Personalized Analytics is becoming essential in healthcare, stemming from the movement from fee-for-service to a value-based market. The need to preempt and prevent disease on a more personal level, rather than merely reacting to symptoms, has created a significant opportunity for machine learning-based applications. This “analytics of one” approach (using advanced mathematical models and artificial intelligence techniques) is already impacting several key areas:
1. Medical imaging is utilizing AI to process images far faster and more efficiently than the human eye. Prime examples include cardiac imaging analysis that aides physicians in assessing conditions, including heart attacks and coronary artery disease, and retinal image analysis to detect diabetic retinopathy.
2. AI can be applied to EHRs – the majority of which have been unstructured – to create predictive models; or assess conditions that may be under-recorded, such as processing radiologist notes or extracting information from summaries.
3. Conversational AI is being used to collect information from patients in an unattended manner and in transcribing, freeing healthcare professionals from tedious typing and other administrative tasks.
4. The democratization and de-centralization of care, growth in networking and an increase in capabilities and computing power at the edge is enabling a new breed of smart automation to improve monitoring, make diagnostics more accessible and improve communications.
The Transition from Reactive to Proactive Care
The anticipated goal for AI in healthcare is to enhance and expand the “four Ps” of care delivery – predictive, preventative, personalized and participatory. The last decade has seen significant progress towards this end, helping to realize the future of medicine:
Predictive: Predictions have existed in healthcare for some decades now, as statistical models based on structured data sources. However, AI is improving analytics and modern information scientists have developed new types of machine learning-based progression analysis to create superior and more personalized predictions. We see this as an “analytics of one” – a new breed of machine learning approaches with more available data that can outperform guidelines and provide smarter insights for clinical decision making.
Preventative: Today’s consensus is that preventative medicine is preferable to reactionary treatment. Prolonging life expectancy is a key societal goal. Therefore, prevention should focus on extending the number of years a person can enjoy a healthy life, not just how long they can live with a managed chronic disease. The transition to proactive healthcare will benefit from predictive models that can aid in the earlier detection of deterioration trajectories alongside effective preventive measures. Advanced AI algorithms address this through clinical risk stratification (or personalized computational analysis of cutting edge marker technology), resulting in more accurate and personalized models based on differential risk and more targeted subpopulations with higher positive predictive value.
Here, progress has already been made with models able to identify high-risk patients expected to cost the most in the upcoming year, up to 50% of expected healthcare costs. More refined work will facilitate greater understanding of the clinical risk trajectories, the most effective interventions and where preventative approaches are likely to be most effective.
Personalization: Guideline-based, “one size fits all” disease management approaches, which fail to consider complex patient-specific characteristics, are giving way to precision medicine. An extreme application of genetics and “omics” to healthcare in general is not a given. Such a radical approach only applies to very specific, well-identified cases that benefit from exceptionally personalized care, or when the odds ratio for outcomes are extremely high. AI and machine learning will instead provide the middle path for risk-stratified medicine – dividing patients into risk-related groups to provide intervention options that will improve care delivery.
Participatory: Patient engagement and participation is crucial. This goes beyond patients agreeing to share their medical data for disease management and compliance to include patient self-monitoring. As use of sensors and wearables increases, providing a wide range of data readings, new opportunities will arise. AI algorithms will help analyze this vast data, improving patient monitoring and self-management.
Machine learning can also benefit personalized engagement methods, communication channels and messaging styles that can drive patients to action, encouraging their interest and commitment to change. Smart automation of these activities will enable broader and lower-cost outreach with effective results.
The primary challenges and costs in healthcare today arise from the late diagnosis of chronic diseases (such as diabetes and hypertension) and resulting complications. AI can “connect the dots” and avoid human biases, such as selective attention, by providing continuous monitoring, analysis and feedback to patients.
Using AI to interpret healthcare data can also provide a clearer understanding of individual risk and a more personalized approach to changing risk factors. With personalized lifestyle recommendations for individual patients, AI can prevent disease or further deterioration. In addition to saving or extending patients’ lives and improving quality of life, this will reduce admission rates and associated costs.
Any reduction in costs immediately benefits patients by reducing the number and amount of their co-pays and additional out-of-pocket costs, as well as lost productivity for self-employed individuals.
AI will influence personalized care by predicting individual responses to treatment methods, helping physicians to tailor treatments according to patients’ expected responses. For example, which patients may benefit more from lifestyle changes than medication, or the relative impact of different drug courses. Instead of relying on crude guidelines or outdated trial-and-error, analytics can recommend treatments based on predicted outcomes.
A primary benefit AI can bring providers is in reducing physician burnout, a leading cause being the administration and documentation of EHRs adding to already heavy workloads. AI can improve these processes while creating a proactive record that can make intelligent connections to even predict what physicians will need to transcribe. This can foster more doctor-patient interaction without constant clicking and data entry. Simultaneously, AI can provide more equitable care, with quality mechanisms that help physicians identify conditions without fatigue causing them to miss important details.
Stratified medicine will therefore improve all of healthcare, enabling organizations to better allocate resources. Meanwhile, automating action lists and prioritizing patients likely to benefit from more urgent outreach will reduce unnecessary screenings and time spent on patients who do not need to be seen.
The Bottom Line
While we have a better view of AI’s capabilities and potential, this is only the beginning. The process of AI becoming part of the standard of care will be evolutionary, showing progress and encountering challenges along the way. Some key issues will include the evolution of professional workflows, how AI will adapt to current tools, and how those tools will evolve to leverage AI’s considerable benefits.
However, AI will be for the benefit of both patients and providers alike. Crucially, it will not replace doctors. While technology may replace 80% of what doctors do, that is not the same as replacing 80% of doctors. Instead, physicians’ roles will change, enabling them to provide enhanced levels of care. In addition to prioritizing patients at immediate risk, it can allow physicians more personal time with those who need it most, working “at the top of their license.”
Ultimately, AI will spur the transition of medicine from reactive to proactive care. Doctors, providers, and payers will be better positioned to care for their patients’ needs, with the tools to delay or prevent the onset of life-threatening conditions. Consequently, patients will benefit from timely and personalized treatment to improve outcomes and potentially increase survival rates.
Ori Geva is the Co-Founder and CEO of Medial EarlySign. The company’s advanced AI-based algorithm platform helps healthcare organizations accurately stratify populations to optimize care for individuals and prevent or delay serious health conditions, by leveraging routine blood test results, and common labs and EHR data. Medial EarlySign creates actionable opportunities for better clinical decision making and early intervention to improve patient outcomes, focus financial resources, and reduce overall costs.