Physiologic closed-loop controlled (PCLC) medical devices have the potential to enhance therapeutic precision and effectiveness. However, their complexity introduces potential failure modes that may pose risks to patients if not properly evaluated. This study assesses the effectiveness of a mathematical model of the cardiovascular system in predicting PCLC performance metrics through in vivo and in silico comparisons of blood pressure responses to automated fluid infusion. A closed-loop control system for regulating mean arterial pressure (MAP) was implemented using a custom Java-based software application on a tablet. The system operated in two control speed modes, ‘slow’ and ‘fast.’ The study was conducted in an animal laboratory using thirteen swine subjects (two were excluded due to hardware-related issues). Following intubation, anesthesia, and splenectomy, hemorrhage was induced until MAP reached 45 mmHg, followed by fluid resuscitation using the closed-loop controller targeting 70 mmHg. Arterial blood pressure waveforms were continuously recorded, and cardiac output and hematocrit measurements were taken every 15 min. The same control algorithm and speed modes were applied to in silico subjects generated with a mathematical model simulating cardiovascular responses to fluid perturbation while the same protocol was simulated. The mathematical model effectively predicted key PCLC performance metrics, including rise time,

This paper presents a virtual patient generator (VPG) intended to be used for pre-clinical in silico evaluation of autonomous vasopressor administration algorithms in the setting of experimentally induced vasoplegia. Our VPG consists of two main components: (i) a mathematical model that replicates physiological responses to experimental vasoplegia (induced by sodium nitroprusside (SNP)) and vasopressor resuscitation via phenylephrine (PHP) and (ii) a parameter vector sample generator in the form of a multi-dimensional probability density function (PDF) using which the parameters characterizing the mathematical model can be sampled. We developed and validated a mathematical model capable of predicting physiological responses to the administration of SNP and PHP. Then, we developed a parameter vector sample generator using a collective variational inference method. In a blind testing, the VPG developed by combining the two could generate a large number of realistic virtual patients (VPs) which could simulate physiological responses observed in all the experiments: on the average, 98.1% and 74.3% of the randomly generated VPs were physiologically legitimate and adequately replicated the test subjects, respectively, and 92.4% of the experimentally observed responses could be covered by the envelope formed by the subject-replicating VPs. In sum, the VPG developed in this paper may be useful for pre-clinical in silico evaluation of autonomous vasopressor administration algorithms.

Physiological closed-loop controlled (PCLC) medical devices, such as those designed for blood pressure regulation, can be tested for safety and efficacy in real-world clinical settings. However, relying solely on limited animal and clinical studies may not capture the diverse range of physiological conditions. Credible mathematical models can complement these studies by allowing the testing of the device against simulated patient scenarios. This research involves the development and validation of a low-order lumped-parameter mathematical model of the cardiovascular system’s response to fluid perturbation. The model takes rates of hemorrhage and fluid infusion as inputs and provides hematocrit and blood volume, heart rate, stroke volume, cardiac output and mean arterial blood pressure as outputs. The model was calibrated using data from 27 sheep subjects, and its predictive capability was evaluated through a leave-one-out cross-validation procedure, followed by independent validation using 12 swine subjects. Our findings showed small model calibration error against the training dataset, with the normalized root-mean-square error (NRMSE) less than 10% across all variables. The mathematical model and virtual patient cohort generation tool demonstrated a high level of predictive capability and successfully generated a sufficient number of subjects that closely resembled the test dataset. The average NRMSE for the best virtual subject, across two distinct samples of virtual subjects, was below 12.7% and 11.9% for the leave-one-out cross-validation and independent validation dataset. These findings suggest that the model and virtual cohort generator are suitable for simulating patient populations under fluid perturbation, indicating their potential value in PCLC medical device evaluation.

