Using Machine Learning Algorithms to Develop a Clinical Decision-Making Tool for COVID-19 Inpatients

Background: Within the UK, COVID-19 has contributed towards over 103,000 deaths. Although multiple risk factors for COVID-19 have been identified, using this data to improve clinical care has proven challenging. The main aim of this study is to develop a reliable, multivariable predictive model for COVID-19 in-patient outcomes, thus enabling risk-stratification and earlier clinical decision-making. Methods: Anonymised data consisting of 44 independent predictor variables from 355 adults diagnosed with COVID-19, at a UK hospital, was manually extracted from electronic patient records for retrospective, case–control analysis. Primary outcomes included inpatient mortality, required ventilatory support, and duration of inpatient treatment. Pulmonary embolism sequala was the only secondary outcome. After balancing data, key variables were feature selected for each outcome using random forests. Predictive models were then learned and constructed using Bayesian networks. Results: The proposed probabilistic models were able to predict, using feature selected risk factors, the probability of the mentioned outcomes. Overall, our findings demonstrate reliable, multivariable, quantitative predictive models for four outcomes, which utilise readily available clinical information for COVID-19 adult inpatients. Further research is required to externally validate our models and demonstrate their utility as risk stratification and clinical decision-making tools.

Download Full-text

Predicting mortality, duration of treatment, pulmonary embolism and required ceiling of ventilatory support for COVID-19 inpatients: A Machine-Learning Approach

10.1101/2021.02.15.21251752 ◽

2021 ◽

Author(s):

Abhinav Vepa ◽

Amer Saleem ◽

Kambiz Rakhshan ◽

Amr Omar ◽

Diana Dharmaraj ◽

...

Keyword(s):

Risk Factors ◽

Decision Making ◽

Pulmonary Embolism ◽

Bayesian Networks ◽

Predictive Models ◽

Inpatient Treatment ◽

Clinical Decision Making ◽

Ventilatory Support ◽

Clinical Decision ◽

Inpatient Mortality

AbstractIntroductionWithin the UK, COVID-19 has contributed towards over 103,000 deaths. Multiple risk factors for COVID-19 have been identified including various demographics, co-morbidities, biochemical parameters, and physical assessment findings. However, using this vast data to improve clinical care has proven challenging.Aimsto develop a reliable, multivariable predictive model for COVID-19 in-patient outcomes, to aid risk-stratification and earlier clinical decision-making.MethodsAnonymized data regarding 44 independent predictor variables of 355 adults diagnosed with COVID-19, at a UK hospital, was manually extracted from electronic patient records for retrospective, case-controlled analysis. Primary outcomes included inpatient mortality, level of ventilatory support and oxygen therapy required, and duration of inpatient treatment. Secondary pulmonary embolism was the only secondary outcome. After balancing data, key variables were feature selected for each outcome using random forests. Predictive models were created using Bayesian Networks, and cross-validated.ResultsOur multivariable models were able to predict, using feature selected risk factors, the probability of inpatient mortality (F1 score 83.7%, PPV 82%, NPV 67.9%); level of ventilatory support required (F1 score varies from 55.8% “High-flow Oxygen level” to 71.5% “ITU-Admission level”); duration of inpatient treatment (varies from 46.7% for “≥ 2 days but < 3 days” to 69.8% “≤ 1 day”); and risk of pulmonary embolism sequelae (F1 score 85.8%, PPV of 83.7%, and NPV of 80.9%).ConclusionOverall, our findings demonstrate reliable, multivariable predictive models for 4 outcomes, that utilize readily available clinical information for COVID-19 adult inpatients. Further research is required to externally validate our models and demonstrate their utility as clinical decision-making tools.HighlightsUsing COVID-19 risk-factor data to assist clinical decision making is a challengeAnonymous data from 355 COVID-19 inpatients was collected & balancedKey independent variables were feature selected for 4 different outcomesAccurate, multi-variable predictive models were computed, using Bayesian NetworksFuture research should externally validate our models & demonstrate clinical utility

Download Full-text

Integrating Patient-Generated Health Data Into Clinical Care Settings or Clinical Decision-Making: Lessons Learned From Project HealthDesign

JMIR Human Factors ◽

10.2196/humanfactors.5919 ◽

2016 ◽

Vol 3 (2) ◽

pp. e26 ◽

Cited By ~ 49

Author(s):

Deborah J Cohen ◽

Sara R Keller ◽

Gillian R Hayes ◽

David A Dorr ◽

Joan S Ash ◽

...

