Designing for 'First-Session Confidence' on a UK MRC-Funded Clinical Trial
First-Session Completion Rate
82%
Area
Research
UX Strategy
Interaction Design
Design System
This UK MRC-funded project (MR/W029421/1) is currently undergoing a Randomised Clinical Trial (RCT).
⏱️
The short version.
Key area
Detail
Project scope
Design and validate the end-to-end therapy experience for the mobile app, with a critical focus on ensuring first-session success.
Design Goal
Achieve 80%+ unaided first-session completion in pre-trial usability testing, the minimum behavioural threshold for the RCT to generate reliable adherence data.
Target group
Adults aged 45–70 with chronic knee osteoarthritis, a group with a high proportion of users with limited digital confidence and reduced dexterity.
Timeline
8 months, including recruitment, research, iterative design and pre-trial usability testing
Constraints
Uncertainty in NHS identity and clinical integration models forced the design to be functional regardless of the onboarding method..
Unpacking the problem: Confidence breaks down before habit has a chance to form
While clinical evidence shows that TENS (electrotherapy) is effective for pain relief, the real-world data reveals a critical flaw in home use:
Adherence to unsupervised home therapy rapidly declines within weeks
Nicolson et al., Osteoarthritis and Cartilage, 2018
1/5 adults over the age 45 affected by Knee OA in the England
Arthritis Research UK and Public Health England
Linda Theobald
Electrotherapy user
"I wasn’t sure if I was doing it right, so it was easier to just stop."
Representative archetype, synthesised from adherence research
Business Challenge
Low adherence (47%) posed a direct risk to Rejoovinii’s clinical trial validity and NHS adoption. If early confidence breakdown during setup and device control persists, inconsistent usage will continue to undermine data reliability and clinical outcomes.
The business challenge was to ensure users could confidently complete their first therapy session independently, as first-run success was critical to long-term adherence and the product's pathway to NHS procurement and wider clinical adoption.
My role
As the sole UX designer on a UK MRC-funded clinical trial, I owned the full design pipeline from research strategy through to usability testing, navigating constraints from clinical, engineering, and product stakeholders simultaneously. When the second designer left the project early, I absorbed the entire design scope solo. The core challenge was not just designing a usable app, but designing for confidence in a high-anxiety, low-digital-literacy clinical context where first-session failure directly threatened trial validity.
Key contributions:
Defined the research strategy, leading secondary analysis across 40+ clinical studies and collaborating on primary concept and usability testing with clinical participants
Mapped the end-to-end user flow across setup, in-session and post-session stages, aligning task sequences and interaction behaviours to confidence checkpoints rather than feature requirements
Pushed back on heavy feature-based onboarding and gamification, redirecting both decisions based on research evidence and clinical context
Aligned clinicians, engineers and product stakeholders around a single design anchor, first-session confidence, across an 8-month multidisciplinary collaboration
Built the foundation of the app's design system for clinical consistency and scalability
Working session with the clinical and engineering team. One of many cross-functional reviews across the 8-month collaboration.
Outcome
Pre-trial usability testing achieved 82% unaided first-session completion across 11 participants, exceeding the 80% minimum behavioural threshold required for clinical viability. The two participants who needed assistance encountered difficulties with parameter adjustment and a missed setup instruction, both of which informed final design refinements.
The design validated a confidence-first approach to clinical UX, demonstrating that elderly users with limited digital confidence could complete a complex home therapy session independently and without external guidance. This was the precise condition the clinical trial required.
🤿
Going deeper.
Discovery
Home electrotherapy for chronic knee OA loses most patients within days. To move beyond assumptions, I worked with clinical heads and defined three research questions that became the backbone of my secondary research, guiding every decision in where and why adherence breaks down:
Where do patients struggle first during therapy setup?
When do they lose confidence in effectiveness?
What motivates or prevents sustained use over time?
Mapping the problem space through evidence
Competitor and analogous product review
Initial discussions with stakeholders focused on reviewing global electrotherapy apps to understand industry standards. I realised that relying on these comparisons risked misaligning our design with our low-digital-confidence user base, as the few existing UK products only tangentially addressed chronic knee pain and most international apps targeted fitter, more confident users.