IntroductionTreating extensive burn injury requires an individually tailored resuscitation protocol that includes hourly-titrated intravenous fluid infusion to avert both hypovolemic shock and edema. Due to the complexity of burn pathophysiology and significant variability in treatment protocols, there is an ongoing effort to optimize burn resuscitation. The goal of this work is to contribute to this effort by developing a mathematical model of burn pathophysiology and resuscitation for in silico testing of burn resuscitation protocols and decision-support systems.MethodsIn our previous work, we developed and validated a mathematical model consisting of volume kinetics, burn-induced perturbations, and kidney function. In this work, we expanded our previous mathematical model to incorporate novel mathematical models of cardiovascular system and hormonal system (renin-angiotensin-aldosterone (RAAS) system and antidiuretic hormone) which affect blood volume and pressure regulation. We also developed a detailed mathematical model of kidney function to regulate blood volume, pressure, and sodium levels, including components for glomerular filtration rate, reabsorption rates in nephron tubules, Tubuglomerular feedback, and myogenic mechanisms. We trained and validated the expanded mathematical model using experimental data from 15 pigs and 9 sheep with extensive burns to quantitatively evaluate its prediction accuracy for hematocrit, cardiac output, mean arterial pressure, central venous pressure, serum sodium levels, and urinary output. We then trained and tested the mathematical model using a clinical dataset of 233 human burn patients with demographic data and urinary output measurements.ResultsThe mathematical model could predict all tested variables very well, while internal variables and estimated parameters were consistent with the literature.DiscussionTo the best of our knowledge, this is the first mathematical model of burn injury and resuscitation which is extensively validated to replicate actual burn patients. Hence, this in silico platform may complement large animal pre-clinical testing of burn resuscitation protocols. Beyond its primary purpose, the mathematical model can be used as a training tool for healthcare providers delivering insight into the pathophysiology of burn shock, and offering novel mathematical models of human physiology which can be independently used for other purposes and contexts.

Objective: To develop a high-fidelity mathematical model intended to replicate the cardiovascular (CV) responses of a critically ill patient to vasoplegic shock-induced hypotension and vasopressor therapy. Methods: The mathematical model consists of a lumped-parameter CV physiology model with baroreflex modulation feedback and a phenomenological dynamic dose-response model of a vasopressor. The adequacy of the proposed mathematical model was investigated using an experimental dataset acquired from 10 pigs receiving phenylephrine (PHP) therapy after vasoplegic shock induced via sodium nitroprusside (SNP). Results: Upon calibration, the mathematical model could (i) faithfully replicate the effects of PHP on dynamic changes in blood pressure (BP), cardiac output (CO), and systemic vascular resistance (SVR) (root-mean-squared errors between measured and calibrated mathematical responses: mean arterial BP 2.5+/-1.0 mmHg, CO 0.2+/-0.1 lpm, SVR 2.4+/-1.5 mmHg/lpm; r value: mean arterial BP 0.96+/-0.01, CO 0.65+/-0.45, TPR 0.92+/-0.10) and (ii) predict physiologically plausible behaviors of unmeasured internal CV variables as well as secondary baroreflex modulation effects. Conclusion: This mathematical model is perhaps the first of its kind that can comprehensively replicate both primary (i.e., direct) and secondary (i.e., baroreflex modulation) effects of a vasopressor drug on an array of CV variables, rendering it ideally suited to pre-clinical virtual evaluation of the safety and efficacy of closed-loop control algorithms for autonomous vasopressor administration once it is extensively validated. Significance: This mathematical model architecture incorporating both direct and baroreflex modulation effects may generalize to serve as part of an effective platform for high-fidelity in silico simulation of CV responses to vasopressors during vasoplegic shock.

This paper concerns the design and rigorous in silico evaluation of a closed-loop hemorrhage resuscitation algorithm with blood pressure (BP) as controlled variable. A lumped-parameter control design model relating volume resuscitation input to blood volume (BV) and BP responses was developed and experimentally validated. Then, three alternative adaptive control algorithms were developed using the control design model: (i) model reference adaptive control (MRAC) with BP feedback, (ii) composite adaptive control (CAC) with BP feedback, and (iii) CAC with BV and BP feedback. To the best of our knowledge, this is the first work to demonstrate model-based control design for hemorrhage resuscitation with readily available BP as feedback. The efficacy of these closed-loop control algorithms was comparatively evaluated as well as compared with an empiric expert knowledge-based algorithm based on 100 realistic virtual patients created using a well-established physiological model of cardiovascular (CV) hemodynamics. The in silico evaluation results suggested that the adaptive control algorithms outperformed the knowledge-based algorithm in terms of both accuracy and robustness in BP set point tracking: the average median performance error (MDPE) and median absolute performance error (MDAPE) were significantly smaller by >99% and >91%, and as well, their interindividual variability was significantly smaller by >88% and >94%. Pending in vivo evaluation, model-based control design may advance the medical autonomy in closed-loop hemorrhage resuscitation.

