Are we blocking innovation by not accepting AI? We talk with Louisa Harvey about how embracing AI in the right way might shape how devices and packaging can be better targeted to patients.
"I think it's really important to see devices and packaging as one system, one unit."
Key takeaway: "There is always an inherent risk when anything is connected. However, I do think that we could be blocking innovation by not accepting AI. I feel very strongly that being able to use AI is already proving to be amazing in the field of medicine. I’ve felt for a long time that market insights and human factors haven't been recruiting the right people to take part in research, no matter how stringent the screening criteria."
How might AI tools enable life science manufacturers to better target the medical device/packaging system for improved patient outcomes?
We talked with Louisa Harvey, principal & founder at Harvey Medical (a ClariMed company), with extensive experience in the pharmaceutical and medical device industry since 2004. Passionate about the usability of medical devices, Harvey founded Harvey Medical in 2012, Healthcare Fieldwork in 2016 and co-founded the Medical Human Factors Network in 2018, with the mission to raise the bar of medical human factors work in the UK.
[Editor’s note: answers have been edited for brevity.]
1. What are some of the ways that a pharmaceutical or medical device manufacturer could target experiences or device/packaging designs to regions or patient populations with help from data and AI?
Everything I speak about today is pertaining to both the devices and the packaging. I think it's really important to see devices and packaging as one system, one unit. A user should instantly feel a continuous energy between the device in question, the packaging, and any other materials. This is not only helped with the use of branding but also design cues. Most of all it should be simple.
Sometimes there's a bit of a disconnect. People might really love the packaging and they feel excited to open the box of their life-saving medication—then the vibe they get when they actually see their medical device and use it isn't quite there.
There is a huge amount that we can do with AI in the areas of research that I work in, both human factors and market insights/UX. One of the big differences is going to be the way that we can mine existing data. Many pharmaceutical and medical device companies have a whole wealth of data contained in final study reports, market insight studies, raw audio/video files and human factors reports. AI looks for patterns, patterns that might not have been immediately obvious if a human was to wade through thousands of pages.
Companies could feed all that existing data into a system, then find out trends and current research topics. This in essence would be classified as synthetic data. But be mindful that synthetic data does not live. It is only as current as the data, so companies should continue to feed the system with new insights. Preferably the raw data files, as emotion tracking could retrospectively be used on the videos to provide greater context to responses. It’s important to remember to get to the informed consent right in all studies going forward to ensure any further analysis is ethical and responsible.
2. Do you think companies are hindered by lack of data or incomplete data? Do they already have access to data sets they can feed in? Does it depend on the company?
It depends on the size of the company. For big pharmaceutical and medical device companies, I'm sure there's a lot that can be harnessed. But for smaller companies, they don't have that and there isn't anything readily available. They could use HE75 standard, which is full of rich information on how to develop your device with a bit on packaging, too. But that's really all that they have access to, and various literature searches. The beauty of AI is that it should be capable of pulling competitor information swiftly from databases such as the MAUDE database and others, so companies (regardless of size) can learn what ‘mistakes’ have been made with competitor devices so they can be ahead of the game.
3. What’s an area that you think device and pharma companies could improve upon?
I think that there’s been a lack of market insight (UX) work done early on in many of the devices that I've seen lately. We like to make sure they are put in front of users right from exploratory— people might not even have a concept. For example, you're speaking to people who live with the condition, who are going to be opening the package and using the device, right through to doing the human factors validation testing at the end stage, which is where we're looking at whether the device is safe, easy, and effective for use.
Because that end stage is regulated, people are very much focused on that part, whereas the answers to create a differentiating trial-blazer is much earlier than that. You're almost glued with the wrong device if you don't speak to non-biased users early on. As researchers, I think we are often not asked to ask the right questions by the companies that we work with. We educate and advise how these questions might look, and I think there's a lot to be done in that respect to make sure you're really able to design and develop things that are truly novel.
It's the same with packaging, too. I'm very rarely asked to look at packaging until the end of the design process, where I’m suddenly given the final packaging and instructions for use with the device. I would’ve liked to have had an input early on to make sure that that energy throughout the system is relayed across all the materials within it. When you're brought in at that last minute, you're under quite a few constraints in terms of making any changes.
Sometimes just not having an idea is better. I would rather people come to me and say, “What is going to be the future device in this area?” and not even give me a concept for use. Because it inhibits us immediately—as soon as a participant that we're testing sees that [concept device], they are completely drawn into that device design, instead of really enabling inventive thinking.
4.That can be counterintuitive. Many people in life sciences feel an invisible pressure to “know the answers” right at the outset, where there’s a lack of curiosity or freedom to admit where they’re not really sure.
I think AI will help with that, too, because it will enable us to pull together all sorts of common themes off huge reams of datasets, which a human just wouldn't be able to do in the amount of time available. So, I’m hoping that we'll get innovation turned around a lot quicker than what we've been seeing.
5.What sorts of regulatory hurdles might there be to using AI for targeted design purposes. Data needs to be in a sanitized form before it enters any models, correct?
