Clinical ambiguity in the intelligent machine era (treats breaks and discharges)

Modern medicine is at the nexus of two unhealthy megatrends-growing administrative cost waste and exploding big data. New technologies are often advanced as systemic solutions for these and other indurate complexities. But technology insertion is never neutral and often has unintended consequences. In this complexity-technology context, clinical ambiguity can emerge, contributing to poor patient outcomes, cost-inefficiencies, and medical errors. The challenges of healthcare system complexity and clinical ambiguity can be purposefully addressed by harnessing three related sciences-probability, data, and cognitive computing. Medical decision-making is an application of probability science to the interrelated processes of testing, diagnosis, and treatment. Decision-makers and healthcare systems seek to learn from available information to achieve faster, more accurate, and reproducible solutions to problems and ideally to avert them. Decision analysis is a mathematical approach helpful under circumstances of partial knowledge and diagnostic ambiguity, by showing decision-makers that a preferred treatment plan depends on knowledge, the care objective, and decision criteria. Data science wrangles and deconvolutes high-dimensional big datasets. Knowledge representation makes querying of diverse data structures by humans and/or machines possible, in order to intelligently model and effectively communicate solutions to complexity. Knowledge creation sorts data complexities in the published evidence so that healthcare providers can make credible assumptions and draw logical conclusions that guide complex medical care. The computing science trend of artificial intelligence (AI) uses algorithms arrayed in neural networks to learn patterns in complex datasets, providing insights otherwise obscure to cognitive humans and generating predictive models opaque to standard statistics. Most approved AI medical applications are diagnostic; such narrow AI cannot disambiguate clinical uncertainties from messy big datasets. Global AI challenges have been met in other data-dense context-uncertain domains like autonomous driving vehicles. These solutions offer salient lessons for healthcare, where the probability of flawed human reasoning and bias is high and where errors are life-threatening. Next wave broad AI technologies may help doctors to disambiguate complex individual patient diagnoses in real time, thereby improving clinical reasoning, mitigating biases, and explaining (or even averting) medical errors. However, adopting difficult to interpret "black box" AI models based on suspect data quality or data provenance for high-stakes medical decisions can worsen clinical ambiguities and add to healthcare inefficiencies.