An interesting thought about Precision Medicine

I was recently in The Netherlands for a few days, at an event for ERIBA. One of several interesting conversations at dinner merited a speculative blog post.

The topic under discussion was precision medicine (a recurring theme on the blog), and accidental motifs that could develop while trying to ensure that new drugs end up in the hands of those most likely to benefit. As a reminder, the purpose of precision medicine is the development of protocols such as biomarkers that predict drug efficacy, with the goal of matching patients to medicines. The use of Trastuzumab in HER2-positive breast cancer is a classic example.

Now all medicines must go through three phases of clinical trials, culminating in demonstration of efficacy in Phase III. Companies often develop biomarkers for use during clinical trials (these can but do not necessarily take the same form as the final precision approach), in the hopes that enrolling the “right” people will improve chances for a successful trial. It’s not kosher to run a trial and then post-hoc decide that responders and non-responders can be separated to yield statistical significance – it’s something you need to do right up front.

But even before the frequent use of molecular diagnostics, patients in trials tend to be biased in an important way – they are often very sick (especially in oncology). Which brings us to the dinner topic at ERIBA: is disease severity an accidental diagnostic in modern precision medicine? Are the marker levels chosen to accompany targeted medicines restrictive to the point where they unintentionally select only the sickest patients? This might be OK on the level of an individual drug, since it strives to only give drugs (expensive and with potentially nasty side effects) to those most likely to respond. But in aggregate across the industry, does it quietly bias away from moderate disease and towards the worst severity? Are we working our way into a corner that appropriately pairs medicines only at the extreme?

PS – Yes, it’s been a long time since the last post. I’ve been juggling a lot of travel with getting three papers out the door. The rest of the year will probably continue to be light on the blog. But teaching next semester will reduce my travel and give me plenty of opportunities to write blog posts while procrastinating lesson plans. 


Jacob Corn

Jacob Corn is the Professor of Genome Biology at ETH Zürich. Follow him on twitter @jcornlab.


  1. Ken Taymor

    Williams, et. al. identify another interesting clinical trial “bias” in drug development. Biopharma firms direct more investment to developing late stage cancer drugs (drugs for patients whose cancer has progressed to Stage 3/4 and who have a very short expected life span) than drugs that treat early stage cancers or that might prevent cancer.

    Two factors create this bias.

    First, drugs to treat late-stage cancers are less costly to develop than drugs for earlier-stage cancers because the clinical trials for late stage drugs get wrapped up more quickly. Patients in the trials have a short life expectancy, and the desired endpoints for positive outcomes are measured in months, not years, of extended life, so there is no need follow the patients for years as would be the case to demonstrate efficacy and safety of early stage or preventive therapies.

    Second, shorter clinical trials means more rapid FDA approval and a longer period between approval and patent expiration. This longer period of patent protection gives the company a longer monopoly period to control the price of the drug.

    The article is: “Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinical Trials.” American Economic Review 2015, 105(7): 2044–2085. I viewed it at

    from Ken Taymor, Berkeley Center to Advance Science in Policy & Regulation (UC Berkeley School of Public Health)

    1. Jacob Corn Post author

      That article is very interesting – thanks for bringing it to my attention.

      There’s also an argument to be made that late-stage patients are in more dire need than early-stage patients. But that might be part of a system-wide lack of focus on prevention. Curing early-stage cancer is hardly prevention, but shades of gray and all that.

      Acceptable safety thresholds set by the FDA are another factor. The agency is more willing to look the other way for side effects in late-stage patients who have no other options than in less dire cases. While still in industry, I once worked on a non-fatal neurodegenerative disease whose patient population was aged >70. A safety expert made it clear to me that, given the indication, the agency wouldn’t tolerate any reproductive side effects. Which seemed a rather optimistic outlook on the lifestyle of patients in that age group…

      PS – (in your comment I initially read “patent expiration” as “patient expiration” and was shocked)

  2. AdaB

    Precision medicine will be the trend for drug discovery and development as they provide tailored treatments toward a disease with less unwanted side effects. A concern I hold for this is the cost, which may be much higher than the common drugs cover all people as deep analysis will be done before the prescription. Hope there will be ways out on this issue.



Your email address will not be published. Required fields are marked *