Mathematical Model Fine-Tunes Disease Management

Researchers at Boston University School of Medicine (BUSM; MA, USA) evaluated multiple paired sets of patient data, rather than a single pair, and found a substantial decrease in the statistical difference between values necessary to identify a physiologic change within a given confidence interval, thereby improving the sensitivity of the evaluation. While the standard test for change compares the difference between 2 points with the 95% confidence limit, given as ±1.96·(2)^½·SD, the researchers found that the effect of increasing the number of multiple data pairs on the confidence limit by using the formula 1.96·(2/n)^½·SD, (where n = number of data pairs), significantly reduces the difference between values necessary to achieve a 95% confidence limit. According to the study, this approach has the potential to contribute to the personalization of health care delivery by redefining common thresholds in clinical patient management. The study was published in the January 2010 issue of Archives of Pathology & Laboratory Medicine.

"The decision to treat a patient with a more aggressive therapeutic regimen may be reconsidered when a test can show that a small but highly statistically significant change has taken place,” said study author Martin Kroll, M.D, a professor of pathology and laboratory medicine at BUSM. "Such a result may lead to the postponement of switching to a new therapeutic regimen until the next testing time, where the initial small but significant change in test values has grown to be both statistically and clinically relevant.”

Related Links:
Boston University School of Medicine