A diagnosis means everything for a rare disease patient. However, to process 200 patient records manually, it would take the most dedicated physician approximately 40 hours. That’s one full work week, assuming they’re doing absolutely nothing else. A new algorithm developed by Stanford researchers can process that same number of patient records in merely 10 minutes.
For conditions such as Cystic Fibrosis, Sickle Cell Anemia, and Huntington’s Disease, an earlier diagnosis can make a difference for patients. This new technology has the potential to improve the speed of diagnosis.
So what exactly is it?
The algorithm was developed by researchers Gill Bejerano, Cole Deisserotha, and Johannes Birgmeier. It examines observable traits, or the patients key phenotypes, listed within their medical record. It condenses long sentences into short phrases, sorting out the most important factors of a patient’s profile. These are then converted into codes used within the Human Phenotype Ontology, a phenotype database.
The algorithm ensures that the codes documented are solely for the individual patient. That is to say, if the patient has family history documented in their medical record, this information will not be included as part of their own traits.
The codes are sorted with the most frequent phenotypes and the ones documented at the earliest date listed first.
This algorithm’s name? CliPhen. The name combines “clinical” and “phenotypes.”
How we know CliPhen is accurate
Many people are apprehensive about leaving important things up to technology. However, CliPhen’s accuracy and effectiveness has been carefully studied, and its implementation could significantly improve patient lives. The shift toward automated systems in healthcare has been slow but Bejerano explains that –
“With 60 million patients to be sequenced in the next several years, we simply have no choice.”
CliPhen was evaluated using six different sets of patient records coming from four different medical centers. Its results were then compared to various other phenotype extraction algorithms. In comparison, CliPhen was more precise. In addition, it was 20 times faster than its counterparts. It’s actually thought to be more accurate than a physician.
The researchers who developed CliPhen hope that it will be implemented within clinical settings soon.
You can read more about this new algorithm and the researchers who created it here.
