Monthly meeting. Normally some of us go on to the pub afterwards.
Dr Gusztav Belteki gave a talk at PyconUK 2016 on "Python in Medicine: ventilator data". He has agreed to talk to us about what he is doing, possibly on further work he has done since the conference.
Mechanical ventilators are widely used in intensive care including Neonatology. Even two decades ago they were be primarily mechanical devices whose “only” task was to inflate the patient’s lung. Recently, however, they have become equipped with powerful computers that provide sophisticated ventilator modes and display many parameters about ventilation and patient-ventilator interaction. Many of these ventilator modes are not routinely used in clinical practice and many of the parameters and reports are disregarded due to time constraints and lack of expertise. Data provided by the ventilator are almost never downloaded, stored or analyzed.
I have become interested in the analysis of the data that modern ventilators generate. My ultimate aim is to provide the clinician with simple and easy-to-interpret parameters that inform him or her about the quality of patient’s mechanical ventilation and the patient’s interaction with the ventilator over the previous hours or days of care. In the shorter term my aim is to analyse ventilation and ventilator-patient interaction over longer periods (hours and days). To get enough details, a high sampling frequency of the parameters is required. The combination of high sampling rate and long recording time means that one needs to handle very large datasets and that require computing for analysis and interpretation.
Having obtained a piece of software from a ventilator company, I have downloaded so far raw data from the ventilators with 100/second sampling frequency over a total of 160 ventilator days so far. I have been using the Python computer language to analyse these data. Python has got several additional data analysis tools available that are also free to use. I am using some of these tools (pandas, numpy, matplotlib, scipy etc) to interpret my data. I have already generated interesting data about ventilator parameters and alarms of sick newborn babies. I am using these data for teaching nurses and junior doctors and also for feedback to the clinical care team. I am also using this method for research projects.