Verification & validation algorithms for data used in critical care decision support systems

A decision support system is only as good as the data generating that decision support system. If the data is incorrect, doesn't relate to the other pieces of data, is missing or is not consistent, the decision support system conclusions may be incorrect and inconsistent. While collecting data from several sites during a multicenter randomized clinical trial, we found that some critical data elements were missing, out of correct ranges, totally illogical, and/or inconsistently recorded. In order to get consistent, correct, and dependable information from a our decision support system, the data elements used in that system had to be checked for completeness, valid values, consistent units of measurement, and relationships to other items. Development of data quality assurance rules and the application of those rules is imperative to using the data to generate daily scores for multiple organ failure, sepsis, and barotrauma.