A new model to evaluate the long-term cost effectiveness of orphan and highly specialised drugs following listing on the Australian Pharmaceutical Benefits Scheme: the Bosentan Patient Registry

Abstract

Background: Pharmaceutical subsidy schemes are under increasing pressure to evaluate the cost effectiveness of new highly specialised and orphan drugs for universal subsidy. In the absence of longer-term outcome data, drug sponsors often present modelled data, which can carry a significant level of uncertainty over longer-term projections. Risk-sharing schemes between drug sponsor and government may provide an acceptable method of balancing the uncertainty of longer-term cost effectiveness with the public demand for equitable and timely access to new drugs.

Methods: The Bosentan Patient Registry (BPR) is an example of a unique risk-sharing model utilised in Australia aiming to provide clinical evidence to support the modelled predictions, with the registry survival outcomes linked to future price. Concomitant medication, health and vital status data was collected from clinicians, government health departments and death registries.

Results: The BPR has identified a number of issues surrounding registry governance, ethics and patient privacy, and the collection of timely and accurate data, which need to be addressed for the development of a generic registry model for systematic evaluation.

Conclusion: The success of a generic drug registry model based on the BPR will be enhanced by addressing a number of operational issues identified during the implementation of this project.

Material in this paper was presented in an oral presentation to the National Medicines Symposium: Quality Use of Medicines, 7–9 June 2006, Canberra, Australia.

Keywords::

• highly specialised drugs
• cost effectiveness
• bosentan
• Pharmaceutical Benefits Scheme
• drug registry

Introduction

Equitable access to effective, highly specialised and orphan drugs through universal subsidy schemes such as the Australian Pharmaceutical Benefits Scheme (PBS) remains a contentious issue for policy makers. The Pharmaceutical Benefits Advisory Committee (PBAC) evaluates the comparative clinical efficacy (compared with alternatives) and cost-effectiveness data submitted by the drug sponsor when determining its potential value to the Australian population and any subsequent recommendation under the PBS2. Orphan drugs, and innovative specialised drugs used in relatively small populations, encounter a disadvantage compared with more widely used drugs, as large-scale clinical trial data are usually unavailable3.

Modelling of clinical outcomes is often employed by drug sponsors in the listing process to estimate the longer-term benefits and cost effectiveness of these specialised drugs. The accuracy of these models is difficult to assess as outcome data for these drugs are not routinely collected after PBS listing. The pressure on the PBAC to evaluate an increasing number of these orphan and highly specialised drugs has led to innovative risk-sharing approaches between the drug sponsor and the Australian Government in recent times. Reimbursement authorities in the UK and Australia are now exploring systematic methods of post-marketing cost-effectiveness assessment for cases where a significant level of uncertainly surrounds the modelled predictions4–6.

Risk-sharing agreements allow governments to balance the uncertainty of long-term cost effectiveness with the public demand for equitable access to effective but high-cost drug therapies. The advantages of risk-sharing agreements for key stakeholders include that they: encourage the drug sponsor to promote responsible prescribing and ensure that healthcare resources are not wasted on ineffective treatments; provide an outcomes guarantee that has the potential to ensure predictable health gains for a given drug expenditure; and allow the drug sponsor and the government to share the identified budgetary risks associated with the listing of high-cost drugs. For patients and the drug sponsor, the risk-sharing agreement may facilitate earlier access of the drug to the market. The disadvantages of such risk-sharing strategies include the requirement for structures to be in place to allow comprehensive data capture of the treatment population to ensure that a representative and unbiased sample are used in post-marketing cost-effectiveness analyses. The advent of new treatments during the monitoring period, which may or may not be subject to the same monitoring requirements, may also disadvantage the risk-sharing agreement.

