The Rise of AI

Awareness of artificial intelligence (AI) is increasing throughout health care. Relative to pharmacy, the American Society of Health-Systems Pharmacists Foundation’s “Pharmacy Forecast 2020: Strategic Planning Advice for Pharmacy Departments in Hospitals and Health Systems” specifically lists the emergence of AI in its report, inspiring industry discussion and guiding the strategic planning processes of health system pharmacy leadership across the country.¹

For pharmacy, AI provides information on drug interactions, drug therapy monitoring, formulary selection, costs, usage trends, and more.

“AI is already transforming health care but will become increasingly valuable as investments in systems that can capture and manage data are made and clinical informatics entities work more collaboratively to address current data shortfalls,” said Doug Zurawski, PharmD, senior vice president of clinical strategy at Kit Check, Inc, maker of radio frequency identification (RFID)/AI technology for hospital pharmacies to help with medication management. “It is incumbent upon us, in this industry and in this field, to take the lead and learn more, invest in systems that support AI and machine learning, and prepare for the future with access to artificial intelligence.”

Deborah Sadowski, MHA, RPh, director of pharmacy services at Deborah Heart and Lung Center in Browns Mills, New Jersey, said IBM estimated back in 2011 that there were over 161 billion gigabytes of data in the entire health care arena, which she projects has easily doubled by now.

“It is impossible to manage that much information safely and efficiently with the human mind, so it is essential that we incorporate AI as a tool to help us better care for patients—be it preventing medication errors, managing pharmacogenomics applications for better medication choices, or doing prospective cost analysis for treatment choices—all by processing billions of bits of information efficiently and quickly,” she said. “It is key to remember however that AI is a tool to supplement knowledge and should be used in conjunction with, not a replacement for, the clinical judgment and human experience that the pharmacist or other health care provider brings to the table.”

Vin Singh, MBA, MS, the CEO of Gaithersburg, Maryland–based BullFrog AI Holdings, Inc, a precision pharmaceutical company built around an AI platform, noted that pharmaceutical companies have been plagued for years with vast amounts of data, but most do not have the proper technologies to analyze it all.

“We are seeing more and more pharmaceutical companies looking to leverage AI and machine learning solutions to translate their data into therapies that have a better chance of benefiting patients,” he said. “By combining experts in machine learning and AI with subject matter experts in pharmaceutical and health industries, we can approach it with an open mind and share strategies from a multifunctional perspective.”

Wolf Ruzicka, MBA, chairman and IT innovator for Washington, DC–based EastBanc Technologies, said AI can be used to search for similar chemical molecules to already known drugs in huge existing databases.

A new application is to use AI technologies to analyze large clinical trial data sets to link certain responses and adverse effects to individual genetic markers.

“The goal is to find new drugs with similar treatment properties but less expensive or with less [adverse effects],” he said. “As a result, we can determine which patients could be treated successfully, in some cases lifesaving, while for other patients the drug may not be useful.”

New AI Tools

AI represents an innovative way in which entrepreneurs and health care professionals are reinventing the US health care system to better serve patients, which is why many in the pharmacy industry are embracing AI offerings today.

Health care technology company DrFirst recently released its next-generation, patented SmartSig AI technology to improve the quality of patient medication history when it is imported into hospitals’ and health systems’ electronic health records (EHRs).

SmartSig 2.0 accurately translates nearly 93% of incoming prescription information, saving up to 30 seconds of work for each drug entered during the medication reconciliation process.

Rebecca Sulfridge, PharmD, an emergency medicine clinical specialist with Covenant HealthCare in Saginaw, Michigan, credits the new technology with helping to minimize additional patient risks and protecting staff from unnecessary face-to-face exposure during the influx of patients with coronavirus disease 2019 (COVID-19) over the past few months.

“We are spending less time reconciling medication histories manually, and in the first month of use, we recaptured an additional 15% productivity per shift, which represents approximately $650,000 per month, while also improving patient safety and outcomes,” she said. 

Tom Knight, MBA, MS, the CEO of Invistics Corp, a Peachtree Corners, Georgia–based software company working with AI, explained that machine learning is being used to create drug diversion prevention software solutions that are self-learning.

