs.gov.au/info/statistics/asm/asm2011 [1 July 2014]. 14. DVA Factsheet IS160. Concession and benefits: overview of cards available to veterans and their dependents. Canberra; 2014. Available at: http://factsheets. dva.gov.au/factsheets/documents/IS160%20Overview%20of%20Cards%20Avai lable%20to%20Veterans%20and%20Their%20Dependents.pdf [1 July 2014].

Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds

australian pharmacoepidemiological research 15. The Pharmaceutical Benefits Scheme: 4. Patient charges. Canberra; 2014. Available at: http://www.pbs.gov.au/info/healthpro/explanatory-notes/section1/Section_1_4_Explanatory_Notes [1 July 2014]. 16. The Department of Health: Pharmaceutical Benefits Scheme collection of under co-payment data. Canberra; 2014. Available at: http://www.health.gov.au/internet/main/publishing.nsf/Content/pbs-under-copayment-data [1 July 2014]. 17. Structure and principles. Oslo; 2011. Available at: http://www.whocc.no/atc/ structure_and_principles/ [15 June 2014]. 18. Hollingworth SA, Nissen LM, Stathis SS, Siskind DJ, Varghese JMN, Scott JG. Australian national trends in stimulant dispensing: 2002–2009. Aust N Z J Psychiatry 2011; 45: 332–336. doi:10.3109/00048674.2010. 543413. 19. Leggat PA. Trends in antimalarial prescriptions in Australia 2002 to 2005. J Travel Med 2008; 15: 302–306. doi:10.1111/j.1708-8305.2008.00227.x. 20. McManus P, Henry DA, Birkett DJ. Recent changes in the profile of prescription NSAID use in Australia. Med J Aust 2000; 172: 188. 21. Stephenson CP, Karanges E, McGregor IS. Trends in the utilisation of psychotropic medications in Australia from 2000 to 2011. Aust N Z J Psychiatry 2013; 47: 74–87. doi:10.1177/0004867412466595. 22. Vitry AI, Thai LP, Lu CY. Time and geographical variations in utilization of endocrine therapy for breast cancer in Australia. Intern Med J 2011; 41: 162–166. doi:10.1111/j.1445–5994.2010.02304.x. 23. Caughey GE, Preiss AK, Vitry AI, Gilbert AL, Ryan P, Shakib S, Esterman A, McDermott RA, Roughead EE. Does antidepressant medication use affect persistence with diabetes medicines? Pharmacoepidemiol Drug Saf 2013; 22: 615–622. doi:10.1002/pds.3424. 24. Caughey GE, Roughead EE, Pratt N, Killer G, Gilbert AL. Stroke risk and NSAIDs: an Australian population-based study. Med J Aust 2011; 195: 525–529. doi:10.5694/mja11.10055. 25. Pratt N, Andersen M, Bergman U, Choi NK, Gerhard T, Huang C, Kimura M, Kimura T, Kubota K, Lai EC, et al. Multi-country rapid adverse drug event assessment: the Asian Pharmacoepidemiology Network (AsPEN) antipsychotic and acute hyperglycaemia study. Pharmacoepidemiol Drug Saf 2013; 22: 915–924. doi:10.1002/pds.3440. 26. Pratt N, Roughead EE, Ramsay E, Salter A, Ryan P. Risk of hospitalization for hip fracture and pneumonia associated with antipsychotic prescribing in the elderly: a self-controlled case-series analysis in an Australian health care claims database. Drug Safety 2011; 34: 567–575. doi:10.2165/11588470-00000000000000. 27. Roughead EE, Kalisch LM, Pratt NL, Killer G, Barnard A, Gilbert AL. Managing glaucoma in those with co-morbidity: not as easy as it seems. Ophthalmic Epidemiol 2012; 19: 74–82. doi:10.3109/09286586.2011.638743. 28. Australian Government Department of Veterans’ Affairs. Treatment population statistics: quarterly report June 2014. Canberra; 2014. Available at: http://www. dva.gov.au/aboutDVA/Statistics/Documents/TPopJun2014.pdf [1 August 2014]. 29. Population Health Research Network. Perth; 2014. Available at: http://www. phrn.org.au/ [20 July 2014]. 30. Begg SVT, Barker B, Stevenson C, Stanley L, Lopez AD. The burden of disease and injury in Australia 2003. Cat. no. PHE 82. Canberra: AIHW; 2007. Available at: http://www.aihw.gov.au/publication-detail/?id=6442467990 [1 August 2014]. 31. Damianovich D, Adena M, Tebbutt NC. Treatment of 5-fluorouracil refractory metastatic colorectal cancer: an Australian population-based analysis. Br J Cancer 2007; 96: 546–550. doi:10.1038/sj.bjc.6603590. 32. Pearson SA, Ringland CL, Ward RL. Trastuzumab and metastatic breast cancer: trastuzumab use in Australia–monitoring the effect of an expensive medicine access program. J Clin Oncol 2007; 25: 3688–3693.

