Navigating the Complexities of Insurance Claims in Behavioral Healthcare
In North Carolina, behavioral healthcare providers are only compensated for services provided to Medicaid recipients and uninsured indigent populations after multiple third-party institutions, Local Management Entities/Managed Care Organizations (LME/MCO), approve the service encounter. I designed and implemented a tool that allowed claims data specialists to more efficiently remediate denials, resulting in approximately $800,000 in recoverable funds being identified for FY 2022.
The Life of an Insurance Claim
There are multiple failure points that contribute to lost revenue. Some stem from process-related mistakes by providers while others may result from LME/MCOs erroneously denying valid claims. Still others are irrecoverable due to higher level systemic changes, such as the appointment of a new official to a relevant regulatory institution. In the context of this project, we focused on LME/MCO denials of proactively funded services with a high likelihood of remediation once appealed.
Of the 4000+ denial codes LME/MCOs are required to provide, a small portion disproportionately represent the opportunity to recover lost revenue. The existing claim-submission process revolved around claims data specialists managing denial codes manually via .xls spreadsheets, so the first step towards a more efficient system was to automate the connection between related claims (those which shared a common client) and visualizing the history of a claim which had been repeatedly denied or rejected due to technical hiccups. A simplified UI followed, reducing the data-burden on claims specialists.
Next, we meticulously mapped out the logic for the top 3-5% of denial codes and implemented queries that ruled out clerical and administrative errors (misspelled family name, incorrect birth year, etc.) and filtered out denial codes where appeals would only be possible if such an error existed — client not enrolled in Medicaid, service not billable for age group, etc. Below is one such piece of logic I wrote to retroactively analyze clients’ participation in a program relative to a prevalent denial code. Italicized text describes redacted table and column names.
Once the scope of denials was narrowed, we were able to identify large-scale patterns among erroneous denials and take action to work with LME/MCOs to identify errors in either their claims processing system, the formatting of data we provided, or third party errors in the software essentially used to track NC Medicaid consumers, called NCTRACKS. Doing so resulted in the immediate identification of approximately $800,000 worth of denied claims that we could target and, ideally, translate to revenue.
