Unleashing the power of data-driven medicine

Genomic data can now be analysed and used to diagnose conditions and tailor treatments to individual patients, but concerns over broader data sharing and security need to be overcome for the full benefits to be realised.

The recent commercial deal between Roche and Foundation Medicine, along with the establishment of the 100,000 Genomes Project in the UK, made one thing clear: the genomics and analytics revolution in medicine is now a reality. Data-driven medicine is being used in a number of European countries, offering patients with chronic conditions, or at risk of developing them, the best and most accurate interpretation of genetic data possible to optimise treatment and outcomes.

Digital technology has enabled bioinformaticians to analyse genes through Next Generation Sequencing (NGS) and extract detailed individual information. This data can be used to diagnose, manage treatment or prevent risk.

It can be used in cancer, where analysis of genetic data can illustrate whether or not people carry mutations of the BRCA1 or 2 genes, which are associated with an elevated risk of developing breast or ovarian cancer, for example. Preventative action can be taken following the discovery of such data, and the same benefits can be delivered to immediate relatives who may also be at risk from a similar mutation.

Hospitals and laboratories collect and manage huge amounts of raw genetic data that help understanding of chronic diseases and their treatment. Effective analysis of this 'dirty' data is now possible thanks to powerful tools like machine learning systems. These develop and deploy advanced algorithms that improve the interpretation data, making sense of the information captured by genetic sequencing and providing doctors with the information they need to better diagnose and treat patients.

"Speedy and highly accurate interpretation of genetic data can now identify hundreds of genetic disorders"

In addition to addressing genetic susceptibility to cancer risk for parents of new-born babies, speedy and highly accurate interpretation of genetic data can now identify hundreds of genetic disorders like cystic fibrosis and define personalised medicines effective for particular genetic profiles. Similarly, for young children with congenital defects such as heart conditions, understanding the specific make up of their genetic profile is critical for doctors in making the right treatment choices.

Unlocking these improvements lies in not only in deploying the technology, but also in a collective approach to interpretation, learning from experiences in different countries and healthcare systems. The revolution in using data in this way has been made possible because laboratories have started to share patients' anonymised data and genetic information, to learn from each other. This collective approach is using real world experience to reduce the number of 'unknowns' – improving understanding of how mutations of genes or variations in genetic data are associated with the course of chronic diseases. As a consequence, data from today's patients might help treat and improve the prospects for the patient of tomorrow.

However, technical, political and some psychological barriers remain to the more widespread adoption of data-driven medicine. The first major issue relates to trust: the safe storage and exchange of data will always be high on the list of concerns. Recent high-profile government and private sector data breaches mean the public and healthcare professionals are right to remain cautious over the right way to handle sensitive information.

"Often, data is siloed in the hospital unit treating patients, or in the laboratory researching a specific disease or complaint, and there is no complete patient picture"

Sharing data is crucial to improving understanding of disease management, but it is not always a straightforward process. Often, data is siloed in the hospital unit treating patients, or in the laboratory researching a specific disease or complaint, and there is no complete patient picture. This can reduce the usefulness of the data in predicting outcomes – or indeed restrict its role in improving overall understanding of disease management.

On occasion, patient information can't be shared because of stringent national regulations. Some European countries are more constrained in what can and cannot be shared. There is cause for optimism, with good progress underway in Switzerland regarding legislation on the use of patient records, for example, but a lot remains to be done. Initiatives like those undertaken by the Global Alliance for Genomics and Health and EuroGentest in Europe are pushing in the right direction, but a change of attitude is needed to ensure the full benefits are realised.

Another issue is the need for agreed standards. This relates to standardisation of data and results and also to encryption standards and storage. Data must be held in a standard way to allow effective comparison and usage. This challenge is more a matter for the private sector than regulators and governments. There is a great opportunity for whoever generates the right standardised solution for capturing, handling and analysing genetic data – becoming the 'Google' of data-driven medicine and establishing the path for others to follow.

As is often the case in healthcare, sufficient investment and funding can be a barrier to the more widespread adoption and use of genetic data. The more that data is used in healthcare, the more obvious it is becoming that governments should ring-fence money in their budgets for this purpose. On the surface this sounds like an additional cost burden, but such investment can frequently lead to more effective use of medicines and more effective preventative action to reduce the burden of ill health, and thus cost, in the first place.

Analysing data relies not only on machines and advanced algorithms but also on the bioinformaticians. Governments and education systems need to ensure sufficient numbers of young people enter training and are inspired by careers in the sector as the requirement for experts in the field grows.

Finally, the proliferation of wearable technologies suggests steps should be taken to harness the data generated by these devices to improve understanding of every individual's lifestyles and genetic make-up to improve treatment and prevention. By combining big data with smart data, healthcare can further personalise treatment and improve outcomes. In the end, genetic data only attains its full potential when pooled with information from other sources.

The data-driven medicine revolution is delivering a step change in the treatment and understanding of disease. However, there are a myriad challenges and barriers still to overcome. Through working together, the private sector, regulators, payers and healthcare providers can help unlock the possibilities. The future for chronic disease management and prediction remains positive, with patients the ultimate beneficiaries of improvements in our understanding.

About the author:

Jurgi Camblong is an entrepreneur and co-founded Sophia Genetics in 2011 with Dr Pierre Hutter and Prof Lars Steinmetz, where he is CEO. He holds a PhD in Life Sciences (University of Geneva) and an EMBA in Management of Technology (EPFL-HEC Lausanne).

Have your say: How can governments and industry build public trust in data-sharing?