Flux Notes

Improve cancer care by allowing easier collection and curation of patient data

Enterprise | Electronic Health Record | Oncology

Role

I conducted an end-to-end design process for Flux Notes, an enterprise application for doctors to capture health data. This led to its pilot testing in dozens of clinics and hospitals around the US, and adoption of a health data standard at the national level.

I was a designer in a core 11 person team of 1 - 2 designers, 7 engineers, 1 SME, and 1 product manager.

I collaborated on user research, and executed on system mapping, interaction design, UI design, prototyping, design validation, design QA, and client management for all examples shown below.

Time

1.5 years | 4/2017 - 10/2018

Client

MITRE Corporation

Background

Clinicians rely on accurate data to create the best treatment plans for their patients. This quality of data is difficult to obtain.

Oncology is a space with an especially high need. Research of new oncology treatments rely heavily on high-quality patient data, making it optimal as a starting point to achieve the goals of Flux Notes.

Only 3% of oncology patients have high-quality data ¹

1.7 million people are diagnosed with cancer in the US each year ²

“Clinical treatment data cannot currently support data driven decisions of care” ³

Goals

Flux Notes product goal was to create an experience that demonstrates collection and curation of high quality patient data.

Design and develop an innovative method of capturing structured health data.
Design and develop a human readable view of the patient record.
Build all the features necessary to demonstrate the end-to-end workflow of Flux Notes.
Show how MITRE health data models and elements can be used within Flux Notes.

Flux Notes business goal was to help push development and adoption of the Standard Health Record (SHR), which was the data model and elements driving the content within Flux Notes.

Meet expectations of MITRE Corporation leadership on the progress being made in Flux Notes development.
Use Flux Notes to help gain additional partnerships for driving development of the SHR.
Use Flux Notes in a clinical setting to demonstrate the applications of the SHR in the real world.

Initial concept created to help gain funding and buy-in for the Flux Notes project from MITRE executives.

Research

“How do clinicians get patient health data now?”

“How do they interact with existing patient data?”

These were key questions for me at the beginning of the project. The users being clinicians meant that I needed to understand how they get and use data from day-to-day. With their time primarily being spent within hospitals or clinics during patient visits, it was important for me to be close to that space to see how they work.

After co-brainstorming with team members on available clinician resources and connections, it led me to engage in several initial research efforts over 2 weeks, in collaboration with an engineer colleague:

I conducted 3 subject matter expert interviews to understand interaction with patient data in an electronic health record (EHR) when creating clinical notes.
I shadowed 4 oncologists at the Dana Farber Cancer Institute (DFCI) to understand patient encounter workflows.
Set up weekly feedback sessions from 1-3 physicians to discuss project strategy and receive feedback on designs.

User Needs Synthesis

A picture of needs began to emerge from these research efforts as I continued to compare and contrast key written notes from my colleague and I after each research session.

I shared insights with additional physicians at MITRE and the Dana Farber Cancer Institute (DFCI) to see if my conclusions were in line with their personal experiences, adding a layer of validation and clinical accuracy to my findings.

Primary users

Oncologists at the point of care, meeting with breast cancer patients.

Secondary users

Physician assistants and nurse practitioners working in the oncology space.

Primary pain points

Holes in the patient record, such as a lack of structured information for family history, make it difficult to make a comprehensive plan of care.
Provider time is wasted because the display of patient information is not prioritized in a way that is in line with the oncologists cognitive workflow.
Understanding the patients health history is difficult because of the high visual density of information presented in the EHR.
Precious patient encounter time (15 minutes max) is wasted while trying to find specific information within the deep, tab based hierarchical navigation.
80% of patient information is in the clinical notes, but was not very useful because of its unstructured form.

An example interface I found while surveying EHRs. There is too much information, organized in a way not suitable for clinical workflows.

Design principles

Using these new insights, I generated a list of principles to help drive designs moving forward, to address key user needs uncovered during research.

Reduce the workload of capturing structured patient data.
Prevent information overload by prioritizing the display of information.
Make patient information human readable.
Reduce the time needed to find specific patient information.

Design

I found that health data capture and viewing was inherent to every single patient visit from start to finish. I created the following designs to fit within this framework of use.

