I read with interest the article by Wilson and Cohen published recently in Value in Health that compared patient access to new cancer drugs in the United States and Australia over a 10-year period (2000–2009) [1]. The motivation for their study is the ongoing debate in the United States on the greater use of comparative effective research (CER) and its effect on patient access to health care and, in particular, medicines. Australia has extensive experience in the use of results from CER to inform decisions on patient access to new medicines via the Pharmaceutical Benefits Scheme (PBS), and so the comparison will be of interest to many. Nonetheless, their academic study raises a number of issues.

Wilson and Cohen chose “new cancer drugs” as their study sample. This is surprising, given that the CER movement in the United States (and Australia) is not confined to interventions for cancer. Many new drugs for a wide array of diseases/conditions have high unit prices, not just those for cancer. One could argue that they should have compared a wider, more representative sample of new drugs; this would also have helped to increase their sample size.

They found 34 Food and Drug Administration (FDA)-approved new molecular entities and biologics for the treatment of patients with cancer in the study period. They did not define cancer; the inclusion of gemtuzumab ozogamicin, which was first approved by the FDA for patients with acute myeloid leukemia, suggests the use of a broad definition that includes hematological cancers and syndromes (e.g., leukemias and lymphomas).

Their sample also includes the bisphosphonate zoledronic acid. The FDA first approved zoledronic acid for patients with hypercalcemia of malignancy. The inclusion of zoledronic acid suggests the use of a very liberal definition of cancer because it does not have any antineoplastic activity. One could argue why other new drugs used by patients with cancer, such as aprepitant, which was approved by the FDA in 2003 to treat chemotherapy-induced nausea and vomiting, were not included [2].

Bacillus Calmette-Guérin (or Bacille Calmette-Guérin) (live) is another interesting inclusion. While Bacillus Calmette-Guérin (or Bacille Calmette-Guérin) (live) vaccine, as the Connaught and Tice strains, has been used for many years in the United States and Australia to treat certain patients with bladder cancer, the Montreal strain (Pacis) is new. Pacis is probably best categorized as a new formulation/presentation of an existing (biological) drug rather than a new (biological) drug. The inclusion of Pacis but not Abraxane (nanoparticle albumin-bound paclitaxel), which was approved by the FDA in 2005 for the treatment of certain patients with breast cancer, is puzzling [2].

For unstated reasons, Wilson and Cohen chose to exclude chemotherapeutic agents and hormonal treatments. If chemotherapeutic means cytotoxic, then oxaliplatin (platinum compound), bendamustine (alkylating agent), ixabepilone (cytotoxic antibiotic), vorinostat and romidepsin (histone deacetylase inhibitors), pemetrexed and pralatexate (folate antimetabolites), and nelarabine and clofarabine (purine nucleoside metabolic inhibitors) should have been excluded because they are all cytotoxic agents. Likewise, the hormonal treatments fulvestrant and degarelix should have been excluded.

The inclusion of everolimus is controversial; while the US FDA first approved it (as Afinitor) in 2009 for use in certain patients with renal cell carcinoma, it was first approved by the Therapeutic Goods Administration (TGA) (as Certican) in Australia in 2005 for the prophylaxis of organ rejection [3].

Their study sample also has errors of commission; azacitidine and lenalinomide, which were approved by the FDA for myelodysplastic syndrome in 2004 and 2005, respectively, should have been included [2]. There may be more examples.

The authors claim that 15 of the 34 FDA-approved new molecular entities and biologics had not been approved by the Australian TGA as of December 31, 2009. One of the 15 is zoledronic acid; it was approved by the TGA on March 15, 2001. Vorinostat was approved by the TGA on December 15, 2009 [3].

To enhance their sample size, the authors could have extended the study period for Australia to include drugs approved by the TGA after 2009. Pazopanib and ofatumumab were approved by the TGA on June 22, 2010, and December 15, 2010, respectively [4,

The authors speculate as to why the remaining 11 drugs have been approved only in the United States. One is that the drugs' sponsors do not have a presence in Australia. Allos Therapeutics, the sponsor of pralatrexate, and Spectrum Pharmaceuticals, the sponsor of ibritumomab tiuxetan, do appear not to have any operations outside the United States [6,7]. Cephalon has only recently entered the Australian market by gaining a controlling interest in Arana Therapeutics [8].

The authors used the date of a drug's first entry into the Australian Register of Therapeutic Goods as its date of approval. This is not always the case. Cetuximab was first registered by the TGA on February 4, 2005 [9].

New drugs for cancer are approved earlier in the United States because the FDA has a fast-track, accelerated-approval, and priority-review process for new drugs to treat serious diseases and fill an unmet medical need [10]. No mention was made of this. Twenty-two of the 34 drugs were registered under the accelerated-approval process. Gemtuzumab ozogamicin has since been withdrawn in the United States because of safety concerns and a lack of therapeutic benefit [11].

Patient access was studied by examining Medicare beneficiaries in the United States and those eligible to access drugs on the PBS in Australia. The latter covers all Australians, whereas the former covers only those in the United States who are aged 65 years and older. Wilson and Cohen have not studied the 44% of US cancer patients aged younger than 65 years who access drugs via other means. Mason and colleagues [12] have shown that patient access to new cancer drugs via private insurance is poorer than with Medicare. One has to question whether those in need of clofarabine, which is FDA approved for treatment of pediatric patients 1 to 21 years old with relapsed or refractory acute lymphoblastic leukemia, can do so via Medicare.

The most important metric is probably time from registration to reimbursement. The time point at which a drug's reimbursement commences under Medicare Part D is not clear. Imatinib was first listed on the PBS on December 1, 2001, but only for two of its three initial TGA-approved indications (accelerated phase and blast crisis); listing for chronic phase after interferon failure commenced on October 21, 2002. Oxaliplatin was first listed on the PBS on December 1, 2001. Zoledronic acid was first listed on the PBS on May 1, 2002. The PBS maximum quantity for bevacizumab (either strength) is one vial [13].

The authors have not explained why they compared prices at a time point outside of the reimbursement study period. It would have been more informative if they had cited each drug's price at reimbursement and on December 31, 2009. International price comparisons for drugs are problematic for a number of reasons, including differences in the way prices are reported and the existence of hidden discounts and rebates [14]. Average wholesale price, average selling price, and Medicare Part D prices do not correspond with PBS prices, which are at the dispensed price (price to patient) level. Three of the 12 PBS-listed drugs in Table 2 are subject to special pricing arrangements [13].

Comparing prices at a single time point also fails to account for currency fluctuations over time. The global price of imatinib was set in mid-2001 at US$2200 per month when AU$1.00 was worth approximately US$0.50 [15,16].

The authors may wish to reconsider their results and conclusions in light of the above. CER would be a useful method to determine whether the 15 [11] drugs that are available in the United States but not in Australia are clinically superior to the treatments they are likely to replace. It is unclear whether the results and conclusions from their study are applicable to all new drugs.

In Australia, all who need to access a new cancer drug are invariably able to do so via the PBS with a modest patient co-payment. Access can be rapid in situations where there is no time to lose (e.g., imatinib), even in the absence of randomized controlled trial evidence. In the United States, access is highly variable depending on one's age and/or income with higher patient co-payments. High patient co-payments associated with the Medicare Part D “donut hole” are reducing access [17]. Which country has the better access to new cancer drugs? Until there is an established evaluation framework, the jury will remain out. While the United States might be ahead in terms of breadth (i.e., more drugs) and timeliness, Australia seems to fare better on equity and sustainability [18].