To mitigate this risk, I shifted the team’s focus from feature benchmarking to behavioural friction. I worked with product, clinical and engineering stakeholders to broaden the scope and realign our efforts around how low-confidence users build trust in unsupervised therapy.
Indirect Competitors: Rehabilitation exercise apps that successfully engage chronic pain patients.
Analogous Competitors: Accessible therapy for people facing physical movement challenges.
View the detailed competitive breakdown
Findings
The competitor review revealed three recurring experience gaps that no existing solution had fully resolved:
No product designed for the moment confidence breaks - Every competitor assumed the user would figure out setup. None had designed specifically for the moment a first-time user hesitates, second-guesses their electrode placement, or simply stops because nothing confirmed they were doing it correctly. That gap was the most consistent failure across the entire landscape.
Therapy adapted to the product, not the patient - Existing solutions offered programmes and presets but no real responsiveness to how a patient was feeling on a given day. Users with fluctuating pain had no way to adjust their experience meaningfully. The result was a fixed experience in a context that demanded flexibility, and users quietly lost trust in whether the therapy was relevant to them.
Data was collected but never made meaningful - Most products tracked session logs. None connected that data to anything the user could understand or act on. Pain trends were separate from session history. Users had no way to see whether their effort was accumulating into something, which made stopping feel rational rather than like a failure.
The competitor gaps raised a deeper question: why do users keep stopping? The clinical evidence answered it.
Clinical studies review
I conducted a structured review of over 40 clinical studies on home-based TENS and adjacent OA self-management. While clinical protocols focused heavily on stimulation parameters and treatment schedules, the research consistently showed that patient drop-off was driven far more by psychological and behavioural friction than by technical inefficiency.
Findings
Setup anxiety is a design problem, not a user problem - Patients did not abandon therapy because the technology failed them. They abandoned it because they were never confident enough to start properly. Clinical evidence consistently linked early drop-off to setup confusion and fear of incorrect use and not to pain, inconvenience or disinterest. This reframed the challenge entirely: the first screen a user sees is as clinically significant as the therapy itself.
Confidence is built through feedback, not instructions - Users who could not tell whether the therapy was working stopped assuming it was. Generic guidance and fixed programmes gave patients no way to correlate their effort with their experience. Without that feedback loop, doubt accumulated quietly until stopping felt more rational than continuing.
Motivation needs a visible reason to persist - The research showed a consistent pattern: initial enthusiasm, followed by rapid decline once sessions began to feel repetitive. Patients did not lose interest in getting better. They lost the ability to see that they were. Progress visibility was not a feature request. It was the missing link between first session and long-term habit.
Based on the clinical evidence, it became clear that adherence failures were not primarily caused by poor protocols or insufficient instructions. They were due to psychological breakdowns in confidence during the therapy.
I therefore realigned the team's focus away from stricter guidance and towards designing for confidence, clarity, and reinforcement during the first therapy experience.
Affinity mapping: From findings to patterns
With data gathered from both the competitive review and the clinical literature, I ran an affinity mapping exercise with the team to synthesise findings across both sources. Rather than treating them separately, we mapped all evidence along the patient therapy journey, from first touch through to long-term use, to find where the patterns converged.
Working across two passes, the data consistently clustered around five recurring breakdown points:
Guidance: Users need reassurance, not just instructions
Feedback: Without clear signals, they start guessing
Customisation: When therapy doesn’t adapt, trust breaks
Tracking: Data without context feels meaningless
Motivation: Repetition without reinforcement leads to drop-off
What was significant was that both data sources pointed to the same five themes independently. The competitive analysis revealed these as market gaps. The clinical evidence confirmed them as psychological mechanisms. This convergence gave us high confidence that these were not assumptions: they were the real barriers to first-session success.
