Genomics can bring significant benefits to healthcare systems, by accelerating clinical research and drug development, personalising treatment. But, the cornerstone for genomic data leadership is a robust data strategy, to support drug R&D, meet increasingly tough regulations and preserve ethical standards. As per a KPMG report, this, along with the use of the most appropriate technologies — to process large data volumes and generate insights — should enable life sciences companies to become more fleet of foot in discovering and taking forward new drugs. Excerpts
As regulators bed down rules on privacy and consent, R&D heads are thinking about how to speed up the pace of innovation. The future research agenda, therefore, should address the commercial, regulatory, technological and ethical challenges of getting new products to market quicker.
Strategies to take control of genomic data: five pillars
All life sciences companies need robust internal policies for keeping data safe, secure and reliable; policies that extend to all customers/patients, suppliers and partners, ensuring anonymity and pseudonymity, and avoiding any data tampering. With fake news, data leaks and breaches hitting the headlines, trust in data has become a hugely important issue. In this section, we outline the five pillars of a robust genomic data strategy.
Standardised sequencing and analysis
The science behind genomics should be reproducible, to ensure that results are reliable and can be compared. Reproducibility is the ‘litmus test’ of robust science. Regrettably, R&D units and sequencing labs often sequence on different platforms, using a variety of software and analysis techniques. Without a common standard, something as simple as the amount of time a blood sample spends in a centrifuge, or the temperature during testing, can distort findings and prevent meaningful comparisons.
With more and more community platforms being used to access biobanks and other repositories, there are literally dozens of strategies available to sequence genomes. According to one study on technical analysis: “Each approach makes trade-offs between the cost of sequencing, time to results, and type and frequency of errors. This means that different approaches may produce different results and these differences may have important clinical implications. To move toward precise genomic medicine, we must be able to reliably sequence and decipher the difficult regions of the genome.”
There is some hope that from a data perspective, use of blockchain technology can help bring greater consistency to sequencing during clinical trials. With its immutable nature, this technology helps to ensure traceability and data integrity, so trials can be automated and standardised.
To maintain and foster individuals’ trust, personal data should be used for limited purposes and in line with the expectations set with the individual at the point of data collection. Privacy notices should be clear and accessible to individuals whose genomic, health and lifestyle data is being collected. Due consideration should also be given to the best format for notice requirements: for instance, ‘Just in Time’ notices can go a long way towards meeting the transparency requirements for individuals using wearable devices. Organisations using genomic data must also consider the most suitable legal basis for processing the data (with attention given to the GDPR requirements and other privacy laws applying to the organisation), and establish their ethical position on using individuals’ data for medical research.
Companies should appoint a Data Protection Officer (DPO) to oversee IT, Legal and other functions involved in processing genomic information, to vouch for reliability, trustworthiness and completeness of data, and confirm legal consent for its use. One of the key governance goals is to balance the desire for new products with a commitment to accurate and credible trial reporting, to avoid any inaccuracies that could backfire in future.How data is processed is not just a regulatory matter — security should also be considered.
Life sciences companies and the industry together will need to commit more resources to managing cyber risk as the threat continues to evolve — with high stakes. European governments have regular meetings with major life sciences companies to participate in cross-industry working groups.
Cyber Security is everyone’s responsibility — and it starts at the top. Leadership and all members of the executive management team should be committed, and that commitment should radiate throughout every level of every department. Best practice for cyber security involves raising awareness, performing training and simulation exercises, monitoring threats,
assessing and detecting vulnerabilities, establishing processes to address weaknesses, adopting disclosure policies, and building systems that mitigate cyber risks.
In safeguarding data, the guiding principle is to constantly assume the possibility of a breach. Cloud technology providers are increasingly expected to provide a high baseline of security. Consequently, Life Sciences companies and their cloud security providers should be carefully assessed to ensure compliance with HIPAA (Health Insurance Portability and Accountability Act of 1996), the GDPR, and other regulations to minimise the impact of data breaches. In assessing the security of cloud technology providers, Life Sciences companies should also check for encryption techniques for transmitting and stationary data, assurance over third party contracts and security vetting of employees and other insiders.
