EXL’s client, a life sciences company, provided a therapy to patients diagnosed with multiple sclerosis. However, some patients would stop using the treatment. The client needed to know what differences separated those who ceased using the therapy and those who continued. The client was specifically interested in understanding the reasons behind the patients dropping off the treatment within three months, a relatively short time period. Understanding this could help the company improve treatment plans, health outcomes, address treatment barriers, and improve the lives of its patients.
Human Ingenuity in Action
The EXL team deployed several statistical regression/machine learning algorithms that helped identify the factors which led patients to drop out of treatment programs. EXL’s solution enabled the design of datadriven short-term and long-term strategies to reduce treatment nonadherence. The approach looked at holistic groups of features which impacted patients’ treatment journeys, including:
- Demographics and social determinants
- Medication usage
- Healthcare service utilization
This strategy included creating a decision tree that visually displayed what patient attributes were correlated to staying on therapy, and which were not. This enabled the client to understand the complex nature of the underlying analysis in an easy-to-understand design.
Using other advanced algorithms, EXL helped the client identify patients at a high risk of stopping treatment, and the factors leading to this result over a longer timeframe of the treatment journey. To help create customized tactics for dealing with the inherent problems leading to this negative outcome, EXL provided a simulator which predicts patient outcomes by taking into account different individual factors.
Armed with the information created using advanced analytics and machine learning, the client is now actively leveraging these insights to improve treatment programs and increase patient adherence to treatment plans. This includes targeting multiple areas, such as physician and provider education, patient assistance programs, optimizing payments with insurers and pharmacy benefits managers, and other tactics.