Horizon scanning: How shoddy clinical trial reporting undermines health policy making

This guest blog is by Ben Porter, a patient registry project manager and curator with the John Walton Muscular Dystrophy Research Centre in England. He previously worked as an Evidence Synthesis and Horizon Scanning Specialist with the NIHR Innovation Observatory. He contributed this guest blog in a private capacity.

Who do clinical trial registries benefit? Most people would immediately answer “patients,” but in fact, their usefulness extends far beyond informing individual treatment choices.

Public health bodies routinely use trial registry data to make plans for future service delivery, and to decide what promising areas of research to prioritise. Access to reliable information on clinical trials is pivotal for the future provision of medicines in the UK.

However, clinical trial registries are currently riddled with incomplete, inaccurate and out-of-date data. I know, because I used to search registries for the NIHR Innovation Observatory, the horizon scanning centre for England.

Horizon scanning is the identification and analysis of new and emerging health technologies to improve policy making. For example, the Innovation Observatory informs NICE about new medicines that are in advanced clinical trials or are likely to enter the market within the next three years. In addition, stakeholders such as the NIHR Biomedical Research Centres, NHS England Specialised Commissioning and the Office for Life Science, often request horizon scanning reports for certain population groups.

Health care providers need this intelligence to be able to plan ahead, as new disruptive technologies may significantly benefit a population or require a change in service delivery. In some cases, horizon scanning is also used to inform funding decisions, notably to accelerate research in a particularly promising field.

Clinical trials are fundamental to horizon scanning because they are the primary source of information on new and innovative medicines, repurposed medicines in new disease areas and patient groups, and new formulations of existing medicines.

In theory, registries should provide a clear picture of what development stage promising innovations are at.

In practice, pharmaceutical companies and universities often fail to update their trial registry entries. Often, a trial is falsely listed as “ongoing” even though it was completed long ago, or it is listed as “completed” on one registry and as “ongoing” on another. If a trial has been completed, its summary results are often missing, and journal articles discussing the trial’s findings have not been linked. At times there are even discrepancies between protocols and various registries regarding the treatment procedures and doses used in a trial, so we cannot even tell with certainty what exactly is being tested.

For example, take a recent BBC news headline: “Radical Parkinson's treatment tested in patients”. About 145,000 people in the UK have been diagnosed with Parkinson’s, which at present cannot be slowed down or reversed, so this new treatment could potentially have huge implications for patients and the NHS in future years.

So let’s look at the information available about this trial on public registries. As usual, the picture is messy. According to the American registry, Clinicaltrials.gov, the trial was completed in April 2016, but on the European registry, the same trial is still listed as ongoing.

According to EU guidelines, this trial should have posted its summary results onto the European registry within 12 months of completion, i.e. by April 2017. However, the trial’s sponsor, North Bristol NHS Trust, never uploaded the results.

Other scientists looking for an effective treatment for Alzheimer’s had to wait until February 2019, when the results were finally published in an academic journal, to learn what their peers had discovered.

Such unnecessary delays in making trial results public considerably slow down the pace of medical progress.

Multiply this single trial by the over 34,256 trials listed on the European registry, and the whopping 236,713 trials on its American counterpart, and you get an idea of what horizon scanners and other registry users are up against.

Registries were established to make information easy to locate and share, but due to the failure of many sponsors to keep their entries up-to-date, horizon scanners and other registry users still waste huge amounts of time contacting sponsors directly for accurate information on their trials.

I was glad to hear that the House of Commons Science and Technology Committee is now writing to trial sponsors – including to North Bristol NHS Trust – to remind them to post missing trial results. Going forward, the Committee and other stakeholders should also develop mechanisms to ensure that sponsors regularly and routinely update their registry entries.

Trial registries can only fulfil their promise if entries are complete, accurate and kept up to date. This would provide clarity for everyone – be it patients, academics, clinicians, industry, or horizon scanners.