.png){kind=logo}
Predict drug adverse effects with Artificial Intelligence, increasing patient health and satisfaction.
Adverse drug reactions (ADR) account for 5% to 10% of all hospital admissions in developed countries [1].
In Spain, ADRs are responsible for approximately 3% of hospitalizations and affect 20% of patients undergoing polypharmacological treatments [2].
However, many patients at risk remain unidentified, leading to severe complications, high costs, and increased mortality.
A predictive AI model could identify patients at high risk of ADRs by analyzing demographic data (age, sex) and clinical history (comorbidities, treatment history).
By using this data, the model could classify risk as low, medium, or high.
The model dataset was structured to reflect clinical realities, based on research on ADRs in older adults by Nair et al., analysis of spontaneous reports by Dubrall et al., impact studies of medication continuation by Weir et al., Pharmacogenetic risk factor piloting by Finkelstein et al., medication appropriateness reviews by Fick, and intervention meta-analyses by Gray et al. Using variables such as age, liver and kidney function, and number of medications, the model simulates patient profiles to help healthcare providers personalize treatment to reduce ADR risks.
Adverse drug reactions (ADR) account for 5% to 10% of all hospital admissions in developed countries [1].
In Spain, ADRs are responsible for approximately 3% of hospitalizations and affect 20% of patients undergoing polypharmacological treatments [2].
However, many patients at risk remain unidentified, leading to severe complications, high costs, and increased mortality.
A predictive AI model could identify patients at high risk of ADRs by analyzing demographic data (age, sex) and clinical history (comorbidities, treatment history).
By using this data, the model could classify risk as low, medium, or high.
A recent study showed that predictive models can reduce severe ADR incidents by up to 30% by adjusting treatments proactively [5].
Without a predictive model, ADRs that are not identified early lead to additional hospitalizations, prolonged treatments, and complications, increasing costs. A model like the one proposed could reduce these costs by 20-25% [2], avoiding unnecessary hospital admissions and improving healthcare system efficiency.
Implementing this AI model would not only enhance patient safety but also reduce preventable deaths related to ADRs. It is estimated that adopting predictive systems could reduce healthcare costs by 5% annually, representing a significant saving for both Spanish and global healthcare systems [1,2].
The model dataset was structured to reflect clinical realities, based on research on ADRs in older adults by Nair et al., analysis of spontaneous reports by Dubrall et al., impact studies of medication continuation by Weir et al., Pharmacogenetic risk factor piloting by Finkelstein et al., medication appropriateness reviews by Fick, and intervention meta-analyses by Gray et al. Using variables such as age, liver and kidney function, and number of medications, the model simulates patient profiles to help healthcare providers personalize treatment to reduce ADR risks.
