Thursday, November 12, 2015

Uncertain Health in an Insecure World – 64

Either Good or Bad

In Act 2 Scene 2 of Hamlet, the Prince of Denmark laments to his spies, Guldenstern and Rosencrantz, that “… there is nothing either good or bad, but that thinking makes it so.” This statement, the fourteenth century equivalent of the modern, “Psych!”, revealed a Hamlet deeply troubled by murder, incest and ghosts in the Danish royal family. He adapts to all these negatives by putting on a mask of “antic disposition” – feigning madness as he walks the corridors of Elsinore castle.

Not so long ago, in his 2012 book Bad Pharma, author Ben Goldacre (below) nobly took on medical journal publication bias towards studies with “positive” results over those with “negative” outcomes. Former U.S. Food & Drug Administration (FDA) official Henry Miller responded to Goldacre’s angry antics, correctly pointing out that many negative studies are simply statistically underpowered – as such, they should best be called “unsuccessful” or “inconclusive”.

A truly negative clinical trial is one that shows a new treatment is actually inferior to standard therapy. In fact, the FDA often approves new drugs on the basis of non-inferiority compared to a standard treatment. Of course, an underpowered study that demonstrates non-inferiority might actually prove to be negative if adequately powered to conclusively disprove the null hypothesis (i.e., that a new drug is the same as a standard drug or a placebo). 

Goldacre’s book and medical literature meta-analyses provoked a lot of soul searching in the research and regulatory sectors.

For example, a 2012 Nature article showed that the results of 47 of 53 published molecular biology papers couldn’t be reproduced by another scientist. The following year, noted that half of all clinical trials go unpublished and many omit key details, despite U.S. FDA rules requiring all results be submitted to the database (above).  A recent New England Journal of Medicine article (March 12, 2015) by Duke University investigators showed that only 13.4% of >200,000 clinical trials reported since 2000 published summary data within one year of completion. Two-thirds of these studies were funded by industry; industry funding increased the likelihood of timely reporting. 

Has any of this existential angst about publications changed practice at the point-of-care?

When medical practitioners use new drugs to treat chronic diseases – infections, diabetes, hypertension, arthritis and cancer – they experience responses in their clinic patients (below) that are not the same as those in clinical trial subjects. Why? Because clinical trials are highly managed protocols carried out on carefully selected cohorts, they may not predict the responses of patients who are less adherent and more heterogeneous.

Before becoming generic in 2012, some 50,000,000 Americans had used the anti-platelet drug clopidogrel (Plavix™) for cardiovascular illnesses at an average cost of $200 per month. Since 2008, the FDA has monitored the value of dual anti-platelet therapy (DAPT) with both aspirin and clopidogrel after drug-eluting stents are placed in blocked coronary arteries. Last week, an FDA meta-analysis of multiple clinical trials concluded that DAPT, “does not save lives, reduce heart attacks, or result in lower risk of stent thrombosis.” Such FDA-mandated post-marketing surveillance in general populations often causes early enthusiasm based on randomized clinical trials (RCT’s) to wane. This late-breaking medical news makes patients, like my mother who nearly bled out on DAPT, MAD!

Today, there is nothing “rotten in the state of Denmark”, a modern Nordic country of some 5,678,348 souls.

Denmark is home to a global diabetes pharmaceutical giant, Novo Nordisk (NVO), headquartered in Bagsværd. Although the prevalence of diabetes in Denmark is low (above) and less than that in Europe (52M) and around the world (387M) – this 90-year old company has intentionally focused its R&D efforts on diabetes care. Its long-acting insulin, Levemir™, has been a big seller. In late September 2015 (below), the FDA approved NVO’s longer-acting Tresiba™, after “head-to-head” studies versus Sanofi’s market leader that showed Tresiba™ was non-inferior to Lantus™.

Another progressive Denmark-based firm, Genomic Expression, is becoming a first-mover in the field of companion diagnostics by making RNA sequencing (OneRNA™) gene expression targets actionable in oncology patients. In addition to high initial cancer therapy failure rates, the inconvenient truth is that only one in four current cancer treatments prolong life. OneRNA™ is a precision medicine (PM) approach that identifies a positive chance of treatment response in triple-negative breast cancer and other diseases.

Full disclosure – I have brilliant professional colleagues who work at both of these highly progressive Danish firms.

On the verge of despair amid negative and/or irreproducible RCT’s, in July 2015 along comes The Economist’s brilliant data shop with a clinical trial simulator  This interactive "run a trial" exercise (above) makes the point that there is a critical need to fix a situation where only regulators see all the good and bad data (below, left). The selective publication of only positive clinical trial results is highly flawed (below, right). When doctors only see what under-reporting Big Pharma and clinical trialists want them to see, there is a 14.1% distortion!

To be or not to be?… that is the question.” Patients need to stop flying blind related to the applicability of RCT’s to their healthcare. 

What Institute for Systems Biology founder Roy Hood calls “dense dynamic personalized data clouds” offer the potential for individualized PM, instead of relying on averaged data responses from RCT patient cohorts. In the wake of 2012’s rotten revelations, it will be good when “N of 1” data-driven PM care trumps that based on notoriously underpowered and inconclusive clinical trials. 

So when it comes to, “the heart-ache, and the thousand natural shocks that flesh is heir to…”, here’s “the rub...

Unlike Hamlet, who was unable to escape “the many confines, wards, and dungeons” of Denmark, we in the Square think it good to face such real-world clinical drug trial ambiguities, then shuffle them off from our medical practice thoughts and troubled treatment failure dreams. 

Anything else is antics, amid a modern Shakespearean tragedy.

No comments:

Post a Comment