Table 1.
The 44 reviewed articles showing the title, authors, year, specialty, population, features, purpose and design of the studies.
| Title | Reference | Year | Specialty | Sample | Features | Design Studio | Purpose | Objective | Analysis |
|---|---|---|---|---|---|---|---|---|---|
| Early hospital mortality prediction of intensive care unit patients using an ensemble learning approach | Awad et al. (2017) [18] | 2017 | IC | 11,722 patients (subgroups) | 20–29 ** | RC | Pg | To highlight the main data challenges in early mortality prediction in ICU patients and introduces a new machine learning based framework for Early (6h) Mortality Prediction for IC Unit patients (EMPICU) | De |
| Prediction of Recurrent Clostridium Difficile Infection (rCDI) Using Comprehensive Electronic Medical Records in an Integrated Healthcare Delivery System | Escobar et al. (2017) [19] | 2017 | ID | 12,706 | 150–23 ** | RC | Pg | To develop and validate rCDI predictive models based on ML in a large and representative sample of adults | De |
| Enhancement of hepatitis virus immunoassay outcome predictions in imbalanced routine pathology data by data balancing and feature selection before the application of support vector machines | Richardson and Lidbury (2017) [20] | 2017 | LM | 16,990 | 5–27 ** | RC | Dg | To use SVMs to identify predictors for the enhanced laboratory diagnosis of hepatitis virus infection, and to identify the type of data balancing and feature selection that best assisted this enhanced classification of HBV/HCV negative or positive |
De |
| Machine Learning Algorithms for Risk Prediction of Severe Hand-Foot-Mouth Disease in Children | Zhang et al. (2017) [21] | 2017 | Pd | 530 | 18 | RC | Pg | To identify clinical and MRI-related predictors for the occurrence of severe HFMD in children and to assess the interaction effects between them using machine learning algorithms | De |
| Novel Risk Assessment Tool for Immunoglobulin Resistance in Kawasaki Disease: Application Using a Random Forest Classifier: Application Using a Random Forest Classifier | Takeuchi et al. (2017) [22] | 2017 | Pd | 767 | 23 | RC | Th | To develop a new risk assessment tool for IVIG resistance using RF | De |
| Supervised learning for infection risk inference using pathology data | Hernandez et al. (2017) [23] | 2017 | ID | >500,000 patients | 6 | RC | Dg | To evaluates the performance of different binary classifiers to detect any type of infection from a reduced set of commonly requested clinical measurements | De |
| Applied Informatics Decision Support Tool for Mortality Predictions in Patients with Cancer | Bertsimas et al. (2018) [24] | 2018 | On | 23,983 patients | 401 | RC | Pg | To develop a predictive tool that estimates the probability of mortality for an individual patient being proposed their next treatment | Re |
| Machine learning model combining features from algorithms with different analytical methodologies to detect laboratory-event-related adverse drug reaction signals | Jeong et al. (2018) [25] | 2018 | Ph | 1674 drug-laboratory event pairs | 48 | RC | Th | To develop a more accurate ADR signal detection algorithm for post-market surveillance using EHR data by integrating the results of existing ADR detection algorithms using ML models | De |
| Using Machine Learning-Based Multianalyte Delta Checks to Detect Wrong Blood in Tube Errors | Rosenbaum and Baron (2018) [26] | 2018 | LM | 20,638 patient collections of 4855 patients | 3 features for each of 11 lab tests | RC | Rch | To test whether machine learning-based multianalyte delta checks could outperform traditional single-analyte ones in identifying WBIT | De |
| An Interpretable ICU Mortality Prediction Model Based on Logistic Regression and Recurrent Neural Networks with LSTM units | Ge et al. (2018) [27] | 2018 | IC | 4896 | NA | RC | Pg | To develop an interpretable ICU mortality prediction model based on Logistic Regression and RNN with LSTM units | De |
| High-density lipoprotein cholesterol levels and pulmonary artery vasoreactivity in patients with idiopathic pulmonary arterial hypertension | Jonas et al. (2018) [28] | 2018 | Ca | 66 | NA | PC | Pg | To investigate the association between cardiometabolic risk factors and vasoreactivity of pulmonary arteries in patients with Idiopathic Pulmonary Arterial Hypertension | NE |
| Development and Validation of Machine Learning Models for Prediction of 1-Year Mortality Utilizing Electronic Medical Record Data Available at the End of Hospitalization in Multi-condition Patients: a Proof-of-Concept Study | Sahni et al. (2018) [29] | 2018 | ID | 59,848 | 4 classes ** | RC | Pg | To construct models that utilize EHR data to prognosticate 1-year mortality in a large, diverse cohort of multi-condition hospitalizations | Re |
| Predicting the risk of emergency admission with machine learning: Development and validation using linked electronic health records | Rahimian et al. (2018) [30] | 2018 | IM | 4,637,297 | 43 + 13 ** | RC | Rch | To improve discrimination and calibration for predicting the risk of emergency admission | Re |
| Analyte Quantity Detection from Lateral Flow Assay Using a Smartphone | Foysal et al. (2019) [31] | 2019 | LM | 15 LFA set for 75 readings | NE | RC | Dg | To propose a robust smartphone-based analyte (albumin) detection method that estimates the amount of analyte on an LFA strip using a smartphone camera | Ch |
