How precision medicine will leverage data, analytics and AI
Precision medicine is sometimes defined as the “application of genotypic and OMICS-biomarkers to determine the most appropriate, outcome-driven treatment of or therapy for individual patients.” The applications of precision medicine result in a customized approach to healthcare, where medical decisions, care pathways and services are tailored for each individual patient to achieve the most beneficial outcomes.
Precision medicine is rapidly evolving as the influences of genomic biomarkers are becoming better understood and next generation sequencing (NGS) is becoming more affordable.
While precision medicine (specifically, precision oncology) will continue to build upon and leverage new genomic datasets and knowledge, it helps to look at a comprehensive definition of precision medicine that addresses important questions on outcomes, access and affordability.
Precision medicine, when defined in that way, becomes the determination and delivery of the right care to the right patient at the right time. This acknowledges a few realities that must be addressed via a multidisciplinary approach that combines people, behaviors, social determinants of health (patients’ ZIP codes have as much influence on their health as their genetic code), and their phenotypic data. Additionally, addressing key concerns on the affordability and generalization of individualized therapies to a broader population will require providers to treat precision medicine and population health as two sides of the same coin.
Generalizing targeted therapies and making them applicable for larger cohorts or stratified populations will be important to scale the latest therapies, medications and patient outcomes and satisfaction to society at large. The applications of precision medicine span diverse specialties like oncology, cardiology, infectious disease and pediatrics, among many others.
Targeted diagnosis and treatments are increasingly becoming more prevalent in the outpatient and chronic care settings, in addition to inpatient acute care settings. To a large extent, this transformation is happening because the efficacies of novel, targeted and precise therapies are being demonstrated to clinicians, patients, and payers through the use of informatics, analytics and data.
Precision medicine that will transform oncology, as well as chronic and acute care must address the increasing prevalence of not just genomics sequencing data and biomarkers but also account for health variables that are constantly being defined.
These variables can be from disparate sources within and outside today’s healthcare facility. They will increasingly include sensor, device and wearables; clinical narratives; new and existing research (clinical research, trials, publications, unified disease registries); reimbursement, cost and financial considerations; family and disease histories; environmental variables; behavior and sentiment data; and, increasingly, income, educational and cognitive disparities.
Data’s importance for precision medicine
With cancer, the predicament is that clinicians must keep running merely to stay in place with the great amounts of data pouring in. Data and analytics can make a difference and help providers stay ahead and make progress, for the patient’s benefit. Some salient points on cancer:
• Cancer is a disease of the genome.
• Cancers are ancient. They survive and evolve continuously.
• To escape targeted therapy, cancers are known to change the target. We need to keep moving just to “keep our position.”
• “Precision Therapy Requires Precise Data” – that combines “little” and big data across multiple sources for actionable insights in varied individual contexts.
• Conversations on outcomes, affordability and patient satisfaction are data driven.
• Complications and patient experience need to be measured continuously.
• Discoveries will result from combining data from multiple sources.
• Informatics can make a difference. Precision oncology (and medicine) require, at the very least, precise measurement based on integrated data that is actionable.
• Analytics can be used to “burn hay to find the needle.”
IT-enabled big data management and healthcare are required tools to manage and leverage the complex data that result from genomic, clinical, financial and behavioral data that brings the biggest benefit to individual patients.
Precision Medicine and FIRE at MD Anderson
Achieving success with precision medicine requires healthcare organizations to take a step back and ensure they take an enterprise-wide, rather than departmental, approach towards analytics.
The FIRE analytics program at the University of Texas MD Anderson Cancer Center highlights how the shift to a data environment capable of processing all data types (clinical, genomic, financial, operational and more) and data formats (structured, unstructured and other variants) requires fresh approaches to enterprise data architecture. “A core goal is to accelerate and improve the translation of scientific knowledge into clinical care.
The effective use of data and technology is integral to the organization’s efforts. MD Anderson faces mounting pressure to sustain more complex and new forms of research while increasingly operationalizing intelligence for better care. In this respect, the extent to which it can use its own and third-party data is becoming more critical.
The FIRE analytics program at MD Anderson has been in operation for more than four years and has seen rapid and continual growth in the number of users, diverse integrated clinical/genomics/operational analytics, and integrated big + little data.
Based on its experience, here are a few recommendations for healthcare organizations to consider for their precision medicine and analytics journeys.
• Keep the end in mind. This is a challenging journey, not a finite project.
• Measure success by outcomes, not the number of analytics or data sources. Well-implemented next-generation analytics programs can deliver substantial benefits in clinical informatics, data analysis and operations.
• Precision medicine requires precise measurement to determine outcomes, access and affordability.
• Little data is ubiquitous and provides important context to big data. Therefore, organizations must focus on all data.
• Use variety in analytics to derive value and to address new users—human and machine.
• Data fidelity is important. Focus on data that is appropriate for the context of its use instead of merely focusing on a “One Size Fits All” data quality approach.
• There is a corresponding need in the context of therapy and procedures to examine the veracity and value of data/analytics on a continual basis.
• Precision medicine requires precise vocabulary and identities. Data governance is not optional.
• Ai and specifically contextual intelligent agents (CIAs) are important to take advantage of vast, new and fast data.