Model-based Optimization of Ventilator Settings for Bedside Application J. Scherer, C. Schranz, A. Knörzer, K. Möller Institute of Technical Medicine, Furtwangen University, Germany jonas.scherer@googlemail.com Abstract: Mechanical ventilation is a life-saving therapy on intensive care units but has the risk of deploying additional pulmonary trauma due to non-personalized ventilator set- tings. Mathematical models can be used to derive individual- ized settings. Such an approach was implemented on a tablet computer receiving real-time data from a mechanical ventilator to compute model-based ventilator settings. Therefore, a mobile application displays real-time data measured by the mechanical ventilator and calculates sug- gestion of optimized ventilator settings in terms of minimal inspiration pressure and sufficient expiration time to be directly applied at the ventilator in pressure controlled venti- lation. With such an application, alveolar stress and the risk of intrinsic PEEP could be reduced directly at the bedside. Keywords: Android Systems, Mechanical Ventilation, Data-Communication Introduction Mechanical ventilation carries the risk of ventilator- induced lung injury (VILI), mainly caused by excessive stress on lung tissue [1]. To minimize the risk of VILI, ventilator settings should be personalized according to the individual patient’s physiology [2]. A recently developed model-based algorithm visualizes the nonlinear patient- specific relation of ventilator settings to meet a defined minute ventilation (MV) [3]. The algorithm requires real- time measurements of respiratory mechanics for model individualization to calculate optimized settings in terms of minimized inspiration pressure (p I ) and sufficient expi- ration time (t E ). Thus, un-physiological stress on lung tissue can be minimized and the risk of intrinsic positive end-expiratory pressure (PEEP) is reduced. To apply this application at the bedside, this paper presents an approach to implement this algorithm on a tablet computer for in- creased practicability. Methods Serial Connection Bluetooth Connection Mechanical Ventilator PC Tablet- Computer Figure 1: Communication concept between the mechanical ventilator and a tablet computer Hardware Setup The mechanical ventilator (EvitaXL, Dräger medical, Lübeck) is connected via a serial connection to a com- puter bridging the connection to a tablet computer (Mo- torola XOOM, Android 4.04) via Bluetooth (Figure 1). Software Concept The computer runs a LabVIEW application and provides a Bluetooth – port (Open_Port) for serial communication between tablet and PC. Secondly, the LabView- Application initiates and terminates the communication between PC and mechanical ventilator (Start_Com- munication, Stop_Communication) and transfers the status of the communication to the tablet. A Get_Data command requests ventilation information being the pa- tient’s resistance (R), compliance (C), expiratory time constant (τ E ) and applied PEEP measured by the mechani- cal ventilator. Get_Data packs the information into a protocol and forwards it to the tablet. The Get_Stream command receives real-time measurements from the me- chanical ventilator of airway pressure, flow rate and vol- ume sampled at 125 Hz. The communication between PC and mechanical ventilator is based on manufacturer spe- cific commands implemented in a provided DLL using the MEDIBUS protocol (Dräger Medical, Lübeck). The Android-based application was developed using Android SKD. When starting the application, the user is asked whether a Bluetooth connection should be estab- lished or an offline mode is preferred. The Bluetooth connection is established by sending the Start_Com- munication command to the PC. Afterwards the user can request ventilation information of R, C, τ E and PEEP for a subsequent optimization of ventilator settings. Secondly, real-time data can be requested to be handled in a service running in the background of the Android system. The received real-time data can be dynamically plotted in a separate tab (Fig. 2). Additionally, the data stream can be locally saved on the device for a subsequent analysis in the offline-mode. In this mode, the file with previously recorded data can be opened to proceed for additional planned offline analysis. This function offers the possibil- ity to simulate an input stream without heaving a Blue- tooth connection. Optimizing ventilator settings: Patient-specific respira- tory mechanics information of R, C and τ E can either be obtained by the mechanical ventilator (Get_Data) or by an online model-identification using the real-time data. R and C can be determined by identifying the 1 st order model of respiratory mechanics (FOM) to measurements of flow rate and airway pressure. τ E is estimated by fitting an exponential function to the flow signal during expira- Biomed Tech 2013; 58 (Suppl. 1) © 2013 by Walter de Gruyter · Berlin · Boston. DOI 10.1515/bmt-2013-4212 Unauthenticated Download Date | 2/10/19 1:57 PM