| Prevalence and Predictability of Low-Yield Inpatient Laboratory Diagnostic Tests | Xu et al. (2019) [32] | 2019 | LM | 10,000 samples per feature | 43 | RC | Rch | To identify inpatient diagnostic laboratory testing with predictable results that are unlikely to yield new information | Re |
| Using artificial intelligence to reduce diagnostic workload without compromising detection of urinary tract infections | Burton et al. (2019) [33] | 2019 | LM | 225,207 | 21 | RC | Dg | To reduce the burden of culturing the large number of culture-negative samples without reducing detection of culture-positive samples | De |
| Interactive Machine Learning for Laboratory Data Integration | Fillmore et al. (2019) [34] | 2019 | LM | 4 ∗ 10^9 records | NE | RC | Rch | To develop a machine learning system to predict whether a lab test type clinically belongs within the concept of interest | Ch |
| Early prediction of acute kidney injury following ICU admission using a multivariate panel of physiological measurements | Zimmerman et al. (2019) [35] | 2019 | IC | 23,950 | NA ** | RC | Dg | To predict AKI (creatinine values in day 2 and 3) using first-day measurements of a multivariate panel of physiologic variables | De |
| A New Insight into Missing Data in IC Unit Patient Profiles: Observational Study | Sharafoddini et al. (2019) [36] | 2019 | IC | 32,618–20,318–13,670 patients (days 1–2–3) | NA ** | RC | Pg | To examine whether the presence or missingness of a variable itself in ICU records can convey information about the patient health status | De |
| Survival outcome prediction in cervical cancer: Cox models vs deep-learning model | Matsuo et al. (2019) [37] | 2019 | On | 768 | 40 ** | RC | Rch | To compare the deep Learning neural network model and the Cox proportional hazard regression model in the prediction of survival in women with cervical cancer | Re |
| Relative criticalness of common laboratory tests for critical value reporting | Yang et al. (2019) [38] | 2019 | IC | 22,174 | 23 | RC | Pg | To evaluate the relative strength of association between 23 most commonly ordered laboratory tests in a CCU setting and the adverse outcome, defined as death during the CCU stay within 24 h of reporting of the laboratory result | NE |
| Enhanced early prediction of clinically relevant neonatal hyperbilirubinemia with machine learning | Daunhawer et al. (2019) [39] | 2019 | Pd | 362 | 44–4 | PC | Dg | To enhance the early detection of clinically relevant hyperbilirubinemia in advance of the first phototherapy treatment | De |
| A clustering approach for detecting implausible observation values in electronic health records data | Estiri et al. (2019) [40] | 2019 | LM | >720 million records, 50 lab tests | NE | RC | Rch | To develop and test an unsupervised clustering-based anomaly/outlier detection approach for detecting implausible observations in EHR data | De |
| Modelling outcomes after paediatric brain injury with admission laboratory values: a machine-learning approach | Kayhanian et al. (2019) [41] | 2019 | Ns | 94 | 14 | RC | Pg | To identify which admission laboratory variables are correlated to outcomes after Traumatic Brain Injury (TBI) in children and to explore prediction of outcomes, using both univariate analysis and supervised learning methods | De |
| Automatic Machine-Learning-Based Outcome Prediction in Patients with Primary Intracerebral Haemorrhage | Wang et al. (2019) [42] | 2019 | Ns | 1-month outcome: 307; 6-month outcome: 243 | 1-month outcome: 26; 6-month outcome: 22 | RC | Pg | To predict the functional outcome in patients with primary intracerebral haemorrhage (ICH) | Ch |
| A Real-Time Early Warning System for Monitoring Inpatient Mortality Risk: Prospective Study Using Electronic Medical Record Data | Ye et al. (2019) [43] | 2019 | IM | 42,484 retrospective, 11,762 prospective |
680 ** | PC | Pg | To build and prospectively validate an Early Warning System-based inpatient mortality Electronic Medical Record | Ch |
| Routine laboratory blood tests predict SARS-CoV-2 infection using machine learning | Yang et al. (2020) [44] | 2020 | ID | 3356 | 33 | RC | Dg | To develop a ML model integrating age, gender, race and routine laboratory blood tests, which are readily available with a short Turnaround Time | De |
| Development and validation of prognosis model of mortality risk in patients with COVID-19 | Ma et al. (2020) [45] | 2020 | ID | 305 | 33 | RC | Pg | Investigate ML to rank clinical features, and multivariate logistic regression method to identify clinical features with statistical significance in prediction of mortality risk in patients with COVID-19 using their clinical data | De |
| Exploration of critical care data by using unsupervised machine learning | Hyun et al. (2020) [46] | 2020 | IC | 1503 | 9 | RC | Rch | To discover subgroups among ICU patients and to examine their clinical characteristics, therapeutic procedures conducted during the ICU stay, and discharge dispositions | NE |
| Predicting Adverse Outcomes for Febrile Patients in the Emergency Department Using Sparse Laboratory Data: Development of a Time Adaptive Model | Lee et al. (2020) [47] | 2020 | EM | 9491 | NA | RC | Pg | To develop time adaptive models that predict adverse outcomes for febrile patients assessing the utility of routine lab tests (only request OSO, and request and value OSR) | De |
