#EAPM: No place for poor science in personalised medicine

In these heady days of the rapid development of personalised medicine, clinical research is undergoing something of a seismic shift,

As Big Data is gathered from various sources right across the planet it is as clear as day that such data from clinical trials can have a profound affect on the lives of the potential 500 million patients here in the EU and, indeed, billions worldwide.

Of course, data sharing is a complex moral, ethical and practical area covering different levels of consent, personal privacy and agreed standards plus issues such as interoperability, costs of gathering and dissemination, silo mentalities and more.

The clinical research community has much to gain from the areas underpinning the growth in personalised medicine. Data-sharing is highly popular among patients as they seek to find cures for their own diseases and conditions as well as thinking ahead for the benefit of the patients that will follow.

Industry is getting heavily involved in Big Data in the field of personalised medicine with giants such as GlaxoSmithKline (GSK) and Intel investing heavily with time and resources.

Indeed, three years ago GSK floated the idea of a single system through which clinical trial data could be easily shared by sponsors. Now, in 2016, there are more than 3,000 trials listed involving 13 companies.

The pharmaceutical giant actively encouraged other sponsors to join in order that all could benefit from the infrastructure put in place. Clearly, among other things, a full commitment to data sharing was vital to its success and several other challenges presented themselves, such as some members calling for an option to deny requests in cases where there could be a potential conflict of interest or, of course, a competitive risk.

Because of this a ‘safety net’ has been put in place but has yet to be used, which is encouraging to say the least.

When it comes to such a system, many believe that it may work very well if all interested sponsors sent data and clinical trial details to an independent party to take care of privacy, scientific review and other matters arising. Among other benefits, this would certainly bring down costs.

The EAPM has watched events unfold with interest and what is certainly clear is that the scientific research community needs to find more (and better) ways of collaborating down the line to fully realize the potential of this new form of diagnosis and treatment.

There are dangers, of course, inherent in the results of clinical research. As long ago as 1962, New York psychologist Jacob Cohen stunned the scientific community. He analysed 70 articles published in a peer-reviewed journal in his particular discipline and came to the conclusion that the effects that the authors sought would only occur around one-in-five times, although most reported major positive results.

Unsurprisingly, Cohen concluded that many of these scientists were failing to record their unsuccessful research. OK, that’s perhaps not a shock, but there were even instances of ‘false positives’, which is a whole new ball game.

It’s not much different more than half-a-century further on. A more recent study put the figure at 24%, not much higher than Cohen’s one-in-five, or 20%, although much wailing and gnashing of teeth has gone on in the interim in a bid to get researchers to perform better when reporting results or lack thereof.

It seems that even the most honest researchers can accidentally produce sub-standard reportage of results, possibly due to incentives. And in one very recent study in psychology (2015) in excess of 200 researchers repeated 100 studies to try and reproduced the original results. They succeeded in only just over one-third of cases.

This clearly needs to be addressed. It would be major tragedy if the incredible potential of personalised medicine were to be undermined by what can only be described as ‘shoddy science’.