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.

A sensor‐mediated strategy was applied to a laboratory‐scale granular sludge reactor (GSR) to demonstrate that energy‐efficient inorganic nitrogen removal is possible with a dilute mainstream wastewater. The GSR was fed a dilute wastewater designed to simulate an A‐stage mainstream anaerobic treatment process. DO, pH, and ammonia/nitrate sensors measured water quality as part of a real‐time control strategy that resulted in low‐energy nitrogen removal. At a low COD (0.2 kg m −3 day −1 ) and ammonia (0.1 kg‐N m −3 day −1 ) load, the average degree of ammonia oxidation was 86.2 ± 3.2% and total inorganic nitrogen removal was 56.7 ± 2.9% over the entire reactor operation. Aeration was controlled using a DO setpoint, with and without residual ammonia control. Under both strategies, maintaining a low bulk oxygen level (0.5 mg/L) and alternating aerobic/anoxic cycles resulted in a higher level of nitrite accumulation and supported shortcut inorganic nitrogen removal by suppressing nitrite oxidizing bacteria. Furthermore, coupling a DO setpoint aeration strategy with residual ammonia control resulted in more stable nitritation and improved aeration efficiency. The results show that sensor‐mediated controls, especially coupled with a DO setpoint and residual ammonia controls, are beneficial for maintaining stable aerobic granular sludge. Practitioner points Tight sensor‐mediated aeration control is need for better PN/A. Low DO intermittent aeration with minimum ammonium residual results in a stable N removal. Low DO aeration results in a stable NOB suppression. Using sensor‐mediated aeration control in a granular sludge reactor reduces aeration cost.

The objective of our study was to reanalyse the Ethiopia STEPwise approach to Surveillance Noncommunicable Disease Risk Factors survey (NCD STEPS), using causal path diagrams constructed using expert subject matter knowledge in conjunction with graphical model theory to map the underlying causal network of modifiable factors associated with prediabetes/diabetes and hypertension. We used data from the 2015 Ethiopia NCD STEPS representative cross‐sectional survey (males; n = 3977 and females; n = 5823 aged 15–69 years) and performed directed acyclic graph‐informed logistic regression analyses. In both sexes, a 1‐unit higher in body mass index (BMI) and waist circumference (WC) were positively associated with prediabetes/diabetes (BMI: males: adjusted odds ratio [aOR]: 1.07 [95% confidence interval: 1.0, 1.1], females aOR: 1.03 [1.0, 1.1]; WC: males: aOR: 1.1 [0.9, 1.2], females: aOR: 1.2 [1.1, 1.3]) and hypertension (BMI: males: aOR: 1.2 [1.1, 1.2], females aOR: 1.1 [1.0, 1.1]; WC: males: aOR: 1.6 [1.4, 1.8], females: aOR: 1.3 [1.2, 1.5]). Although residing in urban settings was associated with higher odds of hypertension in both males (aOR: 1.79 [1.49, 2.16]) and females (aOR: 1.70 [1.49, 1.95]), it was only associated with prediabetes/diabetes in males (aOR: 1.56 [1.25, 1.96]). Males and females in pastoralist areas had lower odds of prediabetes/diabetes compared with their agrarian counterparts (males: aOR: 0.27 [0.14, 0.52], females: aOR: 0.31 [0.16, 0.58]). Physical activity was associated with lower odds of prediabetes/diabetes among females (aOR: 0.75 [0.58, 0.97]). Other diet‐related modifiable factors such as consumption of fruit and vegetable, alcohol or salt were not associated with either prediabetes/diabetes or hypertension. Our findings highlight the need to implement interventions that prevent overweight/obesity and nutrition‐related NCDs, particularly in urban areas.

No abstract