Improved algorithm detects thyroid disease
Researchers in India report developing a more accurate algorithm to screen patient data for thyroid disease than previous versions, according to a study to be published in the International Journal of Computational Science and Engineering.
The algorithm, from the PSNA College of Engineering and Technology in Tamilnadu, India, produced a 93.5 accuracy rate, compared with earlier tests with a 92 percent or less accuracy rate, according to a related announcement. That translates to 15 patients per 1,000 diagnosed correctly, which could mean significant numbers across a healthcare network without requiring blood tests, authors Jaganathan Palanichamy and Rajkumar Nallamuthu wrote.
Thyroid disease--either over- or under-active thyroid--is often undiagnosed and untreated until it causes further problems, they said. The algorithm could help doctors determine whether patients who present a range of symptoms should undergo blood tests and subsequent pharmaceutical intervention.
Computer-aided detection (CAD) has been used in breast-cancer screening, though reviews have been mixed on its effectiveness. A study published in March in the American Journal of Roentgenology found CAD improved the likelihood for radiologists to identify cancer that initially went undetected during a screening. That contrasts with work published in Journal of the National Cancer Institute saying that CAD subjects patients to unnecessary tests.
Last month the Food and Drug Administration released guidelines on the regulation of CAD technology.
Statistical modeling continues to grow more prevalent in healthcare. University of Washington researchers, for example, have created an algorithm that can predict patients' future ailments based on their history and that of people with similar conditions.
FDA publishes final computer-assisted detection technology guidance
Computer-assisted detection aids cancer testing, treatment
CAD use for breast cancer patients leads to unnecessary tests
Algorithm looks to past to predict future health conditions