As of April 2019, health expenses account for 17.9% of the U.S. GDP, the highest of developed countries in the world. Medical costs in the U.S. are the number 1 cause of bankruptcy in U.S. and a federal study conducted in December 2018 shows that 40% of American households can’t afford a $400 financial emergency.
GoFundMe was founded in 2010 as a way to help people raise money for weddings and vacations. In 2019 GoFundMe claims to be the number one medical fundrasing platform with over 250,000 medical campaigns raising more than $650 million dollars.
Using a hybrid approach of structured and unstructured learning, and the philospohy of human-centered design, the data from campaigns dating back to the last election in November 2018 were collected and analyzed using k-means clustering to group the campaigns and tf-idf to find the most relevant unique terms within the corpus and each group. Continue scrolling to reveal the top terms and hover to see the group associations and distributions on the node chart.
Not surprisingly help is number one. Interestingly the non-medical terms work and family had similar rankings to the more medical terms of surgery and cancer. The node colors represent the unique groups that each campaign was assigned in relation to the structured k-means learning, although a larger corpus of data could reveal new trends.
Exploring the campiagns by group shows us patterns of needs as well as who is contributing to what, and maybe more importanly, which campaigns are struggling the most.
Since 2013 average national salaries have stagnated while the out of-pocket costs per individual with employer supported healthcare has risen 17.9%, or around $5,893. For those with a chronic disease? The costs have risen $20,000 on average per individual. Hover over the graphs below to see which campaigns, sorted by topics and terms, raised the most money in the shortest amount of time (success rate), and which had the highest rate of people donating who interacted with the campaign (engagement rate).
The average success rate, amount of goal raised, across all groups is 40.9% and the average engagement rate, the percent of people who donate to a campiagn out of all the users who interact in an explicit way, via sharing or liking or donating.
The words below indicate highest ranking per group, but does not represent exclusivity and each groups top terms will likely show up in the other group's campaigns contents. The group labels were ranked higher by the strcutured learning algorithm and therefore indicate higher significance to campaign.
work
help
service
family
expenses
cancer
treatments
hospital
surgery
kidney
fire
home
family
lost
accident
Digital literacy should not be a prerequisite to accessing affordable healthcare, but with more than 1 in 3 campaigns on the GoFundMe platform being for Medical expenses of all kind, it is slowly becoming one. This is not a sustainable solution for a healthcare substitute, and many are being left out of the conversation particularly due the the stigma of crowdfunding, people with diseases perceived to be “choices” or “blameworthy” diseases like mental health and addiction or “getting my life together as a single mom with a sick kid”, as compared to “faultless” diseases such as cancer or a genetic disorder.
The social element also cannot be ignored, we can in the charts above that the campaigns with higher engagement yield higher success rate, so those without a build in network or community to help promote their campaign are more likely to suffer. “If there’s already a hierarchy of affliction and need in society, then crowdfunding often works to exacerbate it.”
Submitted in partial fulfillment of the requirements for the degree of Master Science in Data Visualization at Parsons School of Design
This project aims to reveal new data and meaningful insights that can be used by healthcare policymakers and civic-minded innovators to create a more integrative healthcare system.