scholarly journals Using Machine Learning to Predict Early Onset Acute Organ Failure in Critically Ill Intensive Care Unit Patients With Sickle Cell Disease: Retrospective Study

10.2196/14693 ◽  
2020 ◽  
Vol 22 (5) ◽  
pp. e14693
Author(s):  
Akram Mohammed ◽  
Pradeep S B Podila ◽  
Robert L Davis ◽  
Kenneth I Ataga ◽  
Jane S Hankins ◽  
...  
Keyword(s):  
Machine Learning ◽  
Intensive Care ◽  
Retrospective Study ◽  
Sickle Cell Disease ◽  
Organ Failure ◽  
Organ Dysfunction ◽  
Sickle Cell ◽  
Machine Learning Techniques ◽  
Cell Disease ◽  
Learning Techniques

Background Sickle cell disease (SCD) is a genetic disorder of the red blood cells, resulting in multiple acute and chronic complications, including pain episodes, stroke, and kidney disease. Patients with SCD develop chronic organ dysfunction, which may progress to organ failure during disease exacerbations. Early detection of acute physiological deterioration leading to organ failure is not always attainable. Machine learning techniques that allow for prediction of organ failure may enable early identification and treatment and potentially reduce mortality. Objective The aim of this study was to test the hypothesis that machine learning physiomarkers can predict the development of organ dysfunction in a sample of adult patients with SCD admitted to intensive care units (ICUs). Methods We applied diverse machine learning methods, statistical methods, and data visualization techniques to develop classification models to distinguish SCD from controls. Results We studied 63 sequential SCD patients admitted to ICUs with 163 patient encounters (mean age 30.7 years, SD 9.8 years). A subset of these patient encounters, 22.7% (37/163), met the sequential organ failure assessment criteria. The other 126 SCD patient encounters served as controls. A set of signal processing features (such as fast Fourier transform, energy, and continuous wavelet transform) derived from heart rate, blood pressure, and respiratory rate was identified to distinguish patients with SCD who developed acute physiological deterioration leading to organ failure from patients with SCD who did not meet the criteria. A multilayer perceptron model accurately predicted organ failure up to 6 hours before onset, with an average sensitivity and specificity of 96% and 98%, respectively. Conclusions This retrospective study demonstrated the viability of using machine learning to predict acute organ failure among hospitalized adults with SCD. The discovery of salient physiomarkers through machine learning techniques has the potential to further accelerate the development and implementation of innovative care delivery protocols and strategies for medically vulnerable patients.

scholarly journals Using Machine Learning to Predict Early Onset Acute Organ Failure in Critically Ill Intensive Care Unit Patients With Sickle Cell Disease: Retrospective Study (Preprint)

2019 ◽  
Author(s):  
Akram Mohammed ◽  
Pradeep S B Podila ◽  
Robert L Davis ◽  
Kenneth I Ataga ◽  
Jane S Hankins ◽  
...  
Keyword(s):  
Machine Learning ◽  
Intensive Care ◽  
Retrospective Study ◽  
Sickle Cell Disease ◽  
Organ Failure ◽  
Organ Dysfunction ◽  
Sickle Cell ◽  
Machine Learning Techniques ◽  
Cell Disease ◽  
Learning Techniques

BACKGROUND Sickle cell disease (SCD) is a genetic disorder of the red blood cells, resulting in multiple acute and chronic complications, including pain episodes, stroke, and kidney disease. Patients with SCD develop chronic organ dysfunction, which may progress to organ failure during disease exacerbations. Early detection of acute physiological deterioration leading to organ failure is not always attainable. Machine learning techniques that allow for prediction of organ failure may enable early identification and treatment and potentially reduce mortality. OBJECTIVE The aim of this study was to test the hypothesis that machine learning physiomarkers can predict the development of organ dysfunction in a sample of adult patients with SCD admitted to intensive care units (ICUs). METHODS We applied diverse machine learning methods, statistical methods, and data visualization techniques to develop classification models to distinguish SCD from controls. RESULTS We studied 63 sequential SCD patients admitted to ICUs with 163 patient encounters (mean age 30.7 years, SD 9.8 years). A subset of these patient encounters, 22.7% (37/163), met the sequential organ failure assessment criteria. The other 126 SCD patient encounters served as controls. A set of signal processing features (such as fast Fourier transform, energy, and continuous wavelet transform) derived from heart rate, blood pressure, and respiratory rate was identified to distinguish patients with SCD who developed acute physiological deterioration leading to organ failure from patients with SCD who did not meet the criteria. A multilayer perceptron model accurately predicted organ failure up to 6 hours before onset, with an average sensitivity and specificity of 96% and 98%, respectively. CONCLUSIONS This retrospective study demonstrated the viability of using machine learning to predict acute organ failure among hospitalized adults with SCD. The discovery of salient physiomarkers through machine learning techniques has the potential to further accelerate the development and implementation of innovative care delivery protocols and strategies for medically vulnerable patients.

