Every hospital we have worked with over the past twelve years starts the conversation the same way. ‘Our patients know what they need to do. They just don’t do it.’
I have heard this from GPs, hospital executives, and mental health service directors. And they are right. The treatment plans are solid. The patient education is thorough. The follow-up systems are in place. And yet half of all patients with chronic conditions do not take their medication as prescribed.[1] Lifestyle changes do not stick. Prevention programmes reach 15% of the people they were designed for.
Healthcare keeps investing in the same solution: more and better information. Clearer brochures, digital portals, multilingual materials, motivational interviewing techniques, reminder apps. That is the equivalent of giving someone a map and expecting them to run a marathon. The information is not the problem. The behaviour is the problem.
The missing variable is human behaviour. And the discipline that addresses it is called behavioural design.
Behavioural design for healthcare applies behavioural science to the gap between medical knowledge and patient action. Instead of assuming patients are rational decision-makers, it maps the psychological forces that drive health behaviour through the SUE | Influence Framework©. By thinking from the patient’s Job-to-be-Done rather than the clinical task, healthcare organisations redesign patient journeys, choice environments, and default options to work with human psychology, not against it.
The numbers behind the behaviour gap
The wrong question healthcare asks
In 2019, we mapped the patient journey for a major diabetes care organisation. The brief was to improve treatment adherence: help patients better follow their self-management protocol. We started where we always start: the patient’s Job-to-be-Done.
Clayton Christensen defines a Job-to-be-Done as “the progress someone is trying to make in a given circumstance.”[5] The standard healthcare question is: how do we get patients to manage their diabetes better? But that is the system’s question, not the patient’s. When we asked patients what they actually wanted, the answer was universal and fundamentally different: to live a normal life. Not to manage their disease. To remain a woman, a father, a colleague. Someone who goes on holiday and does not have to think about their blood sugar six times a day.
That shift changed everything. The care programme designed to optimise treatment adherence was optimising the wrong thing. It imposed the disease identity that patients were actively trying to avoid. When the programme was redesigned around the question “how do we help you live a normal life with diabetes”, enrolment was full within two hours.[6]
As I write in The Art of Designing Behaviour (2024): “What people say and what people do are two fundamentally different things.” Patients say they want to improve their health. They say they want to be adherent. But their actual behaviour reveals a deeper desire: to feel normal, not to be identified with illness, to live the life they had in mind.
Patients don’t fail treatment plans. Treatment plans fail to understand what patients actually want to achieve.
This is the core of what behavioural design does: it starts from the patient’s Job-to-be-Done, not the clinical task. It then maps the environment driving or blocking that behaviour, and redesigns it so the desired behaviour becomes the path of least resistance. Change the environment, not the person.
Three healthcare challenges that behavioural design solves
Challenge 1: the medication that sits in the cabinet
A regional hospital was struggling with post-operative treatment adherence. Patients were discharged with clear instructions, printed materials, and a scheduled follow-up appointment. Three weeks later, nearly 40% had stopped taking the prescribed medication prematurely. The clinical team’s response was to add more information to the discharge conversation: a longer explanation, a more detailed leaflet, an extra nurse follow-up call.
Adherence did not improve. Because information was not the problem.
The real barrier was a combination of present bias (the medication had no immediate felt benefit, so the advantage of taking it did not register), friction (the dosing schedule did not fit into existing daily structure), and simply the absence of a cue. Once home, patients returned to their normal environment where nothing reminded them at the right moment. Research by the Behavioural Insights Team and Imperial College London showed that even redesigning medical prescription forms, pre-printing dosage options rather than leaving them blank, significantly improved prescribing accuracy without any additional clinical training.[7]
Challenge 2: the wellbeing programme nobody uses
A large hospital group launched a comprehensive wellbeing programme for clinical staff. Yoga classes, a mindfulness app, access to a counsellor, discounted gym membership. The programme was well-funded and well-communicated. After six months, participation was below 8%. Leadership concluded that staff simply did not care about their own wellbeing. That was wrong.
Staff cared enormously. But the programme required them to do something extra within a schedule that already had no slack. Taking a yoga class means asking a colleague to cover your shift. Using the mindfulness app means finding ten uninterrupted minutes that do not exist. Visiting the counsellor means showing vulnerability in a professional culture that rewards stoicism. Every option demanded effort from people who had none left. More than 50% of healthcare professionals experience burnout symptoms.[8] A wellbeing programme that requires mental energy from people whose mental energy is depleted is not a bad programme. It is a badly designed one.