We prioritised these themes based on their direct impact on the 47% adherence rate, arriving at three strategic insights:
Insight 1: The first session is a trust-building moment, not a learning moment
If confidence breaks during setup, retry probability drops sharply. The entry experience cannot be designed like an onboarding tutorial. It needs to be designed like a clinical handover: calm, guided, and forgiving of mistakes.
Insight 2: Control and clarity are the mechanism of confidence
Patients lose confidence not when therapy is hard, but when they are unsure what the device is doing or whether they are in control. Real-time feedback and personalised parameters are not comfort features. They are the clinical interface between the patient and their therapy.
Insight 3: Motivation requires visible proof of progress
Sessions feel repetitive without meaningful reinforcement. Patients do not lose interest in recovering. They lose the ability to see that they are. Motivation links to seeing change, not just completing sessions.
🧭
Plan Broken: What I did next
Original plan
Recruit experienced electrotherapy users to reduce trial noise and ensure clean data capture.
What went wrong
Multiple participants dropped out due to illness and availability conflicts, putting our research timeline at risk.
My call
Instead of delaying the study, I proposed including a cohort of complete electrotherapy novices. This decision provided essential usability data and protected the overall project timeline.
Why this mattered
The trial goal was clean data, but most future NHS users would be first-timers. If novices failed, passing the trial would not prove real-world viability.
Designing confidence across the first session
With backend system integration and NHS data sharing policies still under internal review, we couldn't finalise a login strategy. Rather than letting this delay the project, I chose to mitigate the risk of a 'fragile first run' by strategically pivoting the focus. We conducted concept tests to test the user's emotional experience and confidence at three critical points in the application.
Electrotherapy Novices
Electrotherapy Users
Ages
The mix was deliberate. Novices reflected the likely NHS user profile, while experienced users provided a benchmark for expected behaviour during setup and in-session control.
Setup - Bluetooth pairing + sleeve setup
What we tested (Insight 1)
The Trust-Building Moment to ensure first-time users could connect the device and wear the sleeve with confidence, without feeling they were doing something wrong.
What we observed:
1
Participants needed to see if the device was paired while viewing sleeve instructions.
2
Dense text and sleeve positioning instructions slowed first-time users.
3
Participants preferred sequenced setup with visible progress steps (e.g. Step 1 of 5)
4
Some users preferred video instructions
User signal (paraphrased)
Steps like 1 of 4 would reassure me.
Show a clear paired state.
In-session - Control + pain logging
What we tested (Insight 2)
Whether users could clearly understand what the device was doing during a live session and feel they remained in control of it, rather than just passively letting it run.
What we observed:
1
Participants felt unsettled when controls looked dense or overly technical.
2
Pain logging as a meaningful feedback.
3
All participants valued seeing what was happening in real time.
User signal (paraphrased)
Recording pain score will be brilliant.
I would go for a personalised programme, but I’m open to customisation.
Post-session - Result + Progress
What we tested (Insight 3)
Whether showing outcomes and progress after a session helped users make sense of the therapy and feel motivated to continue, instead of treating each session as an isolated event.
What we observed:
1
Rewards were less motivating than seeing actual progress trends.
2
Participants cared more about pain trends than one-off session results.
User signal (paraphrased)
Often the reward is just relief from pain and seeing the information.
It would be about seeing what patterns arise.
I initially explored light gamification to support long-term adherence, but concept testing revealed a mismatch with the clinical context. Participants were not motivated by rewards or badges. They were motivated by relief, visible improvement, and understanding their own progress. Introducing gamification risked trivialising the experience and undermining trust, especially during the first session where anxiety is highest.
So, I made a conscious decision to kill overt rewards and instead design motivation around clinical progress, using pain trends, comparative insights, and subtle reinforcement messaging. This aligned better with both patient psychology and trial goals, strengthening first-session confidence and supporting sustained adherence through meaningful, not artificial, motivation.
From insights to archetypes: Creating a single archetype for first-session confidence
To ensure the design stayed grounded in real behaviour, I collaborated with clinicians, product leads, and engineers to synthesise research findings into a single primary user archetype. This archetype represented the most common behavioural and attitudinal characteristics of the target user for this critical phase.
