Streamlining Referral Processing by 86%

Shipped

Desktop Design

0 to 1

Accessibility

Overview

The referrals team at Perlman Clinic was stuck working around a disorganized, outdated system that made an already complex job way harder than it needed to be. Training took up to 6 months. Placing a single accurate referral could take anywhere from 3 minutes to 13 minutes. And they were doing this thousands of times a day.

I led this project end-to-end as the sole designer, handling everything from research through final prototype.

My Role

As the sole UX designer, I led end-to-end design ownership from staff research through successful implementation. My key responsibilities included:

  • Conducted comprehensive user research and 10+ hours of shadowing the referral coordinators

  • Collaborated with stakeholders across marketing, PM, and development teams

  • Designed the complete experience that directly addressed both patient needs and business objectives.

Results

By consolidating multiple systems into a single, intuitive interface, the custom-built directory reduced referral processing time by 86%.

Role

Perlman Clinic

Skills

4 weeks

Timeline

Product Designer

Tools

Figma

85%

Faster processing

40%

Increase in accuracy

86%

Reduction in time spend

Problem

No central source of truth for specialist info

Staff needed to cross-reference a patient's medical group, specialty, and location to find the right specialist, but there was no single place to do that reliably. Information was scattered, stale, and hard to act on quickly. Errors were piling up, and new hires had no efficient way to learn the system.

Research

10+ hours of shadowing showed exactly where things broke down

Key insight #1

Incomplete Information

Staff had to Google providers due to missing or scattered subspecialty details.

Key insight #2

Geographic Confusion

Unfamiliarity with locations led to inaccurate travel estimates and rejected referrals.

Key insight #3

Training Gaps

Each medical group had different rules and portals, making training difficult.

Workflow mapping

The biggest problem was a repetitive loop in the middle of the flow. Staff would search for a matching specialist, potentially Google subspecialty details, cross-reference Google Maps for distance, and if the distance didn't work, start over. This loop alone cost 5 - 20 minutes per referral.

Mapping this out made the solution direction clear: collapse the middle of the flow into one place where staff could find everything they needed without switching tabs.

Problem Statement

How might we consolidate essential referral information into a single, easy-to-navigate tool?

The problem statement highlighted three key requirements: accurate subspecialty matching, geographical proximity, and formatting the information to send to the patient, all accessible within a unified interface that would reduce cognitive load and training complexity.

Usability Testing

Validated efficiency and accuracy gains through task-based usability testing with referral coordinators

I conducted 7 moderated usability tests with referral coordinators using an interactive Figma prototype. Each session compared the current workflow vs. the directory across realistic referral scenarios.

Method

Task-based evaluation grounded in real workflows

Participants completed timed scenarios using both their existing tools (Epic, portals, Google) and the new directory:

  • Find the closest specialist by zip code + specialty

  • Identify correct subspecialty providers

  • Complete full referral workflow from Epic → portal → MyChart

I captured:

  • Task completion time

  • Referral accuracy (closest + correct provider)

  • Ease of use (1–10 rating)

  • Qualitative feedback on usability + edge cases

Key Metrics

Significant gains in speed and accuracy

  • 86.2% faster referral completion
    3:31 → 0:29 average time

  • Up to 7m 35s → 52s max time reduction

  • +40% improvement in referral accuracy
    (More coordinators selected the closest appropriate provider)

  • 10/10 average ease-of-use rating across participants

Design Implications

Directly informed product decisions

  • Prioritized subspecialty coverage

  • Defaulted to distance-based sorting

  • Introduced provider notes + copy-ready formatting

Challenges

Prioritized for Impact

The most significant challenge was technical feasibility. We couldn't immediately integrate all eight medical groups due to API limitations and data access constraints.
So, I identified the top 3 groups causing 65% of staff pain points using stakeholder interviews and usage analytics.
We prioritized those for launch and created a roadmap for the rest.