Clinicians first view the patient record to understand current health status in preparation for the visit.

Personalize the display of information.

It was important to prioritize the patient record to the clinicians specialty, patient condition, and part of the encounter workflow that the clinician is currently in.

Make exploring the record feel natural.

Once information was personalized, I used interaction design techniques to reduce effort needed to explore that information.

Clinicians can explore using a map to jump or scrub to points of high interest, instead of using a sea of tabs and button clicks as usually seen in most EHR software.

Make exploring the record feel natural.

Make exploring the record human readable.

I found that clinicians described recall of patient information through narrative, not through a list of data points. I designed patient data within the health record to be viewed in narrative format.

After understanding the patient’s health status and needs, doctors then start writing a clinical note. I found that 80% of patient health information was being saved within these notes. With so much data, I felt that note creation would be a key step for my designs to focus on for structured data capture.

In the following early designs, doctors could capture patient data before starting their note by clicking through a point of sale interface using their mouse or finger. This way, when they did start writing the actual note, all necessary health data would already be captured.

Unfortunately, doctors had negative reactions towards using any workflow that forced them to add extra steps before note creation, or, that required them to alternate between a keyboard and mouse. I later changed my approach to allowing data to be captured without changing their existing workflows, and without the doctor needing to move their hands away from their keyboard during note creation and writing.

Similar to dot phrases, hashtags are used while writing a clinical note to quickly capture structured data without disrupting clinical note documentation workflow.

That new structured data could then be used to visualize and show clinical insights.

By showing informational relationships and connections through visuals, it allows easier communication of high density areas of patient information.

This is especially important for the display of longitudinal information like patient history.

Explore the Flux Notes prototype

Recent in-progress build of Flux Notes.

Impact

Flux Notes product goal was to create an experience that demonstrates collection and curation of high quality patient data.

Designed and developed all features necessary to show the end-to-end interaction of capturing and exploring structured health data within a follow-up visit use case.
Positive qualitative feedback for the end-to-end interaction, from testing Flux Notes with 6 clinicians from major hospitals such as Dana Farber Cancer Institute.
Populated all Flux Notes content with data elements from the SHR, allowing it to be a vehicle for clinicians to explore the SHR in a tangible way.
Created a research sandbox for several major MITRE initiatives surrounding standardized health data models.

Flux Notes business goal was to help push development and adoption of the Standard Health Record (SHR), which was the data model and elements driving the content within Flux Notes.

National adoption of the SHR in 2018 into the FHIR standard as a new oncology profile.
Gained an additional million in funding to push the SHR initiative forward for 2018 by exceeding expectations of MITRE leadership in Flux Notes progress.
Flux Notes enabled communication of the value of the SHR at several key health informatics conferences such as HIMSS and Mass Bioinnovation.
The team presented Flux Notes at major healthcare organizations, gaining partnerships at more than 4 major hospitals such as Intermountain Healthcare.
Flux Notes and its health data models (SHR) are being pilot tested at more than a dozen clinics and hospitals around the US.

Next steps

Though not currently feasible, my future vision of Flux Notes would be to provide a self-documenting encounter with a system generated plan of care to further save time within the patient visit and reduce clinician burden.

Takeaways

It was an eye-opening experience to see the time lost in clinic due to poor EHR design, and even more surprising to discover that it is the current standard within medicine.

The benefits of learning more about health informatics was apparent after seeing how much of the design constraints and considerations were driven by the clinical workflow and the flow of patient information through the health system.

That alone prompted me to enter the 2018 - 2019 health informatics certificate program at Northwestern University to expand my understanding in that space.

References

DOI: 10.1200/JOP.2013.001119 Journal of Oncology Practice 9, no. 6 (November 1 2013) 267-276.
Cancer Facts & Figures 2019. (2019). Retrieved from https://www.cancer.org/research/cancer-facts-statistics/all-cancer-facts-figures/cancer-facts-figures-2019.html
DOI: 10.1200/JOP.2013.001119 Journal of Oncology Practice 9, no. 6 (November 1 2013) 267-276.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3545660/

Edwin Choi

Work

About

Resume