Feature Identification Using Interpretability Machine Learning Predicting Risk Factors for Disease Severity of In-Patients with COVID-19 in South Florida

Overview

Journal Diagnostics (Basel)

Specialty Radiology

Date 2024 Sep 14

PMID 39272651

Authors

Debarshi Datta

Subhosit Ray

Laurie Martinez

David Newman

Safiya George Dalmida

Javad Hashemi

Candice Sareli

Paula Eckardt

Affiliations

Soon will be listed here.

Abstract

The objective of the study was to establish an AI-driven decision support system by identifying the most important features in the severity of disease for ntensive are nit (ICU) with echanical entilation (MV) requirement, ICU, and nterediate are nit (IMCU) admission for hospitalized patients with COVID-19 in South Florida. The features implicated in the risk factors identified by the model interpretability can be used to forecast treatment plans faster before critical conditions exacerbate. We analyzed eHR data from 5371 patients diagnosed with COVID-19 from South Florida Memorial Healthcare Systems admitted between March 2020 and January 2021 to predict the need for ICU with MV, ICU, and IMCU admission. A Random Forest classifier was trained on patients' data augmented by SMOTE, collected at hospital admission. We then compared the importance of features utilizing different model interpretability analyses, such as SHAP, MDI, and Permutation Importance. The models for ICU with MV, ICU, and IMCU admission identified the following factors overlapping as the most important predictors among the three outcomes: age, race, sex, BMI, diarrhea, diabetes, hypertension, early stages of kidney disease, and pneumonia. It was observed that individuals over 65 years ('older adults'), males, current smokers, and BMI classified as 'overweight' and 'obese' were at greater risk of severity of illness. The severity was intensified by the co-occurrence of two interacting features (e.g., diarrhea and diabetes). The top features identified by the models' interpretability were from the 'sociodemographic characteristics', 'pre-hospital comorbidities', and 'medications' categories. However, 'pre-hospital comorbidities' played a vital role in different critical conditions. In addition to individual feature importance, the feature interactions also provide crucial information for predicting the most likely outcome of patients' conditions when urgent treatment plans are needed during the surge of patients during the pandemic.

References

Meille G, Decker S, Owens P, Selden T . COVID-19 Admission Rates and Changes in US Hospital Inpatient and Intensive Care Unit Occupancy. JAMA Health Forum. 2023; 4(12):e234206. PMC: 10692846. DOI: 10.1001/jamahealthforum.2023.4206. View

Lundberg S, Erion G, Chen H, DeGrave A, Prutkin J, Nair B . From Local Explanations to Global Understanding with Explainable AI for Trees. Nat Mach Intell. 2020; 2(1):56-67. PMC: 7326367. DOI: 10.1038/s42256-019-0138-9. View

Johnson K, Wei W, Weeraratne D, Frisse M, Misulis K, Rhee K . Precision Medicine, AI, and the Future of Personalized Health Care. Clin Transl Sci. 2020; 14(1):86-93. PMC: 7877825. DOI: 10.1111/cts.12884. View

Miller I, Becker A, Grenfell B, Metcalf C . Disease and healthcare burden of COVID-19 in the United States. Nat Med. 2020; 26(8):1212-1217. DOI: 10.1038/s41591-020-0952-y. View

Ryan C, Minc A, Caceres J, Balsalobre A, Dixit A, Ng B . Predicting severe outcomes in Covid-19 related illness using only patient demographics, comorbidities and symptoms. Am J Emerg Med. 2020; 45:378-384. PMC: 7480533. DOI: 10.1016/j.ajem.2020.09.017. View

Van Lissa C, Stroebe W, VanDellen M, Leander N, Agostini M, Draws T . Using machine learning to identify important predictors of COVID-19 infection prevention behaviors during the early phase of the pandemic. Patterns (N Y). 2022; 3(4):100482. PMC: 8904175. DOI: 10.1016/j.patter.2022.100482. View

Bhatraju P, Ghassemieh B, Nichols M, Kim R, Jerome K, Nalla A . Covid-19 in Critically Ill Patients in the Seattle Region - Case Series. N Engl J Med. 2020; 382(21):2012-2022. PMC: 7143164. DOI: 10.1056/NEJMoa2004500. View

