Big data combined with advanced analytics can help deliver sustainable healthcare when the human body being analyzed and described at levels of detail never previously possible.
A learning healthcare system leveraging integrating data streams and assisted decision making may help the world’s complex and diverse healthcare systems to find cost-effective solutions to meet the evolving needs of the world’s patients, where the growth in health spending is out-pacing the rest of the global economy1. In order for the world to meet the health-related SDGs, a fundamental shift in policies and strategies is required2.Privacy matters when data is everywhere
The growth of healthcare-related data is outstripping every other industry3. Conventional data streams both internal and external (but clinically relevant) to the healthcare system could provide real value in a variety of settings. From wearables to genome sequencing to new imaging methods, technological advances are also providing us with a greater depth of untraditional health and wellness data on individual patients and society as a whole.
But access, authenticity, ownership and use of health data raise fundamental questions around ethics and privacy. Robust security and governance are required to ensure the quality and appropriate use of such sensitive information.Changing relationships and consumerization
Clinicians are keen to leverage medical knowledge from a wider variety of sources to guide their decision making. Patient acceptance levels for health research are also increasing – for example, one in four adults in Norway is enrolled in one or more such projects. Disruption in healthcare delivery models means that patients increasingly have more options on how their healthcare is provided and by whom. This leads to rising competition and quality of care for patients.
Major tech players, including Apple, Google and Microsoft, are attempting to capitalize on the consumerization of the healthcare sector. They are in a unique position as they interact directly with consumers, and could continue to disrupt the healthcare market when national borders no longer represent barriers for testing. Their involvement could also link and unify traditionally disjointed value chains to provide a holistic approach to managing an individual’s health.
The quality of health data is a major stumbling block, where AI algorithms require well-curated and well-annotated data training sets to achieve the best performance. Yet health data is highly fragmented, and 80% is unstructured. However, the combination of electronic health journals, improved computational power and standardized protocols could allow clinicians to harness health data systematically over global populations and cohorts to create new and additional value for both patients and healthcare systems.New social contract requires building trust
A mixed business model may emerge where consumers share their data in exchange for services and other incentives. This raises fundamental questions about ethics and privacy.
Health data may need to be shared and processed across multiple jurisdictions, but each one has its own set of laws, with additional requirements coming from regional legislation, such as the GDPR in the EU. As geographical and industry boundaries erode, partners who share health data need to ensure they are doing so lawfully and assure the quality of this data when it is applied to clinical decision making.Impact on society
Personalized treatments and services could vastly improve the health of patients, who, ideally, should also have control over their data. In addition, humanmachine hybrid workforces enable more effective resource allocation and treatment options to improve organizational efficiencies and patient treatment.
However, privacy violations and the misuse of personalized medical information for unethical reasons pose significant risks, as does unequal access to the benefits of big health data, which could deepen existing inequalities and create new ones.AI in clinical systems and decision making
Advancements in artificial intelligence, or AI, are laying the foundation for this developing technology to be used in clinical decision making.
AI is already prevalent in the interpretation of medical images, with 14 algorithms FDA-approved as Class II medical devices to date. However, patients are being represented in increasingly complex and clinically relevant ways. Some molecular profiling methods, in particular, could inform clinicians’ risk stratification and patient management and/or treatment tailoring strategies.
Health systems also stand to gain greatly from the application of AI, including the development of risk-based outcome predictions. Natural language processing could be used both as information readers and writers (or scribes) for the effective digitization of clinical observations and notes. Large gains in efficiency and productivity alongside improved quality of care could also be realized through AI-improved workflows. This, in turn, could reduce the occurrence of medical errors, which also has positive implications for patient safety.
Finally, AI could reverse the decline in face-to-face time between clinician and patient, where doctors’ and nurses’ time may be freed for increased patient interaction4. This is an important point because, despite the potential of AI in healthcare, the ultimate aim is synergy rather than replacement of human care givers.
- World Health Organization. (2018) Public spending on health: A closer look at global trends.
- Global Burden of Disease Health Financing Collaborator Network. (2019) Past, present, and future of global health financing. The Lancet, DOI:10.1016/S0140-6736(19)30841-4.
- IDC. (2018) The digitization of the world from edge to core.
- Hall et al. (2016) Healthcare staff wellbeing, burnout, and patient safety. PLoS One, DOI:10.1371/journal. pone.0159015.