AbstractPhysiological closed-loop controlled (PCLC) medical devices monitor and automatically adjust the patient’s condition by using physiological variables as feedback, ideally with minimal human intervention to achieve the target levels set by a clinician. PCLC devices present a challenge when it comes to evaluating their performance, where conducting large clinical trials can be expensive. Virtual physiological patients simulated by validated mathematical models can be utilized to obtain pre-clinical evidence of safety and assess the performance of the PCLC medical device during normal and worst-case conditions that are unlikely to happen in a limited clinical trial. A physiological variable that plays a major role during fluid resuscitation is heart rate (HR). For in silico assessment of PCLC medical devices regarding fluid perturbation, there is currently no mathematical model of HR validated in terms of its predictive capability performance. This paper develops and validates a mathematical model of HR response using data collected from sheep subjects undergoing hemorrhage and fluid infusion. The model proved to be accurate in estimating the HR response to fluid perturbation, where averaged between 21 calibration datasets, the fitting performance showed a normalized root mean square error (NRMSE) of $$7.41 \pm 2.8 \%$$ 7.41 ± 2.8 % . The model was also evaluated in terms of model predictive capability performance via a leave-one-out procedure (21 subjects) and an independent validation dataset (6 subjects). Two different virtual cohort generation tools were used in each validation analysis. The generated envelope of virtual subjects robustly met the defined acceptance criteria, in which $$95\%$$ 95 % of the testing datasets presented simulated HR patterns that were within a deviation of 50% from the observed data. In addition, out of 16000 and 18522 simulated subjects for the leave-one-out and independent datasets, the model was able to generate at least one virtual subject that was close to the real subject within an error margin of $$9.56 \pm 3.15\%$$ 9.56 ± 3.15 % and $$11.1 \pm 1.22\%$$ 11.1 ± 1.22 % NRMSE, respectively. In conclusion, the model can generate valid virtual HR physiological responses to fluid perturbation and be incorporated into future non-clinical simulated testing setups for assessing PCLC devices intended for fluid resuscitation.

Background Organ function is known to decline with age. Optimizing cardiac, pulmonary and renal function in older adults has led to significant improvements in perioperative care. However, when substantial blood loss and fluid shifts occur, perioperative outcomes still remains poor, especially in older adults. We suspect that this could be due to age-related changes in endothelial function—an organ controlling the transport of fluid and solutes. The capillary filtration coefficient (CFC) is an important determinant of fluid transport. The CFC can be measured in vivo, which provides a tool to estimate endothelial barrier function. We have previously shown that the CFC increases when giving a fluid bolus resulting in increased vascular and extravascular volume expansion, in young adults. This study aimed to compare the physiologic determinants of fluid distribution in young versus older adults so that clinicians can best optimize perioperative fluid therapy. Methods Ten healthy young volunteers (ages 21–35) and nine healthy older volunteers (ages 60–75) received a 10 mL/kg fluid bolus over the course of twenty minutes. Hemodynamics, systolic and diastolic heart function, fluid volumetrics and microcirculatory determinants were measured before, during, and after the fluid bolus. Results Diastolic function was reduced in older versus younger adults before and after fluid bolus ( P < 0.01). Basal CFC and plasma oncotic pressure were lower in the older versus younger adults. Further, CFC did not increase in older adults following the fluid bolus, whereas it did in younger adults ( p < 0.05). Cumulative urinary output, while lower in older adults, was not significantly different ( p = 0.059). Mean arterial pressure and systemic vascular resistance were elevated in the older versus younger adults ( p < 0.05). Conclusion Older adults show a less reactive CFC to a fluid bolus, which could reduce blood to tissue transport of fluid. Diastolic dysfunction likely contributes to fluid maldistribution in older adults.

