Medicines that work great in a petri dish or a mouse might not be effective when deployed in people.

That’s why clinical trials, which test new therapies in actual patients, are so important.

But all too often, clinical trials are never finished or never published, so their results are lost to the medical community and never revealed to the patients who volunteered themselves for study.

According to a research paper published earlier this month in the journal Pediatrics, this happens just as often in trials involving children as it does in those involving adults.

The study authors combed through two years’ worth of pediatric trials registered on clinicialtrials.gov, a national registry run by the U.S. government.

They found that 19 percent of these trials ended early, and 30 percent of completed trials were never published.

That means that more than 77,000 children were enrolled in trials that went nowhere.

Read more: Lack of mental healthcare for children reaches ‘crisis’ level »

The issues with children’s studies

“I think when we embarked on this study we hypothesized that trial discontinuation and nonpublication would be somewhat common, but we really had no idea how common it would be,” Dr. Natalie Pica, an author of the study, and pediatric resident at Boston Children’s Hospital, told Healthline.

Testing new drugs in children is fraught with issues of consent, safety, and profit.

In fact, drug companies have been so reluctant to do these trials that Congress passed laws in the early 2000s incentivizing pediatric research and giving the Food and Drug Administration (FDA) the authority to require companies to test new drugs in children.

Yet pediatric studies are still relatively uncommon. Right now, there are about 47,000 pediatric trials registered on clinicaltrials.gov, compared with more than 200,000 adult trials.

With so few clinical trials focused on children, it’s all the more important that their results be shared, Dr. Christopher Jones, an assistant professor at the Cooper Medical School of Rowan University, told Healthline.

Jones, who performed a similar analysis, but was not part of the Pediatrics study, said the study’s results are “disturbing,” if not surprising.

His study found that 29 percent of clinical trials, including adults and children, went unpublished, representing nearly 300,000 people.

Read more: Pediatric group cautions against violent video games »

A push for action

In recent years, medical organizations and government officials have begun to push for the dissemination of results.

In 2007, Congress passed legislation requiring that the results of most trials testing FDA-approved drugs and devices be uploaded to the clinicaltrials.gov site, although compliance has been low.

The following year, the Declaration of Helsinki, a set of guidelines put forth by the World Medical Association, called upon scientists to “make publicly available the results of their research” no matter what the results were.

And last year, the European Medicines Agency (EMA), the European analog to the FDA, announced it would begin releasing the clinical trial reports associated with any new drug that it approved.

Read more: Should we believe medical studies? »

Why studies end early

So with all these efforts to make the most of clinical trial data, why is so much of it still being lost by the wayside?

Sometimes the reasons for a trial to end early are legitimate.

The new medicine might be associated with a lot of bad side effects, or new information may come out during the course of the trial that suggests the trial isn’t necessary.

In these cases, ending early actually prevents unnecessary suffering and waste. Pica and her co-author called this “informative termination,” and it accounted for 13 percent of the trials that ended early.

But the reasons for not publishing are harder to defend.

“I can say for the nonpublication rate I really think the rate should be zero,” Pica said.

Sometimes scientists are reluctant to publish data for trials where results were neutral or negative.

Science has a history of favoring studies that show what does work and ignoring studies that show what doesn’t work. But both results are equally important, says Jones.

“Because ‘negative’ trials are generally less likely to be published than ‘positive’ trials, nonpublication tends to distort the entire medical literature to make a given intervention look more effective than it really is,” he said.

Other times, it’s a proprietary issue. The Pediatrics study found that trials funded by industry rather than academia were more likely to be completed, but less likely to be published in a peer-reviewed journal.

Read more: Why some cancer research information isn’t shared »

Like an iceberg

Peter Doshi, associate editor at The BMJ, and an assistant professor at the University of Maryland, compares the wealth of unpublished trial data held by industry and others to an iceberg.

Almost all of it is below the surface and inaccessible to outsiders.

To fix that, he proposed an initiative called Restoring Invisible and Abandoned Trials (RIAT). The initiative is “a call to publish — or be published.”

If researchers and their sponsors don’t attempt to publish the results of their abandoned trials within a year, then outside investigators should do it for them, Doshi and colleagues suggest.

Doshi told Healthline that with newfound access to trial data, such as that provided by the EMA’s recent decision, the scientific community finally has the means to address the problem of nonpublication.

“Rather than just complain about it, because we’ve known about the problem for a long time now, we can actually do something about it,” he said.