Having a deep interest in medicine and computing, I found fever charts to be an amazing starting point to explore. Including the amazing strategy of how the doctors can use fever charts interpretation to predict the possible cause of sickness, also the easy opportunity to collect data or crowdsource and then simple challenge to analyse for the patterns with respect to time.
1) Learnt about the fever charts and use during my BMJ Electives. (My learning portfolio here- https://classworkdecjan.blogspot.com/ )
2) Using google sheets to quickly digitize fever charts, also let other students and patients copy and utilize it to make charts digitally by themself. https://classworkdecjan.blogspot.com/2017/08/using-fever-chart.html
3) Using vision API to explore its application on fever charts. https://classworkdecjan.blogspot.com/2020/03/fever-charts-part-1-extracting-data.html
4) Collecting resources to utilize when sharing and discussing the idea. https://classworkdecjan.blogspot.com/2020/03/resources-for-fever-project.html
5) Discovering the question, "How we know normal body temperature is 37°C" and reading research and even found a small very old dataset and this interesting book which I am yet to read - "On the Temperature in Diseases: A Manual of Medical Thermometry". The 37 °C value was set by German physician Carl Reinhold August Wunderlich in his 1868 book,[34] which put temperature charts into widespread clinical use.[35] https://en.wikipedia.org/wiki/Human_body_temperature
6) Trying to understand what analytics to apply for predicting the possible pattern correlating with disease/diseases.. Time series analysis? Clustering? Here I faced the roadblocks.. I understand the ML concepts but implementing them programmatically on a new challenge is something I haven't tried yet.. Discussed with some interested folks but the idea never took off as a digital health research project. Still hopeful and trying.
7) During my last visit to same electives, I met with his another elective student Mr. Neelankit, who was doing his M. Tech and started researching on IoT device for fever charts and we had our elective days together. Later he came up with a device prototype and progressing with it ahead. There are many devices in the market for this purpose but we need something cheap and usable in the hospital setting in the rural area.
7) Found this today - http://feverprints.com/ , It's inspiring to see this happening in real and excited to see the clinical useful outcomes in future. This is a publication from the project "Using Smartphone Crowdsourcing to Redefine Normal and Febrile Temperatures in Adults: Results from the Feverprints Study" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6258625/
We had been doing tasks like digitizing fever charts, helping patients learn making it, utilize the data to improve clinical accuracy for diagnosis and antibiotic use etc. with a few students, the IoT device maker and prof. Rakesh Biswas always helped us with data access and clinical reasoning but for large scale data we needed more resources.
8) Chat gpt for fever charts https://youtube.com/shorts/Ya_IE23uSyI?si=wV0dHgytVeSlX4T4
which next sign you would suggest to explore? and why?
No comments:
Post a Comment