Physiological closed-loop controlled (PCLC) medical devices are complex systems integrating one or more medical devices with a patient’s physiology through closed-loop control algorithms; introducing many failure modes and parameters that impact performance. These control algorithms should be tested through safety and efficacy trials to compare their performance to the standard of care and determine whether there is sufficient evidence of safety for their use in real care setting. With this aim, credible mathematical models have been constructed and used throughout the development and evaluation phases of a PCLC medical device to support the engineering design and improve safety aspects. Uncertainties about the fidelity of these models and ambiguities about the choice of measures for modeling performance need to be addressed before a reliable PCLC evaluation can be achieved. This research develops tools for evaluating the accuracy of physiological models and establishes fundamental measures for predictive capability assessment across different physiological models. As a case study, we built a refined physiological model of blood volume (BV) response by expanding an original model we developed in our prior work. Using experimental data collected from 16 sheep undergoing hemorrhage and fluid resuscitation, first, we compared the calibration performance of the two candidate physiological models, i.e., original and refined, using root-mean-squared error (RMSE), Akiake information criterion (AIC), and a new multi-dimensional approach utilizing normalized features extracted from the fitting error. Compared to the original model, the refined model demonstrated a significant improvement in calibration performance in terms of RMSE (9%, P = 0.03) and multi-dimensional measure (48%, P = 0.02), while a comparable AIC between the two models verified that the enhanced calibration performance in the refined model is not due to data over-fitting. Second, we compared the physiological predictive capability of the two models under three different scenarios: prediction of subject-specific steady-state BV response, subject-specific transient BV response to hemorrhage perturbation, and leave-one-out inter-subject BV response. Results indicated enhanced accuracy and predictive capability for the refined physiological model with significantly larger proportion of measurements that were within the prediction envelope in the transient and leave-one-out prediction scenarios (P < 0.02). All together, this study helps to identify and merge new methods for credibility assessment and physiological model selection, leading to a more efficient process for PCLC medical device evaluation.

Closed-loop automation of critical care therapy has the potential to reduce the workload of clinical personnel while maintaining the quality of care. In the real-world clinical arena, critically ill patients receive multiple medical treatments. However, existing body of work has predominantly focused on closed-loop automation of isolated individual treatments. How these individual treatment loops interact with each other has not been investigated. The goal of this work is to garner insights on the safety of critical care therapy and potential deleterious conflicts therein when multiple isolated and individually closed-loop controlled medical treatments act upon a patient, using a case study of hemorrhage resuscitation and intravenous propofol sedation. For this purpose, a physiological model of a critically ill patient was developed and experimentally validated to describe the collective cardiovascular and pharmacological effects of these treatments. Then, isolated and individually closed-loop controlled hemorrhage resuscitation and intravenous propofol sedation treatments were simultaneously applied to the physiological model and their interactive behavior was investigated. The results showed that (i) the influence of one treatment on the other must be taken into account in selecting treatment set point to maintain the safety of overall therapy, and that (ii) information sharing between control loops may enhance the efficacy and robustness of individual treatment loops. In sum, it was concluded that hemorrhage resuscitation and intravenous propofol sedation treatments may benefit from coordination both at the set point and the loop levels. The conclusion may generalize to a wide spectrum of multiple closed-loop controlled medical treatments.

Objective: Individual physiological experiments typically provide useful but incomplete information about a studied physiological process. As a result, inferring the unknown parameters of a physiological model from experimental data is often challenging. The objective of this paper is to propose and illustrate the efficacy of a collective variational inference (C-VI) method, intended to reconcile low-information and heterogeneous data from a collection of experiments to produce robust personalized and generative physiological models. Methods: To derive the C-VI method, we utilize a probabilistic graphical model to impose structure on the available physiological data, and algorithmically characterize the graphical model using variational Bayesian inference techniques. To illustrate the efficacy of the C-VI method, we apply it to a case study on the mathematical modeling of hemorrhage resuscitation. Results: In the context of hemorrhage resuscitation modeling, the C-VI method could reconcile heterogeneous combinations of hematocrit, cardiac output, and blood pressure data across multiple experiments to obtain (i) robust personalized models along with associated measures of uncertainty and signal quality, and (ii) a generative model capable of reproducing the physiological behavior of the population. Conclusion: The C-VI method facilitates the personalized and generative modeling of physiological processes in the presence of low-information and heterogeneous data. Significance: The resulting models provide a solid basis for the development and testing of interpretable physiological monitoring, decision-support, and closed-loop control algorithms.

