Principal Software Engineer, Medtronic
Principal Software Engineer, Medtronic
Prior to Cognitive Health in November, we spoke with speaker Vincent Mahfoud, Principal Software Engineer at Medtronic.
Vincent is a Professional Embedder Software Engineer who has benn working in different industries including automotive and financial. In the past 10 years, he was involved in the medical device industry developing embedded software for different tiers of ventilators. His current role is Principal and Lead software Engineer with Medtronic. Vincent has authored and co-authored a number of SAE and IEEE papers in the subjects of Distributed Parallel Processing on CAN networks, Connected Cars (remote diagnostics), and Dynamic Discovery Services between automobiles. He holds a bachelor’s degree in electronics engineering and a Master’s in computer engineering.
Vincent: Computer security is increasing rapidly, it is not unlikely to be able to apply quite secure connectivity to medical devices in the very near future. In addition, connectivity can be selective and controlled, it does not necessarily need a complete open device application or data. Several layers of access could be deployed, where only data of interest could be allowed. Also, private individual data could stay anonymous if it were to be used for medical research and statistical experiments.
Vincent: The medical devices are continuously evolving towards more sophisticated applications, user-interface, presentation layer, etc. example, the early ventilating machines existed in the first half of the 20st century were plain mechanical, then they evolved to electrical with few control knobs (pioneered by Puritan Bennett, now Medtronic), and so on where they are now using advanced GUI with limited web capabilities (for security reasons) nowadays. It is the market demand, user needs, better service, and increased capability along with the dropping cost of electronics what driving this evolution.
Vincent: This is not 100% a vision, there is a small and very limited level of connectivity between certain devices made by different manufacturers but for a very specific application. For example, there are data loggers which were adopted to interface to ventilators from certain brands to collect “certain data” and for specific application. This is still tightly coupled connectivity. The vision is to allow for a standardized connectivity, using a common interface. This idea has gained momentum in the recent years, it seems to be gaining some seriousness from manufacturers.
Vincent: A focused joint effort by specialized cross-functional teams from different medical device manufacturers of different applications and purposes. A top-down approach should be taken, starting by the WHAT questions before the HOW, from the very high level of the foreseeable benefits, use-case scenarios, potential scientific advantages, etc. rather than starting form the deep down technical how-to.
Vincent: The access should be limited to the two beneficiaries of this apparatus: 1st, the medical team directly involved with this cluster of devices (in the ICU or surgical operations room) who would benefit from the combined data from the multiple devices in connecting dots for a condition of a certain patient, for example, in addition to the breath parameters that the ventilator plots and displays, also Co2, O2, ACG, blood pressure, etc. samples for their respective devices could be combined in the same time stamp for better analysis of the patient condition. The second is the “authorised” medical research teams, scientists, etc. who would most benefit from the “mining” concept in this apparatus in possibly “excavating” some unexpected, unknown observations or patterns form the collected data of different patients.
Vincent: Data-mining, in simple, is to find unexpected things in a large collection of database, in contest with the direct queries for “expected” existing data in this database. Data-mining in this connected medical device application might allow the same concept on the large collected database of patient parameters from different measurement and treatment devices and from many patients. When a large database is collected, with time, age, gender, medications, nutrition, maybe genetic data, etc., applying data mining might uncover certain trends, patterns, observations, or form new treatment or diagnostic ideas, etc. based on the study of this data base.
Vincent: The expectation is to get feedback from subject-matter experts in the different disciplines from different companies and fields, which this conference provides, because such an idea requires several pairs of eyes looking at it from different angels. Also, possibly continuing the discussion with the interested partners.