Keyword(s):

Decision Making ◽

Clinical Decision Making ◽

Clinical Care ◽

Clinical Decision ◽

Health Data ◽

Lessons Learned

Download Full-text

Reversibility of Delirium in Palliative Care (DRAFT)

10.1093/med/9780190658618.003.0010 ◽

2018 ◽

Author(s):

Tiffany Shaw ◽

Eric Prommer

Keyword(s):

Risk Factors ◽

Decision Making ◽

Palliative Care ◽

Clinical Case ◽

Clinical Decision Making ◽

Clinical Decision ◽

Factors Associated ◽

Frequent Event ◽

Study Intervention

Delirium is a frequent event in patients with advanced cancer. Untreated delirium affects assessment of symptoms, impairs communication including participation in clinical decision-making. This study used specific diagnostic criteria for delirium and prospectively identified precipitating causes of delirium. The study identified factors associated with reversible and irreversible delirium. Impact of delirium on prognosis was evaluated. This chapter describes the basics of the study, including funding, year study began, year study was published, study location, who was studied, who was excluded, how many patients, study design, study intervention, follow-up, endpoints, results, and criticism and limitations. The chapter briefly reviews other relevant studies and information, gives a summary and discusses implications, and concludes with a relevant clinical case. Topics covered include delirium, neoplasms, palliative care, polypharmacy, risk factors, and therapeutics.

Download Full-text

Clinical Decision Making in the ICU: Dysphagia Screening, Assessment, and Treatment

Seminars in Speech and Language ◽

10.1055/s-0039-1688980 ◽

2019 ◽

Vol 40 (03) ◽

pp. 170-187 ◽

Cited By ~ 4

Author(s):

Martin B. Brodsky ◽

Emily B. Mayfield ◽

Roxann Diez Gross

Keyword(s):

Intensive Care Unit ◽

Decision Making ◽

Intensive Care ◽

Clinical Decision Making ◽

Clinical Care ◽

Clinical Decision ◽

Level Of Care ◽

Assessment And Treatment ◽

Proper Training ◽

Screening Assessment

AbstractClinicians often perceive the intensive care unit as among the most intimidating environments in patient care. With the proper training, acquisition of skill, and approach to clinical care, feelings of intimidation may be overcome with the great rewards this level of care has to offer. This review—spanning the ages of birth to senescence and covering oral/nasal endotracheal intubation and tracheostomy—presents a clinically relevant, directly applicable review of screening, assessment, and treatment of dysphagia in the patients who are critically ill for clinical speech–language pathologists and identifies gaps in the clinical peer-reviewed literature for researchers.

Download Full-text

Predictive Models for Mortality After Ruptured Aortic Aneurysm Repair Do Not Predict Futility and Are Not Useful for Clinical Decision-Making

Journal of Vascular Surgery ◽

10.1016/j.jvs.2015.11.025 ◽

2016 ◽

Vol 63 (2) ◽

pp. 563

Author(s):

Patrick C. Thompson ◽

Ronald L. Dalman ◽

E. John Harris ◽

Venita Chandra ◽

Jason T. Lee ◽

...

Keyword(s):

Decision Making ◽

Aortic Aneurysm ◽

Predictive Models ◽

Clinical Decision Making ◽

Clinical Decision ◽

Ruptured Aortic Aneurysm ◽

Aneurysm Repair ◽

Aortic Aneurysm Repair

Download Full-text

Clinical decision making and cardiovascular risk factors: What about the Heart Team?

Revista Clínica Española ◽

10.1016/j.rce.2014.02.006 ◽

2014 ◽

Vol 214 (4) ◽

pp. 200-201 ◽

Cited By ~ 1

Author(s):

M.-J. de Boer ◽

H.J.G. Bilo

Keyword(s):

Risk Factors ◽

Decision Making ◽

Cardiovascular Risk ◽

Cardiovascular Risk Factors ◽

Clinical Decision Making ◽

Clinical Decision ◽

Heart Team

Download Full-text

Predictive models for mortality after ruptured aortic aneurysm repair do not predict futility and are not useful for clinical decision making

Journal of Vascular Surgery ◽

10.1016/j.jvs.2016.07.121 ◽

2016 ◽

Vol 64 (6) ◽

pp. 1617-1622 ◽

Cited By ~ 8

Author(s):

Patrick C. Thompson ◽

Ronald L. Dalman ◽

E. John Harris ◽

Venita Chandra ◽

Jason T. Lee ◽

...

