.png){kind=logo}
Poor medication adherence causes 125k deaths, 10% hospital admissions, costs $300B yearly. AI can boost adherence, reduce risks.
Poor medication adherence is a major challenge in chronic disease care, with approximately 50% to 70% of patients affected. Although there are written prescriptions, 20% remain unfilled, and even when they are filled, only half are taken as directed (1). This non-compliance is costly, both in human and economic terms, contributing to approximately 125,000 deaths and 10% of hospitalizations annually, resulting in up to $300 billion in costs (2-4). Studies reveal that adherence rates vary by condition, with cancer patients demonstrating the highest levels at 80%, while other chronic diseases have approximately 75% adherence (4).
Addressing adherence can lead to substantial healthcare savings. For example, every dollar spent on prescription medications for certain commercial populations results in a $3 to 10 decrease in medical costs (5), and high adherence correlates with 8 to 26 percent fewer hospitalizations and 3 to 10 percent fewer hospitalizations. 12% fewer emergency room visits. (6). Medication adherence is multifaceted and influenced by various factors specific to each patient, provider, medication, and medical condition (7,8), indicating the need for personalized strategies to improve adherence rates and outcomes.
An AI-based model known as “MediComply AI” has been created to anticipate medication adherence levels. This model evaluates various clinical and demographic factors to detect patients who are more likely to deviate from their treatment plans, allowing healthcare professionals to implement targeted and timely interventions.
The synthetic database for the model emulates real-world conditions and was created with insights from a range of medication adherence literature, including analyzes by Brown and Bussell (2), medication adherence impact studies by NEHI (3), cost and use assessments by Roebuck et al. (4)(5), risk assessments related to cost-related nonadherence by Briesacher et al. (6), and broader reviews of adherence intervention strategies by Viswanathan et al. (8) and Conn et al. (9). These sources guide the range and dynamics of the variables used to predict adherence, ensuring the accuracy and relevance of the model.
Poor medication adherence is a major challenge in chronic disease care, with approximately 50% to 70% of patients affected. Although there are written prescriptions, 20% remain unfilled, and even when they are filled, only half are taken as directed (1). This non-compliance is costly, both in human and economic terms, contributing to approximately 125,000 deaths and 10% of hospitalizations annually, resulting in up to $300 billion in costs (2-4). Studies reveal that adherence rates vary by condition, with cancer patients demonstrating the highest levels at 80%, while other chronic diseases have approximately 75% adherence (4).
Addressing adherence can lead to substantial healthcare savings. For example, every dollar spent on prescription medications for certain commercial populations results in a $3 to 10 decrease in medical costs (5), and high adherence correlates with 8 to 26 percent fewer hospitalizations and 3 to 10 percent fewer hospitalizations. 12% fewer emergency room visits. (6). Medication adherence is multifaceted and influenced by various factors specific to each patient, provider, medication, and medical condition (7,8), indicating the need for personalized strategies to improve adherence rates and outcomes.
An AI-based model known as “MediComply AI” has been created to anticipate medication adherence levels. This model evaluates various clinical and demographic factors to detect patients who are more likely to deviate from their treatment plans, allowing healthcare professionals to implement targeted and timely interventions.
The synthetic database for the model emulates real-world conditions and was created with insights from a range of medication adherence literature, including analyzes by Brown and Bussell (2), medication adherence impact studies by NEHI (3), cost and use assessments by Roebuck et al. (4)(5), risk assessments related to cost-related nonadherence by Briesacher et al. (6), and broader reviews of adherence intervention strategies by Viswanathan et al. (8) and Conn et al. (9). These sources guide the range and dynamics of the variables used to predict adherence, ensuring the accuracy and relevance of the model.
