iHealth Knowledge Base

A dedicated Knowledge Base for care team members to have a centralized resource that offers quick and consistent answers to common questions, improving efficiency and knowledge sharing.

My Role

Product Designer

Team

2 Product Designers
1 Product Manager
1 Content PM
Engineer Team
Data Team (LivexAI)

Time

Oct 2023 - Nov 2023
(5 weeks)

OVERVIEW

01. What is iHealth Unified Care?
iHealth Unified Care is a remote patient monitoring program designed to help healthcare providers manage chronic diseases through digital means, contributing to a more connected and streamlined healthcare experience.
02. Unified Care Knowledge Base
WHAT is this?
Digital documentation that offers quick and consistent answers to common questions.
WHO is this for?
B2B internal users (Care Team members)
HOW to do it?
Design an easy-to-use online platform to solve heavy document problems.
Problem

Care Team members struggle to handle countless paper documents. Document digitization becomes a huge pain point for company's digital transformation.

Real-World Data
120+
Care Team
members
5+
Times to check doc
per day
4+
Types of doc per
day per person
1 hr
Average time looking
for info
65%
Percentage of acquiring the info they need
Document Types
CA Manuals
HC Manuals
RD Manuals
Care Team Mastersheet
Reporting Parameters
Implementation Manuals
Abbreviation List
Official HIPPA Manual
Device Troubleshooting
etc...
Solution

Build a dedicated digital platform for care team members to have a centralized resource that offers quick and consistent answers to common questions, improving efficiency and knowledge sharing.

Final Design Prototypes
Looking for specific articles
Refining Search Results with Advanced Filter
Approaches
Evaluation
Moderated Remote/In-Person Testing
Heuristic Evaluation
A/B Testing
Research
Card Sorting
Personas
User Interviews
Surveys
Design
Wireframes
Hi-fi Design
Prototypes
My Process

RESEARCH

WHO are the users?

All Care Team members:

  • Clinical Assistants (CA)
  • Health Coaches (HC)
  • Registered Dietitian (RD)
User Interviews & Surveys

To better understand Care Team members' workflows, daily responsibilities and their pain points, I conducted the user interviews first to get a basic understanding of their workflows and pain points. I also sent out the surveys to care team members afterwards in order to get better understandings of the pain points. 

We had 2 designers host the interviews, I was responsible for the Care Team Workflow, and she was responsible for Device Troubleshooting. The surveys have two parts as well.

We interviewed 6 Care team members, and received 40 feedbacks from them for the surveys. Through all these research, we've found that,

  • The Care team is overwhelmed by extensive protocols and documents, which often exceeding 60-90 pages, making it challenging to quickly locate specific information.
  • Newly hired Care Team members usually needs to reference the documents at least 15 times a day to become acquainted with their responsibilities.
  • The protocols and documents were written by several people at different time. Determining the most current version of the information in protocols is difficult due to inconsistent updates, creating confusion in daily responsibilities.
  • Care Team members will need to take on others' duties. RDs will need to take on CA and HC's duties occasionally, and HC will need to take on CA's duties sometimes.
  • Clinical Assistants often overlook crucial steps like obtaining consent forms during patient interactions, indicating gaps in current training protocols.
Personas

DESIGN CHALLENGE 01

Define the Information Architecture

Ideation

Informed by our research findings and with our user personas guiding our approach, we gathered our PM, engineers, fellow designer and our data scientists to collaboratively ideate potential solutions in order to solve the problem of document digitization.

- mind map -

Design Challenge 01: Define the Information Architecture

In tackling the challenge of defining an effective Information Architecture for our Knowledge Base, my journey began with two fundamental steps: content auditing and open card sorting.

Recognizing the importance of a solid foundation, I collaborated closely with our content Product Manager during the content audit phase. Our goal was not just to catalog our existing materials but to critically evaluate and update our content. I advocated for the need to rewrite our manuals, ensuring they were not only up-to-date but also structured in a way that best served our users' needs.

The screenshot of a hierarchy diagram of the re-written manuals.
Open Card Sorting

The first strategy to achieve the goal would be Open Card Sorting exercise based on the re-written manuals to uncover how different users conceptualize and categorize information based on their own understandings or mental models. This user-driven approach provided invaluable data on how real users perceived and categorized our information, laying the groundwork for our Information Architecture.

I've identified two distinct categorization strategies from the participants:

By Tasks
──────
This categorization reflects a
process-oriented view, where participants might prefer to see the information in the context of the workflow or task they are performing, such as "how to enroll a new patient".
By Roles
──────
This categorization reflects a
role-based view, where participants might prefer to access information
that is specifically relevant to their
job function, like "Clinical Assistant
daily responsibilities" or "Registered Dietitian's responsibilities". 