Absolutely. Any data that comes out of research studies is always anonymized if you're doing everything correctly. Even at the point of the participant turning up, they're given a participant number. So it's already anonymized to a certain extent, and there's a lot that goes into that to make sure that the person isn't identifiable.
But beware of pitfalls, for example if you've got an orphan indication, where there's only five people in the world with that condition, it could be obvious as to which person was talking about it [based on demographics]. So, you have to be very careful to make sure that it is truly anonymized.
There’s a whole raft of privacy and ethics involved in this particular scenario. I'm sure there'll be more and more documents coming out about how to appropriately use data and ask for people's consent.
In terms of regulatory perspective, a lot of devices we're now seeing contain AI. I was just looking at the FDA website the other day to see how many contain AI and it's a huge proportion of what they're now approving, so we need to be really aware of that. Across the globe, regulators are looking at plans and pathways to work out how devices should be developed responsibly when they contain AI and how that testing is going to take place. There will be quite a lot of regulatory changes that are going to be happening in the near future.
6. With targeted therapies, are we going to need targeted components?
We're looking at a lot of personalized therapies which is fantastic in the world of biologics. I’m hoping that there is enough platform technology research that will allow platforms to be used for multiple personalized medicines.
I know the impact that changes can have on manufacturing lines. It could be extremely costly, for instance, to change the primary drug container. However, I do feel that people have been locked on to the 1-mL and 2.25-mL glass syringes for a long time. At some point, I think more companies will decide to take the risk to address a novel primary drug container to produce a novel device. I don't know at what point that is going to happen.
But on the whole, large pharmaceutical companies have large manufacturing lines, and they don't want to have to change those. At times, I think the solutions are already out there. There are potentially simpler solutions to the problem using existing technologies. Again, it's a matter of not going into market insight studies with a perception of what that end device is going to look like… but instead asking freely, “What might people like?” and seeing whether we can encase existing primary drug containers with the new novel ideas. That's an area which is very, very tricky for drug manufacturers at the moment with the complexity of high volumes. High viscosity is always a big issue, as well.
7. In terms of the ethics questions that you run into, is patient consent the main focus or are there other hurdles?
AI in general can inherently produce bad outcomes, which people have seen all over the news, and TV programs like Bleeding Edge, which has created immense social distrust. There is always an inherent risk when anything is connected.
However, I do think that we could be blocking innovation by not accepting AI. I feel very strongly that being able to use AI is already proving to be amazing in the field of medicine. I’ve felt for a long time that market insights and human factors haven't been recruiting the right people to take part in research, no matter how stringent the screening criteria. My colleague and I, alongside two amazing software developers developed an app—when people are screened to take part in studies, we can detect whether they're likely to adhere to the medication as prescribed or not.
We can really target the materials, packaging, and device to those people who would inherently struggle to take their medication. For a long time, I'd started to see a pattern of people turning up for research that are the sorts of people who would take their medication regardless of what their device or packaging looks like.
At Pharmapack, we were speaking about sensory concepts, as well. I think that's a huge area that's not being looked into. But yet, it is looked at in the consumer world. Particularly with regards to packaging, there are perfect opportunities to enhance the energy of the system by stimulating other sensory factors.
For example, even as you open the box, there’s a certain scent that makes you feel relaxed about taking the medication. I'm sure there's some sensory innovation to be had there as well.
8. On a personal note, any books that you've read recently?
Earlier this year, I was gifted a book from one of my clients. I'd been to present on World Usability Day, and they gave me this book, “Design for a Better World.”
It’s really fantastic, because one thing we haven't touched on today is the world of sustainability. That book is interesting—it's written by Don Norman, who heavily focuses on usability and human interaction.
His book gave me hope that we can design our way out of the hideous situation we have created.
At the end of the day, for devices to be sustainable, we have to have the input from the people who are going to be using that particular device, whatever the system is. We cannot impose recycling strategies on people without getting their input.
It also gave me some thoughts on disposable devices versus reusable, because what you put into the world, you should be able to take out in a circular economy system. I do think that's probably the way forward. There’s been a lot of talk about reusable devices. But actually, I still think there is a place for disposables as long as we can get the devices back. This book was quite interesting in that respect, and I ran some research on that, seeing whether people would be happy to go and take their device and packaging back to the pharmacy to be recycled. In the UK, it was 85% of people said they would have taken the trip to go and take it back. This was surprising but again hopeful.
Cutting-edge packaging technologies are transforming the medical device sector in PMMI’s “Innovation in Medical Device” report, featuring advanced materials, smart solutions, and evolving regulations. Editors share insights from nearly 300 PACK EXPO booth visits—each product deemed new and truly innovative—alongside video demonstrations of the equipment and materials on display.
Discover the latest breakthrough packaging technologies shaping the pharmaceutical sector. This report dives into cutting-edge innovations, from smart containers that enhance patient safety to eco-friendly materials poised to transform the industry’s sustainability practices. All from PACK EXPO. Learn how forward-thinking strategies are driving efficiency and redefining what’s possible in pharma packaging.
