Let the Subjects Rule: the New World of Clinical Trial Data

For my talk at the Digital Health Festival run by the Digital Catapult.  

The Horrific Past

It is essential to conduct clinical research in an ethical way. We must respect each individual through informed consent and the investigator putting their welfare above all else. History show us the importance of ethical research by reminding us where medical professionals have done horrible wrongs in the past. For example, the notorious Auschwitz Nazi physician Josef Mengele and many other Nazi doctors found guilty of crimes against humanity for conducting deeply unethical research. These atrocities led to the Nuremberg Code in 1947 which later led to a more comprehensive set of ethical principles called the Declaration of Helsinki in 1964. The Declaration has been the central moral document guiding research practice ever since. However, the complexity of prescribing how clinical research should be conducted across all clinical settings in all countries around the world, means that the detailed implementation of the moral codes is now being established through Good Clinical Practice.

The Stormy Present

In June 2014, Facebook created a storm of protest by publishing a 'research study' into emotional contagion conducted with academics from Cornell and the University of California. The research was published in the journal of the Proceedings of the National Academy of Sciences in the US. At the time, Facebook said it was carried out "to improve our services and to make the content people see on Facebook as relevant and engaging as possible". The problem was, they deliberately manipulated News Feed content to change the emotional state of Facebook users in the real world without their informed consent. This led to the journal publishing an Editorial Expression of Concern, but the paper was not withdrawn because it was deemed that Facebook's Data Use Policy was sufficient for the participants consenting to the study.

The Ongoing High Cost

According to Medidata, 10 yrs ago in the US the average phase III program for a pharmaceutical put 1,400 people through clinical trials across 67 sites which cost about $86.3 million over an average duration of 2.5 years and each investigational medicine had a 70% probability of reaching market. Now, an average phase III program has at least 2,800 people across more than 100 sites, costing $235 million over the same 2.5 year duration but with less than 60% of the medicines likely to reach market.

The Lack of Interoperability

One of the big issues facing researchers is the need to collect high-quality data on clinical trial participants in a world where getting data from multiple disconnected sources (including paper medical notes) is not simple. But after the NHS Connecting for Health's National Program for IT (NPfIT) (the £10 billion effort to replace all NHS IT systems) was terminated in 2013, it seems the focus of the Health and Social Care Information Centre (HSCIC) on interoperability is a much more sensible idea. The thinking being that rather than building one great big overarching system, we should focus on connecting all the existing systems so data can be just passed from one to be used by another. However, interoperability is actually surprisingly complex because people and the diseases they get aren't that easy to classify. But with HL7 standards, the International Health Terminology Standards Development Organisation's (IHTSDO) SNOMED CT, the World Health Organization's 10th International Classification of Diseases and a few other clinical coding classification systems in use by the NHS, it seems that things are much getting much simpler now.

New Trial Models

Virtual Clinical Trials

In 2011, Pfizer tried to run a trial called REMOTE, also dubbed the trial in a box, entirely remotely to reach a wider population. Unfortunately they failed to recruit sufficient numbers of people with overactive bladder so they were forced to stop the study. But this does show us that companies are trying new approaches to try and drive costs down and research quality up.

Real-World Data

GlaxoSmithKline is successfully conducting the Salford Lung Study with a research team called North West e-Health (NWeH), part of Salford Royal NHS Foundation Trust and the University of Manchester. They have developed an electronic patient record system that enables them to collect detailed clinical information about study participants' episodes of care in the real world, reducing the need to rely on the participants' recall of this information at the next study visit. patientslikeme, a community of patients who share their data over time to give and receive support from each other, is another example of using real-world data. Patients can be invited to participate in clinical trials by simply giving permission to researchers to have access to their data, including .

Big Data

During the 2010 US election, Facebook ran a 61-million-person experiment in social influence and political mobilization where users were shown a non-partisan 'get out the vote' message with useful supporting information. Users were able to click an 'I voted' button, so their vote and their profile picture were shown to their friends within their friends' 'get out to vote' messages. The results were published in Nature (without a big a storm of protest) and showed that clicking the 'I voted' button was more likely to influence close connections in a positive way to vote, compared with 600,000 control users who saw a plain informational message without their friends' 'I voted' responses.

The Future

Apple ResearchKit

In March 2015, Apple announced it was enhancing the previously launched HealthKit and Apple Health with an open framework for clinical research called ResearchKit. ResearchKit enables researchers to develop apps, within the proprietary Apple ecosystem that includes the Apple Watch, for people to download on their phones that enable them to participate in clinical research studies. For example, MyHeart Counts was a trial announced by Stanford Medicine alongside the launch of ResearchKit by Apple. The trial recruited 11,000 patients within the first 24 hours, beating their 10,000 target. A feat that would take most clinical studies a year to accomplish according to the trial medical director.

Google's health-tracking wristband

Last week, Bloomberg announced that Google[x] (the team known for Project Loon, Google Glass and Google Diabetes Contact Lenses) are making a health-tracking wristband specifically for research purposes. An activity tracker with research-grade, continuous measurements of pulse, heart rhythm and skin temperature, and also environmental information like light exposure and noise levels. I suspect this device will enable researchers to develop apps that are dedicated to more critical areas of healthcare, perhaps in hospital or for people under direct medical supervision. Oh and use Google Fit, Google's alternative to HealthKit.

Wearables

Wearables aren't just for the Quantified Selfers, fitness freaks or Apple Fanboys. They are quickly evolving to enable all of us to get fitter, avoid and manage ill health better and provide the data needed for the next generation of medical innovations. Crucially, this evolution isn't being caused by the traditional healthcare providers, it is being caused by the consumer when they choose something as simple as their next phone or watch. It is research subjects (now called 'people') who rule this new world, and we need to help them remember the lessons of the past.

Missed something important? Add your thoughts by commenting below. Or get in touch with the medDigital team.