CoSort Combines Data Transformation and Security for AIM Healthcare

REVIEWER: Wesley Bowden, development team lead, Data Transformation for AIM Healthcare Services, Inc.

BACKGROUND: AIM Healthcare Services, Inc. is a leading provider of claims-cost management services. The company’s mission is to reduce the cost of health care and simplify health care information management. Since AIM’s inception, the company has grown to represent more than 1,800 hospital and health system clients and more than 1,100 payer clients, including the largest indemnity, HMO, Medicaid and self-insured payers in the country. With more than 1,700 team members working in client facilities in all 50 states, AIM identifies, validates and reconciles pre- and post-pay claim and account data errors to deliver more value on a daily basis than any other company in this sector. AIM’s Data Transformation Team (DTT) handles extraction, transformation and loading (ETL) of incoming client data. DTT works with large data extracts in many formats, including COBOL, fixed-width, variable-length, EBCDIC, ASCII and EDI. Translated data is loaded into a database where analysts perform ad hoc and standard queries.

PLATFORMS: Microsoft Windows Server 2003, Microsoft SQL Server 2005.

PROBLEM SOLVED: The wide range of incoming data formats presented a challenge for AIM’s DTT developers. Using Microsoft Visual Studio 2005, the team created a proprietary C# program to manipulate raw client data into flat files that could be processed and organized for direct loading into AIM’s database. A third-party sorting application organized these large files to make manipulation manageable. Given the inherent complexity of the code base, however, this approach proved to be inefficient early on. A second challenge for DTT was obfuscating HIPAA-protected information within client data. Where appropriate, AIM utilized outsourcing to increase its capacity. The ETL development process involved contact with clients’ data extracts, so obfuscating the embedded HIPAA data reliably was critical. The CoSort product suite solved both of these problems. CoSort natively handles every file format and data type that AIM touches. With minimal coding, it also handles sorting, merging and joining data - all of which previously required thousands of lines of code and great effort to achieve. Furthermore, CoSort enables AIM to obfuscate embedded HIPAA-protected data in client extracts without disturbing file structures. As a result, AIM can provide outsourced developers real-to-life data formats without compromising HIPAA standards.

PRODUCT FUNCTIONALITY: CoSort is easy to use and supports many of the native file formats found in today’s ETL world. Its job scripts are easy to prepare, understand and debug. Because CoSort has already done the work, developers can concentrate on the fundamentals instead of the lines of code required to support those fundamentals. The application’s sort function works quickly and, in one-on-one tests, reduced AIM’s processing time dramatically. DTT members are able to call CoSort from within AIM’s managed code, as well as create custom managed functions that CoSort can consume while running. These achievements allow DTT developers to continue working in a language familiar to them and redeploy the team’s current functions. CoSort’s data privacy functions are impressive; they enable AIM to mask confidential data at the field level without disturbing the file layout or other nonconfidential data. Given this functionality, DTT developers can work and code against the exact file format that will run in production.

STRENGTHS: CoSort enables AIM to reduce complexity in the code base, thus preventing errors and cutting the man-hours required for client implementations. Sorting is lightning fast, and the merging, joining and summing features enable AIM to reduce the number of passes over a file. AIM’s DTT has tested multiple solutions for data obfuscation, but CoSort is the only application that has succeeded at both file and field levels.

WEAKNESSES: Implementation of CoSort requires individual licensing for all of DTT’s developer workstations, and the application currently does not support direct interactions with databases without custom user-exit routines.

SELECTION CRITERIA: AIM sought a reliable tool that would quickly sort and compare files. CoSort met this criterion, brought merge and join capabilities, and offered protection of sensitive field data. This combination of functions in a single package made CoSort a clear winner for AIM.

VENDOR SUPPORT: CoSort support has been excellent. IRI consistently responds to DTT’s requests and has supplied needed examples.

DOCUMENTATION: CoSort’s PDF manual is easy to read and understand.