Magunia H, Lederer S, Verbuecheln R, Gilot B, Koeppen M, Haeberle H . Machine learning identifies ICU outcome predictors in a multicenter COVID-19 cohort. Crit Care. 2021; 25(1):295. PMC: 8370055. DOI: 10.1186/s13054-021-03720-4. View

Altmann A, Tolosi L, Sander O, Lengauer T . Permutation importance: a corrected feature importance measure. Bioinformatics. 2010; 26(10):1340-7. DOI: 10.1093/bioinformatics/btq134. View

10.

Patel D, Kher V, Desai B, Lei X, Cen S, Nanda N . Machine learning based predictors for COVID-19 disease severity. Sci Rep. 2021; 11(1):4673. PMC: 7907061. DOI: 10.1038/s41598-021-83967-7. View

11.

Senkal N, Meral R, Medetalibeyoglu A, Konyaoglu H, Kose M, Tukek T . Association between chronic ACE inhibitor exposure and decreased odds of severe disease in patients with COVID-19. Anatol J Cardiol. 2020; 24(1):21-29. PMC: 7414823. DOI: 10.14744/AnatolJCardiol.2020.57431. View

12.

Zhang A, Xing L, Zou J, Wu J . Shifting machine learning for healthcare from development to deployment and from models to data. Nat Biomed Eng. 2022; 6(12):1330-1345. DOI: 10.1038/s41551-022-00898-y. View

13.

Zhang X, Cai H, Hu J, Lian J, Gu J, Zhang S . Epidemiological, clinical characteristics of cases of SARS-CoV-2 infection with abnormal imaging findings. Int J Infect Dis. 2020; 94:81-87. PMC: 7270493. DOI: 10.1016/j.ijid.2020.03.040. View

14.

Dixon D, Sattar H, Moros N, Kesireddy S, Ahsan H, Lakkimsetti M . Unveiling the Influence of AI Predictive Analytics on Patient Outcomes: A Comprehensive Narrative Review. Cureus. 2024; 16(5):e59954. PMC: 11161909. DOI: 10.7759/cureus.59954. View

15.

Liang W, Liang H, Ou L, Chen B, Chen A, Li C . Development and Validation of a Clinical Risk Score to Predict the Occurrence of Critical Illness in Hospitalized Patients With COVID-19. JAMA Intern Med. 2020; 180(8):1081-1089. PMC: 7218676. DOI: 10.1001/jamainternmed.2020.2033. View

16.

Richardson S, Hirsch J, Narasimhan M, Crawford J, McGinn T, Davidson K . Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA. 2020; 323(20):2052-2059. PMC: 7177629. DOI: 10.1001/jama.2020.6775. View

17.

Hong W, Zhou X, Jin S, Lu Y, Pan J, Lin Q . A Comparison of XGBoost, Random Forest, and Nomograph for the Prediction of Disease Severity in Patients With COVID-19 Pneumonia: Implications of Cytokine and Immune Cell Profile. Front Cell Infect Microbiol. 2022; 12:819267. PMC: 9039730. DOI: 10.3389/fcimb.2022.819267. View

18.

Steyerberg E, Eijkemans M, Habbema J . Stepwise selection in small data sets: a simulation study of bias in logistic regression analysis. J Clin Epidemiol. 1999; 52(10):935-42. DOI: 10.1016/s0895-4356(99)00103-1. View

19.

Passarelli-Araujo H, Passarelli-Araujo H, Urbano M, Pescim R . Machine learning and comorbidity network analysis for hospitalized patients with COVID-19 in a city in Southern Brazil. Smart Health (Amst). 2022; 26:100323. PMC: 9485420. DOI: 10.1016/j.smhl.2022.100323. View

20.

Ebinger J, Achamallah N, Ji H, Claggett B, Sun N, Botting P . Pre-existing traits associated with Covid-19 illness severity. PLoS One. 2020; 15(7):e0236240. PMC: 7377468. DOI: 10.1371/journal.pone.0236240. View