May 18, 2022

yudha.xyz

go health

Simplify drug labelling to show benefits clearly

Aducanumab, an Alzheimer’s disease treatment, was approved by the US Food and Drug Administration (FDA) in June this year despite a lack of robust evidence that it actually slows cognitive decline. In protest, three experts quit an advisory panel. Amid all the soul-searching over how FDA approvals should work, something has been largely missing: a clear description of benefits.

The FDA (and the relevant community of scholars) knows that most new drugs offer only modest incremental benefit over drugs already available. But although its thorough analyses of submitted clinical-trial data are disclosed, the relevant documents are impenetrable to nearly everyone. Drug labelling should clearly state what effectiveness was demonstrated and how. The metrics should have real-world relevance, such as how a patient feels, functions or survives.

The FDA’s statute requires “substantial evidence” of efficacy, but that term is flexible. Clinical trials are increasingly using ‘surrogate end points’, or non-clinical measures that are ‘reasonably likely’ to predict clinical benefit. But one study found that 52% of 65 oncology-drug correlations were not strong (V. Prasad et al. JAMA Intern. Med. 175, 1389–1398; 2015). Aducanumab’s label reports “Amyloid Beta PET Composite SUVR”, a confusing measure that is hard to translate into, for example, a patient’s ability to recognize loved ones. If a trial outcome cannot be reasonably explained in patient-relevant terms, the FDA should reconsider whether that end point is acceptable for drug approval.

There is also no requirement for substantial efficacy, leaving patients guessing at a drug’s therapeutic value. I have scoured FDA documents underlying drug approvals that have been described in lay media as “life-saving”, “dramatically effective” or ”curative”. That is rarely what the evidence demonstrates. In 2021, I identified ten reports by multiple authors, each examining the evidence of benefit of a range of newly approved drugs (J. J. Darrow J. Manag. Care Spec. Pharm. 27, 685–688; 2021). Collectively, these papers concluded that only 2–31% had better than modest benefits over existing treatments. Aducanumab might therefore be an extreme example, but it is hardly alone.

The medical community attempted to fix a related labelling problem more than a decade ago. In the 1990s, practitioners protested that drug labelling was so lengthy and complex that they found it difficult to locate crucial information. Now the FDA requires that drug labelling include a half-page box at the top of package inserts, called ‘highlights of prescribing information’. It requires warnings and precautions to be included, as well as dosage and administration, but not, alas, how well the drug actually works.

In a report to Congress in 2010, the FDA concluded that clear efficacy labelling in a space-limited box might often be impossible for drugs because medical information is too complex and variable. Unlike me, few patients or physicians can devote full-time efforts to researching a drug’s value and its regulation, and might therefore be tempted to use price as a proxy for value. In my view, meaningful improvement is possible, and the dramatic increase in drug costs means that it is past time to revisit the issue.

The few drugs that provide understandable statements of efficacy offer a hint of just how useful such information could be. For example, the FDA labelling for Yasmin (drospirenone/ethinyl oestradiol), a birth-control product, explains that “pregnancy rates in the clinical trials were less than one per 100 woman-years of use”. Sunscreen — regulated in the United States as an over-the-counter drug — bears quantitative benefit information in the form of a single number: sun protection factor, or SPF. Although SPF is not a perfect measure and few people understand how it is calculated, repeated use allows consumers to learn over time what SPF they require to avoid burning.

Benefit cannot always be described with a single figure such as SPF or pregnancies per 100 woman-years. But nutrition labels could serve as a model. Their clear columns and uncluttered text show key information, such as calories and vitamin and mineral content, at a glance.

Two randomized controlled trials involving 450 people suggested that an analogous drug-facts box could greatly improve patients’ comprehension of drug benefits and risks (L. M. Schwartz et al. Ann. Intern. Med. 150, 516–527; 2009). Without it, 65% of patients overestimated benefits tenfold or more. For instance, most participants not given the facts box thought that the absolute risk reduction in heart attacks from a hypothetical drug was between 6% and 70%, whereas 72% of participants given the drug-fact box correctly perceived the actual value of 0.8%.

The European Medicines Agency has acknowledged the need to “provide more information on benefits in understandable terms”, but regulators around the world have been slow to deliver. I think part of the problem is that showcasing limited drug benefits would be awkward for regulators, health-care providers and manufacturers while also dashing patient hopes, however unrealistic. But there is an ethical duty to inform patients of benefits and risks.

Clearer labelling could also change drug companies’ incentives. Radical transparency would favour drugs that work best over those approved despite only marginal benefits.

Competing Interests

The author declares no competing interests.

https://www.nature.com/articles/d41586-021-03050-z