Novel algorithm for predicting adverse drug-drug interaction

Abstract

Molecules are the fundamental units of any chemical compound. A drug is a predefined composition of chemical compounds which is expected to produce a target biological behavior. Drug error is a term used to indicate a scenario where one drug increases or decreases the effect of another drug when both are present inside the human body. This may be either due to pharmacodynamics (effect or action of a drug) or pharmacokinetics (movement of a drug within the body) or both. The results are painful as a person can suffer from serious disability or health hazard. The problem is mainly due to human error while prescribing and lack of efficient systems that can aid in automatic diagnosis of interacting drugs. Most of the issues that a patient undergoes or becomes a victim due to polypharmacy can be avoided and prevented. What we lack is an intelligent system that can predict the adverse interaction of poly drugs and raise an alarm to the medical team. This can be a live saver. This paper attempts to propose a model based on machine learning that can predict adverse drug-drug interaction.