Physiological closed-loop controlled (PCLC) medical devices are systems integrating medical devices with a patient's physiology through closed-loop control algorithms. Mathematical models are used throughout the development and evaluation of PCLC devices. Uncertainties about the fidelity of these models need to be addressed before achieving a reliable PCLC evaluation. To identify the best candidate model for in silico evaluation of PCLC devices, this research develops tools for assessing and comparing accuracy and predictive capability performance across multiple models. Clinical Relevance— This research develops tools to support the safety and effectiveness of PCLC medical devices, thus promoting public health.

Physiological measurements are an integral part of many established and emerging engineering and biomedical applications that involve physiological modeling, physiological state estimation, and physiological closed loop control. In practice, such measurements exhibit a large degree of variability, which is apparent at multiple levels, including disturbances acting on measured signals and unexpected physiological behavior in certain individuals. In this short paper, we present an inference-based approach to estimating the atypicality of an individual's physiological data both at the level of measurement and physiological behavior. For this purpose, we use data from a cohort of subjects to infer, simultaneously, model representations for measurement disturbances and atypicality of physiological behavior. Using a case study on hematocrit (HCT), cardiac output (CO), and mean arterial pressure (MAP) measurements in response to hemorrhage and colloid infusions, we discuss the merits of the presented approach in deriving reliable subject atypicality and signal quality indicators for physiological data.

Objective: Existing burn resuscitation protocols exhibit a large variability in treatment efficacy. Hence, they must be further optimized based on comprehensive knowledge of burn pathophysiology. A physics-based mathematical model that can replicate physiological responses in diverse burn patients can serve as an attractive basis to perform non-clinical testing of burn resuscitation protocols and to expand knowledge on burn pathophysiology. We intend to develop, optimize, validate, and analyze a mathematical model to replicate physiological responses in burn patients. Methods: Using clinical datasets collected from 233 burn patients receiving burn resuscitation, we developed and validated a mathematical model applicable to computer-aided in-human burn resuscitation trial and knowledge expansion. Using the validated mathematical model, we examined possible physiological mechanisms responsible for the cohort-dependent differences in burn pathophysiology between younger versus older patients, female versus male patients, and patients with versus without inhalational injury. Results: We demonstrated that the mathematical model can replicate physiological responses in burn patients associated with wide demographic characteristics and injury severity, and that an increased inflammatory response to injury may be a key contributing factor in increasing the mortality risk of older patients and patients with inhalation injury via an increase in the fluid retention. Conclusion: We developed and validated a physiologically plausible mathematical model of volume kinetic and kidney function after burn injury and resuscitation suited to in-human application. Significance: The mathematical model may provide an attractive platform to conduct non-clinical testing of burn resuscitation protocols and test new hypotheses on burn pathophysiology.

Introduction: 'Fluid creep' or excessive fluid delivered to burn patients during early resuscitation has been suggested by several studies from individual burn centers. Methods: We performed a Medline search from 1980 to 2015 in order to identify studies of burn patients predominantly resuscitated with lactated Ringers with infusion adjusted per urinary output. Data was abstracted for 48 publications (3196 patients) that met entry criteria. Results: Higher resuscitation volumes compared to Parkland estimates were reported, but the trend of increasing resuscitation volumes over the last 30 years is not supported by regression of total fluid infused versus year of study. Mean 24h fluid infused for all studies was 5.2 +/- 1.1 mL/kg per %TBSA. The mean 24 h urinary output reported in 30 studies was 1.2 +/- 0.5 mL/kg per hr. Burns with inhalation injuries (5 studies) received significantly more fluid than non-inhalation injured burn patients (5.0 +/- 1.3 versus 3.9 +/- 0.9 mL/kg per %TBSA). Fluid infused and urinary outputs were similar for adults and pediatric patients. The most striking finding of our analyses was the great ranges of the means and high standard deviations of volumes infused compared to the original Baxter publication that introduced the Parkland formula Conclusions: These analyses suggest that burn units currently administer volumes larger than Parkland formula with great patient variability. Individual patient hourly data is needed to better understand the record of burn resuscitation and Fluid Creep. (C) 2019 Elsevier Ltd and ISBI. All rights reserved.

