“AI is fascinating. I hear about it all the time. But how do I really apply it to my patient support program?”
It’s a question I’m asked a lot.
I lead business development for AppianRx, a healthcare technology firm that uses AI solutions to help manufacturers increase patient adherence and decrease abandonment for their specialty medications. In the course of meeting with manufacturers and patient support services leaders around the country, I have found that there’s plenty of awareness of AI in general but less understanding for how machine learning and other tools translates into specific bottom line advantages.
Part of the challenge in explaining the value proposition of AI is this: every company is unique. When it comes to patient support service programs, specialty medication manufacturers all face different landscapes of opportunity and difficulties keeping patients adherent to their drug protocols. Machine learning will therefore serve each business in unique ways that are particular to the company’s specific situation.
For example, companies collect myriad data points about their patients, but they may not be extracting that data and using it to their greatest advantage. AI tools can ingest those myriad data points and create predictive algorithms to better analyze the data. With deep intelligence about the historical patterns of their programs and patients, companies can more precisely allocate resources, develop laser-targeted interventions, and improve patients’ success with their medications.
So back to the question: “How can AI help my patient support program?” The challenges that AI solves for your patient program can't be answered until you fully understand the challenges your patients face.
And without machine learning tools, you don’t know what you don’t know.
86% of healthcare organizations use AI
AI is a well-established technology that has long been transforming healthcare through applications for drug discovery, robotic surgery, diagnosis of certain conditions, and myriad other medical innovations. But machine learning has been notably absent from patient support service programs.
Pharma and biotech leaders know that AI can deepen insights, improve operations, and dramatically scale capabilities. A CB Insights report found that approximately 86% of healthcare organizations, life science companies and med tech firms were using AI technology in 2016. According to a study by TechEmergence, more than 50% of healthcare industry executives currently using AI anticipate broad-scale adoption of AI by 2025. Big pharma names announcing deals and applications related to AI include Bayer, J&J, Merck, Sanofi, Genentech, Novartis, and Pfizer.
There’s even an industry group called the Alliance for Artificial Intelligence in Healthcare (AAIH). Formed by BenevolentAI, GE Healthcare, and Insilico Medicine, amongst others, AAIH advocates for public policy, regulation, and market access for AI-developed products.
But as noted in PharmaTimes, a Harvard Business Review report found that only 8% of chief executives in pharma and biotech successfully led enterprise-wide AI initiatives, indicating the pace of adoption is slower than the speed of AI/ML advancement.
One pharma and biotech niche where AI adoption is lagging is in patient support services. Other industries are decisively stepping up to integrate machine learning technologies into their contact centers and customer service operations. It's an obvious area of excitement, as improving efficiencies in customer service increases customer loyalty and brand retention and allows employees to focus on other areas that provide greater returns.
But patient support services are suffering from their own inertia when it comes to modernizing their programs.
Real world data creates real world solutions
Perhaps as pharma and biotech leaders see the tangible results of predictive algorithms on their patient support services, they’ll hasten the adoption of machine learning initiatives.
Let’s consider some of the practical benefits of AI for patient support programs.
When manufacturers roll out new medications, they face numerous challenges with increasing access and identifying potential non-adherence. This can include, for example, complex enrollment forms to meet HIPAA compliance; the need to send free month-long starter kits while the prior authorization process chugs along; and heavy resource allocation to train caretakers how to use new medications.
Companies collect plentiful data about every patient. But raw data alone doesn’t offer insights. It’s difficult for the humans leading patient support to connect the dots and understand which events are triggering what actions related to adherence and abandonment.
How can manufacturers move away from a one-size fits all approach? How can companies customize interventions and connect that activity to proven outcomes? How can they optimize programs based on past knowledge learning and apply it to ongoing patient/ caretaker outreach in real time?
What manufacturers need are usable, action-driven models that guide their programs on how, where, and when to deploy resources.
Machine learning tools can automatically create this sort of model in ways that are impossible for humans to replicate at scale, creating specific risk scores for each patient and giving manufacturers the information they need to operationalize patient insights.
With this intelligence in hand, companies can make smart decisions that save money, time, and resources, such as with the following key areas:Financial Assistance
- E.g., co-pay cards and free products
- Predictive modeling around authorization timing can guide companies on how much free starter medication each patient should receive
- E.g., nurse educator programs, skilled advisors, and other clinical support
- Instead of a one size fits all approach to this significant clinical investment, AI can help companies identify where these skills are best deployed
- Identifying which patients are low risk helps companies avoid the costly, high-touch support of nurse educators where they’re not needed
- E.g., patient outreach via text, email, and phone; refill reminders.
- Instead of a generic welcome call sequence (one size fits all greeting, three call attempts, etc.), companies can tailor this initial critical touchpoint to each patient’s specific circumstances, thereby increasing patient connection and engagement
- Companies can determine which patients are at high risk of not refilling their medication and create tailored interventions to keep those patients adherent
- Companies can also better understand when to start, stop, or continue interventions for each patient
As someone who built an early-days hub when patient support service programs were just getting started, I can tell you that I would have jumped at the opportunity to integrate AI into our platform. Back in the day, we were throwing darts at a dartboard, guessing about what actions to take to keep our patients adherent.
Not much has changed in the past decade. Most patient support service programs are still playing darts, and the success of specialty medications is suffering as a result. Machine learning changes that equation and allows programs to take intelligent actions based on real world, real time information that can translate into bottom line impact.