Project Name: Substance Use Disorder Claims in the CO APCD
Project Purpose: We wanted to know how many Substance Use Disorder (SUD) Claims we had in the Colorado All Payer Claims Database (CO APCD) data warehouse, and how we could separate them out from analyses.
What were your first steps when beginning this analysis?
First it was imperative to determine what codes fell into the SUD category. We obtained lists of SUD codes from the Centers for Medicaid & Medicare Services (CMS) and ended up with a list of around 980 codes that we determined to be SUD-related. Then we created a manual ‘flag’ using Standard Query Language (SQL) to separate claims with these codes from the rest of the data warehouse.
Were there specific considerations you needed to consider based on the data?
SUD data is highly protected and regulated. It is governed by the Health Insurance Portability and Accountability Act (HIPAA), as well as by rules in part two of item 42 in the Code of Federal Regulations. Historically, SUD data could only be shared in very specific circumstances. However, the Substance Abuse and Mental Health Services Administration (SAMHSA) has recently begun expanding the ways that SUD data can be used and while CIVHC worked to understand the new rules, we created the flag to identify these claims. We will continue to share as CIVHC learns more about SUD claim utilization.
What challenges did you encounter while performing the analysis? How did you overcome them?
The biggest challenge during this project was generating the ‘flag’. We created a true/false flag that would be manually added to all analyses that would make sure SUD data wasn’t being pulled into a report. We also had to start fresh with this data and learn all about what SUD data we had.
Without delving into results, did anything surprise you about this analysis or the process of executing it?
CIVHC definitely learned a lot through this process. After the first manual flag was created, we decided to automate the process instead. Moving the process to an automatic one was helpful to our analysts and made everything more streamlined. I’ve been surprised at how confusing SUD claims can be to understand, and how they can connect to so many different projects (Behavioral Health, Emergency Department Use, etc.).
What did you learn while performing this analysis?
I was able to work on my SQL skills with this project while making the first iteration of the manual flag. I learned a ton about what is included in SUD claims and how CIVHC can make sure that our reporting is as accurate and meaningful as possible. I am excited for the future of SUD data use and what we could potentially do with this data on an aggregate level.