<p style="text-align: justify;">Healthcare facilities have been improving their information systems over the past few&nbsp;<span style="font-size: 12px;">years. Such improvements led to the creation of a multitude of dierent applications&nbsp;</span><span style="font-size: 12px;">essential to the facilities services. Associated with the various applications, there's&nbsp;</span><span style="font-size: 12px;">also a considerable amount of information being produced and stored throughout&nbsp;</span><span style="font-size: 12px;">the facility. Such data constitutes a privileged way of inferring past and current&nbsp;</span><span style="font-size: 12px;">performance metrics of a given healthcare facility for it's dierent activity domains.&nbsp;</span><span style="font-size: 12px;">However, complex challenges arise when trying to gather all the dierent data from&nbsp;</span><span style="font-size: 12px;">all the systems scattered throughout the facility.</span></p><p style="text-align: justify;">We present a proposal for a system capable of displaying production metrics in a&nbsp;<span style="font-size: 12px;">healthcare facility by passively extracting IP packets from the network and reconstruction&nbsp;</span><span style="font-size: 12px;">TCP streams containing HL7 compliant messages and other eHealth relevant&nbsp;</span><span style="font-size: 12px;">network protocols. Based on those messages our system is able to extract meaningful&nbsp;</span><span style="font-size: 12px;">data and with it, it is possible to produce a knowledge database for a given healthcare&nbsp;</span><span style="font-size: 12px;">facility.&nbsp;</span><span style="font-size: 12px;">The HL7 messages moving over the network contain information that can be used to&nbsp;</span><span style="font-size: 12px;">assess many relevant production metrics for a given infrastructure. The challenge of&nbsp;</span><span style="font-size: 12px;">having to query a considerable amount of dierent systems in order to gather such data&nbsp;</span><span style="font-size: 12px;">can be solved by passively extracting packets containing HL7 standardized messages&nbsp;</span><span style="font-size: 12px;">or other eHealth related protocols directly from the network.</span></p><p style="text-align: justify;">We have deployed our system in a large healthcare facility located in Porto, Portugal&nbsp;<span style="font-size: 12px;">where we've been passively extracting HL7 messages from their network infrastructure.&nbsp;</span><span style="font-size: 12px;">Our system extracts and analyses a daily average of 44,000 HL7 messages with several&nbsp;</span><span style="font-size: 12px;">peaks of 1,100 messages per minute. Based on such network trac, our system has&nbsp;</span><span style="font-size: 12px;">been able to infer the daily distribution of healthcare related activities such as lab&nbsp;</span><span style="font-size: 12px;">orders, appointment scheduling and also billing information, among other relevant&nbsp;</span><span style="font-size: 12px;">business metrics.</span></p>