[ad_1]
Researchers from the University of Copenhagen and Rigshospitalet have used data from more than 230,000 intensive care patients to develop a new algorithm. In particular, he uses the disease history of the last 23 years to predict the chances of survival of patients in intensive care units.
Every year, tens of thousands of patients are admitted to intensive care units across Denmark. Determining which treatment is best for each patient is a challenge. To make this decision, doctors and nurses use various methods to try to predict the patient's chances of survival and mortality. However, existing methods can be significantly improved.
Therefore, researchers from the Faculty of Health and Medical Sciences of the University of Copenhagen and Rigshospitalet have come up with a new algorithm that predicts much more accurately the chances of survival of a patient in care intensive. Their research was published in the scientific journal Lancet Digital Health.
We used Danish health data in a new way, using an algorithm to badyze the file data from the patient's disease history. The Danish National Register of Patients contains data on the history of the disease of millions of Danes and, in principle, the algorithm is able to take advantage of the history of the citizen who benefited the patient in relation to treatment .
Professor Søren Brunak from the Protein Research Center of the Novo Nordisk Foundation
Analyze 23 years of the history of the disease
In developing the algorithm, researchers used data from more than 230,000 patients admitted to intensive care units in Denmark during the period 2004-2016. In the study, the algorithm badyzed the history of the patient's illness over a period of up to 23 years.
At the same time, they included in their calculations the measurements and tests performed during the first 24 hours of the admission in question. The result was a much more accurate prediction of the patient's mortality risk than that offered by existing methods.
"Excessive treatment is a serious risk for terminally ill patients treated in Danish intensive care units. Doctors and nurses have not found a support tool that can tell them who will receive intensive care. With these results, we have taken an important step to test such tools and directly improve the treatment of the sickest patients, "said Professor Anders Perner of Rigshospitalet's Department of Clinical Medicine and Critical Care Department.
Significant with respect to death and survival
The algorithm predicts three predictions: the risk of death of the patient at the hospital (which can be any number of days after admission), within 30 days of admission and in the 90 days after admission.
For example, researchers could say that diagnoses of less than 10 years influenced the forecast and that young age reduced the risk of death, even when other values were critical, while aging increased the risk of death. . The algorithm is not just a useful tool in daily practice in intensive care units across the country. This can also tell us what are the important factors when it comes to the death or survival of a person.
"We train the algorithm to remind us which previous diagnoses had the greatest impact on the patient's chances of survival. It does not matter that they are one, five or ten years old. This is possible when we also have data on actual admission, such as heart rate or blood test responses. By badyzing the method, we are able to understand the importance it attaches to the different parameters of death and survival, "says Søren Brunak.
Researchers at the origin of the study hope to be able to use the algorithm in clinical tests within a few years. At the same time, the next step is to try to further develop the algorithm, making it able to make predictions at the time.
Source:
University of Copenhagen Faculty of Health and Medical Sciences
Journal reference:
Nielsen, A.B. et al. (2019) Prediction of survival in intensive care units based on the synthesis of long-term disease history and acute physiology: retrospective study of the Danish National Register of Patients and Electronic Patient Records. The Lancet Digital Health. doi.org/10.1016/S2589-7500(19)30024-X.
[ad_2]
Source link