scholarly journals Machine learning predicts early-onset acute organ failure in critically ill patients with sickle cell disease

2019 ◽  
Author(s):  
Akram Mohammed ◽  
Pradeep S. B. Podila ◽  
Robert L. Davis ◽  
Kenneth I. Ataga ◽  
Jane S. Hankins ◽  
...  
Keyword(s):  
Machine Learning ◽  
Intensive Care ◽  
Sickle Cell Disease ◽  
Kidney Disease ◽  
Organ Failure ◽  
Organ Dysfunction ◽  
Sickle Cell ◽  
Machine Learning Techniques ◽  
Cell Disease ◽  
Learning Techniques

AbstractBackgroundSickle cell disease (SCD) is a genetic disorder of the red blood cells, resulting in multiple acute and chronic complications including pain episodes, stroke, and kidney disease. Patients with SCD develop chronic organ dysfunction, which may progress to organ failure during disease exacerbations. Early detection of acute physiological deterioration leading to organ failure is not always attainable. Machine learning techniques that allow for prediction of organ failure may enable earlier identification and treatment, and potentially reduce mortality. We tested the hypothesis that machine learning physiomarkers could predict the development of organ dysfunction in an adult sample of patients with SCD admitted to intensive care units.Methods and FindingsWe studied 63 sequential SCD patients with 163 patient encounters, mean age 33.0±11.0 years, admitted to intensive care units, some of whom (6.7%) had pre-existing cardiovascular or kidney disease. A subset of these patient encounters (37; 23%) met sequential organ failure assessment (SOFA) criteria. The site of organ failure included: central nervous system (32), cardiovascular (11), renal (10), liver (7), respiratory (5) and coagulation (2) systems. Most (81.5%) of the patient encounters who experienced organ failure had single organ failure. The other 126 SCD patient encounters served as controls. A set of signal processing features (such as fast fourier transform, energy, continuous wavelet transform, etc.) derived from heart rate, blood pressure, and respiratory rate were identified to distinguish patients with SCD who developed acute physiological deterioration leading to organ failure, from SCD patients who did not meet the criteria. A random forest model accurately predicted organ failure up to six hours prior to onset, with a five-fold cross-validation accuracy of 94.57% (average sensitivity and specificity of 90.24% and 98.9% respectively).ConclusionsThis study demonstrates the viability of using machine learning to predict acute physiological deterioration heralded organ failure among hospitalized adults with SCD. The discovery of salient physiomarkers through machine learning techniques has the potential to further accelerate the development and implementation of innovative care delivery protocols and strategies for medically vulnerable patients.

scholarly journals Recent Advances and Machine Learning Techniques on Sickle Cell Disease

2020 ◽  
Vol 5 (1) ◽  
pp. 46-59
Author(s):  
Noorh H. Alharbi ◽  
◽  
Rana O. Bameer ◽  
Shahad S. Geddan ◽  
, Hajar M. Alharbi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Sickle Cell Disease ◽  
Sickle Cell ◽  
Missing Values ◽  
Hereditary Disease ◽  
Machine Learning Techniques ◽  
Cell Disease ◽  
Clinical Complications ◽  
Learning Techniques ◽  
Therapeutic Decision Making

Sickle cell disease is a severe hereditary disease caused by an abnormality of the red blood cells. The current therapeutic decision-making process applied to sickle cell disease includes monitoring a patient’s symptoms and complications and then adjusting the treatment accordingly. This process is time-consuming, which might result in serious consequences for patients’ lives and could lead to irreversible disease complications. Artificial intelligence, specifically machine learning, is a powerful technique that has been used to support medical decisions. This paper aims to review the recently developed machine learning models designed to interpret medical data regarding sickle cell disease. To propose an intelligence model, the suggested framework has to be performed in the following sequence. First, the data is preprocessed by imputing missing values and balancing them. Then, suitable feature selection methods are applied, and different classifiers are trained and tested. Finally, the performing model with the highest predefined performance metric over all experiments conducted is nominated. Thus, the aim of developing such a model is to predict the severity of a patient’s case, to determine the clinical complications of the disease, and to suggest the correct dosage of the treatment(s).

