Out With the Old, in With the New: The Role of Novelty Bias

Tejas P. Desai, MD


July 12, 2019

Bias is present in many aspects of a healthcare professional's work. Whether working as a clinician, educator, investigator, administrator, or learner, the physician must constantly mitigate both conscious and unconscious biases. Cognitive biases come in many forms and include selection and reporting biases. I'd like to call your attention to novelty bias. This bias plays an important role in the way clinicians receive and operationalize scientific data. I'm defining novelty bias as a proclivity to exalt a new treatment at the expense of successful historical interventions. For me, this bias became increasingly apparent in two recent situations.

Nephrology Game-Changers

The CREDENCE[1] and SONAR[2] clinical trials focused on novel treatments for diabetic kidney disease—SGLT2 inhibitors and endothelin receptor antagonists, respectively. I admit that both are game-changers in our field and will make a positive impact on the lives of our patients (CREDENCE more so than SONAR). I wondered how many of the individuals who participated in the many discussions about the data had a novelty bias.

As a "control," I looked back at the PREDIAN trial,[2] published in 2015. Similar to CREDENCE and SONAR, PREDIAN showed a renoprotective effect of pentoxifylline (when combined with renin-angiotensin-aldosterone [RAS] inhibition) over a 24-month period versus RAS inhibition alone. The primary outcome of estimated glomerular filtration rate (eGFR) over time was also measured and heralded in CREDENCE and SONAR.

Despite the positive results, PREDIAN did not bring pentoxifylline to the fore in the way that CREDENCE and SONAR have done for SGLT2 inhibitors and endothelin receptor antagonists, respectively. This novelty bias is clear in the social media conversations among nephrologists, with less than 1% of all conversations discussing pentoxifylline, despite PREDIAN's 4-year head start (Figure 1).

My second encounter with novelty bias came from a study of veverimer,[3] a novel agent that binds gastric hydrochloric acid. It is theorized to be renoprotective in chronic kidney disease (CKD) patients who suffer from acidosis. Such patients have been shown to have a greater likelihood of kidney function decline (whether by eGFR or creatinine clearance) if their acidosis is not corrected (bicarbonate > 23 mEq/L). Two earlier trials on patients with CKD and acidosis showed a slower decline in kidney function with sodium bicarbonate and made a strong argument for the treatment of acidosis with sodium bicarbonate.[4,5] Indeed some of the greatest fears of sodium bicarbonate treatment—ie, an increase in blood pressure or episodes of heart failure—were allayed by these studies.

I was surprised when the veverimer paper crossed my desk because I did not think an agent other than sodium bicarbonate was needed to treat acidosis in patients with CKD. Admittedly, sodium bicarbonate is difficult to ingest and adherence may be less than desired, but the earlier trials convincingly showed kidney protection with minimal adverse effects. In the contemporary study, patients with acidotic CKD were randomly assigned to an oral suspension of veverimer twice daily versus placebo. The primary outcome was the change in or absolute value of serum bicarbonate during a follow-up period of only 12 weeks. Given its mechanism of action, it should not come as a surprise that veverimer outperformed placebo (Figure 2).

Veverimer will need to be further investigated to determine whether it has any direct renoprotective effect. I wonder, however, if future veverimer studies will cultivate and spread novelty bias. PREDIAN and the two sodium bicarbonate studies were investigator-initiated and assessed generic and relatively cheap therapeutics. While the study designs mitigated a number of biases that commonly affect researchers, the nature of the therapeutics may have, inadvertently, promoted a novelty bias.

Don't get me wrong: I am excited about the data in all three contemporary trials (CREDENCE, SONAR, veverimer). It's not the enthusiasm for the new agents that concerns me but rather the limited awareness of historical successes with "older" agents that fuels novelty bias. Should it be the responsibility of writers (such as myself) to identify and quell novelty bias? Should the investigators be expected to predict and then shoulder this burden? Or should the knowledge consumer (ie, you, dear reader) evolve into a knowledge contributor and offer historical perspective on reports of novel agents? If you believe it's the latter, the comments section below is open for you.

Follow Tejas P. Desai, MD, on Twitter: @nephondemand

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.