By Gina Shaw

Artificial intelligence programs may be the key to solving a major concern that emerged in a large survey of pharmacy leaders released at the 2024 NASP Annual Meeting & Expo, in Nashville, Tenn.: how to relieve the staffing pressures that are leading to burnout and other workplace ills.

In a survey of leaders of pharmacy patient support programs conducted between December 2023 and March 2024, the healthcare AI platform developer Infinitus found that 83% of respondents listed “doing more with fewer resources” as one of their top three challenges, and more than one-third estimated that their staff spend more than half their time on the phone with payors. Implementing AI solutions was one of the top strategies these leaders planned to use to address these challenges, with more than 80% reporting that they either already use AI to improve their organization’s services or that they plan to do so in the near future.

AI has multiple potential applications in patient support, said Rachel Clifton, the national account manager for Infinitus, including determining optimal outreach methods for medication delivery scheduling, analyzing patient data to identify the best times and channels for communication, leveraging automated outreach to reduce the need for manual calls, and automating routine back-office functions.

“You can convert a frustrating outbound call process into an automated 24/7 inbound process,” Ms. Clifton said. “Imagine texting a patient to say that you need to confirm a couple of delivery questions, and they can call this number 24/7. There’s no wait, no phone tree to navigate. You could be off the call in 30 seconds.”

She characterized the leveraging of AI in patient support as “repurposing your most precious resource, human beings. Think about what are the most laborious and time-consuming problems processes in your pharmacy and how do you address those.”

However, she advised caution in building an AI plan and working with vendors, particularly with regard to patient privacy and data security. Currently, the HIPAA and Service Organization Control Type 2 (SOC 2), a cybersecurity compliance framework developed by the American Institute of Certified Public Accountants, are the most well-known industry standards for information security compliance in healthcare AI. But they aren’t the only ones—and there is no single governing body at the moment, Ms. Clifton said.

“Ask your vendors what security measures are in place and how protected health information is handled,” she said. “These models should be trained with de-identified data, and if the vendors are not HIPAA or SOC 2 compliant, what do they have in place instead? There have to be guardrails.”

 ‘Conversions’ Will Drive AI Choice

AI is not magic; it’s a tool like any other, said Natalie Park, the CEO of Pharmesol, a pharmacy AI developer. “It’s a potential solution to solve your problems, and there are considerations for its use just like any other solution you would consider.”

The first step in problem-solving is understanding what the problem is and assessing different ways to achieve the goal. “Depending on what your problem is, if you can solve it in a simpler way, that may be a more cost-effective way to go,” Ms. Park said. “If, for example, you want to automate conversations and you know that the responses typically will be very straightforward yes-and-no exchanges, then you might not want to use AI; a simple decision tree may be a better approach.”

If the conversations are complex “and you want to capture the rich context and nuance from these conversations, you may want to leverage AI,” she said. “Even when we think about identifying high-risk patients, yes, you can build a machine learning model, but you can also do it using simpler rules-based criteria, and you have to assess the difference in performance and return on investment [ROI].”

That said, Ms. Park noted that there are many possible ways in which AI might deliver that needed ROI for the specialty pharmacy. “One thing technology is really good at is scaling,” she said. “It can make multiple concurring calls at a time, and you can keep it on 24/7 to enable that concierge experience to increase touchpoints and patient satisfaction. Other applications can also automate delivery routes and planning, not only based on location and traffic but also cold chain packout times to determine the best timing of deliveries, uncovering any issues and missed deliveries in real time.”

Active involvement of pharmacy stakeholders in AI application development is critical, she concluded. “You want to ensure that the innovators deeply understand the pharmacy operations and their unique needs. Only then the direction and design of solutions will truly benefit pharmacies.”

The sources reported no relevant financial disclosures.