Recent Studies and Research on Sickle Cell Disease

2020 ◽  
pp. 237-243
Author(s):  
Bikesh Kumar Singh ◽  
Hardik Thakkar
Keyword(s):  
Machine Learning ◽  
Sickle Cell Disease ◽  
Risk Stratification ◽  
Sickle Cell ◽  
Statistical Significance ◽  
Machine Learning Techniques ◽  
Cell Disease ◽  
Learning Techniques ◽  
Future Challenges ◽  
Stratification System

Machine learning techniques have been successfully applied in various domains of healthcare such as medical imaging, bio-signal processing, pathological data analysis, etc. This chapter discusses the recent studies on sickle cell disease (SCD) based on risk stratification system, predicting the severity of disease, prediction of dosage requirement, prediction of clinical complications of the disease, etc. The blood attributes of SCD patients, which are obtained by high performance liquid chromatography (HPLC) test or complete blood count (CBC) test have been used by many researchers for improving clinical outcomes and therapeutic intervention in SCD. Statistical significance analysis has been reported to determine the correlation and association of pathological attributes with clinical symptoms. Machine learning techniques have been employed for risk stratification and dosage prediction. This chapter summarizes these techniques and suggests research gaps and future challenges.

scholarly journals Hyperhaemolysis in Sickle Cell Disease Is Not Necessarily a Transfusion Reaction

Blood ◽  
2015 ◽  
Vol 126 (23) ◽  
pp. 4595-4595 ◽  
Author(s):  
Elizabeth A Jones ◽  
Louise Smith ◽  
Russell D Keenan
Keyword(s):  
Intensive Care ◽  
Sickle Cell Disease ◽  
Blood Transfusion ◽  
Organ Failure ◽  
Sickle Cell ◽  
Red Cell ◽  
Cell Disease ◽  
Sickle Cell Crisis ◽  
Life Threatening ◽  
Red Cell Transfusion

Abstract Hyperhaemolysis is a rare life threatening complication in sickle cell disease with rapidly dropping haemoglobin, intravascular haemolysis and haemoglobinuria leading to multi organ failure and death. The literature reports that hyperhaemolysis in sickle cell disease is a complication of red cell transfusion (Aragona et al., 2014 J. Pediatr. Hematol. Oncol.) and suggests management based on with holding further transfusion to avoid aggravating the haemolysis and using immunosuppression (Win 2009 Expert Rev. Hematol.). In the literature, all cases of hyperhaemolysis in addition to a recent blood transfusion, were in or had had a recent sickle cell crisis. We report a case of life threatening Hyperhaemolysis in a 5 year old child following a sickle cell crisis who had never previously been transfused. We suggest that, at least in this case, the hyperhaemolysis cannot be transfusion related. The theoretical case for management of withholding transfusion may not be sound and potentially dangerous. A female child with known sickle cell disease presented with temperature and chest pains, she had a Hb 72g/L (stable over a few years). She initially improved with oxygen, fluids and antibiotics. 36 hours after admission she acutely deteriorated with increasing pallor and dropping oxygen saturations. She started passing frank red urine which initially was considered to be haematuria but on investigation was haemoglobinuria. Her Hb dropped to 47g/L with no evidence of blood loss. Within hours of developing haemoglobinuria she required intensive care for respiratory support. She rapidly developed multi-organ failure requiring oscillatory ventilation, inotropes, and haemofiltration for renal support. She was managed with emergency red cell transfusion (her first ever) and within 12 hours of haemoglobinuria received a full red cell exchange transfusion. There were ongoing antibiotics for clinical respiratory infection and she was later confirmed to have influenza B. No steroids or other immune suppression were given. There was no evidence of acute bleeding to explain a drop in haemoglobin at any point. With maximum intensive care support including further transfusions she gradually improved and has made a full recovery. No deterioration was observed following transfusion. She has remained well since. She is now 13 years old and following such a dramatic episode she has remained on a transfusion programme with successful oral iron chelation. She has not experienced any further episodes of hyperhaemolysis and no red cell antibody has been detected at any time. This case demonstrates that hyperhaemolysis in sickle cell disease does not require a previous transfusion. We suggest that it is possible the previous reported cases are also not due to blood transfusion but are an acute form of haemolysis seen on the background of a chronic haemolytic disease. An increase in the rate of haemolysis may be related to other acute complications of sickle cell disease. We propose that the optimum management of hyperhaemolysis should include full supportive care including maintaining haemoglobin by transfusion. Immunosuppression in this case could have led to a worse outcome as influenza pneumonia was the likely initial trigger of the episode. Disclosures No relevant conflicts of interest to declare.

Sign in / Sign up

Export Citation Format

Share Document