ePROMS Portal – a digital health technical implementation and Case Study

Project: HINF 597 Final Field Project at University of Victoria

Author: Ann Chou (2020)

Overview

This project explores the design and prototyping of an electronic Patient-Reported Outcome Measures (ePROMS) portal to support patient agency and participatory care. Built with HL7 FHIR and LOINC standards, the system allows patients to self-report their symptoms digitally using validated clinical tools.

Problem Space

In many health systems, patient feedback is underutilized in electronic health records (EHRs). This project addresses that gap by creating a portal that collects structured symptom data directly from patients, allowing for improved self-assessment, care planning, and shared decision-making.

Patient Agency and the “What Matters to Me” Movement

Why Patient Voice Matters

Until recently, patients in British Columbia had limited opportunity to enter their own health notes. In March 2020, tools like the BC Health Gateway and Telus Health’s Babylon app began allowing people to input medical history and reflections directly.

This shift reflects growing support for patient-as-partner approaches (Chow et al. 2009; Armstrong, 2015). Previously, clinicians largely controlled what entered an electronic health record (EHR)—often prioritizing professional goals over patient needs.

Initiatives like Alberta Health and Fraser Health’s “What Matters to You?” campaign signal a new paradigm: strategic listening and shared decision-making, where personal health priorities can shape care delivery.

Understanding PROMs and ePROMs

Patient-Reported Outcome Measures (PROMs) are validated questionnaires used to capture a patient’s perceptions of their health. They’re tested for reliability, sensitivity, specificity, and predictive value.

PROMs come in two types:

Generic — e.g., EQ-5D for quality-of-life
Condition-specific — e.g., PHQ-9 for depression

Typical formats include Likert-type scales (e.g. strongly agree → disagree) or visual analog lines between endpoints.

In health informatics, ePROMs digitize this process—bridging subjective patient insight with structured clinical records. Organizations often prescribe ePROMs before and after surgery to monitor progress or trigger follow-up. Patients can also self-administer ePROMs for disease management or visit preparation.

Research shows that when patients deliberately reflect and report their data, they make better-informed health choices (William et al., 2015).

Limitations and Design Considerations

While PROMs and ePROMs are valuable, they present challenges. Many tools feel too long or impersonal, with limited cultural or linguistic sensitivity. Questions may not capture the full depth of lived experience.

Psychiatrist and anthropologist Arthur Kleinman (1978) proposed the Explanatory Model, emphasizing that patients and cultural groups hold varied beliefs about health and illness. Open-ended prompts like “What matters to you?” can reveal far richer stories than standardized checklists.

Some PROMs are available as Computerized Adaptive Testing (CAT)—such as PROMIS—but they still face accessibility barriers for self-administering patients.

Reviews have also shown that mismatched PROM content can reduce patient satisfaction, with symptoms either underrepresented or framed too generically.

For example, while EQ-5D-5L is used by CIHI as a benchmark for hip and knee arthroplasty, patients often desire tools that feel more human and less survey-like.

Goals

Enable patients to self-administer PROMs digitally
Bridge pictorial symptom expression with standards (LOINC, SNOMED-CT)
Prototype a mobile-friendly user interface using LHC-Forms tools

Technologies Used

FHIR SDC SMART App – mobile-first interface integration
LForm Builder & Widget – questionnaire generation and rendering
HL7 FHIR STU4 – interoperability standard
LOINC & SNOMED-CT – terminology standards

Design Process

Due to constraints such as limited time, COVID-19 restrictions, and lack of patient partners, a mockup-driven approach was used. Validated PROMs like PHQ-9 and PROMIS were selected and encoded using LOINC panels. The portal interface was built to simulate a patient’s symptom-reporting flow using standard tools from the U.S. National Library of Medicine.

Technical Implementation: Capturing Mental Health PROMs

As part of the ePROMS Portal prototype, this project explores how patient-reported outcomes (PROMs) for mental health can be captured using interoperable standards. After comparing several options, the PHQ-9 questionnaire was selected for its structured format and clinical relevance.

Selected Question

"Little interest or pleasure in doing things" — a key PHQ-9 item used to assess mood and engagement.

Standards Mapping

Element	Standard	Code	Description
Question	LOINC	74250-9	Little interest or pleasure in doing things
Clinical Concept	SNOMED CT	41723004	Loss of interest in previously enjoyable activity (finding)

Answer Options

LA6568-5 – Not at all (Score: 0)
LA6569-3 – Several days (Score: 1)
LA6570-1 – More than half the days (Score: 2)
LA6571-9 – Nearly every day (Score: 3)

Implementation Files

phq9-questionnaire.json – FHIR Questionnaire resource
phq9-questionnaire-response.json – Sample patient data
phq9-snomed-mapping.json – Concept mapping overlays
sample-sample.json – Custom LHC-Forms section

These files collectively demonstrate how PROMs can be digitized and semantically enriched using HL7 FHIR, SNOMED CT, and LOINC standards.

🧠 Mental Health EHR Prototype: PROM Questionnaire

Explore available sections and project reflection

This experimental prototype models how patients might complete mental health–related forms inside a standards-ready personal health record (PHR) system. It uses LHC-Forms widgets and a JSON structure compatible with HL7 FHIR and LOINC.

Design logic: The form includes 9 distinct sections, but the patient is only expected to choose and complete one section per session based on relevance, mood, or clinician suggestion.