Keyword(s):

Decision Making ◽

Aortic Aneurysm ◽

Predictive Models ◽

Clinical Decision Making ◽

Clinical Decision ◽

Ruptured Aortic Aneurysm ◽

Aneurysm Repair ◽

Aortic Aneurysm Repair

Download Full-text

Clinical decision-making in older adults following emergency admission to hospital. Derivation and validation of a risk stratification score: OPERA

PLoS ONE ◽

10.1371/journal.pone.0248477 ◽

2021 ◽

Vol 16 (3) ◽

pp. e0248477

Author(s):

Khushal Arjan ◽

Lui G. Forni ◽

Richard M. Venn ◽

David Hunt ◽

Luke Eliot Hodgson

Keyword(s):

Older Adults ◽

Decision Making ◽

Risk Stratification ◽

Hospital Mortality ◽

Hospital Stay ◽

Risk Score ◽

Clinical Decision Making ◽

External Validation ◽

Clinical Decision ◽

Early Warning Score

Objectives of the study Demographic changes alongside medical advances have resulted in older adults accounting for an increasing proportion of emergency hospital admissions. Current measures of illness severity, limited to physiological parameters, have shortcomings in this cohort, partly due to patient complexity. This study aimed to derive and validate a risk score for acutely unwell older adults which may enhance risk stratification and support clinical decision-making. Methods Data was collected from emergency admissions in patients ≥65 years from two UK general hospitals (April 2017- April 2018). Variables underwent regression analysis for in-hospital mortality and independent predictors were used to create a risk score. Performance was assessed on external validation. Secondary outcomes included seven-day mortality and extended hospital stay. Results Derivation (n = 8,974) and validation (n = 8,391) cohorts were analysed. The model included the National Early Warning Score 2 (NEWS2), clinical frailty scale (CFS), acute kidney injury, age, sex, and Malnutrition Universal Screening Tool. For mortality, area under the curve for the model was 0.79 (95% CI 0.78–0.80), superior to NEWS2 0.65 (0.62–0.67) and CFS 0.76 (0.74–0.77) (P<0.0001). Risk groups predicted prolonged hospital stay: the highest risk group had an odds ratio of 9.7 (5.8–16.1) to stay >30 days. Conclusions Our simple validated model (Older Persons’ Emergency Risk Assessment [OPERA] score) predicts in-hospital mortality and prolonged length of stay and could be easily integrated into electronic hospital systems, enabling automatic digital generation of risk stratification within hours of admission. Future studies may validate the OPERA score in external populations and consider an impact analysis.

Download Full-text

Assessment of violence risk in 440 psychiatric patients in China: examining the feasibility and acceptability of a novel and scalable approach (FoVOx)

BMC Psychiatry ◽

10.1186/s12888-021-03115-3 ◽

2021 ◽

Vol 21 (1) ◽

Cited By ~ 1

Author(s):

Shaoling Zhong ◽

◽

Rongqin Yu ◽

Robert Cornish ◽

Xiaoping Wang ◽

...

Keyword(s):

Mental Health ◽

Risk Factors ◽

Risk Assessment ◽

Decision Making ◽

Clinical Decision Making ◽

Psychiatric Patients ◽

Clinical Decision ◽

Violence Risk ◽

Violence Risk Assessment ◽

Risk Categories

Abstract Background Violence risk assessment is a routine part of clinical services in mental health, and in particular secure psychiatric hospitals. The use of prediction models and risk tools can assist clinical decision-making on risk management, including decisions about further assessments, referral, hospitalization and treatment. In recent years, scalable evidence-based tools, such as Forensic Psychiatry and Violent Oxford (FoVOx), have been developed and validated for patients with mental illness. However, their acceptability and utility in clinical settings is not known. Therefore, we conducted a clinical impact study in multiple institutions that provided specialist mental health service. Methods We followed a two-step mixed-methods design. In phase one, we examined baseline risk factors on 330 psychiatric patients from seven forensic psychiatric institutes in China. In phase two, we conducted semi-structured interviews with 11 clinicians regarding violence risk assessment from ten mental health centres. We compared the FoVOx score on each admission (n = 110) to unstructured clinical risk assessment and used a thematic analysis to assess clinician views on the accuracy and utility of this tool. Results The median estimated probability of violent reoffending (FoVOx score) within 1 year was 7% (range 1–40%). There was fair agreement (72/99, 73% agreement) on the risk categories between FoVOx and clinicians’ assessment on risk categories, and moderate agreement (10/12, 83% agreement) when examining low and high risk categories. In a majority of cases (56/101, 55%), clinicians thought the FoVOx score was an accurate representation of the violent risk of an individual patient. Clinicians suggested some additional clinical, social and criminal risk factors should be considered during any comprehensive assessment. In addition, FoVOx was considered to be helpful in assisting clinical decision-making and individual risk assessment. Ten out of 11 clinicians reported that FoVOx was easy to use, eight out of 11 was practical, and all clinicians would consider using it in the future. Conclusions Clinicians found that violence risk assessment could be improved by using a simple, scalable tool, and that FoVOx was feasible and practical to use.

Download Full-text