Challenge 3: the prevention programme with 12% uptake
A regional health network developed an evidence-based prevention programme for cardiovascular risk. Target group: men aged 45-65 with elevated risk factors. The programme was free, clinically validated, and available through their GP. Participation rate: 12%.
The programme was designed by clinicians who assumed that a free, evidence-based offer would be sufficient motivation. It was not. The men in the target group did not feel ill. Their risk factors were invisible to them. Enrolling required calling the GP practice during working hours, making an appointment, and committing to a series of sessions. Every step added friction. And optimism bias meant most believed they were healthier than the statistics suggested.[9] Research on loss framing in mammography screening showed that women who received a loss-framed message made significantly more screening appointments than those presented with the same information as a gain, with the effect measurable at 12 months.[10]
Influence Framework analysis: what drives and blocks health behaviour
The SUE | Influence Framework© maps the four forces that determine whether patients, staff, or health consumers actually change their behaviour. It always starts from the Job-to-be-Done: what progress is this person trying to make? Only once you know that can you diagnose the forces driving and blocking that behaviour. In healthcare, a consistent pattern emerges: the blocking forces are stronger than the driving ones, and they are almost entirely unaddressed by standard clinical communication.
Why patients don’t follow medical advice, despite understanding it
The JTBD of the average patient is never “manage my chronic condition well.” It is: live a normal life with my health condition. Be a parent, a partner, a colleague. Go on holiday, ride a bike, have a glass of wine. Once you understand that, you understand why medical treatment protocols systematically fail when they centre the disease and build the person around it. Clinical communication addresses the driving forces effectively. Patients understand that they are ill, understand the treatment, and understand the consequences of non-compliance. What remains unaddressed are the blocking forces: the deep comfort of existing habits and the immediate anxieties that make treatment feel worse than the illness.
Symptoms are visible and worsening: For patients with acute conditions, pain, discomfort, or functional impairment creates genuine urgency. The status quo is physically uncomfortable. This is the strongest driver in healthcare, but it only exists for symptomatic patients. For prevention and chronic disease management this force is largely absent. A diabetes patient feels no extra pain when they miss their medication. That is precisely what makes it dangerous.
Medical authority: A recommendation from a doctor carries significant weight. Patients generally trust clinical advice and want to follow it. The intention is real. The gap between intention and action is where the problem lives, and that gap is a design problem.
Social pressure from family: Partners, children, and concerned friends create external pressure. This can be a genuine driver, particularly for lifestyle changes. But social pressure operates episodically, not continuously, and fades quickly after an initial conversation.
Better quality of life: The promised outcome of adherence is real and meaningful: less pain, more energy, longer life. But these gains are future-oriented. Present bias means people systematically discount future benefits in favour of present comfort.[11] The pill you take now for a benefit you feel in ten years always loses to the comfort of not having to think about it.
Normalcy and identity reinforcement: For patients whose JTBD is to live normally, treatment adherence is actually a gain: it enables them to do what they want to do. But this is rarely communicated that way. Healthcare communication reminds people of their illness, not of the normal life that recovery makes possible.
Sense of control: Taking agency over your own health can be empowering. This is a genuine motivator for some patients, but it needs to be activated by design, not just communicated.
Existing routines are automatic and effortless: Patients have deeply embedded daily routines that contain no medication, no exercise, no dietary change. Every deviation requires conscious effort. Conscious effort is unsustainable over weeks and months. Habit research shows it takes an average of 66 days for a new behaviour to become automatic.[12] Healthcare asks patients to change routines but does not design the conditions for habit formation.
“I feel fine”: For chronic and preventive conditions, patients often feel no immediate discomfort. The status quo feels perfectly acceptable. Why change something that does not feel broken? This is especially powerful for hypertension, elevated cholesterol, and early diabetes, all conditions with no symptoms until it is too late.
Healthcare as safety net: The knowledge that medical care is available if things go wrong can reduce urgency. “If it gets worse, I’ll see the doctor” is a comfort that systematically delays action.
