Contact
Healthcare platform

OutcomeMD

Designing intuitive clinical tools for better care.

A redesign of OutcomeMD's clinical outcomes platform — dashboard, assessments and reporting — so care teams find the patients who need them without decoding the interface first.

View case study
Role
Senior UX/UI Designer
Scope
Dashboard redesign, assessment flow, reporting, design system
Focus
Hierarchy, task flow, consistency, clinical trust
Scroll to explore
Mission

Redesign the EHR dashboard around orientation.

Make OutcomeMD more intuitive and efficient so healthcare professionals can navigate patient data and assessments with less friction.

My contribution

Reduce UI complexity without losing clinical depth.

I focused on insight visibility, data-driven workflows, navigation logic, and faster decision-making across the healthcare product.

Electronic Health Records

User-centric design for dense clinical work.

OutcomeMD needed to feel less like a static record system and more like a working surface for care teams. The redesign focused on simplifying complex processes, reducing frustration, and helping clinicians manage patient tasks with fewer second-guesses.

Redesigned OutcomeMD dashboard with cohort filters, scope counts, a 25% average improvement gauge, survey engagement, and a patient improvement breakdown.
92%Task success in the final validation round
6Clinicians in moderated validation sessions
10–12Discovery interviews that set the dashboard's reading order
3Product areas rebuilt on one system: dashboard, assessments, reporting
Refined direction

A dashboard that reads in the order clinicians think.

Every zone answers one question a care team asks, in the order they ask it: which patients, how they are doing, whether follow-up is happening, and what to do next. Here is each component, and the UX reasoning behind it.

Refined OutcomeMD dashboard: a patient-search header, a cohort filters rail, scope cards, the improvement summary, survey engagement, an improvement breakdown and next-step actions.
Refined directionOutcome results overview

  1. Where do I start?

    Search-first header

    Finding a patient is the most frequent job, so search sits in the middle of the header instead of behind a menu, and the same field adds a new patient. The five destinations stay visible as plain-text tabs, and the active group lives top-right as a switcher.

    • Frequent task first
    • Recognition over recall
    • Primary navy#103B63
    • Active tab#EAF1FF

  2. Which patients am I looking at?

    Cohort filters rail

    Filters left the canvas for a dedicated rail, grouped as Organizational, Cohort and Confounding factors. Only the group in use stays open, so the Applied summary stays in view however long the filter set gets, and active filters show up there as removable chips: auditable at a glance, one click to undo.

    • Progressive disclosure
    • Visibility of status
    • Easy to undo

  3. Is it safe to share this screen?

    Page context & privacy

    An eyebrow confirms where you are and the title names the job of the page. PHI view sits apart as its own toggle, so revealing protected health information is always a deliberate choice, and Export is the only filled button up top: one primary action, not five.

    • Privacy by design
    • One primary action

  4. How big is this cohort?

    Scope cards

    Six counts read from the widest level to the narrowest, centers down to patients, mirroring how a health organization is actually structured. Large numerals over quiet labels make the row scannable in one sweep, and the outlined patient card marks who the numbers below are about.

    • Visual hierarchy
    • Match the real world
    • Selected outline#274E72
    • Selected label#2553D9

  5. Are patients getting better?

    Improvement summary

    The answer comes first. Start and latest follow-up scores frame a half-ring, and the change, an average improvement of 25%, is the largest text on the card, paired with an arrow so it never relies on green alone. The time range stays explicit in the corner.

    • Answer first
    • Not color alone

  6. Can I trust these numbers?

    Survey engagement

    Outcomes are only as reliable as survey completion, so the follow-up funnel sits right beside the result. Each step pairs a bar with a right-aligned rate; steps on track stay navy and the ones slipping turn orange, so the eye lands on the gap before reading a single row.

    • Data quality in view
    • Pre-attentive color
    • On track#103B63
    • Needs follow-up#F79009

  7. Who is slipping?

    Improvement breakdown

    An average can hide the patients who are getting worse. The breakdown splits the cohort into three plain-language statements, each a large rate over its patient count, and red is reserved for one thing only: patients doing worse than their first assessment.