Linda Theobald: The face of our NHS trial
Linda is a 58-year-old with chronic knee OA and moderate digital confidence. She wants to manage her chronic knee pain independently, but her biggest barrier is confidence.
She uses everyday apps like messages and banking but feels anxious around medical technology, especially when she doesn’t understand whether she’s using it “correctly”.
Every critical design decision mapped back to one question:
Would Linda feel confident completing her first session alone, without external help?
The Confidence Checkpoints
Design
Using Linda’s anxiety and the above three design bets, I mapped how Linda would actually move through the product across key stages of the therapy experience. They were used to structure the product around behaviour and decision points, not the screens.
The Constraint Shift: From Ideal to Complex (V1 - V2)
The initial flow (V1) relied on pre-personalised devices. This allowed a clean, low-friction path into therapy, with only minimal setup and identity confirmation.
However, this ideal scenario broke once we modelled the real deployment context. In reality, devices would ship unregistered and unpersonalised, requiring full account creation, profile setup and NHS verification inside the app. This created critical user risks delayed gratification and maximised drop-off risk, severely jeopardising first-session success.
The Strategic Solution Post-Value Registration (V3)
Working cross-functionally with engineering and clinical leads, I reframed NHS login from a product dependency into a verification layer.
Instead of full NHS integration, users were verified externally and allowed to proceed with therapy setup immediately. Account creation and data linking were postponed until after first-session value was delivered. This protected first-session confidence while still meeting governance and compliance requirements.
There was a push to introduce a detailed onboarding sequence explaining all features and system behaviours upfront. I challenged this strategy.
Our user, Linda, was already anxious about "doing it wrong." Too much upfront information increases hesitation and cognitive load, making her feel overwhelmed before she has even begun therapy. Instead, I deliberately avoided a long onboarding flow.
I designed confidence to be built in-context through progressive guidance, just-in-time instructions, and real-time feedback during her first session. This strategic approach achieved a rapid Time to Value (TTV) by prioritising immediate success for better adherence.
Designing for the patient’s physical and mental context
Rather than designing screens in isolation, I focused on how each moment would feel across the entire therapy journey. This is where usability, semantics, ergonomics and emotional reassurance converged.
A therapy session is not just a digital interaction. It happens in a physical, vulnerable and often cognitively loaded context. Linda is wearing a medical garment, managing pain, and using a device that she fears getting wrong. Designing for this meant optimising not just for usability, but for confidence and error prevention.
Core page layout and interaction
Designing for a medical device requires touch targets that are easy to use, especially for patients who may have reduced dexterity. My research found that a major barrier to therapy adherence is the difficulty users have in accurately selecting small targets. Adequate spacing between touch targets is also critical, as it prevents accidental taps on nearby buttons, making the interface reliable and trustworthy.
From design decisions to testable reality
These design decisions shaped the working version of the interface. The next step was to validate whether these ideas actually reduced hesitation, improved confidence, and made the therapy experience easier for Linda. This led into multiple moderated usability sessions and iterative refinement across multiple versions.
In the initial phase, Rejoovinii was designed for a dedicated 7-inch tablet. During technical planning, this approach became unfeasible due to hardware cost, firmware reliability and long-term support risks, and the product pivoted to a mobile phones.
When this shift was made, the product was already mid-build and timelines were tight.
I focused on minimising disruption for both product and the engineering team. Rather than redesigning end-to-end, I identified high-impact pressure points where mobile constraints would increase cognitive load and technical risk, and restructured only those flows.
This helped stabilise development, reduced implementation complexity, and preserved first-session confidence for low-digital users.
In this phase, I focused on how the design held up under real use. These decisions were evaluated through moderated usability sessions, aligned against three criteria setup hesitation, in-session confidence and clarity of post-session results
From raw data to personal progress: Making results emotionally meaningful
These iterations address Insight 3: motivation comes from seeing change, not just completing sessions. Early versions exposed raw data, but users struggled to interpret it. I shifted toward clearer progress summaries and comparisons that helped users quickly understand whether therapy was working for them. This reduces doubt after each session and directly supports my core anchor of moving first-session confidence into long-term adherence.
