Back to work

HEALTHCARE OPERATIONS

Registered dietitians were logging daily work in shared Excel files. I designed the Elevra workflow that turned those records into a focused daily form.

Industry

Healthcare ops

Surface

RD daily workflow

Pattern

Log + manager grid

Role
Lead Product Designer
Organization
Compass Healthcare Digital
Team
Product, RD operations, engineering
Status
Production · recreated for this site

TL;DR

Replacing the workbook was the easy framing. The stronger product move was giving each role a better decision: one focused daily record for the RD, and the exact missing day for the manager.

See it

Two surfaces, live.

Recreated, not screenshotted. Switch surfaces, then expand either one full screen.

Elevra

Sacred Heart Medical care team

Team Submissions

March 2026

Total assessments

On track

Assessments per RD

Above target

Team compliance

Below target (85%)

Care setting coverage

Full coverage

14 RDs on this roster11 below the 85% compliance targetCompliance flags reflect a rule, not a model score
NameDays sub.AssessmentsMTWTFCompliance

Working demo, recreated interface. Stand-in for the production application.

THE WORKING BRIEF

A monthly percentage does not name the day or the person to follow up with.

A day-level workflow replaced fragmented workbook records with one role-owned record and directly inspectable exceptions.

The reframe that scoped the build

Evolution

One record, from a shared file to a system of truth.

Three moments in the same record’s life, recreated from the shipped product. Every name and number shown is demo data.

Edit Productivity Log
Quick EntryAI-POWERED

Date of service

Mar 24, 2026

Patient care setting

Outpatient adult

Care settingComp.LimitedTotal timeGroup intv.
Outpatient adult464.52

One RD, one day, captured inside Elevra's existing Assessments engine, with the AI field sitting above the fields it actually feeds.

Mandate

Design owned the workflow, never the whole delivery system.

I designed the module end to end and documented the gaps. Product, RD operations, engineering, identity, and BI each retained a different decision boundary.

The delivery systemDesignWorkflowStatesInteraction modelGap inventoryProduct and RD operationsEdit limitsTime unitsCare-setting rulesPriorityEngineering and BIForm.io configurationRBACAPIsReporting datasetsExisting sourceWorkbook field definitionsOperational conventions

The evidence layer

Three calls that shaped the system.

Use the platform that already owns assessments.

Clinical Productivity became a workflow inside Elevra's existing Assessments infrastructure, not a new platform.

The pain was shared files, fragmented records, and manual aggregation.

EXISTING PLATFORMNET-NEW ENGINEERINGAssessments engineClinical Productivitybuilt inside the engineForm.io form engineExisting design libraryNet-new engineeringdata model · RBAC · Power BI
Clinical Productivity. The fastest path ran through what Assessments already provided. The near-term experience inherited its current design library and limits, the tradeoff for not rebuilding the shell from zero.

The near-term experience inherited the current design library and its limits.

The production data model, RBAC, and Power BI datasets remained engineering work.

Let the day reveal its own form.

Care-category checkboxes reveal only the relevant activity fields, one care setting at a time.

Most of the source schema is irrelevant on any single day.

Clinical Productivity

Daily log

Draft

Care delivered today

Selected category

Inpatient adult
Progressive disclosure keeps the other three categories out of the way, and the unresolved time-unit rule stays visible on whichever one has hours logged, because the interface cannot settle a policy decision.

A shorter-looking form can still be incomplete: fields were missing from the polished prototype.

Design exposed a unit label and validation state; the field rules stayed a product decision.

Show the missing day, then flag only what a rule can explain.

The manager surface uses day-level status, with explicit rules for data-quality exceptions.

A monthly percentage does not name the day or the person to follow up with.

Manager view

Submission status

SubmittedMissingReviewDraft
Team memberMonTueWedThuFriAction
Rachel DempseyMissing Wed
Yuki TanakaDraft Tue
Blessing NwosuReview Wed

9 submitted, 1 missing, 1 held for review across 3 RDs this week.

A percentage summarizes the month. A day-level grid tells a manager where to act. The sample is reconstructed and contains no employee or facility data.

Day-level detail creates a wider table and needs careful weekend and PTO handling.

Notification delivery and production enforcement stayed outside the prototype.

Problem

No lock, no audit trail, no single source of truth.

Daily productivity lived in shared workbooks, with one file per cost center and month and one sheet per RD. Concurrent edits collided, records fragmented across sites, and managers had to reconcile the result by hand.

Two RDs editing the same file on the same day would silently overwrite each other.

No lockNo audit trailNo single source

Outcomes

Production status is real. Product impact remains mostly open.

Validated

The module reached production, covering daily logging, the manager grid, drill-down, and edit states.

Directional

Centralized records and deterministic exception flags offer a credible, legible path away from concurrent workbook edits.

Not measured

Adoption, reconciliation time, notifications, Power BI, and cross-module behavior all stayed outside the prototype's evidence.

What I'd do differently

In .

  1. Field gaps, earlierI would put the field-gap inventory beside the prototype earlier: the interface already looked coherent while required fields were still absent, and that gap is a delivery risk precisely because the surface looks ready.
  2. Policy before polishI would force the time-unit, edit-window, and role-switching decisions into a policy workshop before visual polish. The UI can display the answer; it cannot responsibly invent one.

More work

Continue with another case study.