Managing access to genomics data
Identity and access management technology is increasingly important for managing general and privileged access to assets for employees, customers and other third parties. This technology is important not only for security, but also to streamline digital transactions across the Life Sciences ecosystem.
Think ahead and make practice a priority
No system can ever be completely secure, so continuous threat monitoring and regular testing of organisational cyber security practices is a must, to stay on one’s digital toes and avoid any security or privacy lapses. A crisis response plan is another important safeguard, to build the capability to respond swiftly and appropriately should a breach occur, to stop any further leaks, and communicate openly to stakeholders and the media.
Storage and data transfer
Research is typically conducted in multiple locations, so a central team of data managers should continuously monitor how clinical data is captured, to help verify the accuracy and reliability of source data, with alerts triggered when inconsistent data patterns are spotted.
The global nature of research means that data is often collected, processed, and transmitted in multiple countries, and then integrated into a global clinical trial database. With the GDPR’s restrictions on international data transfers, organisations need to ensure that they have adequate safeguards for the protection of personal data that is sent outside the EEA (European Economic Area), paying particular attention to model clauses, ad hoc contractual clauses, etc. And as genomic data volumes expand, storage becomes a bigger issue. With trials typically using petabytes of data, the bandwidth required to move such huge amounts from remote cloud-based data centers — often inlower-cost countries — is impractical.
In essence, it has become harder to move sensitive personal genomic data across borders, for both regulatory and cost reasons, so Life Sciences companies have rethought the footprint of their data centers and located them closer to the R&D facilities. Realistically, most, if not all will be shifted to global cloud providers, who can offer secure storage closer to the point of use, enabling safe transfer of massive volumes of data.
Harnessing analytical technology
With rare diseases and oncology set to benefit greatly from genomics, the role of technology cannot be underestimated, with vast and increasing volumes of genomic data involved in clinical trials, decision support during treatment, and payments. R&D is likely to require fast, powerful and incisive algorithmic capabilities involving specialised tools.
For example, deep learning algorithms have already demonstrated revolutionary achievements in the field of AI (e.g. image recognition, object detection, audio recognition and natural language processing). The intersection of deep learning and genomic research offers huge promise in understanding human disease, particularly with the introduction of high-throughput sequencing. Life Sciences companies need to acquire and/or gain access to such skills — as well as huge computing power — which is arguably only realistically available via the cloud.
A second example includes Microsoft’s Project InnerEye, a research project that uses machine learning technology to analyse 3D radiological images, including those containing cancerous tumours. The process is designed to help radiologists save time and cost. In a further example of AI, SOPHiA GENETICS has developed a universal technology for genomic data analysis. Its aim is to share knowledge and enable genomic testing worldwide through its software-as-a-service platform, enabling processing and analysing of raw genomic data to help diagnoses in oncology, metabolism, pediatrics, cardiology and hereditary cancer. This so-called “democratisation of data” encourages collaboration from clinicians to improve diagnosis and treatment.
Finally, we have the emerging ‘digital twin’ technology, which attempts to simulate a whole human through genomics, physiology, lifestyle and environment, using genetic information overlaid with other data from IoT. Potentially, we can then simulate and predict healthcare outcomes and evaluate which drugs may be most appropriate for particular conditions. Not surprisingly, this is highly complex and currently very expensive.
Key steps for life sciences companies
This is a very exciting time for R&D in life sciences. Genomic data can take the sector into an era of highly personalised medicine, where patients get treatment tailored to their genetic make-up, with a greater chance of success, delivering value to the healthcare system. While R&D heads ponder how to make the most of genomic data while remaining compliant, they could also enter into an open dialogue with patients on how genomic data is used, and the extent to which patients want their genomic and medical data to remain private.
One challenge the industry needs to resolve is patient numbers for trials. As genomic targeting gets more precise, the volume of potential trial participants continues to fall, to the point where it can be very difficult to reach a significant sample size. More innovative trial design and ways of comparing groups of patients are required. Data sharing between companies that are traditionally competitors may be a step change in mindset, but could open the door to faster trials and subsequent product approval. And, as we have mentioned, genomics is a great step forward, but should be supplemented by wider clinical and lifestyle data, supported by data analytics to achieve pattern recognition, as well as diagnostic technologies like radiology.