VizHealth Methods and Procedures
Related Work and Ideation
Something that we found surprising when beginning this project was the lack of existing tools to compare plans and healthcare. No visual or interactive tools were available to the general public and plan information was often strictly limited to federal and state websites. In order to compare plans, users are required to enter in highly personal information proving you lived in a certain state. In addition, users are required to create accounts using an email address and certain sites even required social security information and personal addresses.
Even after completing all these additional steps, the plans were laid out by text and generally one or two at a time. There were also no methods of comparison between plans at a high level. You could organize them based on filters such as price and coverage, but with how difficult it was to even see the plans in the first place, our feeling was that many users would feel intimidated and reluctant to proceed. In other words, because the initial forms were so long and felt like a breach of privacy, users would not feel comfortable with even gaining access to the traditional marketplace websites.
Personas
Moving to our personas, the intended users were 18-30 years old and low to middle class.
This demographic was chosen based on three primary reasons.
Most young adults start college at the age of 18, and may move to a state with a significantly different healthcare system than their home state, requiring them to get an insurance with better coverage in their particular state (compared to their parent's coverage).
Many young professionals are completely reliant on their employer for health insurance, and if they are fired or need to leave their job for any particular reason, they will be without insurance.
We target low to middle class as upper class young adults often rely on their parents for funding their insurance and have little personal involvement in the process.
Data Sources
The data of VizHealth was acquired from the Centers for Medicare and Medicaid Services and contains two primary important limitations.
First, the data has comparisons between states that are reliant on the federal marketplace, and does not compare states that have their own state-run marketplace.
Second, this dataset is only for the year 2020. Besides the above primary data source, we also included a dataset from US News from 2019, which is the same year that informed the insurance plans of 2020. It contains data comparing states in the quality of care, healthcare access, and public health. The dataset is accessible to the public online with CC0 1.0 Universal (CC0 1.0) Public Domain Dedication License
Data Cleaning
The original dataset contained over 11 gb of data and more than 25 million rows. We then narrowed the fields down to 20 dimensions that we thought were important for our users. To serve our target users, we first decided to only look at plans for ages 18-30, and that are considered bronze or expanded bronze level plans.
Our reasoning for only looking at these two metal levels was that low to middle income young adults would be looking for the most inexpensive plans, and would likely not have the means to pay much extra for silver or gold level plans. By adding these two filters into the cleaning process, we were able to narrow the rows down to roughly 200,000.
(Under Construction)