“This is very important to pharmacies and health care facilities as they work to detect and prevent employees from stealing medications, also known as drug diversion,” he said. “Machine learning allows solutions to learn from past diversion incidents, so they become better and better at detecting potential diversion incidents.”

For example, machine learning technology can pick up on isolated incidents that are aligned with known patterns of diversion, such as if a pharmacist is filling prescriptions for oxycodone 3 times more frequently than their colleagues, and synthesize that information with other data, such as reports of suspicious activities from pharmaceutical technicians, and create alerts for managers, or drug diversion specialists. Therefore, drug diversion incidents that might have taken weeks to discover before can be found in hours.

Kara Earle, MBA, MS, senior drug diversion specialist at FairWarning, which works with hospital pharmacies on their drug diversion programs by using AI to help prevent new diversion cases, says she’s seeing a strong push from pharmacies looking to replace older technologies that require manual processes with new, AI-powered technology to streamline workflows.

For example, Mount Sinai in New York, New York, employs FairWarning’s Drug Diversion Intelligence to monitor all daily transactions and detect irregular behavior such as unusual access of controlled substances, medication access after termination, and inventory discrepancies.

“AI has the ability to handle large quantities of data and cross-reference data from multiple systems, so it helps to automate manual steps and increase visibility into the movement of medications to create a broader monitoring picture,” Earle said. “These benefits are big drivers in why today’s pharmacy should be investing in AI-powered technology like drug diversion monitoring.”

A Look Ahead

Although the impact of AI is growing, Zurawski noted that a lack of standardization in EHRs and regulatory hurdles sometimes hinder access to sufficient volumes of reliable data to effectively inform advanced computing applications.

In the years ahead, he believes clinical informatics, vendor cooperation, and collaboration will help address some of the gaps in data access.

Earle noted AI will continue to evolve, and models will become more effective at identifying diversion before major problems arise.

“The goal is to prevent diversion from happening at all, and employing AI-powered technology is a huge step toward this goal—it can catch policy violations and program loopholes before a diversion even occurs,” she said. “For instance, incorrectly provisioned security settings on dispensing cabinets, users not following proper wasting procedures, or nurse managers failing to follow up on open discrepancies are all typical policy violations that may indicate diversion. Identifying those policy gaps early can prevent diversion and save lives.”

AI in the COVID Era

During the coronavirus disease 2019 (COVID-19) pandemic, it is very likely that decisions are being driven by vast amounts of data being accessed from around the world. Health care AI gives providers, administrators and others access to data sets of information (eg, patient demographics, length of stay, treatments and outcomes, costs, geography, survival rates).

“Because AI can identify and analyze trends from data and predict, via machine learning from vast amounts of data, it is highly plausible that AI is already being used to strategically plan for redeployment of resources and space—an application of AI extremely useful during crises like the COVID-19 pandemic,” Zurawski said.

Singh noted that the pandemic requires quick solutions that only AI can bring. “AI can help speed up the drug development process without jeopardizing safety and accuracy,” he said. “The benefits that AI can bring to help tackle the COVID pandemic have yet to be fully realized, and it has the potential to serve as a valuable and irreplaceable tool. Global infectious disease challenges will continue to spawn, and having AI tools that can prevent or minimize the damage to public health is essential.”

Additionally, the change in workflows during COVID-19, such as staff assignments and the addition of contract, travel, and float nurses, have made manual monitoring methods almost obsolete.

“Pharmacies need the ability to track medication usage across departments, and even system-wide and near to real time as possible, which AI is capable of accomplishing,” Earle said. “One of our Drug Diversion Intelligence customers was able to identify diversion and medication tampering by a travel nurse who was hired during COVID-19 within 10 days. This allowed the hospital to minimize any impact to patients and fully investigate and remediate the incident in a fraction of the time it would have taken them with manual processes.”

Reference:

1. Vermeulen LC, Swarthout MD, Alexander GC, et al. ASHP Foundation Pharmacy Forecast 2020: strategic planning advice for pharmacy departments in hospitals and health systems. Am J Health Syst Pharm. 2020;77(2):84-112.doi:10.1093/ajhp/zxz283