Coding terminology for regulatory submissions is a complex business, and for data management provider SigmaSoft International and Netrics, a maker of intelligent database record matching software, this spells opportunity. They have announced a partnership to produce an automated encoder/browser for adverse event and concomitant medication terminology based on standards put forth by the Medical Dictionary for Regulatory Activities (MedDRA) and the World Health Organization Drug Dictionary (WHODrug).

The driving force behind the autoencoder/browser is Netrics’ “fuzzy search” technology, which SigmaSoft CEO Steve Colville says is the best around. “Fuzzy search,” “fuzzy matching,” and “error-tolerant matching” all refer to technology that compensates for input typing errors as well as errors introduced by optical character recognition. Using the technology, a search returns a list of results based on likely matches.

“We looked at a lot of different search engines to build our new product in hopes of finding one that would ignore spelling mistakes and still deliver the correct medical and drug term. Netrics technology was unbelievable in terms of how well it could find and match terms that were badly misspelled,” Colville says. A simple example is the software’s ability to recognize “Tielinol” as Tylenol.

The ability to match misspelled terms is becoming increasingly important as more clinical trial data are coming from many countries, representing many languages and a variety of names for the same medical condition and adverse event.

But the SigmaSoft/Netrics autoencoder/browser has a broader functionality than matching and correcting spelling errors. The autoencoder can be used to code data for an entire data table all at once, sometimes returning several results, assigning weight to the most likely suspects. “This is a huge time saver,” says Colville. He explains that the free browser included with MedDRA subscriptions requires users to enter information one piece of data at a time, making for a slow and tedious process.

Netrics CEO Stef Damianakis provides further insight into the value of the autoencoder/browser. He explains that Netrics’ technology, like the clinical trial business, is narrowly focused on database records, and has the ability to match data among fields even if there are slight mistakes throughout. “Our approach is based on the fact that human beings are good at looking at data and matching them even if there are small discrepancies, so we worked toward mathematically modeling the human’s ability to see similarities,” says Damianakis. This creates a robust solution whereby the computer can compare data at the core level instead of relying on structured query language (SQL) exact matching techniques or spellchecking.

SigmaSoft also selected Netrics because of the pricing arrangement it was able to negotiate, a key consideration for serving its market of small to mid-sized biopharmaceutical companies and contract research organizations. According to Colville, this market segment is frequently shut out of costly regulatory solutions designed for the largest users, so his company was determined to serve its market by offering a top-notch 21 CFR Part 11-compliant search engine for thousands of dollars instead of hundreds of thousands. “One of the things we were able to work out with Netrics was a pricing policy that fit very well within the range that we wanted our pricing to be,” Colville explains.

Damianakis adds that the company was looking to enter the clinical trials market as a way to expand its core health care capabilities beyond insurance claims processing and hospital information systems. “Our core technology has the building blocks for this sector and we were looking for a partner,” says Damianakis

The autoencoder/browser is slated for release in December 2007.

-----------

This article first appears in Bio-IT World’s eCliniqua newsletter. Click here for a free subscription to eCliniqua.