| Temporal Pattern Detection to Predict Adverse Events in Critical Care: Case study With Acute Kidney Injury | Morid et al. (2020) [48] | 2020 | IC | 22,542 | 17 | RC | Pg | To evaluate approaches to predict Adverse Events in ICU settings using structural temporal pattern detection methods for both local (within each time window) and global (across time windows) trends, derived from first 48 h of ICU | NE |
| Predict or draw blood: An integrated method to reduce lab tests | Yu et al. (2020) [49] | 2020 | LM | 41,113 | 20 | RC | Rch | To propose a novel deep learning method to jointly predict future lab test events to be omitted and the values of the omitted events based on observed testing values | Re |
| Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone | Chicco and Jurman (2020) [50] | 2020 | Ca | 299 | 13–2 ** | RC | Pg | To use several data mining techniques first to predict survival of the patients, and to rank the most important features included in the medical records | De |
| Comparison of Machine Learning Methods and Conventional Logistic Regressions for Predicting Gestational Diabetes Using Routine Clinical Data: A Retrospective Cohort Study | Ye et al. (2020) [51] | 2020 | Ob | 22,242 | 104 | RC | Dg | To use machine learning methods to predict GDM (Gestational Diabetes) and compare their performance with that of logistic regressions | De |
| Mortality prediction enhancement in end-stage renal disease (ESRD): A machine learning approach | Macias et al. (2020) [52] | 2020 | Ne | 261 | NA | RC | Pg | To assess the potential of the massive use of variables together with machine learning techniques for the improvement of mortality predictive models in ESRD | De |
| A recurrent neural network approach to predicting haemoglobin trajectories in patients with End-Stage Renal Disease | Lobo et al. (2020) [53] | 2020 | Ne | 1972 patients ** | NA | RC | Dg | To develop a RNN approach that uses historical data together with future ESA and iron dosing data to predict the 1-, 2-, and 3-month Hgb levels of patients with ESRD-induced anaemia | Re |
| Early diagnosis of bloodstream infections in the intensive care unit using machine-learning algorithms | Roimi et al. (2020) [54] | 2020 | IC | 2351 + 1021 | NA | RC | Dg | To develop a machine-learning (ML) algorithm that can predict intensive care unit (ICU)-acquired bloodstream infections (BSI) among patients suspected of infection in the ICU | De |
| Dynamic readmission prediction using routine postoperative laboratory results after radical cystectomy | Kirk et al. (2020) [55] | 2020 | Ur | 996 | 15 | RC | Pg | To determine if the addition of electronic health record data enables better risk stratification and 30-day readmission prediction after radical cystectomy | De |
| A Machine Learning–Based Model to Predict Acute Traumatic Coagulopathy (ATC) in Trauma Patients Upon Emergency Hospitalization | Li et al. (2020) [56] | 2020 | EM | 818 retrospective, 578 prospective |
6 ** | PC | Dg | To develop and validate a prediction model for ATC that is based on objective indicators which are already routinely obtained as patients are admitted at the hospital | De |
| Improved prediction of dengue outbreak using combinatorial feature selector and classifier based on entropy weighted score based optimal ranking | Balamurugan et al. (2020) [57] | 2020 | ID | 480 | 20 ** | RC | Dg | To analyse the performance of the proposed EWSORA Feature Selector, detailed experimentation is conducted on various ML classifiers | De |
| Using a machine learning approach to predict mortality in critically ill influenza patients: a cross-sectional retrospective multicentre study in Taiwan | Hu et al. (2020) [58] | 2020 | IC | 336 | 76 | RC | Pg | To establish an explainable ML model for predicting mortality in critically ill influenza patients using a real-world severe influenza data set (first 7 days) | De |
| A novel and simple machine learning algorithm for preoperative diagnosis of acute appendicitis in children | Aydin et al. (2020) [59] | 2020 | PS | 7244 | NA | RC | Dg | Provide an easily interpretable model to understand the relationship between blood variables and appendicitis to create an automated decision support tool in the future | De |
| Identification of risk factors for patients with diabetes: diabetic polyneuropathy case study | Metsker et al. (2020) [60] | 2020 | En | 5425 | 29–31 | RC | Pg | Early identification of the risk of diabetes polyneuropathy based on structured electronic medical records | De |
| Feasibility of machine learning based predictive modelling of postoperative hyponatremia after pituitary surgery | Voglis et al. (2020) [61] | 2020 | Ns | 207 | 26 | RC | Pg | Evaluate the feasibility of predictive modelling of postoperative hyponatremia after pituitary tumour surgery using preoperative available variables | Re |
* It was chosen as the most useful, although it was not the best performer; ** Different models were trained with a different number of features; *** A comparison of the ML models was not made; NA: Not available; NE: Not evaluable (meaning not pertinent). For all the other abbreviations, see Appendix B.