Every time we board an airplane, we assume our pilots aren’t going on instinct alone. They’ve worked their way through a checklist—fuel levels reviewed, the flap setting confirmed—to ensure the plane is ready for takeoff. They may be pros, but they’re human, and the complexity of the job makes it easy to forget a step.
Our checklist made all the difference when the Wellcome Trust tasked us with designing a way to cut down on unnecessary antibiotics use in middle-income countries. The world’s antibiotic misuse has staggering implications: By 2050, antimicrobial resistance is expected to kill 10 million people every year. That’s more people than we lose annually to cancer.
What became known as Project Marvelous took us to Kenya and India, where we encountered patients who self-diagnose and prescribe for themselves and their friends, community health workers with very little training, and pharmacies that dole out antibiotics without a prescription in order to make a sale.
As far as complex systems challenges go, it was an extreme one, but after testing several rounds of concepts rooted in behavior change, we found an intervention with real potential. Smart Prescriptions, as we called it, was an endpoint that varied drastically from where we began. The original hypothesis was that the intervention should be designed for pharmacists. But after following the checklist, we found we had to start with doctors.
Here are the five steps we followed to address this complex systems challenge:
We can’t claim to have reached an understanding of the whole ecosystem by a longshot—the sheer number of stakeholders involved is what makes it such a complex problem. So at the outset, we defined our ecosystem, drawing a line at pharmaceutical companies, which seemed beyond the scope of the design challenge Wellcome had set. We then set out to interview antibiotic users, doctors, and pharmacists. In doing so, we realized there was a category we had overlooked: India’s tens of thousands of registered medical practitioners (RMPs). They typically aren’t licensed to prescribe antibiotics but have studied Ayurvedic medicine. Being willing to pivot and expand (or contract) the ecosystem is key: We began interviewing RMPs after realizing the huge impact their population has on antibiotics consumption.
In interviews, we positioned ourselves as people who didn’t know much about the topic. We were there to learn about them as people—their motivations, their struggles, their day-to-day. We listened. We made sure our questions were free of judgement and left our assumptions at the door. What we came to understand is that a primary motivator for doctors is to succeed as a business. They depend on patients coming back.
This insight made it through to the final design. Smart Prescriptions is, at its core, a tool to improve the doctor-patient relationship. Features like automated advice for patients and medication reminders increase the personal interactions doctors have with patients. The antibiotics recommendation is a Trojan horse, tucked inside the features doctors were really clamoring for.
Continually put yourself in your user’s shoes and make sure you’re seeing the situation from their perspective. We quickly came to understand, for instance, that many people working in pharmacies aren’t actually pharmacists, and that their financial motivations eclipse any duty of care. They dole out antibiotics without a prescription because refraining from doing so means sacrificing a sale and possibly losing a customer to a rival. Rather than slapping a “bad” behavior label on this, we accepted it as their reality and used it as a design constraint.
Systems challenges are often as sensitive as they are complex (consider homelessness or mental illness). As designers we are outsiders, so in designing Smart Prescriptions, we needed to make sure doctors were upheld as the experts. We weren’t creating a diagnostic tool, but rather an intervention that offers a recommendation based on their diagnosis. The tool doesn’t actually change the way they prescribe, it just offers guidance. It’s tough to remember every antibiotics guideline, especially as they update. The language that was used was consciously designed to be light and supportive of their expertise.
Don’t fall in love with your first good idea. Every assumption we made was challenged in our first trip, which ended up being more exploratory than we expected. We made a point to talk to every stakeholder and actor about our sacrificial concepts—early ideas we put in front of people to spark reactions. We gauged whether they had a positive or negative response, then iterated accordingly. After the discovery phase, we came up with seven concepts to test with participants and gauged their reactions. Smart Prescriptions was the clear standout, generating the most interest and excitement. We refined the concept back home, then returned to India and Kenya to gather more feedback. This allowed us to design something that spoke to doctors’ primary motivation, and to add features that the doctors felt would make it even more useful.
Unlike pilots, we don’t have the benefit of a known destination. Our checklist doesn’t just guarantee a smoother ride, it plays a major role in mapping the route itself. When the ecosystem is unfamiliar and the complexity is vast, it is a must.