Institutional providers, such as inpatient facilities, are required to report clinical trial information through the Claim Value Code and Claim Value Amount variables. However, Claim Value Codes were created to report monetary conditions related to claims. This means that Claim Value Amount is formatted as dollars and cents. When analyzing CMS data for the Identifying Clinical Trials using CMS FFS Claims Data article, this accounting-style formatting created unexpected challenges when validating and identifying clinical trial numbers.

When working with clinical trial numbers in Claim Value Amount, odd patterns appear. A single clinical trial number could present itself in different ways. For example, if a clinical trial number is 1234560, one might find 12345.60, 1234560.00, or 123456.00. At the time of writing, there are no instructions on how to manage clinical trial numbers with the accounting-style format, so it appears that interpretation plays a major role in how clinical trial numbers are reported. This poses challenges when trying to pull out a valid list of clinical trial numbers.

For the article, I ultimately decided to pivot away from quantifying valid clinical trial numbers and, instead, quantify invalid clinical trial numbers. I was able to find a few invalid values were either zero or 9-filled (Invalid value list: 0, 999,999.99, and 9,999,999). With the seemingly infinite variations of clinical trial numbers, I can’t say with confidence that I captured every invalid clinical trial number. These were the ones I could find through trial and error.

Most researchers are looking into a specific set of clinical trial numbers. Knowing this, I can’t quite justify the time and effort it would take to validate every clinical trial number in a single year of claims data, much less the entirety of claims data. However, I do think that approaching these data with a specific set of clinical trial numbers in mind and knowing the formatting concerns makes the task easier to approach.