Sofware Driven Analysis

At the core of MARA's programs is a complex, clinically derived set of data mining rules. The rules have been developed by MARA physicians, and span hundreds of diagnoses. The rules are designed to identify patient charts where HCC diagnoses have been overlooked or not documented according to ICD-9 guidelines.

The software incorporates diagnostic, procedure, and visit data from medical claims and encounter systems, prescription drug data, and CMS data. Pharmacology reference tables developed by PhD pharmacologists identifies over 100,000 drug types, and assures virtually complete recognition of prescribed drugs. The software filters out redundant diagnoses and those already submitted to CMS.

The rules-based software programs, HCCFIND and MDINFORM, now in their 4th generation, identify the relative likelihood or "signal strength" of finding the diagnoses in the medical chart. The software prioritizes members according to likely yield. Both programs have been tested against MARA's extensive clinical database.

HCCFIND identifies on average 40% of members' charts for retrospective review; MDINFORM is designed for prospective review, and identifies 50% of the charts. In recent studies, the software-predicted diagnoses have been verified through manual review in over two-thirds of the charts reviewed.

This outstanding predictive modeling result yields higher revenue, lower costs, and significantly higher return on investment for MARA's clients.