Live blogging session #2: How to produce data visualizations that motivate citizens to solve problems

18 April 2013

The Govlab

SESSION #2 THEME: If Only We Knew…how to produce data visualizations that motivate citizens to solve problems.

  • CHANGE AGENT: Tim Kelsey, National Director for Patients and Information in the UK National Health Service
  • FACILITATORS: Aaron Cohen and John Seely Brown

The following is a liveblog of one of six sessions at The GovLab Experiment: Making Engagement Work. It will be updated as the sessions progress from 10:00 AM – 3:00 PM. For the full schedule, please see: http://thegovlab.org/events/making-engagement-work/. For more, read the session description at the bottom of the post.
Note: Some live blog notes were inadvertently included here instead of for Session 5. Content has since been corrected.
(Live blog updates provided by N.Y.U. students Ryan Norton, Viral Tarpara and William Heselton-Clements)
—
2:45 PM: Lots of refining and re-hashing original points among the group. Looking forward to the presentation!
2:04 PM: Group identifies three key problems to solve:

  1. Unnecessary duplications of care
  2. Gaps in care
  3. Inappropriate places to receive care
1:49 PM: Point made on thinking of underlying data as a way to drill down and look at past history– “what is the value of the theogram?” — A way for the doctor and the patient to look at and understand their documented health history, and make predictive recommendations for the future.
12:05 PM: The group is discussing making the NHS patient database more like an interactive, tabulated mint.com. Patients can log-in and view their medical history, browse through tabs, and explore the meaning of medical events or issues they face through links to webMD or other medical databases. The consensus is that the medical research has been completed and data collected, but too often a patient’s health is boiled down to the top-line bullets like cholesterol, blood pressure, body mass, etc. A patient should be able to research and understand the issues they face.
A powerful example was a participant’s mother has bone cancer and his family’s struggle to understand that in relation to her progress. The tabulated record could be flipped from patient view (layman’s) to doctor view (expert) for easily referenced background data on a patient. The ability to flip the data saves on redundant pages/tabs, and can be a useful service tool for doctor’s to communicate jargon with patients.
In line with the Mint comparison, other recommended features included:
  • A recent patient history homepage (like Mint)
  • Slidable icon-browsing of content for your medical history (like iTunes album covers)
  • A recommendation feature so that patients can focus on what they need to do to be healthy. NHS officials say they have the data for what good health is, but better services are a matter of linking that knowledge with patient understanding and activity.
11:41 AM: What if every time a child was born, a parent was provided with a health guidebook, personalized using the genomic data of the child? For example, if you knew your child was born with a gene that increased a person’s likelihood for breast cancer, then could knowing this allow the parent, child, and eventual adult make better health decisions? What if the systems could properly explain risks and statistics and a meaningful way that balances the propensity for alarm with the rational change in behavior?

11:27 AM: The conversation is starting to shift to development.
11:21 AM: So much of the change that is happening today is a result of top-down change or grassroots movement. What is needed is a more nuanced change agent who meets somewhere in the middle.
11:19 AM: Non-traditional idea: what if there was a radio station that would educate people about health?
11:13 AM: Understanding context is very important. Connecting with community leaders using the right media channels leads to the best outcomes and foster engagement.
11:11 AM: “How do you get the data to tell a story?” Visualizing things you cannot see is very important.  This is known as “soundifying.”
 10:56 AM: Interesting comment stating that there is little correlation between poor restaurant grades and food-borne illness.
10:55 AM: Convey complex data using simplified visualization models. Example, in New York City, Restaurant grading of “A-F” was available information, but it wasn’t until they mandated restaurants post their grades publicly that impact grew.
10:47 AM:– N.H.S. “111” service compared to a live digital complaint feedback system.
10:45 AM: Net promoter score is the topic of discussion. The challenge with net promoter score is how to effectively visualize. Key insight about big data and data visualization is that rendering data so that people understand the data isn’t enough. One must render the data in such a way that people understand the data AND act.
10:34 AM: U.K. cardiac surgeons opened their own data on their personal surgical death rates with names. Surgeon collaboration reduced death rates by one-third. Lives were saved due to data visualization.
10:29 AM: – Participant says that the N.H.S. has the same problems that every other health system has, which is that the system was designed for the producers.
10:19 AM: – Noted that healthcare in the U.S. is on an unsustainable path. The U.S. spends more on healthcare than the next 10 countries combined. “Targeted Data Liberation” is the buzzword of the day. Government has 90,000 smart people in health and human services but cannot do much because of the “command-control” structure that keeps them idle.
 10:10 AM: People are introducing themselves. Wide set of fields represented: health, mathematics, civic engagement, film production, politics, technology, web. One participant is an electrical engineer who is also a music research professor.
—
SESSION DESCRIPTION:  
With big data, informatics and predictive analytics, we have an unprecedented opportunity to revolutionize policy-making. Yet big data by itself can be overwhelming. We need tools to create visualizations that will engage the public more effectively in data-driven problem-solving. Big data outputs need to harness the explanatory power of graphics and visuals to elicit fact-based insights: from citizens to decision-makers alike.
This conversation will bring together individuals from data science, statistics, design and the arts, citizen engagement and other disciplines with deep insights into how to present data, media and information to the public in order to make their participation better informed and more productive. We will explore successful case studies of visualizations that have increased citizen engagement and awareness. Specifically, we will use health policy as the frame to design tools and methods for incorporating real-time data visualizations into citizen engagement practices in healthcare policy-making. For example, we will discuss what quantified self-movements, home health care professionals, and private urban clinics can teach us about informing and engaging the public. 
During this session we want to explore how to use data visualization to make 21st century, data-driven citizen engagement truly work. The National Health Service of the United Kingdom is in the process of creating an extraordinary new database. Called Care.Data, the project will: 

  • Collate data from all health and social care providers for the whole population of England
  • Record data at the individual patient level
  • Include hospital data, primary care data, community health services data, social care data, and specialist audit data
  • In addition, patients may choose to donate their genomic data

The data will be made available in three formats:

  • Aggregate data that will be published
  • Identifiable data that will be available to patient and their clinicians
  • Pseudonymous data that will be able to health service planners, and to researchers under strict legal contracts 

Some displays of this data have been created and will be demonstrated. Tim and his colleagues are focused on three separate problems. First are patients who have the legal right to download all of their personal medical data. NHS would like patients to better understand their healthcare needs by interacting with their own data. Current solutions would not be acceptable for the vast majority of Britons. Second, and more actionable, are the tools that could be created to enable clinicians, particularly primary care physicians, to see the entire history of their individual patients in a snapshot with wide. (Just as GP’s show patients their x-rays and interpret them, they should also be able to show them their data records and explain them) Finally, in order to provide timely and accurate data to patients and practitioners, medical data must be collected and re-distributed in a timely way; and data must be cleaned and categorized to remove regional variances, data import errors and other data quality problems.
 

Creative Commons License This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License