Bosentan (Tracleer®) is a dual endothelin receptor antagonist, acting to block the binding of endothelin to both of its receptors, ET_A and ET_B. Endothelin is a powerful vasoconstrictor peptide produced by the endothelial lining of blood vessels, and blocking the action of endothelin can achieve a marked lowering of blood pressure. Bosentan was listed on the PBS in March 2004 for the treatment of idiopathic pulmonary arterial hypertension (PAH) (previously known as primary pulmonary hypertension) and PAH secondary to scleroderma7. Patients with PAH are generally considered to have a poor prognosis, however, newer agents, including bosentan, have been shown to improve symptoms and quality of life, to delay clinical worsening and to improve survival8–11. The recommendation to include bosentan on the PBS was based on modelled improvements in life expectancy, however, this has not yet been confirmed in long-term controlled clinical trials. To facilitate timely access for patients with PAH to potentially life-saving therapy, the drug sponsor, Actelion Pharmaceuticals Australia Pty Limited, proposed a unique risk-sharing agreement where the future price of bosentan will be directly linked to the observed survival of patients treated with bosentan under the initial listing on the PBS12. The proposal included the establishment of a registry, known as the Bosentan Patient Registry (BPR), to collect survival outcome data related to the use of bosentan in order to assess the accuracy of modelled and overseas clinical trial data.

The BPR represents a unique and precedent-setting collaboration between the key stakeholders including the Australian Government (through the PBAC and Medicare Australia), the drug sponsor and a Steering Committee comprised mostly of clinicians involved in the management of PAH patients. The drug sponsor bears the establishment and operational costs of the registry, which is developed and maintained by an independent party, the Centre of Clinical Research Excellence in Therapeutics (CCRE Therapeutics) at Monash University (Melbourne, Australia).

Methods

Arrangements for access to PBS-subsidised bosentan include an application under the Section 100 restriction when initiating treatment, and then further prescription renewal applications every 6 months to ensure that patients have met response criteria to be eligible for continuing bosentan treatment. In addition, prescribing bosentan is restricted to specialist clinicians from hospitals approved by the Department of Health and Ageing (DOHA) as having the expertise to evaluate and care for patients with PAH (termed ‘designated centres’). These arrangements were incorporated in the BPR model.

Ethical aspects

Ethics approval for the BPR was sought at each of the 15 original bosentan designated centres, in addition to approval from the Monash University Standing Committee on Ethical Research in Humans. Additional centres were approved by the DOHA throughout the project, resulting in further ethics applications. Ethics approval included informed consent documents, all written material distributed to patients and data collection forms. Preliminary consultation with Medicare Australia (previously known as the Health Insurance Commission) was undertaken to ensure that appropriate wording was contained in informed consent documents to allow access to government health data sources. For minors, the guidelines of Medicare Australia and individual ethics committees were followed regarding consent to participate being given either by the legal guardian/parent or the patient directly depending on their age.

Participation in the BPR was voluntary, with each patient being invited to enrol and to provide written informed consent during a routine consultation with their treating physician. Refusal to participate did not affect a patient's eligibility to be prescribed PBS-subsidised bosentan. The PBS requirement for 6-monthly visits provided regular opportunities for patients to be invited to participate in the BPR at any time during treatment with bosentan (including after discontinuation) as the consent allowed for retrospective data collection.

Data management and flow

Data collection and data management services were co-ordinated by the BPR Data Centre, based at Monash University. At enrolment and at each 6-monthly visit, a one-page data collection form with contact details of the patient and physician, some limited information on functional status and details on concomitant therapy was completed.

Using fax-back electronic data capture methods, the information was transferred to the BPR Data Centre. A reminder system and follow-up procedures were in place to ensure that patients were not lost to follow-up during the course of the project. Treating physicians notified the registry if patients discontinued bosentan, either due to a change in therapy, transplantation/intervention or death. Survival data were verified against the Australian Institute of Health and Welfare National Death Index and adjusted accordingly. Dispensing information related to bosentan was to be transferred from Medicare Australia to the BPR database after review and approval of individual patient consent by the Privacy Officer at Medicare Australia.