This paper presents a data-driven approach to generating virtual patients using mathematical models of physiological processes. Such models often contain a large number of tunable parameters that must be calibrated to capture the observed characteristics of each real patient in a dataset. By sampling from this parameter space, potentially new virtual patients can be generated. However, it is often the case that the resulting set of virtual patients contains members that exhibit physiologically unrealistic behavior. In the present work, we employ a practically important case study on the modeling of cardiovascular responses to hemorrhage and fluid resuscitation in order to demonstrate that subject-specific characteristics observed in a dataset can be alternatively represented within a highly compressed latent parameter space without significant losses in calibration error for each real patient. Then, we show that by sampling from this latent parameter space, it is possible to generate new virtual patients that also exhibit physiologically realistic behavior.

Background: Interest in incorporating research into the medical school curriculum has grown over the years. One of the challenges involved with providing research to medical students is developing programs that allow a large number of students to perform research. This involves securing faculty to mentor students in the design of research projects. In order to accommodate students with research interests, well-established research programs must be implemented. Objective: This article describes the design and implementation of a curriculum-based research program for medical students at the University of Texas Medical Branch (UTMB) at Galveston. The main objective of this article is to describe the program for the purpose of assisting other medical schools to develop a similar student research program. Design: At UTMB we established a Medical Student Summer Research Program (MSSRP) that occurred between the first year and the second year of medical school. Between the years 2000-2017, MSSRP accommodated a minimum of 39 and a maximum of 90 students during an 8 week period. Two surveys were conducted to collect students' views on how MSSRP affected their interest in research. We performed a proportion statistical analysis on the data from both surveys in order to determine the significance of the responses. Results: The benefit of MSSRP is that it provided medical students with an exposure to research. According to the proportions test, the responses were statistically significant with 85% of 26 third and fourth year students stating they would continue to incorporate research into their medical careers; 75% stating that MSSRP increased their interest in research; and 85% responding that MSSRP helped them to understand research methodology. Conclusions: MSSRP is a curriculum-based program that provides a framework to other medical institutions interested in the development of similar student research programs and provides students the exposure and option to continue with research as a component of their medical profession.

This paper investigates a novel regularized system identification approach to physiological modeling using limited data. The proposed approach operates in two steps: 1) limited data from individual subjects are consolidated and leveraged to determine a population-average physiological model; then, 2) a subject-specific model for an individual subject is derived from a regularized system identification procedure whose objective is to reconcile the model's capability to predict individual-specific behavior and to retain typical population-representative trends. This is achieved by embedding a regularizing condition into the cost function for system identification that enforces parsimony in parametric deviation from the population-average model. A few unique advantages of the proposed approach are that 1) it offers superior predictive accuracy in both measured as well as unmeasured physiological system responses when compared to a standard system identification approach; and 2) it provides high-sensitivity parameters in the model associated with each individual subject, thus potentially eliminating the necessity for post-hoc parametric sensitivity analysis. Merits and limitations of the proposed regularized approach are illustrated with a real world case study on physiological modeling of hemodynamics in response to burn injury and resuscitation.

A 1st-principles-inspired physiological model was developed that can predict the redistribution of fluid and albumin in the patient after thermal injury. The key components of the model include volume kinetics, renal function, and burn-induced perturbations in circulatory physiology. The model was developed in such a way that it can be fully characterized by the patient weight, height, and total burn surface area (TBSA) as well as a small set of model parameters to be individualized to predict the patient's response to thermal injury and resuscitation. Based on the underlying physics and constraints of organs, new mathematical functions were developed to describe the renal function and lymph return. A novel approach was employed to implement the effect of burn on the pertinent physiology of patients' body. The model was identified and tested using two groups of sheep undergoing burn injury and resuscitation. The results suggested that the model can boast acceptable accuracy in reproducing experimentally observed plasma volume and urinary output despite its simplicity. Such a model, after meeting the requirements on accuracy and physiological credibility, may facilitate the validation and testing of burn resuscitation protocols and closed-loop decision support systems.