Personalized statin treatment plan using counterfactual approach with multi-objective optimization over benefits and...

Area of Science:

• Cardiovascular medicine
• Pharmacology
• Biostatistics

Background:

• Statins effectively reduce cardiovascular disease (ASCVD) risk but can cause adverse effects, leading to therapy discontinuation and increased mortality.
• Existing personalized statin treatment plan (PSTP) frameworks have limitations including lack of counterfactual predictions, inadequate handling of multi-objective optimization, and retrospective evaluation.
• Addressing these limitations is crucial for improving patient adherence and cardiovascular outcomes.

Purpose of the Study:

• To enhance the robustness and usability of the PSTP framework for statin prescription.
• To develop a proactive strategy that maximizes low-density lipoprotein cholesterol (LDL-C) reduction while minimizing statin-associated symptoms (SAS) and therapy discontinuation.
• To incorporate counterfactual predictions, multi-objective optimization, and robust evaluation methods into the PSTP.

Main Methods:

• Applied overlapping weighting counterfactual survival risk prediction (CP) to account for confounding bias.
• Utilized multiple objective optimization (MOO) to balance competing prescribing objectives, such as benefits versus risks.
• Employed clinical trial simulation (CTS) with various arms (Random, Clinical Guideline, PSTP, Practical) for comprehensive framework evaluation.

Main Results:

• The revised PSTP demonstrated significant improvements in lowering SAS risks across all time points compared to other arms in CTS.
• The PSTP exhibited enhanced flexibility in identifying optimal statin choices within one year.
• Clinical trial simulations confirmed the feasibility of robust counterfactual risk prediction and personalized benefit-risk balancing.

Conclusions:

• The enhanced PSTP framework provides a robust and trustworthy method for personalized statin prescription.
• The study successfully demonstrated the ability of PSTP with Pareto optimization to achieve an optimal balance between statin benefits and risks.
• The improved PSTP framework offers a promising solution to mitigate statin-associated risks and improve long-term cardiovascular health.