Probably not the pharmacy AI article you’ve been waiting for

Pharmacy, like many other professions, could very well be significantly disrupted by AI in future.Perhaps uncomfortably so.But if we figure out how to pre-empt the incoming changes, I think we could cowabunga over the digital tsunami and become stronger than ever.Remember our old pal Bloom, infamous for his learning objective taxonomy? You know, the one presented in the colourful pyramid?No? Yeah, it’s OK. I fell asleep in that class too.In brief, it classifies learning objectives by complexity and specificity, ranging from the relatively easy (i.e., remembering) to the tough (i.e., create). See figure 1 below. It also throws cold water on any strange sense of pride one gets from simply memorizing and recalling stuff to a patient.My theory is that as AI models become more sophisticated, the complexity and specificity of the things they can learn and subsequently iterate with precision will improve. These are simply my theories, mostly fuelled by too much LinkedIn doom-scrolling, high school computer science classes of yesteryear, and a lot of coffee. Take my ideas with a boulder of salt, then check your blood pressure afterward.So let’s map our unique learned knowledge and skills to the pyramid, and see where we stand. This isn’t an exhaustive list, rather, illustrative. And it comes with healthy consideration given to the fact that when the stakes are high (like in healthcare), we demand the highest degree of precision. Read: 10 technnology innovations changing pharmacyLet’s go:RememberThis would include side effects, contraindications, dosing, and labelled and unlabelled indications. These can all easily be sourced from monographs, resources and guidelines and essentially restated. UnderstandAs we get a bit more complex, my take is that this might involve comparative details of drugs, clinical presentations, surrogate markers and general appropriateness of treatments. This data could conceivably be similarly sourced from various data sources like RxTx and UpToDate.ApplyI imagine that this would implicate uncomplicated treatment planning, clinical monitoring, and standard counselling points. It also might involve leveraging a minor ailments algorithm to figure out a treatment. Sourcing this data probably involves similar usage of references mentioned above.AnalyzeNow we start getting a bit more complex, as tasks like managing drug-drug, drug-nutrient and drug-condition interactions begin to push on our deeper cognitive functions despite often being originally flagged by software alerts. Given the level of personalization needed in these decisions, you can see how a pharmacist could start elbowing out the bots.EvaluateThis is all about selecting or suggesting treatment options to allied treatment providers, particularly when having to balance multiple factors (e.g., contraindications, organ function, emerging evidence, established treatment guidelines, and patient preferences). And then course correcting if goals aren’t being met or new challenges are seen, like adverse events. It could also mean deprescribing. Sure, algorithms might help here, but pharmacists would clearly reign supreme.CreateThis could be literal - like designing a customized compounded product to meet unique patient needs, investigating and determining causes of rare, paradoxical or strange adverse effects, or appraising and advocating for 3rd line or experimental treatment options when all else fails, often leveraging findings from published case reports.So where does this leave us? By focusing on analyzing, evaluating and creating, I would surmise that we have a rough sketch of where I think the profession should consider starting to place more of its eggs. You know, doing complicated, specialized pharmacotherapy-focused work in order to benefit patients, with algorithmic and AI-based tools riding in the sidecar as they become available. Despite what our past (and present) looks like, change is within us, inertia be danged.At least until the algorithms get better and replace the entire healthcare system at large. But for now, it’s time for us to jump to the top of the pyramid.

Probably not the pharmacy AI article you’ve been waiting for

Register for free access or login.

Related Topics