Analyzing neonatal ven.lator data with Python

Presentation given on the Cambridge Python User Group evening on the 7th of February, 2017

Transcript

1. Analyzing neonatal ven.lator data with Python Gusztav Belteki CamPUG Meetup, 07/02/2017
2. •  Full /me clinician (neonatologist, doctor looking a9er sick newborn babies) •  Interested in the neonatal lung and mechanical ven/la/on •  Received no formal training in compu/ng or programming •  Learned Python from books and YouTube videos About me hGps://github.com/Ven/lator-Python hGps://twiGer.com/gbelteki hGps://speakerdeck.com/belteki
3. What will I talk about ?¶ •  How I have been using Python to analyse neonatal mechanical ven/la/on •  My experience and learning curve of Python
4. Why do some newborn babies require mechanical ven.la.on ?¶ •  Prematurity: lung, muscles and brain are too immature to support adequate gas exchange •  Some/mes full term babies also require intensive care (e.g. infec/on, a9er an opera/on, birth depression etc.) - In Cambridge NICU we have >1200 “ven/lator days” yearly
5. 1987 2017
7. What is the problem with the ven2la2on and ven2lators? •  Ven/la/on is only one part of a neonatologist’s job •  Ven/la/on is just one of the determinants of neonatal outcome 1 2 •  Ven/lators and the concepts of ven/la/on are complex
8. Why are ven2lators and ven2la2on complex ? “Ar/ficial lung” Fully sedated and muscle-relaxed pa/ent Newborn with spontaneous breathing effort
9. The reality… •  Most data the ven/lators are providing are not considered during clinical decision making •  Ven/lator data are not rou/nely archived at a high sampling rate and not analyzed retrospec/vely
10. Aims •  The collect ven/lator data at high sampling rate and analyse them computa/onally •  To provide the clinician with SIMPLE indicators of ven/la/on and ven/lator-pa/ent interac/on •  Eventually to move closer to automa/on of mechanical ven/la/on
15. Pa.ent not breathing Pa.ent beathing
16. “Clean ven/lator breath” Pa/ent-ven/lator interac/on: inspira/on during expiratory phase
17. “Clean ven/lator breath” Pa/ent-ven/lator interac/on: inspira/on during expiratory phase
18. Flow (L/min) Pressure (mBar) Ar/ficial lung Sedated pa/ent Breathing pa/ent fig = sns.jointplot(x="paw", y="flow", data=fast_2a, s = 5) Two-dimensional scaGerplots using the seaborn library 3 hours of ven/la/on (~1,000,000 data points)
19. plt.hexbin(fast_3a.paw, fast_1.flow, bins='log', cmap=plt.cm.Blues) Hexbin plots Flow (L/min) Pressure (mBar) 3 hours of ven/la/on
20. Acknowledgments:¶ •  Professor Colin Morley •  Ian Ozsvald and Giles Weaver •  Dräger Medical¶ •  All the doctors and nurses of Cambridge Neonatal Intensive Care Unit