Payers and personalised medicine: considerations for success

Dr. Andree K Bates

Eularis

Gleevec and Herceptin (and their clinical and commercial success), were pioneering drugs in terms of embarking us into an age of personalised medicine. Following them Avastin, Tarceva, Iressa and Erbitux. Now, many companies are developing, or have developed, drugs with the aim of identifying patients based on the existence of specific biomarkers, and these subpopulations that respond optimally to treatment are identified. In fact, a 2007 McKinsey report indicated that 30-50% of compounds in development have an associated biomarker program and that number was expected to increase. Having this clearly generates more favourable risk-benefit profiles than with the traditional 'all-comers' approach and enables a personalised approach with the potential to reduce the cost of cancer care. For these drugs with a biomarker, a companion diagnostic test is needed. However, clinical adoption of personalised medicine diagnostic tests remains relatively low. Where is the problem? There are many potential influencers on this result but they appear to fall into three main areas, scientific, economic, and operational although out of these the economic appear to be the biggest hurdle.

"...30-50% of compounds in development have an associated biomarker program and that number was expected to increase."

Payors and possible barriers to widespread adoption

The approach payers take with personalized medicine is important, as payer reimbursement will obviously impact the pharma and diagnostics companies business models. It is widely assumed that using biomarkers and diagnostic tests has the potential to reduce the cost of cancer care and yet payers appear to be slow to approve these. According to Price Waterhouse Coopers, currently, less than 5% of all US private companies reimburse for genetic tests, indicating that we may not be able to deliver personalized medicine. Although there is of course a strong possibility that by targeting patients accurately the costs will go down in the long term, this has yet to be realised and more outcomes studies are probably required to assess this fully. Some innovative payers are enthusiastic but there are things that need to be considered:

1. Uncovering which tests specifically will save costs. Payors need to know the economics of the costs savings to make informed decisions. McKinsey conducted a study analysing per patient savings of tests (so the difference between the cost of treating the illness versus the cost of the treatment suggested by the diagnostic test) as well as the probability that the diagnostic test recommends a treatment for several tests costing between $100-$3000. They found that the cost savings varied from $600 to $28,000 per patient. It was seen that if a test was able to ensure avoidance of a costly drug, decrease adverse events, or delay procedures were found to be very strong on cost savings for payors.

2. Long term results data. Due to many of these tests being relatively new, there is less long term outcomes data available upon which to base decisions. The more this comes to light, the more easily the payer can make decisions based on this.

3. Analysis models used. The actuarial models many payers use were historically based on large, stable, predictable populations whereas the populations targeted for personalised medicine are small, niche, dynamic and unpredictable.

4. Sustainability. The payers no doubt have a nagging worry that costs of personalised medicine have the potential to be unsustainable given the high price premiums of new targeted diagnostic tests and drugs and the potential to be additive rather than replace existing diagnostics and drugs.

"The actuarial models many payers use were historically based on large, stable, predictable populations whereas the populations targeted for personalised medicine are small, niche, dynamic and unpredictable."

5. Plan churn. Yet another prickly issue could be that one payer is paying for something that will not benefit them but one of their competitors in the long term. A significant proportion of people change health plans regularly given much of them are based on the employer so if you change your job you often change your health plan. So, you invest now in something that benefits costs in the long term by eliminating the need for surgery or care in the future but when that benefit is realised, the company that paid for it may not be the beneficiary of the saving. Alternatively people may begin by being in a costly plan that covers diagnostic tests and once they are diagnosed they may switch to a cheaper plan. But these issues would be common to most payers…

Some of the larger players are reimbursing these types of tests and drugs and shining positive examples can be seen at Aetna and Kaiser Permanente (although it should be noted that Kaiser Permanente has low patient turnover compared to many so is less subject to some of these risks), and Geisinger Health System. These innovators may lead others to follow.

Where things could go

An innovation that began in Italy, and is gaining more acceptance in several countries is the pay for performance model. This is also increasing in the USA and growing at an annual rate of around 26% in the US. This model could increase the reimbursement and adoption of personalised medicine when more evidence mounts to show that targeted diagnostics and drugs reduce payers costs. This was the case with Oncotype DX which Genomic Health initially did not cover but once evidence mounted that the overall cost was significantly reduced it gained reimbursement. It was found that if 50% of the eligible patients got the test the overall saving per patient (reduced chemotherapy, adverse events management and supportive care) was $1930 per patient. Although it should be noted that it took some time to gain reimbursement as evidence was collected. Launch was in 2004 and it is only recently becoming a routinely covered test.

"It was found that if 50% of the eligible patients got the test the overall saving per patient (reduced chemotherapy, adverse events management and supportive care) was $1930 per patient."

Yet another study in 2009 showed that $604 million could be saved annually if the use of Vectibix and Erbitux was restricted to those patients with mCRC whose KRAS gene was not mutated since they are the only patients who would benefit from it. Numbers like these do make a compelling case but more studies to show this are needed to convince more payers.

Conclusions for pharmaceutical and biotech considerations

Clearly reimbursement can be a slow moving beast for personalised medicine and companies need to take a long term view of the environment and really stay focused on the area that will impact access including strong outcomes data. Given the move by several countries towards more risk share and outcomes based pricing where possible a partial reimbursement is given until the outcomes data is available, and if positive the company is then paid for the remainder, this is becoming one of the most important factors to increase the value to pharmaceutical companies. This seems a win:win for good drugs, early coverage for the pharmaceutical and diagnostic companies and reduced financial exposure for the payers. This has to be the key – win:win focus and working together to ensure that happens.

About the author

For a discussion of these topics, contact Dr Andree K Bates at Eularis http://www.eularis.com.

How can we increase adoption of personalised medicine diagnostic tests?