This guest blog is one in a series by sponsors of the 2024 PQA Leadership Summit on how MedWatchers is using AI to empower better patient care. PQA does not endorse, recommend or favor any product, service or organization that is a sponsor.

At MedWatchers, we have realized that the quality of the intervention is the most impactful aspect of the services we provide. Our commitment to delivering exceptional care drives every decision we make. While we leverage technology to enhance our services, AI is not the driver of our business objectives - it is the enabler. We embed AI into our workflows to complement and enhance our products and services, ensuring they align with our goals of delivering better patient outcomes and higher-quality care.

Our last PQA update highlighted how AI has sharpened our quality assurance processes. By integrating AI technology, we've significantly reduced manual labor in quality assurance, enabling a tenfold increase in the number of calls reviewed and a sixfold acceleration in identifying quality assurance issues. Our new QA process allows our team to focus on individual training and compliance, ensuring our clinicians feel prepared and empowered to deliver top-tier care. We are now embarking on our next AI enhancement – utilizing an AI-powered call simulator to enhance our overall training process. A call simulator mimics real-world phone interactions to improve our pharmacists' clinical and conversational skills. By providing realistic training scenarios, the simulator equips our team with the necessary skills to handle complex patient interactions efficiently while maintaining compassionate and effective care that defines MedWatchers.

The Quest for the Perfect Training Tool

As we set out to find the ideal AI simulation tool to support our clinicians, we focused on solutions that could meet our unique requirements. We prioritized features like customizable role-play scenarios, a flexible AI interface, and strong reporting tools. Some vendors offered platforms that needed to be more developed or were prone to glitches, while others required excessive effort to create realistic scenarios, often causing training interruptions and frustration. The technologies varied, some used NLP models with custom interfaces for prompts, while others relied on LLM models without corresponding interfaces.

Our first attempt was Vendor 1, which used an LLM model. While it provided the realistic simulation experience, we sought, it lacked an external interface compatible with our curriculum. This limitation prevented staff from using the physical prompts critical to our training programs. Moreover, the AI required rigid adherence to the script, as even slight deviations could crash the simulation, leading us to seek alternatives.

Next, we tried Vendor 2, which featured an NLP model with a custom interface. Although it allowed for the incorporation of action steps, the time required to develop and maintain simulations, limited customization options, and frequent AI prompting issues made it impractical and unscalable for our needs.

Ultimately, these challenges prompted us to leverage our in-house AI expertise to create a fully customized, proprietary platform that enhances team growth, improves conversion rates, and boosts client and patient satisfaction, while ensuring a seamless and effective training experience.

Building a Simulator That Fits Just Right

We designed our in-house call simulator to seamlessly integrate with our systems while mirroring the daily real-world scenarios our clinicians’ encounter. By leveraging advanced AI voice interface technology, we have incorporated customized scripting to build an AI simulation model powered by LLM technology. We carefully developed our scripts using programming rubrics, realistic conversation models, and our years of experience in direct patient care. We carefully crafted over 40 unique scenarios, each to depict challenging patient situations and designed it to help clinicians practice de-escalation to achieve effective resolutions.

To enhance our training experience, we also leveraged our proprietary Care Management Application (CMAPP), hosted on a dedicated training server. This platform integrates patient charts directly tied to the scripts, requiring clinicians to engage with the system actively to complete their simulations, ensuring a comprehensive and practical learning process.

This newfound flexibility and autonomy meant we can keep our training relevant and impactful, equipping our clinicians for real-world situations that demand expertise and empathy, while also allowing us to significantly reduce training costs and address technical issues immediately.

This custom simulator also lets us build on insights from our AI and automation initiatives, further supporting our quality goals of delivering a high standard of care to drive adherence and improve patient outcomes.

What's Next?

We're excited to update our community on our progress with the call simulator. Stay tuned as we refine this tool, set new standards in quality, and continue to build stronger patient connections—one call at a time.

MedWatchers Authors and Collaborators:

Christina Voskanian, PharmD
Manager, Pharmacy Excellence and Operations