    • Averages hide outliers
    • Plain language
    • Improved#027A48
    • Worse#B42318

  8. What do I do next?

    Next-step actions

    The screen ends where the work continues. View patient list is the one filled button, the shortest path from an insight to the people behind it. Details & reports goes deeper, and Add to report sits apart on the right as a collect-for-later action.

    • Clear next step
    • Action hierarchy
Move the cursor across the dashboard to compare the previous direction with the redesign.
UX cases

What a clinician asks, and how OutcomeMD answers.

Pick a question. See the screen that answers it.

Pick or add the patient first.
Receiver panel of the send-assessment screen with Quick Add, Add patient with full info and a search field for patients already added.
Concepts · not shipped

Three I would design next. Try them.

    Pick who gets a reminder.

    Nudge everyone who skipped follow-up.

    Pick who

  • 3 of 12 patients would alert you today.

    Preview who an alert would catch.

    Drag the line
    1. Baseline
    2. 6 wk
    3. 3 mo
    4. 6 mo

    Up 22 points since baseline.

    Show each patient's scores over time.

    Tap a patient
Assessment flow

Assessment work became a guided clinical path.

Sending an outcome assessment is three decisions: who receives it, what to send, and when it goes. Here is that screen rebuilt in the refined design — live, so you can pin, bundle, add to the cart and send it yourself. I built this rebuild in React and TypeScript for this case study.

Assessments · Mimik Demo Group

Send an outcome assessment

Receiver

Choose a patient outcome assessment

  • Try it: pin an assessment and it is saved in the Pins tab.

    • Inguinal Hernia

      • Core Outcome Measures

      16 questions about 4 min

    • Pelvic Organ Prolapse

      • Pelvic Organ Prolapse
      • Urinary Incontinence Sexual Questionnaire – Short Form

      10 questions about 22 min

    Male

    • Pelvic Organ Prolapse

      • Pelvic Organ Prolapse
      • Urinary Incontinence Sexual Questionnaire – Short Form

      10 questions about 22 min

    Female

    • Female Incontinence

      • Urogenital Distress Inventory – Short Form

      10 questions about 22 min

The assessment screen rebuilt in the refined design — live, not a screenshot. Pin, bundle, add to the cart and send. Patients and bundles are sample data.

Every part of the send, explained.

Layout

One screen, three decisions. Who receives it sits at the top, what to send fills the middle, and the cart holds the send. The header is the same one the dashboard uses, so moving between outcomes and assessments never changes the frame.

Receiver

The patient comes first, not last. Quick add takes only what is needed to send — a name and a way to reach them — so a new patient never blocks the clinic. Everyone already in the system is one search away.

The library

Grouped by body system, priced in patient time. Sixteen categories open one at a time, and every assessment shows what it contains, how many questions it asks and how long it takes — so the burden on the patient is visible before anything is sent.

Pins

Pin it once, send it forever. The assessments a clinic sends every day do not belong behind a search. Pin one and it is saved in the Pins tab, waiting for the next patient — which is exactly what happens in the rebuild above.

Bundles

Protocols travel together. Assessments that always go out as a set are grouped once and sent as one — a follow-up protocol becomes a single action instead of separate searches.

Preview

See the patient's side before sending. Preview opens the questionnaire exactly as the patient receives it, so a clinician can judge the wording and the length without sending themselves a test link.

Cart and send

One review, one send. Pins, bundles and search all feed the same cart. Before it goes, the cart adds up what the patient is being asked for — assessments, questions and minutes — and then sends in a single action.

Research archive

A guided carousel of research boards, flow maps, personas, tooling, and solution framing that explains how the UX direction was shaped.

OutcomeMD user research and data collection board with patient, interview, testing, and task metrics.
Research inputs

User research and data collection

Clinician interviews, task analysis, and platform data gave the redesign a measurable starting point.

