Modeling and prediction of pressure injury in hospitalized patients using artificial intelligence

Abstract

Background Hospital-acquired pressure injuries (PIs) induce significant patient suffering, inflate healthcare costs, and increase clinical co-morbidities. PIs are mostly due to bed-immobility, sensory impairment, bed positioning, and length of hospital stay. In this study, we use electronic health records and administrative data to examine the contributing factors to PI development using artificial intelligence (AI). Methods We used advanced data science techniques to first preprocess the data and then train machine learning classifiers to predict the probability of developing PIs. The AI training was based on large, incongruent, incomplete, heterogeneous, and time-varying data of hospitalized patients. Both model-based statistical methods and model-free AI strategies were used to forecast PI outcomes and determine the salient features that are highly predictive of the outcomes. Results Our findings reveal that PI prediction by model-free techniques outperform model-based forecasts. The performance of all AI methods is improved by rebalancing the training data and by including the Braden in the model learning phase. Compared to neural networks and linear modeling, with and without rebalancing or using Braden scores, Random forest consistently generated the optimal PI forecasts. Conclusions AI techniques show promise to automatically identify patients at risk for hospital acquired PIs in different surgical services. Our PI prediction model provide a first generation of AI guidance to prescreen patients at risk for developing PIs. Clinical impact This study provides a foundation for designing, implementing, and assessing novel interventions addressing specific healthcare needs. Specifically, this approach allows examining the impact of various dynamic, personalized, and clinical-environment effects on PI prevention for hospital patients receiving care from various surgical services.

An Integrated System of Braden Scale and Random Forest Using Real-Time Diagnoses to Predict When Hospital-Acquired Pressure Injuries (Bedsores) Occur

“…Machine Learning (ML) has been utilized to predict HAPI patients before occurrence using patients’ Electronic Health Records (EHR) as an adjunct to clinical assessment, which can further narrow down which patients are at risk [ 6 , 7 , 8 , 9 ]. In the last 15 years, 30 studies have answered who will develop HAPIs by utilizing classic machine and deep learning algorithms [ 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 ]. Two studies adopted Grid Search (GS) to optimize the hyperparameters of ML [ 10 , 22 ].…”

“…Two studies adopted Grid Search (GS) to optimize the hyperparameters of ML [ 10 , 22 ]. In most studies, random oversampling, undersampling, and Synthetic Minority Oversampling Techniques (SMOTE) were adopted as oversampling techniques [ 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 38 ]. Nevertheless, one study adopted cost-sensitive learning [ 22 ] to deal with the unbalanced dataset.…”

Estratégias de pronação em pacientes com síndromes respiratórias agudas em unidades de terapia intensiva

“…Pacientes hospitalizados acometidos por SDRA tendem a desenvolver problemas de saúde associados ao imobilismo no leito, que por sua vez é causa de alterações fisiológicas no corpo, sendo fator de risco para trombose venosa profunda (TVP), encurtamento e atrofia muscular, rigidez articular, perda de peso, lesões por pressão, dentre outros, (Parola et al, 2021) (Sartori et al, 2021); (Anderson et al, 2021). Dalmedico et al (2017) afirma que a pronação em conjunto com técnicas ventilatórias protetoras por tempo entre 16 e 20 horas em pacientes com síndrome do desconforto respiratório agudo, com relação PaO2 /FiO2 inferior à 150 mm/Hg, resulta em redução significativa da taxa de mortalidade.…”

Risk profiling in the prevention and treatment of chronic wounds using artificial intelligence

“…6 Therefore, this complex web of factors includes other health information outside of the wound itself. [15][16][17] In today's fast-moving technologically enhanced world, diagnosis by machines is still in its evolution, although, growing rapidly in many facets of clinical care. 12,18 Wound care needs to adapt to this changing world to develop highly accurate AI-based decisionsupport applications to improve patient care.…”

“…This complexity arises from the complex interaction of many interrelated factors, both wound and patient related. As often stated by those who care for patients with wounds, “it's not about the hole in the patient but rather the whole patient.” 6 Therefore, this complex web of factors includes other health information outside of the wound itself 15‐17 …”