Impact
Users moved from struggling to interpret raw pain graphs to being able to say immediately whether their therapy was working. The shift from averaged data to personal comparison language, showing 'pain down 1.4 points since last week', gave Linda a reference point that felt relevant to her journey rather than abstract clinical data. This directly addressed the confidence gap that caused early drop-off after the first few sessions.
From blind controls to guided decision-making: Reducing in-session anxiety
This iteration directly addressed Insight 2 by shifting from raw control to guided control. Instead of letting users guess or override blindly, I broke the flow into pain input followed by explained recommendations, making the system’s logic visible. This reduced anxiety around settings, prevented unsafe adjustments, and built trust in the device during the first session, which is essential for improving long-term adherence.
Impact
Users who previously felt unsettled by dense controls were able to move through the session setup without hesitation. Making the system's logic visible, by connecting pain input to recommended intensity before asking users to confirm, reduced the fear of doing it wrong. Participants reported feeling in control rather than guessing, which is the precise condition required for first-session confidence to hold.
From session closure to continuity: Designing the moment that drives return
In the first version, the summary only confirmed that a session was completed. It showed the outcome, but didn’t help users understand what it meant or what to do next. That gap weakened both perceived control and motivation.
In the second version, I reframed the summary as a confidence and continuity moment, not a closing screen. By comparing today’s outcome with the previous session and inviting light contextual reflection through notes, the interface gives users both clarity on what happened and agency over what happens next. This directly supports Insight 2 by reducing uncertainty, and reinforces first-session success as a foundation for continued adherence rather than a one-off event.
Impact
Users moved from 'I finished' to 'I understand what happened and what to do next.' By reframing the post-session screen as a confidence and continuity moment rather than a closing screen, the design gave Linda both clarity on what happened and agency over what comes next. This reduced early drop-off risk and reinforced the habit loop that turns a first session into a second.
What the testing confirmed
9 of 11 participants completed the full therapy session unaided, exceeding the 80% minimum threshold required for the clinical trial. The two who needed assistance encountered difficulties with parameter adjustment and a missed setup instruction, both of which informed final refinements.
The result validated a specific hypothesis: designing for confidence in the first session is more clinically significant than designing for feature completeness. Users did not fail because of missing functionality. They hesitated because of uncertainty.
Building for clinical consistency: The design system foundation
Designing for a clinical trial meant every visual and interactive decision carried responsibility. A colour that failed contrast requirements or a touch target too small for arthritic hands could directly undermine confidence at the exact moment it needed to hold. Building a design system from scratch was not optional but it was a clinical necessity.
The foundation layer covers colour tokens with documented 4.5:1 contrast ratios, a typography scale optimised for legibility under physical strain, a spacing system built around minimum touch target sizes for users with reduced dexterity, and a custom icon set for communicating device states without relying on text.
The component library spans eleven categories: Buttons, Input Fields, Steppers, Headers, Tabs, Timers, Status Indicators, Notifications, Checkboxes, Sliders and Data Visualisation, and each designed with error prevention and confidence as the primary criteria. Steppers replaced sliders for intensity control specifically because sliders created accidental adjustment risk for users with reduced fine motor control.
What I learned
Anxiety is a design material. The most significant barriers were not usability failures in the traditional sense. They were moments where Linda did not know whether she was doing it correctly. Designing for that feeling required a different kind of attention to every transition, label and feedback moment in the interface.
Constraints are a design tool. The novice cohort pivot, the tablet to smartphone shift, the NHS login constraint and each restriction produced a better, more focused product than the original plan would have.
Clinical context demands a different standard. Guardrails are not limitations in a medical context. They are part of the therapeutic experience. A patient who feels safe is a patient who keeps going.
What I would do differently
The post-session summary screens were validated through concept testing rather than moderated usability sessions. Given that this is where long-term motivation is either built or lost, it deserved the same rigour as setup and in-session flows. That would be the first thing I revisited.
