Data from more than half of cancer trials underpinning drug approvals remains inaccessible, according to a shocking new study that audited over 300 clinical trials backing anti-cancer medicines approved by the US drug regulator over the last 10 years.

Of the 304 industry-sponsored clinical trials that provided data for 115 anti-cancer drugs approved by the US Food and Drug Administration (FDA) between 2011 and 2021, only 45 percent had publicly shared patient-level data – or said they would do so when the researchers enquired about accessing data.

What's more, less than one-tenth of trials cited on product labels of three of the top-selling anti-cancer drugs made individual patient data available, the study found.

Sharing anonymized individual patient data – in secure data portals and deidentified to protect participant privacy – is vital for studies known as meta-analyses that pool data from published trials to assess the balance of evidence for new and existing treatments.

"Pharmaceutical companies are not going to provide this data to other pharmaceutical companies because of competition, so this work has to be done by independent researchers," pharmacist and study author Natansh Modi of Flinders University told The Guardian.

"But it can't be done without the data and transparency."

Aside from meta-analyses, data sharing also enables researchers to make use of existing data, replicate research studies, and validate published results – all of which helps strengthen public trust in science.

And it almost goes without saying that rapid data sharing has been priceless during the COVID-19 pandemic, informing public health responses and expediting research, treatments and vaccines.

"Although inroads have been made toward improving [individual patient data] transparency over the past five years, these findings suggest that a substantial portion of pivotal oncology trials that support the FDA registration of modern anti-cancer medicines remain unavailable to qualified researchers," writes Modi and colleagues.

This is certainly not the first time the US FDA, which approves new drugs and medical devices, has come under scrutiny.

Recently, its controversial decision to approve a new drug for Alzheimer's disease sparked an outcry from health professionals who claimed industry-led clinical trials had not yet demonstrated the antibody therapy – now on the market for around US$56,000 per year – would slow memory loss or cognitive decline.

The drug had been approved through the FDA's accelerated approval pathway whereby, if an innovative new drug is considered safe and the therapeutic need for improved treatments for a given disease is considered great, the agency may approve it based on limited evidence (although pharmaceutical companies must supply more efficacy data from real-world studies shortly thereafter).

Although there are many reasons why clinical trial investigators might hold off sharing data from individual trials, the researchers behind this new study argue that the general public expects greater transparency from the billion-dollar pharmaceutical industry that has a responsibility, along with drug regulators, to ensure safety and efficacy of new medicines.

The most common reason industry sponsors did not share trial data was that long-term follow-up of study participants was ongoing, the study found.

"Ongoing follow-up is, of course, needed, but it should not hinder the release of the key data that is implicated in the global release of medicines to tens of thousands of people," says pharmacist and study author Ashley Hopkins, also of Flinders University.

With only summary-level data from trials, meta-analyses are inherently unreliable because researchers cannot interrogate raw data, which might – as it did in the case of ivermectin – inadvertently let flawed studies slip through the cracks and skew the results of data-crunching, sense-making meta-analyses.

"A poorly scrutinized evidence base supported the administration of millions of doses of a potentially ineffective drug globally, and yet when this evidence was subjected to a very basic numerical scrutiny it collapsed in a matter of weeks," Jack Lawrence and four colleagues wrote in Nature Medicine after exposing the flawed studies.

Based on their study, Modi, Hopkins, and colleagues are calling for greater data transparency on pivotal trials "to protect and maximize public health and ensure the contributions of trial participants and their families reach their full potential."

"If the data isn't made available, it can't be put to good use," Modi adds.

In recent years, scientific journals and research funding bodies have introduced policies to require or mandate data sharing amongst researchers. In 2014, the American and European pharmaceutical associations also made a commitment to share anonymized trial data upon request.

Although some researchers have reported a noticeable shift in researchers' intentions to share data during the pandemic, it appears that issues of data transparency persist – at a time when public trust in science is needed more than ever.

The study was published in JAMA Oncology.