Effectively Calculating Model Systems that Interact

Verónica Quintanilla^*
*Correspondence: Verónica Quintanilla, Department of Didactics of Mathematics, University of Malaga, Malaga, Spain, Email:

Introduction

Mathematical models can be used to predict human physiological processes. These predictions might be used by Medical Decision Support Systems (MDSS) to improve therapy settings. These predictions ought to take into account other organ systems of the human body when treating critically ill patients who require mechanical ventilation. We combine components from three model families into a previously presented framework: Gas exchange, cardiovascular dynamics, and respiratory mechanics. The ability to use model combinations of moderately complex sub models in an MDSS is constrained by the computational cost of doing so. As a result, a decoupled computing strategy was developed that enables individual evaluation of each sub model. Separate calculations cannot allow for direct model interaction. As a result, estimates must take the place of interface signals. Utilizing the hierarchical structure of the implemented model families, the iterative improvement of these estimates involves increasing model detail in each iteration.

Description

Re-enactment mistake united to a base after three emphasess. When compared to the original common coupled computing method, the maximum simulation error was found to be 1.44 percent. The measurement noise typically found in clinical data was found to be lower than the simulation error. One iteration reduced simulation time by 34 percent, while three iterations reduced it by 13 percent. Model-based decision support appears to be applicable to moderately complex model combinations following the proposed calculation scheme. A fundamental strategy for gaining insight into the human body's physiological processes is the application of mathematical models. Clinicians may be able to optimize therapy settings for each individual patient with the assistance of this knowledge. Mechanical ventilation is a good illustration of this process of optimization. Most of the time, artificial respiration is a life-saving intervention. However, if the ventilator settings are chosen incorrectly, this therapy can result in secondary lung injury, also known as "ventilator induced" lung injury (VILI).The clinician must strike the best balance between the patient's benefits and risks in order to avoid this harm. Sadly, the medical professional in charge does not have the time to investigate the intricate nonlinear interactions between the respiratory system and the ventilator or to keep a constant eye on the patient. Without knowing the specific pathophysiology of each patient, it is impossible to adjust the ventilator settings to their highest potential [1,2].

Management of ventilators isn't the only field that benefits from physiological simulation. Medical Decision Support Systems (MDSS) may utilize the results of simulations provided by mathematical models in conjunction with active exploration and artificial intelligence to provide deeper insights or recommend optimized therapy settings. For example, researchers have successfully demonstrated the benefit of such simulations in glucose control in diabetic patients or drug dosing in anesthesia. However, a broader context is required in addition to human respiratory mechanics in order to accurately predict a patient's response to changes in the ventilation regimen. At the very least, cardiovascular dynamics and gas exchange should be included. Different simulation focuses may be required depending on the clinical situation. As a result, a system that permits dynamic reconfiguration of the model system or an all-encompassing, demanding model could be used to create a patient simulation of the interaction between physiological processes. Consider the following scenario as evidence of the significance of a system with multiple physiological processes that can be dynamically reconfigured: A patient with ARDS (Acute Respiratory Distress Syndrome) who is mechanically ventilated has a low oxygen saturation and a high etCO₂ (end-tidal CO₂) level. In ventilator settings, the clinician may increase respiration rate to reduce CO₂ levels in the blood of the patients. Inspired oxygen or tidal volume can be increased to improve oxygen saturation.

However, the clinician can only modify ventilator settings through trial and error and experience. For the patient, this could result in undesirable side effects like intrinsic PEEP, VILI, and alveolar cycling. The clinician may be able to calculate the outcome of potential new ventilator settings or even query optimized settings in order to achieve a predetermined objective, such as lower et CO2 and higher oxygen saturation, by simulating the patient's reaction. After ventilator settings have changed, dynamic models formulated in differential equations capture patient state transitions. If the new equilibrium state does not meet expectations, corrective actions can be taken before stabilization into the subsequent steady states. An automated controller's reaction time would be reduced, and medical staff would receive early feedback. By simulating the results of various settings in relation to pre-defined penalty functions, decision support systems can determine the optimal settings. Model parameters must be tailored to each individual patient in order to simulate them. If this identification procedure yields insufficient results, it could be because the patient's specific physiology was not captured by the model. The simulation results would therefore improve if an appropriate respiratory mechanics model was chosen. As a result, a model family should be made up of models with different levels of complexity and focus in an interconnected system of models [3-5].

Conclusion

By allowing the dynamic generation of complex interacting model systems, a framework that facilitates the aforementioned adaptations was proposed in earlier work. It connects submodels from three distinct model families: cardiovascular dynamics, gas exchange, and respiratory mechanics. The various models in each of the implemented model families vary in complexity. We have created a number of interfaces that allow the employed submodels to exchange parameter values, making interaction possible. The model blends tried with this structure showed physiologically conceivable outcomes.

References

1. Fleissner, Gerta, Elke Holtkamp‐Rötzler, Marianne Hanzlik and Michael Winklhofer et al. "Ultrastructural analysis of a putative magnetoreceptor in the beak of homing pigeons." J Compar Neurol 458 (2003): 350-360.

Google Scholar, Crossref, Indexed at

2. Alerstam, Thomas. "Conflicting evidence about long-distance animal navigation." Science 313 (2006): 791-794.

Google Scholar, Crossref, Indexed at

3. Biskup, T., Schleicher, E., Okafuji, A., and Link, G., Hitomi et al (2009). Direct observation of a photoinduced radical pair in a cryptochrome blue-light photoreceptor. Angew Chem Int Ed 48, 404– 407.

Google Scholar, Crossref, Indexed at

4. Bonadonna, F., C. Bajzak, S. Benhamou, K. Igloi, P. Jouventin, H. P. Lipp, and G. Dell'Omo. "Orientation in the wandering albatross: interfering with magnetic perception does not affect orientation performance." Proc Royal Soc B: Biological Sciences 272 (2005): 489-495.

Google Scholar, Crossref, Indexed at

5. Åkesson, Susanne, Jens Morin, Rachel Muheim, and Ulf Ottosson. "Dramatic orientation shift of white-crowned sparrows displaced across longitudes in the high arctic." Current Biology 15 (2005): 1591-1597.

Google Scholar, Crossref, Indexed at