The life sciences industry has waited a long time for big data to transform the potential of personalised medicine. With AI and machine learning now coming of age, R&D teams can finally seize the opportunity — as long as their data is also clean, standardised, interoperable, and secure.
To understand a treatment’s potential for a specific patient, biopharma companies have to layer together data from multiple disparate sources. Some of these data sources will be common to all disease areas: for example, patient demographics, electronic medical records, and quality-of-life scores. However, the majority (including genetic information, imaging, and activity data from wearable devices) will be unique to each individual. Since the clinical effectiveness and safety profile of a personalised treatment will be different from patient to patient, all relevant stakeholders must be able to trust the data to make medical and business decisions confidently.
Reappraising the approach to quality, ownership, and interoperability will bring usable data to the core of their strategy, even when working with potentially millions of relevant data points. Leading biopharma companies are also rethinking their existing ways of working and systems to get to first-time-right submissions. With access to a clean data foundation, they can identify which functional areas are most critical to speed time to market so that patients aren’t left waiting for innovative new treatments.
Accelerating time to patient
Historically, data collection initiatives have been broad in ambition and scope. These ranged from sequencing, imaging, and electronic health record data to text-based information such as interactions with health authorities and conference abstracts. The main objective was data completeness, and the scale of data points collected made it challenging to spot patterns or identify the most effective uses.
Today, the go-to-market and approval requirements of personalised treatments are more complex than anything seen before. As a result, biopharma companies seek to make appropriate use of their study data far sooner, which shifts the focus from data collection to governance and ownership. Gaining more control and oversight will change the dynamics in their relationships and contracts with third parties. Connected systems are becoming critical so relevant stakeholders can view the data at any time rather than waiting for data to be sent back in meta-data formats, or final text-based documents.
It is also becoming easier for sponsors to pinpoint the most impactful inefficiencies during the clinical development phase — vital to compressing time to market so that personalised medicines remain commercially viable. Analysing data on the cycle times between two critical clinical milestones could indicate whether inefficiencies and operational challenges typically arise during protocol design, site selection, initiation, or elsewhere. These insights can support the whole organisation to become more productive. A single source of accurate data can create competitive advantages by driving better decision-making on patent filings or patient recruitment, or efficiency gains in outsourcing, procurement, or portfolio rationalisation.
Analytical and data science capabilities have improved but limitations remain. Raw data is not standardised and limited industry (or even intra-company) reference models exist. If common pain points around cleaning, ownership, and standards can be resolved, the volume and frequency of access to study data will increase. This will require a transparent data model with stringent user access controls to address privacy and cyber-security concerns.