After conducting a thorough analysis through open card sorting and content auditing, we crafted two comprehensive sitemaps.
In the process, we recognized the need for clearer communication, leading us to revamp the terminology used across our platform. This involved renaming certain content categories and labels to align more closely with the language and mental models of our end-users, ensuring the terms are intuitive and the content easily navigable.

Wireframes

Leveraging the insights from our open card sorting exercise, I crafted two distinct design versions for our Knowledge Base.

Task-focused IA
Role-focused IA

EVALUATION & ITERATIONS

User Testing

The goals of our evaluation sessions are: 1.To determine the most effective categorization strategy for the Knowledge Base. 2.To identify any usability issues within our current design. The methods I used for evaluating the designs are A/B Testing and Think-aloud testing. We gathered 6 participants to do the test. They are newly hired team members and existing members.
Here's a brief overview of the process:

  • Explain the purpose of the test, the tasks they will perform, and the think-aloud protocol.
  • A/B Testing setup: Divided them into two groups, and randomly assign participants to one of two versions of the designs.
  • Instruct participants to complete a set of tasks.
Outcomes
Task-focused IA:
Task Completion Rate
Participants say that the task-focused IA is not intuitive enough.
Error Rate
Participants say that it is hard for them to decide if it is online or offline.
Role-focused IA:
Task Completion Rate
The participants finished all tasks.
Error Rate
Participants made no mistakes on Role-based IA.
User Feedback & Iterations
  • The role-focused IA is much intuitive. Compared to Task-focused IA, it does not cost users much time to think.
  • Users think that they were easily get lost if they need to go back and select other contents to read. What I added is a breadcrumbs navigation.
  • It is confusing for users who's the target audience of the articles.
Based on the user feedback, I made several improvements as following:
  • Added the number of the search results.
  • Added the filters on the side of the page, helping users find the info immediately.

DESIGN CHALLENGE 02

Improving Information Discovery

Evaluations

  • The goal of the evaluation: to evaluate if the search function is effective to use.
  • The methods used for evaluation here: Heuristic Evaluation & User Testing
  • 2 product designers: heuristic evaluation; 5 Care team members: user testing.
Heuristic Evaluation

The goal of this session is to assess how easily users can interact with the search function based on the first design version, provide actionable feedback based on usability principles to enhance the usability of the search function and effectiveness;

I gathered some of the UX designers in our company to review my design based on established usability principles. They assessed my design for common usability issues and provide feedback based on their knowledge and experience.

Solutions & Impact

Drawing from a wealth of data gleaned from our evaluations, l've distilled the key insights that will help our information discovery enhancements.

Throughout the decision-making journey, I spearheaded analyzing and synthesizing the data, and distilled the data into key insights; I also helped prioritize issues based on their impact on the user experience and the product's goals. With these insights in hand, I gathered our Product Management and Engineering teams to navigate through my findings and engage in a dialogue about potential solutions and their technical viability.

Here are the final three decisions we've made:

Implementing "recent"
and "top searches"
──────
Can be helpful to users in reducing cognitive load, saving time, social proof and discoverability.



Implementing content labels
and a filter features
──────
Labels can include content types (article, FAQ topics, etc.), audience (CA, RD, HC), etc; Incorporate a filter feature on the search results page that allows users to refine their search based on specific criteria.
Implementing
spelling correction
──────
Can be helpful to users in
saving time and increasing their work efficiency.
Now, it comes with this question:
How might we...

...Leverage the insights above to improve information discovery, reduce cognitive load, save time and energy for users...

… enabling them to swiftly locate the precise information they need and enhance their work efficiency?
Organizing Filter Tags

In order to build a robust filter feature, I decided to implement content labels before the design.
Based on all the manuals, protocols and documents Care team has provided, the filter tags could be divided into four types:
Content type, Source, Topic, Clinic, and Role.

Iterations 1/4: Filter

Based on the user feedback, I made several improvements as following:
  • Added the number of the search results.
  • Added the filters on the side of the page, helping users find the info immediately.

Iterations 2/4: Top Searches

This is the first time using search function. When clicking onto the search bar, it will give you a list of the top searches that other users has recently searched.
  • Added the number of the search results.
  • Added the filters on the side of the page, helping users find the info immediately.

Iterations 3/4: Recent + Top Searches

Iterations 4/4: Spelling Correction

Tencent Healthcare Projects