Telemetry is rapidly becoming standard practice for clinical studies. Remote monitoring may be performed at multiple locations of subjects over WiFi or 3G cell phone networks. Nevertheless, the accuracy and reliability of monitoring systems have still to be improved for clinical acceptability. Main issues are to enhance the battery lifetime of IoT-enabled devices and to allow the transmission of acceptable quality ECG signals. The objective is to propose an in silico-based method to optimize ECG telemetry systems by identifying critical telecommunication parameters and optimizing both battery lifetime and quality of transmitted ECG. 15 simulation parameters have been selected to test their criticality and three quality attributes have been examined: battery lifetime, number of received packets and ECG distortion. The methodology framework relies on the Quality-by-Design guidelines (ICH Q8-Q12). Its partial implementation was based on two consecutive sets of simulations. In a first step, a screening design of numerical experimentations (Plackett-Burman design) was carried out to identify the critical network parameters. In a second step, an optimization campaign, based on a central composite design, was implemented to identify the design space, i.e. the region of interest in which all the telecommunication parameters have to be kept fulfilling the quality specifications. Simulations were carried out in Omnet++ and the statistical analysis in the software environment R. For the number of packets received, two critical parameters were identified: message length and bit rate. For ECG distortion, the first two most critical factors are background noise power and energy detection of the radio receiver. For the battery lifetime, preliminary results tend to show that background noise power and bit rate are critical. This study demonstrates the relevance to use numerical simulations to design but also check the compatibility of a telemetry system with predefined specifications on quality of services in a given and constrained context. Such an approach can be implemented in the early steps of development and can save of lot of time and money by preventing malfunctions, scraps and posteriori restructurations of the telemetry networks.