Quantum lattice learning and explainable artificial intelligence for maternal and child healthcare

Abstract

The current approach to maternal and child healthcare is extremely patient-centred, it requires costly, risky surveillance and testing before diagnosis besides treatment accompanied with uncertainty despite the essential combination of healthcare expertise, skills and experience in medical care and public health for medical practitioners to support maternal and child health. With the recent maternity and prenatal engagement besides the availability of health data and information, we interpretably revolutionize advances in maternal medicine by turning massive amounts of data into proactive, predictive, preventive, personalized and participatory optimal treatment plans through predictive and preventive medicine for maternal and child well being. This work focuses on interpretable predictive and Machine Learning (ML) modelling of Artificial Intelligence (AI) algorithms to be used in predictive analytics of health data for maternal precision medicine and explainable preventive insights for physicians and patients’ medical decision making. We also introduced the concept of Quantum Lattice Learning for building Explainable Machine Learning models in Quantum Space. Due to the uncertainty caused by abstracted black-box AI and ML models (algorithms) used to support the maternal-child medical decisions, there is ambiguity of safety and trust of all the existing and proposed AI models. That hinders reliability and trust in adoption of the developed models by physicians and patients. We, therefore, implemented Explainable Artificial Intelligence (XAI) and feature interpretability analysis to allow clinicians like obstetricians, perinatologists, gynecologists and midwives to understandably trust, comprehensively assess connections and transparently analyze and use the important derived features for strategic maternal and child predictive, preventive and precision medicine. The adoption of the proposed XAI approaches (models) on health data usage could potentially strengthen health systems, public health, primary and surgical care for mothers and children globally. They can significantly improve accountability, reliability and adoption of safe and trusted artificial intelligence applications for improved maternal-fetal medicine besides global health. Moreover, our transparent models provide useful insights for healthcare management and policy-making to improve the health and well-being of patients and physicians.