By using a combination of predictive and prescriptive next best action insights, providers can close the gap in care for individual patients by leveraging a combination of data sources — clinical data, patient surveys, SDOH data, and consumer and behavioral data sets — and applying artificial intelligence techniques to create those insights.
Healthcare in the United States is often better at treating disease than preventing it. Chronic diseases are the leading causes of illness, disability, death and rising healthcare costs in the country. Lifestyle health issues like obesity, high blood pressure and blood sugar, poor diet, and smoking are linked to more than $730 Billion in healthcare spending in the U.S. As many as six in 10 U.S. adults live with a chronic disease.
But these lifestyle diseases don’t exist in a bubble. The circumstances in which a person lives has a direct impact on health outcomes. Social determinants of health metrics (SDOH) like income levels, food insecurity, education and housing status accounts for 30%–55% of health outcomes. In addition to being the root of a patient’s health challenges, these SDOH can also be a barrier to receiving the cultural and contextually-appropriate care that a person needs.
It’s in the healthcare system’s best interest to close the care gap. When providers can identify patients who may be at risk for stopping disease treatment, the impact could be a matter of life or death. When Medicare providers identify patients with potential negative health outcomes and flags them to providers, it saves both taxpayer dollars and patient lives.
By using a combination of predictive and prescriptive next best action insights, providers can close the gap in care for individual patients. To do this, healthcare systems leverage a combination of data sources — clinical data, patient surveys, SDOH data, and consumer and behavioral data sets — and apply artificial intelligence (AI) techniques to create those insights.
Predictive and prescriptive insights defined
Predictive analytics use historical data to create static models that predict future outcomes. In a real world example, predictive models can be used to prepare a hospital for an influx of new disease; something we experienced often during the peak months of the COVID-19 pandemic. These predictive models analyze past and current infection numbers to identify when a new wave of infection will begin, giving a hospital system a head start to prepare with adequate PPE, staffing plans and protocols.
Arguably the most important aspect of using predictive analytics is knowing what to do with the information. It’s one thing to understand which patients are at higher risk of developing type 2 diabetes based on their bloodwork. It’s another to know what the next best actions are for that specific patient and their lifestyle. Predictive insights identify which patients are at risk and prescriptive insights recommend a set of interactions specifically for those patients.
For instance, a predictive insight might tell us that Member X is at risk for not complying with their treatment based on a few critical factors, such as lack of reliable transport to their doctor, their distrust in the medical system or because English isn’t their primary language. Based on this information, a specific set of next best actions are recommended to target these specific factors to encourage care compliance.
Generating insights through AI and data
AI techniques like machine learning take predictive analytics to the next level. Machine learning algorithms build predictive models using sample data. As the model is fed more data, the algorithms learn and improve on their own from patterns in the data. Self-reported patient data is often combined with third-party data sets, such as publicly-available national SDOH data or from companies that collect consumer data. Each of these data sets are never 100% accurate on their own, but when combined, they create a more complete picture in which to uncover patterns and generate insights.
We believe that it is critical to include a wide range of data sets including clinical, SDOH, consumer data, behavioral data and self-reported data when building these predictive models. This allows us to obtain a better whole-person view of each patient, understand what could potentially prevent them from having positive health outcomes (e.g. they are living in a rural area, are not fluent in English, have transportation issues and are distrustful of the health system) and close the care gap by prescribing a set of interactions specific to that patient’s circumstances (e.g. to ensure that they stay on their medications or find a provider who speaks their native language).
Using predictive and prescriptive insights to close the care gap
I’ve seen how predictive models and next best action insights can make a difference in some of the most at-risk communities. One of the predictive machine learning models we run in Africa tracks patients who are on HIV antiretroviral therapy (ART) in order to identify which patients are at high risk for dropping out of the healthcare system. All of the predictive analytics models are built on large multi-year longitudinal patient data sets (e.g. 500,000 patients): data that identifies who dropped out of the healthcare system, what their CD4 count was, where they lived, etc.
Based on the model which builds the correlations and identifies patients at increased risk for dropping out of care, the providers are notified and take proactive actions with the high risk patients. These interactions are fed back into the model, which uses this information to continually improve upon itself. The results are impressive. One program experienced a 36% increase in retention of high-risk patients.
Looking ahead: Using data to create more personalized next best actions
We are living in a data-driven and data-rich world — from our smartphones in our pockets, to our Apple Watches and FitBits on our wrists, to the purchases we make online. As we look ahead 10 to 15 years, I predict that this type of personal data will be used more often to gain an even more personalized picture of individual patients’ health. If doctors had access to the objective data on a patient’s resting heart rate, sleep patterns, how many steps they’re taking in a day — which is all currently taken via subjective self-reporting surveys — we could build more accurate predictive models to identify at-risk patients and the next best actions needed to meet that patient’s personal needs in the moment.
Like all conversations around big data, especially healthcare data, privacy is paramount. Apple is advancing this technology with their recent update to iOS 15, which allows users to share their health data with others, such as family members or their doctor via the Health app. It’s a step toward closing the gap between patients and their providers. Combining this data with SDOH, we can not only create models that predict illness or disease progression, but create a personalized action plan for each patient, delivered to them in the context of their real life. Equitable healthcare is possible and data and AI will get us there.
Dr. John Sargent is founder of BroadReach Group.