Predicting the Severity of Lockdown-Induced Psychiatric Symptoms with Machine Learning

Area of Science:

• Psychiatry
• Machine Learning
• Public Health

Background:

• The COVID-19 pandemic and associated lockdowns exacerbated psychiatric disorders, increasing anxiety and depression.
• Early identification of at-risk individuals is crucial for preventing symptom worsening and optimizing healthcare resources.

Purpose of the Study:

• To identify predictors of psychiatric symptom severity during the COVID-19 lockdown using a supervised machine learning approach.
• To develop a predictive model for psychiatric prognosis in individuals facing large-scale restrictive measures.

Main Methods:

• A supervised machine learning methodology was applied to a case study.
• The study analyzed a small sample of healthy individuals, obsessive-compulsive disorder patients, and adjustment disorder patients.
• Predictors were identified using demographic and clinical characteristics collected pre-pandemic.

Main Results:

• Machine learning models achieved up to 92% accuracy in predicting depression, anxiety, and obsessive-compulsive symptoms during lockdown.
• The models successfully identified key demographic and clinical predictors of symptom severity.

Conclusions:

• The developed methodology can accurately predict psychiatric symptom severity during large-scale lockdowns.
• This approach supports clinical decision-making and the development of targeted public health policies for vulnerable populations.