Medical Affairs: Support Function or Competitive Advantage? How AI & NLP Can Ensure It is Both

April 28, 2021
Paul Riley, Colin Baughman and Jeff Catlin

Paul Riley is Medical Affairs director at Glasshouse Health, a specialist Medical Affairs consulting practice based in Alderley Park, Macclesfield, UK.

Medical affairs is the voice of the patient in pharmaceutical companies, ensuring that patient health and well-being are at the forefront of marketing decisions. This role can put medical affairs somewhat at odds with the sales and marketing side of the house, but this thinking is outmoded in the social data world. Medical affairs has become a company’s best way to monitor the world at large and keep tabs on its competitors. Sales and marketing have numbers to meet and never want anything but great news, but it’s medical affairs that keeps an eye on drug safety and efficacy and prevents business missteps.

As a trusted partner, medical affairs teams carry the critical mission of informing clinical practice patterns to ultimately improve patient outcomes. So as the title of this piece suggests is this mission a support function or competitive advantage? Support functions aim to maintain service levels while continually reducing costs, often through outsourcing and leverage. In contrast, functions viewed as a competitive advantage see investment, innovation, and technological advances to improve costs, outcomes, and timing. So which one is medical affairs?

The simple answer is that it can be either, but in this socially aware world where a single viral tweet can change the course of a product’s future, it’s the group that can help maintain a billion-dollar revenue stream. Treating medical affairs solely as a support function is penny-wise and pound foolish, as the risks of not innovating in this fast-moving world jeopardize a business. With the advent of machine learning, the medical affairs team can innovate and perform their many functions better and more efficiently. This isn’t to say that AI is a magic bullet for all of your medical affairs issues; like any other part of business, it’s better suited to some problems than others.

The best way to think about this is from a return on investment (ROI) perspective. Where can you spend the least money and get the biggest bang for your buck out of deploying machine learning and AI? Here are a few such examples:

1. Medical information requests (MedInfo AI)

2. Medical Science Liaison (MSL) insights

3. Content meta tagging.

MedInfo AI: As most reading this probably know, the goal in MedInfo is to get the right material to the requester in the shortest amount of time. The longer it takes to answer a question or information request, the more expensive that request gets. Add into this the growing complexity of information requests due to the ever-expanding platforms for asking a question (phone, email, SMS, and eventually Tweets) plus the breadth of spoken languages serviced, and you have a regulatory requirement whose costs are going to accelerate over time. Given the heavy regulatory requirements around MedInfo, it’s not feasible or advisable to remove humans from this loop. Still, if you can make humans more efficient, you can keep the costs in line. An AI-powered MedInfo system built into your company’s workflow can accomplish this. We’ve built and deployed such a system for a client in Japan who actively uses it to manage costs and provide auditability of their work.

MSL insight: The Medical Science Liaison role has the potential to dramatically and positively affect the perception of a drug in the marketplace by leveraging the data collected through the MSL role. Too often, the data gets lost in a company’s CRM system, such as Veeva or Salesforce. Actively integrating with these CRMs and mining the key voiced concepts, areas of concern, or usable quotes by the Key Opinion Leaders (KOLs) means that you are getting the maximum value out of the money you are spending on MSL. Effectively mining this data requires a Natural Language Processing (NLP) engine tuned to the condition being covered by the MSL. Generic NLP won’t do; you need an engine that can be adjusted and tailored to the relevant drugs and conditions.

Content meta tagging or content aggregation: This is another common problem in most larger pharmaceutical companies. The more data you collect, the more likely it is to get siloed and leveraged only by the group that gathers it. This particular issue is a mix of AI and human expertise, where you can design and build taxonomies and tagging schemes that work across various functions within medical affairs and beyond. These taxonomies coupled with NLP and AI can access and enhance the content within a CRM system to allow companies to get as much value possible from the data they are collecting.

The point here is that the medical affairs’ responsibilities aren’t getting reduced, and the advent of social media adds new risks to a pharma company. The judicious use of machine learning and AI can save money and time in the performance of medical affairs and, in some cases, can help protect revenue streams. Thinking of medical affairs as a competitive advantage allows both innovation and a de-risking of many key responsibilities.

Paul Riley, Ph.D, is Medical Affairs director at Glasshouse Health (UK); Colin Baughman is founder and managing partner at True North Solutions Inc.; and Jeff Catlin is co-founder and CEO of Lexalytics.