My Role

UX Research & Design

Timeline

6 months

Collaborated With

Product Owner, Design Systems, Front & Back End Developers, QA

Tools & Methods

Tools: Miro, Figma, Notion

Methods: user interviews, affinity diagraming, customer experience map


Context

When ASH (American Specialty Health) expanded national PT/OT (Physical Therapy/Occupational Therapy) coverage through Cigna, MNR (Medical Necessity Review) volume surged and the process behind every review was held together with spreadsheets, Teams messages, and Post-it notes. CQE (Clinical Quality Evaluator) managers manually assigned MNR chunks in Excel and posted them to a Teams chat. CQEs then cross-referenced Tableau to find their specific cases, copied MNR numbers into IHIS (Integrated Health Internal System) one at a time, and tracked their own completions however they could. Some kept sticky notes. Some built their own spreadsheets. The official closed daily workbook was always too delayed to trust. Scaling headcount wasn't sustainable. The business needed the existing team to handle more volume.

Business Goal

Reduce operational costs to sustain revenue gains.

What was the end product?

Based on my user research, I designed a workload management tool for CQE Managers and CQE Staff to assign and track MNRs. It included three interfaces: 

  • Manager Admin View for automatically distributing MNRs

  • Profile Management View for managers to maintain CQE profiles 

  • My Work Screen for CQEs to track their assigned reviews by status (open, on-hold, closed). 


Where did I start?

Research Problem

What is causing inefficiencies in the MNR review process?

What bottlenecks are preventing CQEs from completing more reviews each day?

User Interviews

  • 8 CQEs across the 3 specialties (Rehab, Chiro, Acu)

  • 3 Directors of CQE (2 Rehab, 1 Chiro)

  • Director of Medical Necessity/Benefit Administration

Process Map

At the core of the problem wasn't one broken thing. It was a workflow stitched together across five disconnected tools. Every CQE started their day by checking a Teams chat, opening Tableau to translate their assignment into actual MNR numbers, manually copying those into IHIS one at a time, and then hunting through F4 notes, ASHCore call logs, pend items, and the MNR itself to piece together enough context to do the review. When errors sent a case back to the MNA team, it sat for up to 3 hours before returning. With CQEs averaging 70 reviews a day at 5 to 15 minutes each, even small inefficiencies compounded into significant lost capacity, and nobody had a reliable way to know what was actually done each day. Without visibility into who’s working on what, two CQEs could unknowingly review the same MNR simultaneously and only discover the conflict when they tried to save it.

After presenting our findings to stakeholders, we collectively decided to focus on 3 opportunity spaces:

  1. Improve MNR assignment and visibility

  2. Centralize CQE communication

  3. Automatically track daily MNR completions.

These offered measurable improvements quickly while leaving space for future enhancements to the broader MNR process.

Who was I designing for?

Research revealed two fundamentally different relationships with the same workflow. Managers thought in batches and owned assignment. CQEs thought in individual cases and owned execution. Designing for both meant the system had to speak two different languages.



Problem Statement

CQE Managers and CQEs need a streamlined way to assign, track, and communicate about MNRs because the current process is manual, fragmented, and lacks visibility — slowing down reviews and driving up operational costs, which impacts the company’s ability to sustain revenue.

Prioritizing and Decision Making

Working alongside the product team, we mapped our findings against effort and impact. Given the scale of the CQE department, we prioritized the Workload Balancer as the highest-impact solution we could ship quickly. The Rehab team represented 70% of the CQE department, so we designed for their workflows first with plans to scale to Chiro and Acu later.

Final Product

I collaborated closely with the Front End Developer, the Product Owner, and the Directors of CQE to ensure feasibility and usability. 

We decided to create a Workload Balancer including three main views: the Manager Admin View, the Profile Management View, and the My Work Screen.


Manager Admin

The Manager Admin screen automatically assigns MNRs to CQEs and gives managers an at-a-glance view of their team’s workload.

Managers can now:

  • View the real-time status of every MNR (assigned, unassigned, on hold, closed)

  • See which CQEs are available for another MNR

  • Filter CQEs by specialty, hours, licensing, health plan qualifications, or state

  • Directly transfer an MNR to a different CQE

Profile Management

The Profile Management screen gives managers a centralized place to view, edit, and create CQE profiles.

Managers can now:

  • View all key CQE information in one place

  • Edit or update profile details directly without involving the backend systems team

  • Create new CQE profiles as needed

My Work

The My Work screen provides CQEs with a clear, structured way to receive and manage their assigned MNRs.

CQEs had been handpicking MNRs, gravitating toward shorter reviews so that longer, more complex cases kept getting pushed to the bottom of the queue. Automated assignment removed that option, ensuring every MNR got equal attention regardless of complexity.

CQEs can now:

  • Receive only one active MNR assignment at a time, ensuring focus and preventing overlap

  • Transfer MNRs directly within Workload Balancer instead of using IHIS

  • View all on-hold MNRs that still need completion

  • See a running total of their completed MNRs in the “Closed” tab

Impact

The Workload Balancer wasn't implemented before I was laid off in August 2025, but the MVP was fully developed with a planned launch of January 2026. The solution directly addressed each root cause identified in research:

  • automated assignment replaced the Excel and Teams manual process

  • a unified work queue eliminated the Tableau lookup step

  • real-time status tracking made the delayed closed daily workbook obsolete

Stakeholder feedback throughout development confirmed the design would meaningfully reduce the coordination overhead that was pulling CQEs away from actual review work. With 83 CQEs averaging 70 reviews per day, even a 10-minute daily reduction in administrative time would recover over 580 person-hours per week across the team. To validate this, I would track hours saved on assignment, duplicate review rates, and average completion time per MNR.



Reflection

This project taught me how much damage invisible problems can cause. The duplicate review issue, where two CQEs could work on the same MNR simultaneously with no way of knowing until they both tried to save it, wasn't in any process doc. It only surfaced through talking to the people actually doing the work. That's what convinced me that the core problem wasn't just assignment efficiency. It was the complete lack of a shared source of truth.

A key constraint was limited access to CQEs directly. They were too busy to sit down for formal interviews, so I relied on directors to fill in gaps my initial CQE conversations left open. In hindsight, I would have found more informal moments to observe CQEs in the flow of actual work, because the details that mattered most, like the duplicate problem and the personal tracking workarounds, only came out when people were showing me rather than telling me.