In Australia, patients who want to access medicines that are not yet approved have only two options: to enrol in a clinical trial if they are eligible, or obtain their medicine through 'compassionate supply', which is provided at the discretion of the manufacturer. In this article, we explore ethical issues associated with the provision of oncology medicines that are still in development, either prior to regulatory approval or government reimbursement.

Although randomized controlled trials remain the scientific ideal for determining the efficacy and safety of new treatments, they are sometimes insufficient to address the evidentiary requirements of regulators and payers. This is particularly the case when it comes to precision medicines because trials are often small, deliver incomplete insights into outcomes of most interest to policymakers (eg, overall survival), and may fail to address other complex diagnostic and treatment-related questions. Additional methods, both experimental and observational, are increasingly being used to fill critical evidentiary gaps. A number of modified early- and late-phase trial designs have been proposed to better support earlier biomarker validation, patient identification, and selection for regulatory studies, but there is still a need for confirmatory evidence from real-world data sources. These data are usually provided through observational, postapproval, phase IIIB and IV studies, which rely heavily on registries and other electronic data sets—most notably data from electronic health records. It is, therefore, crucial to understand what ethical, practical, and scientific challenges are raised by the use of electronic health records to generate evidence about precision medicines.

Personalized medicines hold promise for many diseases. However, demonstrating the clinical efficacy and cost effectiveness of these medicines can be difficult. It is essential that decision-making processes for funding new medicines, including personalized medicines, are both robust and fit for purpose. We will argue that randomized trials of personalized medicines should be routinely supplemented with other research methods, such as observational research and single-arm studies, and that managed-entry funding programs, such as coverage with evidence development, may offer a means of providing early access to technologies where there is uncertainty about efficacy, safety, and cost effectiveness. These programs, however, raise a number of practical and ethical challenges that need to be worked through and resolved.

Summary Insights into the molecular drivers of cancer are providing opportunities for the development of new targeted treatments and more personalised approaches to cancer management Drugs targeting mutant epidermal growth factor receptors, such as erlotinib and gefitinib, may provide more effective, safer and better tolerated treatment options compared with chemotherapy among appropriately selected patients with advanced non‐small cell lung cancer (NSCLC). First‐line access to these newer treatments remains unfunded after several considerations by the Pharmaceutical Benefits Advisory Committee and their assessment that these are not cost‐effective treatments. We suggest that there may be evidentiary and ethical challenges associated with the assessment of the cost‐effectiveness of personalised oncology medicines in Australia, and that a new approach is needed to determine the value and cost‐effectiveness of personalised medicine.

Despite enthusiastic advocacy for what personalized medicine might be able to deliver and major investments into the development of this, there remain disappointingly few examples of personalized medicine in routine clinical practice today, particularly in high areas of unmet need such as cancer. We believe that this is because personalized medicine challenges the moral, economic and epistemological foundations of medicine. In this article, we briefly describe the scientific premises underpinning personalized medicine, contrast these with traditional paradigms of drug development, and then consider the ethical, economic and epistemological implications of this approach to medicine.