The Promise of Clinical Decision Support Systems Targetting Low-Resource Settings

Abstract

Low-resource clinical settings are plagued by low physician-to-patient ratios and a shortage of highquality medical expertise and infrastructure. Together, these phenomena lead to over-burdened healthcare systems that under-serve the needs of the community. Alleviating this burden can be undertaken by the introduction of clinical decision support systems (CDSSs); systems that support stakeholders (ranging from physicians to patients) within the clinical setting in their day-today activities. Such systems, which have proven to be effective in the developed world, remain to be under-explored in low-resource settings. This review attempts to summarize the research focused on clinical decision support systems that either target stakeholders within low-resource clinical settings or diseases commonly found in such environments. When categorizing our findings according to disease applications, we find that CDSSs are predominantly focused on dealing with bacterial infections and maternal care, do not leverage deep learning, and have not been evaluated prospectively. Together, these highlight the need for increased research in this domain in order to impact a diverse set of medical conditions and ultimately improve patient outcomes.

Artificial-Intelligence-Based Clinical Decision Support Systems in Primary Care: A Scoping Review of Current Clinical Implementations

“…This is unfeasible for PHC because of data and dataset scarcity, as most people do not have access to it [ 19 , 22 ]. On top of that, most healthcare organizations lack the data infrastructure to collect the data required to adequately train an algorithm tailored to the local population and practice patterns and to guarantee the absence of bias [ 6 , 15 , 35 , 48 ].…”

“…In 2007, the US government encouraged the introduction of Clinical Decision Support Systems (CDSSs) into Electronic Health Records (EHR), and by 2017, 40.2% of US hospitals had advanced CDSS capabilities [ 4 ]. CDSSs aid physicians in diagnosis, disease management, prescription, and drug control, often through alarm systems [ 5 , 6 ]. They have been especially effective in increasing adherence to clinical guidelines, applying prevention and public health strategies, and improving patient safety [ 3 , 7 ].…”

A comprehensive review on efficient artificial intelligence models for classification of abnormal cardiac rhythms using electrocardiograms

“…A summary of such data annotation efficient models for ECG cardiac arrhythmia classification has been presented in Table 4 . While most of the studies in contrastive learning [143] , [144] , [145] , [146] have focused on proposing efficient transformation as it has considerable potential to improve the efficacy of existing approaches, some have developed novel patient-specific contrastive loss functions [143] to boost classification accuracy. Kiyasseh et al [143] surpass the well-known contrastive learning framework SimCLR by 15.8% in terms of area under the curve (AUC) on Chapman dataset [132] by introducing a patient-specific noise contrastive loss and three transformations namely: Contrastive Multi-segment Coding (CMSC), Contrastive Multi-lead Coding (CMLC) & Contrastive Multi-segment Multi-lead Coding (CMSMLC).…”

“…While most of the studies in contrastive learning [143] , [144] , [145] , [146] have focused on proposing efficient transformation as it has considerable potential to improve the efficacy of existing approaches, some have developed novel patient-specific contrastive loss functions [143] to boost classification accuracy. Kiyasseh et al [143] surpass the well-known contrastive learning framework SimCLR by 15.8% in terms of area under the curve (AUC) on Chapman dataset [132] by introducing a patient-specific noise contrastive loss and three transformations namely: Contrastive Multi-segment Coding (CMSC), Contrastive Multi-lead Coding (CMLC) & Contrastive Multi-segment Multi-lead Coding (CMSMLC). Another state-of-the-art (SOTA) method self-supervised contrastive learning approach called sCL-ST [147] for multi-label classification of 12-lead ECGs introduced two novel transformations - split-join and semantic weighted peak noise smoothing, which enabled a robust method insensitive to changes in ECG signal.…”

Sepsis Mortality Prediction Using Wearable Monitoring in Low–Middle Income Countries

“…This requires careful clinical evaluation and monitoring of patients’ vital signs, with or without supporting laboratory data. Lack of laboratory access, monitoring equipment, and skilled health care staff are serious impediments to achieving this in LMICs [ 6 , 7 , 8 ]. To assist triage and prognostication, various scores have been developed.…”