Privacy issues

Data collected as part of the BPR are managed by the CCRE in accordance with State and Federal Privacy Laws as well as those laws governing data provided by Medicare Australia. Only the relevant staff involved in management of the BPR at the CCRE have access to individual patient records. Any other data provided to third parties, including the sponsor or government agencies, are in aggregated, de-identified form to ensure that confidentiality is maintained.

Future data analysis

Analysis of the BPR data will focus on the survival of patients taking bosentan (compared with clinical trial and modelled data) for the purposes of a cost–benefit assessment by the PBAC according to the risk-sharing agreement12. The analysis will be done according to a pre-designed analysis plan overseen by the Steering Committee. The outcomes of the registry can only influence movement in the price of the drug to the government in a downward direction (or the price can remain the same). Price reductions are triggered once the lower 95% confidence limit for observed mortality in bosentan-treated patients exceeds that of the predicted mortality rate.

Discussion

Development of the BPR has identified a number of challenges to the establishment and management of data registries. Issues surrounding registry governance, patient privacy, the jurisdiction of Human Research Ethics Committees (HRECs) and the collection of timely and accurate data need to be addressed if a generic version of the BPR is to be employed to evaluate these drugs systematically.

Ethics

The requirement to gain individual centre ethics approval for the BPR in Australia was slow, costly and administratively complex. Ethics committees required site-specific changes to informed consent wording as well as stipulating differing ages of consent for minors, which were at times in conflict with the requirements of Medicare Australia. Others13 have noted this situation, and change is required for the continued development of health registries in Australia. Ideally, a national ‘umbrella’ HREC, with appropriate expertise in reviewing health registry applications, would be able to provide national approval for all sites and federal governmental agencies. The recent development of the National Ethics Application Form by the Australian Government's National Health and Medical Research Council14 should assist in streamlining the ethical application process for multicentre registry projects.

Governance

Establishment of the BPR has been a collaborative approach between the PBAC, the drug sponsor and the clinician-led BPR Steering Committee, with data collection and analysis services being contracted to an academic institution. As each of these stakeholders has different requirements for the BPR, a clear set of nationally recognised guidelines similar to those governing clinical research needs to be established. Such guidelines should provide sufficient detail regarding the establishment, development and evaluation of health registries. Both in Australia and overseas, national bodies involved in research governance are aware of the need to develop uniform standards for the management of registry activities15,16 to ensure best practice in their governance. The current lack of guidelines may slow the adaptation of the BPR to a generic model.

Accurate and timely data

In a recent article, it was noted that ‘accurate and timely data’ are important in the evaluation of risk-sharing schemes6. Ensuring that data collected for the registry from all sources are able to be used and analysed effectively is one of the primary concerns of health registry activities.

It is recognised that registries should be designed to reflect normal clinical practice and limit additional data collection to increase registry compliance and avoid putting unnecessary strain on hospital resources4. The success of the BPR is highly dependent on clinician participation and will largely be determined by the attitudes and practices of this group.

The registry model aims to capture 100% of the PBS-subsidised bosentan-treated cohort. However, the voluntary nature of participation means that not all patients will be included. It is imperative to promote the existence and purpose of the BPR to ensure the cohort of patients is sufficiently large to generate meaningful data. An alternative approach that has been used is to enrol all patients in a registry but allow them to opt out17. Further qualitative research is required to determine whether consumers will value the privacy trade-off of an opt-out system in order to provide greater academic rigour in health outcomes research.

