A comprehensive scoping review of Bayesian networks in healthcare: Past, present and future

No comprehensive review of Bayesian networks (BNs) in healthcare has been published the past, making it difficult to organize research contributions present and identify challenges neglected areas that need be addressed future. This unique novel scoping BNs provides an analytical framework for comprehensively characterizing domain its current state. A literature search health informatics databases using relevant keywords found 3810 articles were reduced 123. was after screening out those presenting statistics, meta-analysis or neural networks, as opposed describing predictive performance multiple machine learning algorithms, which simply one type. Using framework, we show that: (1) are not used their full potential; (2) a generic BN development process is lacking; (3) limitations exist way presented literature, impacts understanding, consensus towards systematic methodologies, practice adoption; (4) gap exists between having accurate useful clinical practice. highlights several issues, such restricted aims BNs, ad hoc methods, lack adoption reveals researchers clinicians address these problems. To map forward, paper proposes future directions makes recommendations regarding methods