Copyright © 2015 John Wiley & Sons, Ltd.

455

33. Stavrou EP, Lu CY, Buckley N, Pearson S. The role of comorbidities on the uptake of systemic treatment and 3-year survival in older cancer patients. Ann Oncol 2012; 23: 2422–2428. doi:10.1093/annonc/mdr618. 34. Mealing NM, Dobbins TA, Pearson SA. Validation and application of a death proxy in adult cancer patients. Pharmacoepidemiol Drug Saf 2012; 21: 742–748. doi:10.1002/pds.2257. 35. Cohen SA, Lawson JA, Graudins LV, Pearson S-A, Gazarian M. Changes in anticonvulsant prescribing for Australian children: implications for quality use of medicines. J Paediatr Child Health 2012; 48: 490–495. doi:10.1111/j.1440– 1754.2011.02223.x. 36. Hollingworth S, Duhig M, Hall W, Scott J. National trends in the community prescribing of second-generation antipsychotic medications in Australian children and youth: the incomplete story. Australas Psychiatry 2013; 21: 442–445. doi:10.1177/1039856213497809. 37. Kemp A. Hypoallergenic formula prescribing practices in Australia. J Paediatr Child Health 2006; 42: 191–195. 38. Wilson EJ, Nasrin D, Dear KBG, Douglas RM. Changing GPs’ antibiotic prescribing: a randomised controlled trial. Commun Dis Intell 2003; 27(Suppl): 32–38. 39. Colvin L, Gill AW, Slack-Smith L, Stanley FJ, Bower C. Off-label use of ondansetron in pregnancy in Western Australia. BioMed Research International 2013; 2013: 909860. DOI: 10.1155/2013/909860. 40. Colvin L, Slack-Smith L, Stanley FJ, Bower C. Pharmacovigilance in pregnancy using population-based linked datasets. Pharmacoepidemiol Drug Saf 2009; 18: 211–225. doi:10.1002/pds.1705. 41. Colvin L, Slack-Smith L, Stanley FJ, Bower C. Linking a pharmaceutical claims database with a birth defects registry to investigate birth defect rates of suspected teratogens. Pharmacoepidemiol Drug Saf 2010; 19: 1137–1150. doi:10.1002/pds.1995. 42. Colvin L, Slack-Smith L, Stanley FJ, Bower C. Dispensing patterns and pregnancy outcomes for women dispensed selective serotonin reuptake inhibitors in pregnancy. [Erratum appears in Birth Defects Res A Clin Mol Teratol. 2011 Apr;91(4):268]. Birth Defects Res A Clin Mol Teratol 2011; 91: 142–152. doi:10.1002/bdra.20773. 43. Colvin L, Slack-Smith L, Stanley FJ, Bower C. Early morbidity and mortality following in utero exposure to selective serotonin reuptake inhibitors: a population-based study in Western Australia. CNS Drugs 2012; 26: 1–14. doi:10.2165/11634190–000000000–00000. 44. Colvin L, Slack-Smith L, Stanley FJ, Bower C. Are women with major depression in pregnancy identifiable in population health data? BMC Pregnancy Childbirth 2013; 13: 63. DOI: 10.1186/1471–2393–13–63. 45. Vajda FJE, Horgan D, Hollingworth S, Graham J, Hitchcock AA, Roten A, O’Brien TJ, Lander CM, Eadie MJ. The prescribing of antiepileptic drugs for pregnant Australian women. Aust N Z J Obstet Gynaecol 2012; 52: 49–53. DOI: 10.1111/j.1479–828X.2011.01359.x. 46. National Statistical Service. A guide for data integration projects involving commonwealth data for statistical and research projects. Canberra; 2014. Available at: https://statistical-data-integration.govspace.gov.au/roles-and-responsibilities/ integrating-authorities/ [1 June 2014].

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article at the publisher’s web site.

Pharmacoepidemiology and Drug Safety, 2015; 24: 447–455 DOI: 10.1002/pds