OutcomeMD search and discovery flow board with filter mechanisms and discovery pain points.
Discovery flow

How clinicians find the data they need

The search path was mapped from login to dashboard, filters, patient list, assessment review, and return actions.

OutcomeMD design process board showing research, synthesis, sitemap, wireframe, testing, and delivery.
Design process

From research signal to shipped interface

The process moved through synthesis, sitemap, wireframe, experiment, high-fidelity design, testing, and delivery.

OutcomeMD tools and tech stack board showing design, project management, collaboration, and research tools.
Tooling

The execution system behind the work

Design, project management, collaboration, and research tools were organized around handoff and decision tracking.

OutcomeMD project management approach board with sprint cycle, responsibilities, rituals, and risks.
Project rhythm

A managed cadence for clinical product work

Sprint rituals, responsibilities, risk management, and communication practices kept research and delivery aligned.

OutcomeMD user personas board with goals, pains, needs, and cross-persona themes.
User personas

Who the product had to support

Personas clarified the needs of primary care, specialist, administrator, and patient-facing workflows.

OutcomeMD key solutions board showing filter dashboard, assessment interface, color system, and navigation improvements.
Solutions

What changed in the product experience

The redesign focused on persistent filters, assessment flow, navigation that surfaces next steps, and a more usable layout system.

Challenge

Make clinical complexity legible, structured, and actionable.

In clinical software, every extra moment of hesitation matters. The redesign reduced scanning effort, clarified the hierarchy of patient data, and made long assessment workflows feel less mentally expensive.

Dashboard hierarchy

Put the clinical signal in the first scan path

The dashboard was reorganized so outcomes, patient scope, and follow-up status could be understood before users dug into details.

Assessment workbench

Turn dense form work into a guided task

Assessment selection, long forms, and record detail views were structured around progress, state, and next-action visibility.

Reporting continuity

Keep evidence connected from summary to patient

Filters, result summaries, export actions, and patient lists share the same logic so users do not rebuild context at every step.

SiteMap

Structure came before surface polish.

Before moving into detailed UI, I mapped the dashboard ecosystem so the product could support real clinical movement instead of isolated screens.

OutcomeMD sitemap and workflow map connecting dashboard, assessments, reports, and patient details.
System mapThe product was mapped before polishing screens.

Dashboard entry points, assessments, reports, filters, and patient details were connected first so the interface could behave like one system.

01

Map the dashboard before wireframes

The sitemap clarified how dashboards, assessments, reports, and patient details needed to connect before visual design started.

02

Identify the core clinical objects

Centers, providers, patients, assessments, diagnoses, filters, and reports became the organizing structure for the product.

03

Design navigation around real tasks

The final structure supported easier movement between scanning outcomes, filtering cohorts, and reviewing patient-level evidence.

Design system

The foundations and components powering the redesign.

Explore the interactive library of tokens, foundations, components, and clinical patterns. The original documentation boards sit alongside the expanded system.

Experimentation

Testing how clinical data should be presented.

We explored multiple dashboard directions and used research feedback to understand which patterns made the data easier to find, compare, and trust.

OutcomeMD dashboard showing organizational filters, patient outcomes, and improvement metrics.
OutcomeMD experimental dashboard showing provider layout, filters, and score widgets.
The data was visible, but the visual language competed with itself.

Research showed that mixed colors, gradients, and competing panels made some users work harder to understand the hierarchy.

OutcomeMD experimental dashboard showing result cards and filtered categories.
Navigation and icon weight were too dominant.

The top navigation and left icon system needed to become less dominant so clinicians could find the data faster.

Behind OutcomeMD filters

Filters became the operating system for the product.

OutcomeMD needed clinicians to segment patient data, move into reporting, and return to detail views without losing trust in the state of the system. The filter model was redesigned to stay persistent, visible, and easier to reason through.

UX principle

Persistent context lowered the mental cost of clinical analysis.