📆 Session Information (Date, Location, Context)
🩺 Medical Conditions (Diagnosis Timeline & Notes)
❓ Questions to Ask Your Doctor
🗒️ Notes & Journaling
🧠 PHQ-9 Depression Assessment
📋 Mental Residual Functional Capacity (RFC)
⚡ FACIT-F Fatigue & Functional Health (v4)
🧩 PROMIS Applied Cognition (General Concerns 4a)
📊 Scoring Panels (PHQ & PROMIS raw + T-scores)

My reflection: While the form successfully models semantic richness and covers diverse dimensions, I quickly noticed how overwhelming it felt when all sections loaded at once. Through this exploration, I learned that minimalism and emotional resonance matter more than completeness—and that patient choice should guide data entry. I also saw how my custom field design allowed softness and adaptability, even within structured formats.

👉 Try a live demo via miniportal.html →

🛠️ Technology Note

View technical details about the PROM JSON & LHC-Forms setup

This prototype uses 9-sections.lforms.json — a structured LHC-Forms–compatible file that defines a semantic PROM questionnaire for mental health domains. You can explore it directly inside the GitHub repository:

📁 View 9-sections.lforms.json on GitHub →

📦 How It Works

To load this JSON as an LHC widget inside a browser-based form:

Include LHC-Forms JS & CSS in your HTML (via CDN)
Create a container (e.g. <div id="lforms-container">)

Use JavaScript to fetch and render the form:

fetch('9-sections.lforms.json')
  .then(response => response.json())
  .then(formData => {
    LForms.Util.addFormToPage(formData, 'lforms-container');
  });

When rendered, the form allows the patient to complete one selected section based on their needs or clinical instruction. Submitted responses are returned as a FHIR QuestionnaireResponse object for processing or integration.

Evaluation

The portal prototype was assessed using heuristic inspection and cognitive walkthrough methods. These evaluations showed good usability in terms of navigation and visual clarity but highlighted length and cultural sensitivity of PROMs as areas for future improvement.

Patient Journey Walkthrough

The following screens demonstrate the ePROMS Portal experience—from referral input to PROM completion and prototype evaluation. Screens are organized by stage, using visual mockups and forms created with LForm Builder & Widget and FHIR-compliant JSON.

🧭 Phase 1: Initial Mockups

Prompt asking patient to enter referral number — epr_2_ref: Referral number prompt

Input field to enter referral number — epr_3_keyinref: Referral entry interface

Body silhouette interface for selecting symptom location — epr_4_body_sil: Body outline for selecting areas of concern

Screen showing available PROM forms for selection — epr_5_select: PROMs selection screen

📝 Phase 2: Symptom Reporting Forms

The screens below show implemented data fields using LHC-Forms and customized JSON. Each interface supports accessible form input and semantic encoding.

Form with demographic fields and a free-text field for patient questions — epr_6_phr: Demographics and patient-entered question

Start screen for PHQ-9 depression questionnaire — epr_7_PHQ-9: PHQ-9 questionnaire interface

Start screen for alternate applied cognition questionnaire — epr_8_appl_cogn: Applied cognition questionnaire

Question with dropdown menu for 'Little interest or pleasure in doing things' — epr_9_little_interest: PHQ-9 dropdown for mood response

📊 Phase 3: Prototype Evaluation

Evaluation screen summarizing usability inspection — epr_10_eval: Heuristic evaluation summary (limited testing)

Next Steps

Refine PROMs for brevity and accessibility
Incorporate visual symptom capture (e.g., drawing interface)
Recruit patient partners for participatory testing
Explore mappings to Human Phenotype Ontology (HPO)

📘 Academic Report (By Request)

Why it's not shared publicly

Visual placeholder for academic report (blurred cover)

This ePROMS Portal prototype was originally developed for a graduate health informatics course during a period of illness and time constraints. The technical report includes early implementation notes and instructor feedback, but remains an unedited academic draft.

To preserve its value while honoring its rawness, the report isn’t published directly on this page.

📨 If you're interested in reviewing the original material for teaching, research, or documentation purposes:
Please reach out via annchou.github.io or DM @annreflection on Instagram.

🧪 Developer Repository

View Technical Repository and JSON Resources

For health informatics developers, clinical coders, or students exploring HL7 FHIR integrations—this repository includes implementation code, JSON mappings, and documentation for the ePROMS Portal prototype.

It hosts everything from PHQ-9 form logic to SNOMED overlays, coded symptom fields, and semantic infrastructure for building patient-centered health apps.

          📁 /json/
          ├── phq9-questionnaire.json           # Structured FHIR Questionnaire
          ├── phq9-questionnaire-response.json  # Sample response with coded answers
          ├── phq9-snomed-mapping.json          # SNOMED CT overlays
          ├── sample-sample.json                # LHC-Forms custom PHQ-9 field
          ├──  9-sections.lforms.json           # full mental health focused PROMS prototype and Personal Health Records demographics section
          
          📁 /rss/
          ├── ePR-screenshots/                  # UI walkthroughs and evaluations
          ├── json-phq9-LOINC-SNOMED-map.png    # Visual mapping diagram

Open GitHub Repository →

🪡 Project Update: Exploring Visual Symptom Capture

After outlining potential mappings to the Human Phenotype Ontology (HPO), I’ve chosen to pause further technical schema work and instead explore a different modality: draw-aloud symptom research.

This approach centers patient expression through sketching or annotation—allowing them to identify discomfort or concern on body silhouettes or spatial canvases. It's especially useful for those whose symptoms may be difficult to articulate with clinical language.

These visual methods open up inclusive design possibilities for health reporting tools. Long term, they may be layered with structured standards like LOINC or SNOMED CT to create hybrid visual-semantic interfaces that reflect the felt experience of care.