30-second blood analysis with Raspberry Pi

A portable, affordable Raspberry Pi-powered blood analyser is helping to “establish a robust healthcare ecosystem” in remote parts of India. Samples can be tested in just 30 seconds, and the cost and size of the parts make it an attractive solution for rural and resource-strapped areas.

It is the work of researchers Sangeeta Palekar and Jayu Kalambe from the Department of Electronics Engineering at Shri Ramdeobaba College of Engineering and Management.

blood analyser
(Image credit: Shri Ramdeobaba / College of Engineering And Management)

Tiny computer — massive processing power

Regular blood tests are vital in the tracking and elimination of many diseases, but there is a huge time and monetary cost currently tied to this type of laboratory work.

The researchers’s device measures light absorbance through a blood sample, a common type of analysis, and they harnessed the processing capability of Raspberry Pi 4 Model B to analyse the absorbance data. Their Raspberry Pi-powered solution was found to perform on a par with the kind of expensive lab-based blood test typically used.

Quick and easy

Sangeeta and Jayu’s analyser is not only cheaper to build and maintain than the lab-based version, it also does the job better. Using the lab-based method means that samples from patients in rural areas must be sent away for analysis, with results communicated back to patients at a much later date. In contrast, Sangeeta and Jayu’s device can process blood samples there and then. All you need is an electricity source. Patients get their results immediately, and there is no need to transport delicate samples across rural terrain.

Shri Ramdeobaba College of Engineering and Management

Incorporating an IoT element into their design, which would allow for remote monitoring, is the next step for the researchers. They also intend to develop their invention to allow it to carry out different types of blood analyses.

Read more about the science behind the creation

The full research paper is behind a paywall, but the abstract does a great job succinctly explaining all the science. Sangeeta herself also explains a lot of the magic behind her creation in this interview with IEEE Spectrum.


Harry Hardjono avatar

I’m curious what kind of analysis is involved. There are many kinds of blood analysis, some more involved than others. Otherwise, Theranos comes to mind.

Joseph Alway avatar

It’s hard not to make comparisons with Theranos considering the ongoing trial. Yet, this is likely using more traditional methods, just powered by a Raspberry Pi. Whereas Theranos was trying to make accurate analysis using inadequate blood samples. They were literally trying to examine a drop of blood and provide accurate results.

Amit avatar

SO true. Theranos was the FIRST thing that came to my mind. The worst part of Theranos was that they could have done it! If they had focused on CBC & CMP over “doing every test there is”, listened to Ian Gibbons, been more receptive of employees’ concerns, they could’ve done it!

Nic avatar

The system described in the academic paper suggests blood glucose level measurement based on light transmission in combination with a reagent. The authors expect the concept to be compatible with other common diagnostic tests simply by adding the relevant processing logic and using the appropriate reagent. They also refer to the advantage of plenty processing power which reduces the diagnostic time.
It appears to operate basically on the same principles as established laboratory equipment, only implemented with cheap non-proprietary hardware.

Hitesh Katara avatar

I was thinking we are so much ahead now

Andy (guy who works on blood analyzers everyday) avatar

This looks like a basic CBC type of analyzer. A simple lysing reagent to get the WBC and HGB count and using light absorbance for the Differential. Using the Coulter principle is a simple resistance counting process. We have really old instruments running on original pentium processors, 64MB of memory and Windows NT. A raspberry PI 4 would be a rocketship comparatively. An interesting project but certainly not groundbreaking. I’d like to see a Raspi 4 review and analyze 60 microscope captured high def images of a blood-stained slide and use AI to determine cell type and counts. Now THAT would be REALLY IMPRESSIVE

crumble avatar

The impressive impact will be social not technical, if such devices can be used legal.

They are cheap, because they are produced and regulated like official medical devices. So they may be affordable for smaler town medical centers. If cooling and transport time of the samples have an huge impact on the result, these devices may be more trustworthy than the better products. Rural India has fantastic roads. But you cannot travel fast on them.

Its the same with the 02 blood measurement. Pis were used to collect data of cheap devices. So the personal was able to keep track of more patient during the pandemic. A data logger is technical not fancy at all.

And don’t mention the split. Yesterday the blog should show young people in which areas computer are used and how simple DIY can be. Here is the right place for such articles. But IEEE?

mrsean2k avatar

Legality and regulation are interesting philosophical questions; if a Pi-based system has commonly available parts, who’s going to stop you?
If you have *no alternative* to such a system, are false positives then the main issue? And what up to what ratio of false positives still constitute a net gain?
(mrsean2k: a guy who works on the orchestration software for blood analysers every day :))

Victor avatar

nice test application but they only do glucose. I think the pharmacy glucose checkers come in cheaper and with zero build time.
would like to see a paper that does at least the basic CBC with “homemade” equipment

someone avatar

I didn’t think that RPis were so popular in India too!

Comments are closed