The Ask the Analyst series is a deep dive into the data by those most familiar with the CO APCD – the analysts themselves. We’ll hear about their experiences with recent analyses and answer any pressing questions that come up. Have a question for the CIVHC Analyst Team? Email it to firstname.lastname@example.org.
Featured Project: The Community Dashboard is a new publicly available report that helps communities understand how their health care cost, utilization, access and quality of care compares to the state, other counties and regions. This information can be used to help identify opportunities to improve care and lower cost, and track trends towards positive change.
Analyst Name: Ioana Crisan, Data Scientist on the Population Health Team at HSRI, a team responsible for managing the CO APCD Data Warehouse including data quality, data enhancements, measure development, building reports and visualizations
- What were your first steps when beginning this analysis?
I have been involved with the development of this report since 2017 (starting with the previous iteration). I started by developing the SQL code for the health care cost per person per year (PPPY) measures and creating the first prototype of the dashboard in Tableau.
For the current iteration, the first steps were to plan the development of access to care and quality of care measures. Soon after the development work for these measures took off, we met with the Public Reporting team at CIVHC to plan the full scope of work for the current iteration: selection of measures, the design changes requested for the dashboard, and determining the preliminary timeline for completion.
- Were there specific considerations you needed to consider based on the data?
As analysts, we are often too close to the data to have the perspective of those who are the end users – especially those who will browse the report directly on the website. It is challenging to keep things simple and exercise restraint from wanting to present all or most of the data points you’ve produced.
The CIVHC team proposed to simplify the design considerably for this iteration, which I think has improved the user experience for report viewers. Those who are inclined to dig further or in a different view can still access the full set of available data points in separate data files (contact email@example.com to download), but the dashboard is a clear overview into each topic area for the geography of interest.
One important component of developing this report or any analysis is to ensure that every step is thoroughly checked for errors. Given the many facets of the report, this work had to be carefully assessed by several team members at CIVHC and HSRI – more than once! I worked with each person to ensure that we have their complete feedback and then to address it. I am confident that the result is as accurate as possible because of this rigorous quality control process and due to everyone who reviewed various parts of the report.
- What challenges did you encounter while performing the analysis? How did you overcome them?
Some of the measure methodology specifications were somewhat challenging to implement with CO APCD data. In most cases, these challenges were addressed by collaborating with other HSRI and CIVHC team members and making decisions together based on similar analyses or based on what we agreed would be the best approach for this report.
- Without delving into results, did anything surprise you about this analysis or the process of executing it?
For this report, we worked with a very large share of the CO APCD database – we included almost everyone represented in the database for a total of seven calendar years’ worth of records! We performed measure calculations from the very simple to the very complex. I did not think much about this while performing these analyses, however thinking back at the massive amounts of data being processed, I think it all went relatively smoothly.
- What did you learn while performing this analysis?
I gained a lot of experience with APCD-based measure creation methodology during the development of this report. Also, I used multiple tools and expanded my knowledge of each tool (Vertica SQL, R, Tableau).