Lifelong Analytics: Equity, Data Viz & Precision Medicine

About the Author: Emily Chu is a Data Visualization Developer and Motion Designer currently focusing on data visualization engineering, machine learning and interactivity. Her background spans program management, research, business, design and technology. (MS Data Visualization, Parsons School of Design) | Website: 3milychu.github.io

 

The Spotlight on Healthcare Data

The tidal wave of new healthcare technologies is dizzying. Telehealth, artificial intelligence, AR/VR, 3D printing and patient monitoring promise that the future of medicine will be more efficient, affordable, secure and personal. Most advancements spotlight healthcare data as the foundation: we are capturing it, sharing it, making use of it and disseminating it in ways we’ve never done before.

Healthcare Data’s Rising Value and Impending Inequity

Consider this year’s Economic Forum’s meeting in Davos, where key industry leaders stated that using global healthcare data available to them, Machine Learning will be able to uniquely pinpoint the most effective treatment for an individual. At the current rate of data representation, however, health systems will be much poorer at offering efficient and highly accurate treatment for individuals that are not of European or East-Asian descent.

Meanwhile, the momentum behind capturing healthcare information is the heightening awareness of its value and security. Companies like Nebula Genomics, for instance, are offering to sequence your genome and secure it on a blockchain, wherein you will have control over the transfer of this packet of information and it will only go where you send it. In a consumer-driven healthcare world, what types of customers will have the privilege to understand what this even means?

What we can do with healthcare data can level the playing field.

We can make it secure and affordable for everyone, regardless of condition, race, or socioeconomic background, to receive the most effective treatment available. Looking at the typical health system infrastructure, where do we start?

Enter Electronic Health Records

Electronic Health Records or Electronic Medical Record (EHR/EMRs) are now a standard method of healthcare data storage and exchange. Patients are able to view an electronic copy of their medical records and physicians are able to share test results with patients. It can be thought of as the start of healthcare data consumerization. It is perhaps the perfect training ground to help the most vulnerable populations understand –

  1. how valuable their healthcare data is and
  2. how to harness it to improve their health and receive the most affordable, effective treatments in the future.

Since its inception, we now know that approximately half of the U.S. population encounter difficulties in comprehending and utilizing their health information, ushering in the need for a “visual vocabulary of electronic medical information to improve health literacy”. In 2014, a study revealed that 63.9% of of EMR survey respondents complained that note-writing took longer, and as of 2017, 94% of physicians in a survey were overwhelmed by what they believe to be “useless data”.

Visualizing Healthcare Data for the Most Vulnerable: From Collection and Adoption to Accuracy and Feedback

One important step is to get the most vulnerable populations – lower literacy individuals, patients with chronic or debilitating conditions, the elderly – to find a real use in capturing data and finding an enjoyment in doing so. The following demonstrates an example of how this updated electronic health record might function.

From Integrated Treatment Adherence to Responsive Feedback to Lifelong Analytics

In Visual 1.0: Simple Gamification of Healthcare Activities (below), for example, the patient is first shown how medications and healthcare tasks such as “take your blood pressure” can be gamified in a simple user experience to encourage data collection.  

Visual 1.1: Progress Over Time (below) shows how collecting vitals and treatment plan adherence might then be synced and displayed in the record shared with physicians. 

In Visual 1.2 Breakout view of healthcare activity or Biometric Marker (below), consider that the main dashboard view can be broken down and analyzed easily by physicians.  

Visual 1.3 Condensed Progress Summary and Feedback for the Patient (below) then illustrates closing the feedback and health comprehension gap that is often left open after treatment, by condensing the analytics into a simple progress view over time. Recommendations for the medical aspect (i.e. treatment plans) or maintenance behaviors (i.e. exercise) are adaptive. For example, at predetermined check-in intervals or when tracking metrics trigger a certain threshold, the treatment plan adapts based on level of adherence or other care plans that were implemented. Finally, consider that patients should be able to view future states assigned to them by predictive analytics (not pictured). In this scenario, what I would call Lifelong Analytics, individuals securely own all their healthcare information and are able to compare how predictive analytic models place them in the future.

By using the electronic health record as a catalyst to drive data collection and adoption among the most vulnerable, we are securing a pool of representative information for groups that may otherwise be left behind in the race for precise treatment at near-to-no cost. Along the way, through digestible habits and user-friendly actions, patients will be exposed to the power behind documenting their healthcare information. Once individuals are empowered with their data and what it really means, we can imagine a future where people are quick to stand up for the ownership of their data – and ensure that advancements that are made considering their footprint.

Takeaways

The poor, the elderly, the sick and the underrepresented have much to offer to the future of medical practice. They offer interesting challenges and high payoffs in cost efficiencies. When we consider a future where data will be dynamically classified and trends predicted, it is important to concentrate adoption among these groups. Some methods we discussed in this article:

Making treatment plans easy to track and adaptable

Treatment plans should be easy to track. Monitoring can be easily integrated into our routines, or in the future – automatically reported back to us. Providers should be able to understand what adaptive measures need to be taken should we miss a dose, or life interferes with our rehabilitation plan.

Making our medical history secure, transparent and shareable

Technologies currently exist to ensure our healthcare information belongs to us, and we have ownership over where it is transferred virtually. Visualizing healthcare information using a visual vocabulary that demystifies our health history, and shared among all providers in our care network can strengthen this transparency.

From responsive feedback to lifelong analytics

Consider a future where individuals with secure ownership of their healthcare data can access not only responsive feedback from their care providers, but see how their lifelong analytics are affected with each stint of perfect treatment plan adherence or alternative care plan. In other words, imagine what predictive analytics has to say about us is eventually comprehensible and accessible to us as individuals.

By visualizing and making healthcare information for the most vulnerable readily accessible and comprehensible, we make it possible to access the most difficult treatments responsively and potentially risky treatments with transparency. In the end, this can teach an entire population how to better develop an infrastructure that prepares and cares for us when we too age, get sick or fall into disadvantaged circumstances.