YASHA LIM
PRODUCT MANAGER
PRODUCT MANAGER
"Clinical trials still ask patients, clinicians, research sites, sponsors and CROs to hold a fragmented system together by hand. Eligibility gets checked on paper. Referrals lose momentum at every hand-off. A patient who would have been a perfect match never hears the study exists."
I treated the brief as a hypothesis and used desktop research to test whether the underlying problem was real. The research validated the core problem: clinical-trial recruitment in Australia is fragmented and heavily reliant on manual coordination. There has been reviews conducted in which fewer than 30% of eligible patients were offered trials.
The research gave me enough confidence that there is a meaningful problem space around discovery → eligibility → referral, while also identifying the assumptions I would want to validate with clinicians, trial coordinators and patients before committing to a solution.
*Illustrative inferred range. All confidence ratings in the source are Low; several values are single-context observations or evidence gaps rather than national estimates.
Research Nurses and Clinical Research Coordinators need to determine which patients are likely to meet complex trial eligibility criteria, but they often rely on incomplete clinical information and manual interpretation. This makes screening time-consuming, increases unnecessary review effort, and risks suitable patients being missed.
From the brief, I identified two core hypotheses that needed to be tested: one around user value and one around commercial viability.
If the platform surfaces likely trial matches from existing clinical information and supports eligibility review, research teams can reduce screening effort and move more suitable patients toward enrolment.
If the platform measurably improves screening efficiency and recruitment outcomes, research sites, sponsors and CROs should see sufficient value to adopt it as a paid clinical trials product.
As a Research Nurse or Clinical Research Coordinator, I want the platform to surface likely eligible patients with clear supporting evidence so I can review them efficiently and make the final screening decision with confidence.
Engagement with engineering and key stakeholders is completed as early as possible through ideation.
Each idea is compared across customer and company value, implementation effort, pros and cons, key dependencies and risks, then use that discussion to narrow the prototype scope.
Prioritisation Limitations: These placements and scores are illustrative due to the two-day timeframe.
They are included to show the prioritisation thought process I would use with engineering and stakeholders; in practice I would refine value, effort, dependencies and risks collaboratively before committing to scope.
Primary Outcome
More truly eligible patients progress toward enrolment
Improvement in the proportion of platform-surfaced patients who are ultimately confirmed as eligible
Success metric: e.g. +15% vs baseline eligibility conversion rate
Reduce screening time
Lower median time from eligibility review start to screening decision
Success metric: e.g. 25% reduction vs baseline
Reduce unsuitable patients progressing
Reduce the proportion of surfaced patients later rejected as ineligible
Success metric: e.g. 20% reduction vs baseline rejection rate
Make recommendations easier to verify
Recommendations include traceable criteria and supporting clinical evidence
Success metric: e.g. ≥95% include traceable evidence
Metric Limitations: These targets are illustrative because there was no access to current product, operational or customer baseline data. I would first establish those baselines using screening conversion rates, time-to-screening data, false-positive and rejection rates, product analytics and CRM data. I would then refine the targets with input from research sites and internal stakeholders.
Trust
Users can understand and verify why a patient was surfaced
Guardrail metric: e.g. ≥80% correctly verify supporting evidence without assistance
Workflow burden
Administrative effort does not increase versus the existing process
Guardrail metric: e.g. ≤10% increase in clicks while total task time decreases by ≥20% vs baseline
Accuracy
Faster screening does not come at the cost of excessive false positives
Guardrail metric: false-positive rate does not exceed baseline while eligibility conversion improves
The platform may recommend and explain, but it must provide supporting evidence and the final eligibility decision remains with the Research Nurse or Clinical Research Coordinator. e.g. 100% of final eligibility decisions require explicit coordinator confirmation.
How do I best showcase my ownership of the prioritisation, trade-offs and final product decisions whilst accelerating the process with AI?
Below are three moments where I used AI to accelerate work without giving up ownership of the prioritisation, trade-offs or final product decisions.
Task: Explore the clinical-trials workflow and evidence.
Why AI: With limited time to research and knowledge of the space, to move quickly across a broad problem space.
My decision: Use the output to identify the broad user flow in AU and identify most common pain point, while preserving evidence caveats.
Task: Challenge my product thinking, proposed outcomes and guardrails.
Why AI: There are plenty of things that I might not consider and I wanted to be comprehensive, using it to augment rather than replace.
My decision: Set instructions for it to challenge my assumptions for any task or question I ask.
Task: Generate and iterate the prototype flow.
Why AI: Shorten implementation cycles.
My decision: Provide flow and rationale, asking it to ask questions for every step and refine prompts and flow until the interaction matched the product intent.
I like to challenge myself and learn new things. One of my main hobbies is indoor rock climbing.
When I first tried it, the height and free-falling (even safely on a harness!) was petrifying and I could barely go higher than 5 metres. Now I climb over 25 metres and outdoors too.
Always happy to chat over a coffee or matcha, so feel free to reach out to me on LinkedIn!
I enjoy supporting creators and giving credit where it’s due. The icons used throughout this website were illustrated by LittleFoxDigital. You can check out more of their work here.