Algorithms designed to detect potentially transmissible infectious diseases in patients in the emergency room were effective in identifying possible diseases, according to a study from France. But false positives grew as the algorithms increased in specificity.
Because patients coming in to the hospital can spread diseases, it's important for clinicians to spot these potential problems early, according to research published at BMC Medical Informatics and Decision Making.
This system was designed to evaluate both structured and unstructured data in computerized emergency department records in real time. The researchers developed algorithms for three types of illnesses: infectious respiratory, which included pneumonia, influenza and influenza-like illnesses; cutaneous, which included measles, rubella, scabies and the like; and gastrointestinal syndromes, including viral gastroenteritis and diarrhea due to bacterial infection.
At a base 75 percent sensitivity rate, the algorithms for infectious respiratory and cutaneous illnesses were found to be acceptable, striking a reasonable balance between true positives and false positives. However, the number of false positives for the algorithm on gastrointestinal diseases prompted the researchers to say it was not specific enough for routine use.
Hospitals are using an array of technologies to thwart the spread of dangerous, preventable infections. One of the most interesting is a robot that that uses flashes of ultraviolet light to sterilize and kill germs.
While one of the lowest-tech programs to increase hand-washing among staff is to encourage patients to ask about it, nurses at the Toronto Rehabilitation Institute wore badges that buzzed when they failed to wash their hands.
Meanwhile, a hand-washing compliance system based on touchscreen technology can determine whether a staff member has clean hands before physical contact with a patient.
To learn more:
- read the research