Artificial intelligence to predict mental health crises

[banner-DFP_0]

The Covid-19 pandemic has had a noticeable impact on the mental health of a large part of the population, causing Increasing the number and severity of cases. This is why health systems are facing an increasing demand and need solutions that allow preventive measures to be taken.


In this context, the provider of digital mental health solutions, KoaHealthdesigned a file Machine learning algorithm can predict mental health crisis. This research – which was co-authored with the Birmingham and Solihull NHS Foundation Trust – highlights the usefulness of this technology as a potential tool for early intervention for patients, thereby improving their prognosis.[banner-DFP_1]


This machine learning model uses Anonymous electronic medical records To continuously monitor patients for mental health crisis risk patterns. Tests showed that this model More than half of crises anticipate 28 days in advance, without a significant rate of false positives. In addition, a subsequent six-month study evaluated the use of said algorithm in clinical practice, determining that in up to 64% of cases, the model proved to be of clinical value, both in case management and crisis risk reduction.


“This research demonstrates how collaboration between data scientists and healthcare professionals can lead to preventive care, having a significant impact on patients’ lives.”


The first author of this research published in the journal nature medicinePhD student in the Department of Information and Communication Technology Pompeu Fabra University, Roger Garrigaexplains that the electronic medical records of more than 17,000 mental health patients were analyzed “to develop a machine learning model that predicts the next mental health crisis in the following four weeks”.

See also  Iberdrola starts a media campaign for its customers to reduce consumption and save bills


Our algorithm is designed to provide Weekly updated risk score for all patients With a history of relapse to support clinical decisions with an adequate time window for clinicians to apply preventive measures. This research demonstrates how collaboration between data scientists and health professionals can lead to preventive care, which has a significant impact on patients’ lives,” says Garriga.[banner-DFP_4]


On his part, the CEO of Koa Health said, Oliver Harrison, remembers his career as an NHS psychiatrist. “I saw a lot of people coming into our services through the emergency room. By the time we saw them, they were mostly very sick and often had problems with their relationships, finances, employment and housing. Putting the pieces together has always been expensive, and time-consuming. And distressing for the people involved.”


“We believe this is the future of mental health care in the UK and around the world”


for this reason, Harrison celebrates participating in the study to check the algorithm’s usefulness. “We have achieved great results in this experiment and look forward to continuing development work. We believe this is the future of mental health care in the UK and around the world.”


To the thread, I+D Director of Koa Health, Alexander Maticadds that “ Prevention is the holy grail of cure of mental health problems”, which is why “using AI to predict mental health crises can allow timely, proactive, rather than reactive, interventions to mitigate or prevent mental health crises and truly help both patients and health systems.” “We were very excited when the results of the clinical application of our model suggested that our predictions It could have prevented more severe symptoms from developingMatic concludes.

See also  Portugal will reopen to British tourism from May 17th


The next step for researchers is Model validation in other health systems and larger-scale studies to obtain regulatory approval and the tool to become a part of the everyday life of mental health services.

Because health we all need… ConSalud.es

Leave a Reply

Your email address will not be published. Required fields are marked *