Article

The Future of Care Management in the Age of Artificial Intelligence

Author(s):

We need to eliminate this data crunch situation. Healthcare as an industry generates mounds of critical data, but our inability to utilize it and make the most out of it serves as a crucial problem.

I have always been intrigued by weather forecasts. Once while I was going through the newsfeed on my phone, 1 specific line caught my attention: “There’s 56% probability that it might rain today.” That was my awe-struck moment.

Such statements not just influence the way people think, but also their future course of action. That particular statement above from the weather department definitely made me re-plan my Taco Tuesday!

However, it made me wonder, “Can we make such claims in healthcare? Can we predict such outcomes for patients? Can we ever say to a patient you have a 30% probability of getting a heart attack in the next 3 months?”

What can enable us to find these answers? Artificial intelligence (AI)! Let’s dive into the depths of exploring this.

We have everything, don’t we? Then why are we failing?

When I look into this situation, I think that we, as healthcare stakeholders, are just living under the impression that we have everything. Actually, we are still working in silos. Success in the value-based care environment cannot be achieved based solely on clinical insights.

Social determinants of health (SDOH), factors that are responsible for 80% of health outcomes, are still a mystery for a lot of us. Although numerous research articles demonstrate that SDOH substantially contribute to a patient’s health condition and well-being, major complications in understanding them include:

  • How do we address these complex challenges?
  • Who is the best-positioned stakeholder to do so in a clinical environment?

We need to eliminate this data crunch situation. Healthcare as an industry generates mounds of critical data, but our inability to utilize it and make the most out of it serves as a crucial problem.

What is the right way in addressing SDOH?

Well, this article actually fuels up from here! AI, machine learning (ML), and predictive analytics—this trio holds the answer to most of the problems in healthcare.

Working with SDOH requires a thorough, drilled-down approach and the detailed use of predictive analytics to measure the at-risk population accurately and advance preventive care methods in a care ecosystem.

The effects of social determinants vary in accordance with a very small region. There is a high possibility that all the zip codes in a county will have different susceptibility to a particular social determinant. The best way to address SDOH is to start the analysis from the national level and drill it down to the zip code level to efficiently measure the impact of every social factor accurately and precisely.

We now have the data. What’s next? We are still nowhere close to predicting the care journey of patients.

It all starts with the right data. The data from such sources including the clinical data, claims data, and SDOH, among many, allows us to gain a comprehensive view of the patient’s care history and apply ML algorithms to predict the future.

Imagine what we could have achieved if we have both:

  • Right clinical and claims data about the patient
  • The dollar amount that a patient might spend on the illness

A major role is played by "time" in such analysis. For predicting the right outcomes in the upcoming year, we need to make sure that we know what has already happened to the patient in the past. The predictive model should take the patient’s past years’ cost, utilization, and diagnosis data into consideration to understand the patient’s health in depth.

Once we have all these handy, just 1 final ingredient is left, which is the inflation rate. Not just the financial aspect but some demographic adjustment is also required to understand the cost of care that will be incurred for a patient in the upcoming year.

What new innovation can we bring to the table with this approach?

Once the organizations are able to leverage these advanced analytics and AI tools, they can easily stratify their patients and identify the ones who are in dire need of immediate care interventions and can make the cost-cutting process more streamlined. Additionally, proper insights into the non-clinical data sets such as the SDOH data can assist the care teams to overcome the adverse outcomes of the citizens of that region.

The analysis of social determinants can be applied for multiple use cases such as:

  • Identifying the role of behavioral health providers, social workers, and health coaches;
  • Increasing the efficiency of the care coordination team;
  • Forging better partnerships with community resources and social improvement funding agencies, and many more.

What does the future with AI hold for us?

Most people see technology as the replacement of humans. Wrong! They can never be the replacements; they are the support systems. We are stepping into the age of predicting and preventing diseases instead of curing them. That was the traditional approach, with non-clinical data and resources such as SDOH, we can change the future of US healthcare. From assisting providers to use data-driven insights to target patients, to projecting the cost of care-predictive analytics, this can change the dynamics of healthcare. The future with AI and ML is here, all we need is to just embrace it.

Related Videos
Roberto Salgado, MD.
1 KOL is featured in this series.
1 expert is featured in this series.
5 experts are featured in this series
Keith Ferdinand, MD, professor of medicine, Gerald S. Berenson chair in preventative cardiology, Tulane University School of Medicine
Related Content
AJMC Managed Markets Network Logo
CH LogoCenter for Biosimilars Logo