Registry activities reliant on government data sources are faced with the challenge that the purpose of data collection by those agencies differs from the objectives set by health registries. For example, dispensing information, which can be used as a proxy of drug exposure, is collated by Medicare Australia for the purposes of reimbursing hospital and community pharmacies. Variability in the submission of claims from these sources can lead to a potential delay in the availability of a complete dataset. In addition, the access and use of government data is usually subject to additional legislation as well as the standard privacy principles enshrined in law. In Australia, the privacy section of Medicare Australia is required to review all individual written patient consent forms prior to authorising transfer of health data to an external source. This labour-intensive process can add further delays to accessing data from government sources. The timeliness of information sourced from external databases needs to be considered when planning the evaluation of outcome data for risk-sharing schemes. Better linkage of government data sets, while ensuring public confidence in the privacy of their health information through rigorous informed consent processes, is important for the future of registry activities in Australia and worldwide.

With the constant evolution of pharmacotherapy, new treatment options may arise during the life of the registry. The registry requires the flexibility to capture new treatment options that might impact on cost effectiveness, whilst limiting the workload on contributors. The extent to which the introduction of new treatments will bias survival estimates will depend on the uptake of the new treatments and whether it is reasonable to assume they impact on survival in the rescue setting to an extent that they provide an advantage over continued treatment with bosentan. There are also other factors that might bias the registry results, such as patient aetiology and disease severity, or recruitment if rates of participation in the registry are low. Importantly, there will be no ascertainment bias as the National Death Index provides complete mortality information on the registry. These issues of analysis and interpretation are common to all epidemiological studies and pose a challenge for the stakeholders. Nevertheless, the information gathered through the registry will provide best- and worst-case estimates that are not currently available. These will remain valuable and relevant, as few long-term data are available on this patient group, and the pricing of new drugs has been, and is likely to continue to be, based on the pricing of bosentan.

Benefits

Balancing the technical and logistical challenges faced during the development of the BPR has a number of substantial benefits. As part of the risk-sharing arrangement there is an incentive for the drug sponsor to maximise survival (and thus price) by supporting expert centres providing opportunities for education and communication between local opinion leaders. This incentive is clearly also beneficial for patients. The information collected by the registry provides the drug regulatory authority (PBAC) with an opportunity to evaluate the uncertainties surrounding longer-term cost effectiveness of a highly specialised drug in the relevant population. This approach may ultimately enhance the decision-making process of the PBAC for the PBS and allow more timely access to drugs for patients suffering conditions that have limited alternative therapies. High-quality observational data for a well defined cohort provides researchers and clinicians with information to better evaluate clinical prescribing guidelines in the management of rare diseases. Such databases may in the future be integrated into existing national pharmacovigilance systems providing routine monitoring of adverse events for specialised and orphan drugs.

Conclusions

Innovative research will continue to produce more targeted, highly specialised, high-cost/high-value medicines for the treatment of rare diseases. As public pressure for access to these medicines increases, government reimbursement schemes will continue to look for novel ways to assess and manage cost effectiveness. The dynamic nature of drug development means that additional agents will become available for use in therapeutic areas that previously had few treatment options. This constantly changing treatment paradigm requires that the role of the registry be constantly assessed to determine whether it is still able to meet the original objectives. The success of risk-sharing schemes such as the BPR is reliant on finding an acceptable method of dealing with the issues of governance, privacy, ethics review, and timely and accurate data. Alliances between government, clinicians, researchers and drug companies must continue if such initiatives are to realise fully their public health potential.

Acknowledgements

Declaration of interest: The registry described in this paper was funded by a research grant from Actelion Pharmaceuticals.

AJO, AM and CMR belong to a research centre established by an Australian Government National Health & Medical Research Council Grant. JS belongs to a research centre that provides consulting research for the Australian Government drug regulatory authority. TR, MH and DK are employees of Actelion Pharmaceuticals, and JW acts as a statistical consultant for Actelion Pharmaceuticals.

Notes

*Drugs and agents used to treat, prevent or diagnose rare diseases in Australia (i.e. not affecting > 2,000 Australians at any given time) are eligible to apply for orphan drug status under the Therapeutic Goods Act (1989). The registration fees are waived to encourage sponsors to develop drugs and agents that may not be considered otherwise commercially viable under regular conditions. Further information is available on the Therapeutic Goods Administration website1.