Author: danlagalante
The end of prototyping… for now
Yesterday, we reached the pinnacle of what we could achieve in our limited amount of time. After running our second prototype past our patient, they provided us with potential additional features that a user would want on the finished device, and we delivered.
Our first point that we wanted to address was the alerting mechanism. Our second prototype had only an LED to alert the user that the insulin was pooling. Since the insertion site of the pump would be under clothing, a less visual method of alerting would be desirable. A few quick edits to our code and circuit added a vibrating motor which activated when the circuit was completed with insulin. We kept the LED as well, because Neeraj noted that the FDA requires a visual and physical alert system in this sort of device.
Secondly, our patient expressed a desire to turn off the alert if they were in public and did not want to go through the hassle of changing the insertion site at the moment when they were alerted. This turned out to be a little bit more involved than the first point that we addressed, but with some clever wiring and coding, we added a non-momentary switch to our circuit that would turn off the light and motor when activated.
We have plenty of ideas for our device in the future, but this seems to be a smart stopping point in terms of prototyping. Our device does what we originally set out to do, which is detect insulin and alert the user. That seems pretty successful to me
-Dan
Let there be light!
“Our patient needs a way to be alerted to when their insulin pump is not effectively delivering insulin due to tunneling.”
The key part of this need statement is the “alert.” Our device is relatively useless if it can detect the presence of insulin, but would not do anything as a result. Yesterday, we solved this issue.
Detecting the insulin was no easy task, and differentiating the insulin from other liquids required a series of tests with our first prototype (described in the previous blog post). Based on the results of these voltage tests, though, we were able to produce a second prototype that will do nothing if the circuit is completed by ordinary water or sweat (modeled as diluted saline solution), but will activate a red LED if it is completed by insulin.
This is obviously a huge step forward for us, as our device does both of the necessary functions that we originally set out to do. Our next prototype will address the issues of optimizing space, a smaller power supply, and a less visual way of alerting the patient to any infusion issues.
Successful first idea
Today was a landmark day in the development of the development of our product. We obtained a vial of over-the-counter insulin (who knew?) and ran a bunch of tests and collected a lot of usable data for the project. After a bunch of trials, it was clear that the presence of insulin on the skin caused a significant effect on the resistance and conductivity of the sample. The main problem was finding a way to use that information. Constructing a galvanic skin response (GSR) sensor to sense this change was the next logical step, but we ran into some problems…
After many attempts at constructing the allegedly simple sensor to no avail, we began to ponder some second or third ideas about how to create the patch that would alert a patient to a defective insulin pump insertion. The color changing patch was reemerging as a potential solution to our problem. While researching this, though, a partner (s/o Neeraj) fixed our problem and finished our first prototype of an insulin sensor for use in our patch.
While there were many successes today, not all of our tests produced usable results, as our attempts to quantify the amount of insulin lost to tunneling was a big failure. The silicone sample that we attempted to inject into was too firm to accurately represent tissue, so we now have to pivot and find some other material to help us gather this important information.
-Dan
Week in review / next steps
This week was a whirlwind; so much stuff was accomplished and I am very excited about the possible success of our patient project. We met with our patient multiple times this week, both in a public setting and privately to talk about some more intimate details. When we revealed that our idea would alert a user if their insulin pump was tunneling (not properly delivering insulin to the body), our patient responded that such a device would be “frickin’ awesome.” I think that may have been one of the highlights of my week, knowing that the project I am working on could have such a positive reception. There is much more work to be done, though.
In this coming week, we plan on testing the conductance of insulin on the skin, so that we will be able to quantify a value that will be used as a sensory threshold for our device. We also need to find a way to account for the presence of sweat on the skin and prevent that from causing false positives. This is our main roadblock in our design, and I hope that it doesn’t cause the wheels to fall off of this project.
-Dan
Patient Health Project
Looking at the directions for writing this blog post, the first thing I noticed was the order that the two questions were written: “What am I doing?” was first, and “Why am I doing it?” was second. Whether it was by coincidence or on purpose, it is presented in the opposite order of importance to me. More important to me is why I am working on something, because the true passion I have for this field is to help a particular person or group, which is why this class intrigued me so much. The “why” in this case is that we as a group are helping our patient deal with a condition that, frankly, really sucks to deal with. The “what”, I hope, would be to create a device that senses when our patient’s blood sugar is lowering by certain chemicals released when blood sugar is dropping. Hopefully, we would be able to provide a non-intrusive way to monitor dangerous blood sugar drops during the day, qualitatively. Currently, I am researching what chemicals are produced when a blood sugar drop is occurring. The Diabetes Alert Dogs are trained to react to blood sugar drops by smell, so some sort of chemical must induce this change (“what”). I am doing this because, once the chemical is identified, we would in theory design a wearable device that could sense the presence of that chemical in close proximity to the diabetic patient to alert them when a dangerous drop is happening, rather than relaying on checking with glucose monitors.
-dan
BME Design 