Side effects and loss of control: Medication side effects, even mild ones, create immediate negative reinforcement. Patients associate treatment with feeling worse, not better. This is a powerful anxiety that rational explanation cannot overcome. Radiology research showed that adding a patient photograph to their imaging file gave radiologists significantly more empathy and attention during diagnosis, resulting in more findings and fewer errors. The personal dimension changes the quality of care.[13]
Identity disruption: “Being a patient” or “being someone who takes daily medication” conflicts with how many people see themselves. Especially for chronic conditions, accepting the patient identity feels like a loss. This operates well below conscious reasoning and explains why patients who have all the information still discontinue treatment.
Complexity anxiety: Multi-drug schedules, dietary restrictions combined with exercise protocols, or rehabilitation programmes requiring daily effort trigger anxiety about competence. “Can I actually keep up with this?” is a question patients rarely voice but frequently feel. When that anxiety is large enough, not starting is more comfortable than starting and failing.
The key insight: Healthcare communication overwhelmingly targets the driving forces: explain the Pains, promise the Gains, present the evidence. But patient behaviour is determined by the blocking forces, which operate below the level of rational argument. The comfort of existing routines and the fear of side effects and identity loss are not overcome by better brochures. They are overcome by redesigning the patient journey so the desired behaviour is easier, more familiar, and more directly connected to the patient’s JTBD than the alternative. As I write in The Art of Designing Behaviour: “Simplicity eats willpower for breakfast.” Every step you remove is more powerful than any information you add.
Five behavioural interventions for healthcare
-
Redesign the discharge moment as a behavioural mechanism (SPARK + AGAIN)
The discharge conversation is the most impactful moment in the patient journey, and it is almost universally badly designed. It happens when patients are cognitively depleted, emotionally relieved, and focused on getting home. Information delivered at this moment does not stick. Redesign the moment: couple medication to an existing daily habit (piggybacking, SWAC C25), give a visual 30-day tracker rather than a text-heavy leaflet, and create a concrete implementation intention: not “take your medication every day” but “tell me, when do you make your coffee tomorrow morning?” That is the moment of medication. Recorded, concretised, coupled to an existing behaviour.
-
Default-optimise appointments and medication schedules (CAN)
Every friction point between a patient and the desired behaviour is a drop-off point. Scheduling that requires a phone call loses patients who do not call. Medication that requires a complex dosing schedule loses patients who cannot integrate it into their routine. Design the defaults: pre-scheduled follow-up appointments with opt-out rather than opt-in, simplified once-daily schedules where clinically possible, automatic prescription renewals. At Toddington Medical Centre, a simple intervention where patients wrote their own appointment cards, instead of the receptionist, resulted in an 18% drop in no-shows.[14] People attend appointments they have committed to themselves.
-
Use social norms and loss framing in patient communication (WANT)
Patients are powerfully influenced by what they believe other patients do. “73% of patients in this practice complete their screening” is more effective than any clinical argument for screening. And loss framing works better than gain framing for preventive health behaviour. “Patients who skip this check miss their earliest chance at treatment” moves more people than “this check helps you improve your health.” Research on opioid prescribing showed that physicians prescribed significantly fewer opioids after learning that a patient had died of a fatal overdose, a loss framing that changed colleague behaviour without any direct instruction.[15]
-
Reduce clinical decision fatigue through environmental design (SPARK + CAN)
Clinicians make hundreds of decisions per shift. Research shows that the quality of medical decisions measurably declines as the day progresses, and that doctors at the end of a shift are significantly more likely to prescribe antibiotics than at the start, not because they know less but because deciding costs energy.[16] This is not personal failure; it is decision fatigue. Design the clinical environment to protect decision quality: schedule complex diagnostic decisions in the morning, use standardised protocols for routine decisions, redesign the EHR interface so the most critical information appears first.
-
Build repetition and habit into the patient journey (AGAIN)
One consultation does not change a lifelong health behaviour. Habit formation requires repetition in a consistent context. Design micro-touchpoints into the patient journey: a weekly two-question check-in via SMS, a monthly automated review of adherence data, a quarterly brief conversation with a practice nurse. The power of digital reminders is empirically established: in a randomised trial in Kenya, weekly SMS support for HIV patients produced 13 percentage points better adherence versus standard care, assessed through pharmacological monitoring.[4] Each touchpoint reinforces the desired behaviour and catches non-adherence before it becomes permanent.
