Insight, Impact, Innovation
HomeOur WorkCase StudiesProblem-finding and machine-learning allows healthcare payer to save $5M from medication non-adherence

Problem-finding and machine-learning allows healthcare payer to save $5M from medication non-adherence

Download PDF


The Business Challenge

Non-adherence of medication is one of the most important problems when treating patients with chronic conditions. It’s not surprising that patients who don’t follow the prescribed drug regimen are more likely to suffer poor health outcomes, significantly contributing to the total cost of care. According to a study published in the Annals of Internal Medicine, medication non-adherence costs the US healthcare system $100-$289 billion annually (or 13% of all healthcare costs). According to the study, factors behind non-adherence include: 20%-30% of medication prescriptions are never filled, and 50% of medications for chronic disease are not taken as prescribed, resulting in over 125,000 avoidable deaths each year and 10% of all hospitalizations. Yet according to Q1 Productions, roughly 75% of adults over 40 with a chronic condition admit to at least one non-adherent behavior in the past 12 months.

A top 5 health insurance company, wanted to improve health outcomes at a lower cost by improving medication adherence, as well as improve customer engagement. This payer needed insights to understand and anticipate medication non-adherence to drive and measure more effective intervention strategies.

The Solution

The solution involved developing a medication non-adherence framework to identify individual patients less likely to adhere to their prescribed drug regimen during one year. A structured problem-finding framework was designed to identify the types of non-adherence, model predicted behavioral drivers, and then implement the right intervention plan to decrease medication non-adherence.

Stage 1 – Problem finding: We identified over 200 potential hypotheses by integrating data from claims, medication utilization history, medication history, member demographics and consumer data, and classified medication adherence based on key drivers.

Stage 2 – Predictive behavior modeling: We evaluated various drivers of medication non-adherence along patient factors, medication therapy and zoned in on five major types of non-adherence that could be measured and acted upon. These are:

  1. no prescription or refill filled,
  2. incorrect dosage,
  3. medication at the wrong time,
  4. forgetting to take doses and
  5. stopping therapy too soon. We used advanced machine-learning models to identify and predict past medication adherence, medical utilization and demographics. 

Stage 3 – Assigning action plans by segment: We mapped out recommended strategies for three key patient segments that behaved differently in their medication adherence for chronic diseases. 

The Results

Armed with more accurate prediction and classification, the payer was able to intervene and improve current medication adherence that improved health outcomes, resulting in $5 million dollars in cost savings.

Insight

Advanced analytics can help healthcare insurance companies improve patient outcomes and reduce cost by increasing patient medication adherence

Impact

By creating targeted intervention and outreach programs, can improve medication adherence for patients with chronic disease, saving $5 million in annual care related savings for the provider

Innovation

  • Ensemble machine learning models effectively predict medication adherence
  • Targeted action plans by segment predicted adherence levels