David Sellars, CEng, Principal, Product Innovations at DrFirst, has over 20 years of healthcare experience with a deep emphasis in Big Data, AI, healthcare interoperability, programming, and systems optimization.
ComputerTalk: What do I need to consider about AI for my pharmacy?
David Sellars: News about AI [artificial intelligence], especially generative AI, is bombarding us from every channel these days. It’s important to look past the hype and evaluate how AI can help your pharmacy. While it’s certainly not a panacea, it’s remarkably good at solving some problems that traditional programming practices don’t address well. There is a lot that it can do to help pharmacies operate more efficiently, help improve productivity, and avoid errors and adverse events.
First, I want to clarify that at DrFirst, we use the term “augmented intelligence” because our AI solutions support clinicians’ decision-making rather than replace it. This makes it easier and faster for pharmacists and pharmacy technicians to do their jobs while keeping control in their hands. The AI handles the data entry, and the pharmacy staff verifies it.
With augmented intelligence, humans are always part of the equation. AI serves the role of “co-pilot,” which allows the pharmacy staff to make the final call whenever needed.
It’s important to remember that not just any type of AI is appropriate for use in healthcare, just as a dentist wouldn’t use a cordless drill intended for home repairs to prepare a cavity for a filling. Clinical-grade AI, vetted by clinicians for use in healthcare and used in an augmented intelligence role, is the critical element here. It has a better safety rate than humans, and when you couple the AI with the expertise of healthcare providers, the safety rate is even stronger. In addition, the AI improves over time with machine learning that can identify the root causes of mistakes and make changes to avoid them in the future. Working within carefully considered boundaries, the system alerts pharmacy staff if there isn’t enough information to safely infer clinical intent, preventing “hallucination errors.”
With augmented intelligence, humans are always part of the equation. AI serves the role of “co-pilot,” which allows the pharmacy staff to make the final call whenever needed.
An AI co-pilot can do a great deal of good for pharmacies and patients, especially considering ongoing staffing challenges and the time it takes to review and edit e-prescriptions. Nearly 84% of e-prescriptions require manual edits by pharmacy staff, and half have errors in patient instructions, according to a recent study published by BMJ Quality & Safety.
ComputerTalk: How can AI make my pharmacy workflow more efficient, safer for patients, and reduce staff burnout?
Sellars: Efficiency is AI’s superpower. Saving time on menial tasks like data entry means that pharmacy staff can process prescriptions more quickly with less stress. Managing stress matters, especially when so many pharmacies are short-staffed and stretched thin. This also creates an opportunity for pharmacists to have more time to counsel patients about their medications, which should lead to fewer ADEs and the best outcomes for each patient.
Click On The Image To Download A PDF Of This Content
Think of the elements that go into every step of a pharmacy transaction: reading and confirming the prescription, entering missing information into the pharmacy management system (PMS), and matching the drug on your shelves to the NDC the prescriber used. Our clinical-grade AI automates these tasks in the pharmacy management system; it converts the sigs to the preferred nomenclature, safely fills in missing data, and matches the NDCs to the drugs in your inventory. If needed at any point, the AI will flag issues needing staff input. AI handles tedious, mundane tasks and avoids keyboard errors so pharmacy staff can spend more time on patient care.
Here are some ways that AI improves efficiency and safety:
- Structures free text. Much of the information in an e-prescription arrives as free text because the structure as sent doesn’t match your PMS, or it contains notes from the prescriber. Free text adds friction because staff must read through it, make decisions, and manually add it to the correct fields in the PMS.
- Fills in missing data. Our AI safely infers missing sig data and adds patient allergies to the PMS. It also transcribes the prescriber’s NDC to what the pharmacy actually stocks, which may seem like only a matter of efficiency but significantly improves safety.
- Documents allergies. Patients’ allergies are not always noted on prescriptions, and when they are, it’s in free text, which would require staff to enter the information manually. If the allergy is not relevant to the prescribed medication, pharmacy staff may not take this step. That is a missed opportunity to capture data that could avoid an adverse drug reaction in the future when a different prescriber may not be aware of the allergy. Just as it does with sigs, our AI can recognize the allergy and add it to the appropriate fi eld in the PMS for decision support.
- Matches drugs. If the doctor prescribes a drug and the NDC is missing or doesn’t match the pharmacy’s system, pharmacy staff need to look it up manually, which is an opportunity to introduce errors. For example, drugs may have similar spelling or pronunciations. Or they are offered in a multitude of strengths. Misplacing a mouse click by one line can result in a dispensing error. AI can transcribe the NDC sent with the prescription so your staff doesn’t spend time hunting down what your pharmacy stocks.
These may not be headline-grabbing examples of AI, but they bring value by delivering meaningful time-savings hundreds of times a day, which adds up!
Increased use of clinical-grade AI as a part of augmented intelligence will free up your pharmacy staff to do more of what drew them to the profession in the first place, including counseling patients and collaborating with other clinicians so patients achieve the best results from their medications
ComputerTalk: What distinguishes automation and AI? And how can AI improve automated processes?
Sellars: People often conflate automation with AI, but there is a difference. Automation doesn’t require the “I” for intelligence; it’s simply an “if this, then that” situation. The same thing will happen every time.
AI is different from automation. With AI, the “inputs” can cause a change in the process in real time. Automation is greatly enhanced when AI is in place, especially in the context of augmented intelligence.
For example, let’s consider dispensing automation. What happens if the prescriber adds a note to the e-prescription? Should it change anything about the downstream process or not? Maybe yes, maybe no. The message might instruct the patient to finish a previously prescribed drug before starting this new one. Or may indicate the patient should take both prescriptions on alternate days. These types of treatment instructions can only be shared in a note. Automation cannot assess the impact of this instruction and make any changes to the prescribing process. But AI can.
Increased use of clinical-grade AI as a part of augmented intelligence will free up your pharmacy staff to do more of what drew them to the profession in the first place, including counseling patients and collaborating with other clinicians so patients achieve the best results from their medications. CT