Which cognitive biases matter most in healthcare
Health decisions are shaped by the same cognitive biases that operate in every other domain, but the stakes are higher. Here are the five biases with the greatest impact on patient behaviour, clinical decisions, and healthcare system design.
Status quo bias
Patients cling to existing routines not because they have evaluated the alternatives and rejected them, but because the current behaviour requires no effort. In chronic disease management, this is the single biggest barrier to lifestyle change.
Read the full analysis → Behavioural DesignPresent bias
The benefits of treatment adherence lie in the future. The costs are immediate. Patients systematically choose present comfort over future health. This is not irrationality. It is how the brain is wired.
Read the full analysis → Behavioural DesignFraming effect
“This procedure has a 90% survival rate” and “this procedure has a 10% mortality rate” are the same fact. Patients respond entirely differently to each. How treatment options are framed determines which option patients choose.
Read the full analysis → Behavioural DesignOptimism bias
Most people believe they are healthier than average and less likely to develop serious conditions. This makes preventive care feel irrelevant. “That won’t happen to me” is optimism bias operating at population scale.
Read the full analysis → Behavioural DesignDecision fatigue
Clinicians making decisions at the end of a long shift are measurably less accurate. This is not about competence. It is about the depletion of cognitive reserves. Designing clinical rosters that protect decision quality is a patient safety intervention.
Read the full analysis →Frequently asked questions
How does behavioural design work in healthcare?
Behavioural design for healthcare maps the psychological forces that determine whether patients follow treatment plans, attend appointments, and change health behaviours. It always starts with the patient’s JTBD: what progress is this person trying to make? Not “manage my condition”, but “live a normal life.” The SUE | Influence Framework© then maps the forces driving and blocking that behaviour, and SWAC interventions are designed to redesign the clinical environment, appointment defaults, medication packaging, and discharge moments.
What results can healthcare organisations expect from behavioural design?
The results are substantial and empirically documented. Opt-out organ donation registration increases consent from 20-27% to near 100%. SMS reminders for HIV treatment adherence produced 13 percentage points improvement in Kenya. An opt-out appointment system more than doubled prevention programme participation. Patients who write their own appointment cards are 18% more likely to attend. A hospital that redesigned the discharge moment saw 28% better adherence in three months. These are not exceptions. They are the predictable result of designing for behaviour rather than knowledge.
Is behavioural design in healthcare the same as nudging?
Nudging is one tool within behavioural design, but behavioural design is broader. Nudging typically changes one choice environment, such as defaulting patients into appointment reminders. Behavioural design starts with JTBD thinking and a diagnostic phase using the Influence Framework to map all four forces, then designs a system of interventions across the entire patient or staff journey. The SWAC codes (Spark, Want, Again, Can) ensure every intervention is tied to a specific behavioural mechanism.
How does the SUE Influence Framework work in a healthcare setting?
The SUE | Influence Framework© maps four forces for any health behaviour: Pains (what makes the current situation unsustainable), Gains (what patients stand to win), Comforts (what makes current behaviour feel safe), and Anxieties (what makes change feel risky). In healthcare, the blocking forces almost always dominate. The comfort of familiar routines and the fear of side effects are stronger than rational medical arguments. Understanding these forces allows you to design interventions that address them rather than argue around them.
Can behavioural design improve treatment adherence?
Yes. Non-adherence is primarily a behavioural problem, not a knowledge problem. Patients know they should take their medication. They do not, due to present bias, friction in complex dosing schedules, and status quo bias (existing routines do not include the medication). Behavioural design addresses this by simplifying schedules, coupling medication to existing habits via piggybacking (SWAC C25), using commitment devices that convert intention into action, and redesigning packaging so adherence becomes visible and effortless.
PS
I have worked with healthcare organisations for years and the pattern is always the same. Brilliant clinical teams, strong evidence, thoughtful strategies. And then patients who do not do what works. The frustration is real. But the cause is not patient irrationality. The cause is a system designed for rational actors who do not exist and for patients whose actual Job-to-be-Done was never asked. The moment you stop blaming the patient, start designing for the person rather than the condition, and adjust the environment rather than the information, everything changes. That is the core of what we do at SUE, and it is what I describe in The Art of Designing Behaviour (2024). Healthcare does not need better information. Healthcare needs better design.