  • Filter state stays visible in the sidebar, tabs, and report outputs.
  • Include and exclude logic supports real clinical segmentation instead of a generic search box.
  • The UI moves from population-level signal to patient-level evidence without resetting context.
OutcomeMD diagnosis filter panel with include or exclude rules and long diagnosis lists
02 / Clinical taxonomyDiagnosis logic is visible, not hidden.

And/or, include/exclude, select-all, and long diagnosis lists are exposed as deliberate choices for cohort building.

OutcomeMD filter sidebar showing confounding factors, compliance range, and applied chips
03 / Persistent contextConfounding factors remain editable.

Life events, injuries, laterality, compliance ranges, and selected chips live beside the data they affect.

OutcomeMD provider dashboard with a dedicated filters module and progress charts
04 / Dashboard structureThe overview clarifies who and what the data represents.

Provider, center, team, and patient selectors establish scope before completion gauges and outcome cards appear.

OutcomeMD dashboard with follow-up metrics, department filters, and summary results
05 / Outcome summaryInitial and follow-up scores become the main story.

The dashboard compares baseline, follow-up, improved, and worsened values without making users decode raw tables first.

OutcomeMD results grid showing applied filters and progress measures across assessment categories
06 / Filtered resultsFilter choices become readable result cards.

Outcome categories inherit the same scope, making it easier to compare clinical signals across a selected cohort.

OutcomeMD results overview with filters, patient list tab, and export actions
07 / From analysis to actionExport and patient lists stay attached to the same state.

Reporting actions remain beside the filtered view, so evidence can move directly into review or follow-up.

OutcomeMD detailed results cards and patient list under the active filter state
08 / Patient evidenceDetail cards bridge the macro and micro view.

Patient names, status, scores, and movement are grouped so teams can inspect the evidence behind a trend.

OutcomeMD patient list and reporting table paired with active filter controls
09 / Report auditabilityThe table keeps the workflow accountable.

Filtered rows, report fields, and notification setup give teams a traceable path from cohort to patient follow-up.

Exploration and system thinking

The interface was designed from the flow up.

Beyond the shipped screens, the project included workflow maps, care-management explorations, and monitoring concepts that helped align the broader product behavior.

OutcomeMD care management flow leading into patient details
Care continuityConnect service history to patient detail.

The flow shows where clinicians need to move from care management context into individual patient review without losing the thread.

OutcomeMD notification and recent results monitoring screen with scatter chart
MonitoringTurn recent results into action triggers.

Alerts, result movement, and patient lists transform outcome tracking from passive reporting into a follow-up workflow.

Outcome

Clinicians found patients without decoding the UI.

What shipped, what I measured, and the decision I would make differently.

What shipped

  • Redesigned outcomes dashboard: cohort filters, scope counts, improvement gauge, survey engagement and the patient improvement breakdown
  • Assessment sending flow — receiver, assessment library, pins, bundles and cart
  • Reporting and patient-list views that carry the filter state from summary to patient
  • A documented design system: type, colour tokens, spacing and a focused component library

What was measured

task success in the final validation round
92%
6 clinicians, moderated sessions on the redesigned dashboard and assessment flow
clinician discovery interviews
10–12
They set the dashboard's reading order before any wireframe

The numbers above are the ones I measured directly during the project; I don't quote post-launch platform metrics I didn't own.

The call I’d revisit

I'd ship fewer filter controls on day one.

The filter model is the product's engine — include/exclude logic, confounders, compliance ranges — and I designed all of it to be visible at once. Looking at it now, the persistent sidebar earns its space for a clinic leader and costs a clinician who only wants one provider's list. If I did it again I'd ship the organizational filters by default, hold the clinical taxonomy behind one “Refine cohort” step, and let usage decide which controls graduate to the sidebar.

What I’d instrument next

  • Time from login to first filtered cohort
  • Share of assessments sent as bundles rather than one at a time
  • Follow-up completion within 30 days of a flagged result