Overview

The benefits of a pump with the flexibility and simplicity of an insulin pen.

Objective

The initial idea of this project is to develop a convenient discreet way for a user to deliver an insulin bolus. So no insulin pen and needles to carry, and no self-conscious hiding in corners when your out and you need to inject yourself; and therefore no funny looks from passers-by. Just get out your mobile phone and instruct the pump to deliver how ever many units you need, or press a button on the pump, once for each unit required.

That's my initial aim, so you would still use your long acting insulin (Lantus or Levemir). However the potential for further development is extensive. Here are some ideas, many of which will be easy to implement once the initial prototype is done.

0. Timed bolus delivery. Program the pump to deliver a bolus of insulin automatically. For my daughter this would be very useful in the morning to offset the dawn phenomena. She's always a little too high in the mornings and we can not offset it with insulin because we would have to give her a bolus just before she wakes.
0. Record all your boluses. The pump could easily record each bolus and store it internally and then transmit it to a PC. These values could then be correlated with blood glucose readings enabling the user to better understand, and therefore manage their diabetes.
0. Send reading to a central server via the mobile phone network. The bolus sizes could be automatically sent to a central server for monitoring by a health care professional.
0. Automatic calculation of bolus. The mobile phone program could easily be extended so that you could input you blood glucose reading, and also, either your carbohydrates, or what you have eaten, so that the software can automatically calculcate the correct bolus and deliver it.
0. Remote delivery of bolus. Adults could deliver a bolus for their child over the mobile phone network.

There are probably lots of other ideas I haven't thought of. Once the pump has proved itself it could easily be extended to deliver insulin over the entire day like the presently available insulin pumps, thereby removing the need for the long acting insulin injection.

Operation

I'm going to design the i-Pump to work in two ways.

0. Via a mobile phone. You'll start an application on your phone (it may automatically start) and simply enter the number of units you want delivered and press a button labeled "pump" and it will deliver the specified amount of insulin.
0. Or, if you forgot you mobile or you don't want to carry it with you, you'll be able to press a button on the i-Pump to specify the amount of insulin you want delivered and the i-Pump will deliver it.

Design

Here I'll set out the big picture of how the i-Pump works, a block diagram is shown below, and the main parts are listed . For details of each part use the side bar menus.

0. The Micropump - Pumps the insulin from the reservoir through the infusion set to the user. The one I found is tiny; it's only 14mm square and 4mm thick.
0. The Power - Power to drive the pump. This consists of a circuit normally used for photoflashes in digital cameras or mobile phones.
0. The Controller - The heart of the system that controls the pump and interfaces with the user.
0. The Infusion Set - Means by which insulin is passed into the body.
0. The Insulin Reservoir - Contains the insulin that the pump delivers through the infusion set.

Now here's how it works, again just a high level description, there's a lot more detail in the sub-menu on the left.

0. The user informs the controller of the number of units of insulin that they want delivered. They can do this in one of two ways; either manually through a button on the pump or via a mobile phone.
0. The controller charges the power circuit and once charged this is detected by the controller which then routes the power to the micropump.
0. The micropump pumps a tiny amount of insulin from the reservoir and pushes it trouble the infusion set into the user. Note that a single pump-action is not equivalent to 1 unit of insulin, so a number of pump-actions are needed to deliver 1 unit.

It's as simple as that. Actually it's not quite. I nearly gave up on more than a few occasions. You can now look at each part in detail using the menus on the left and/or read on and find out about why it wasn't that simple and why I nearly gave up.

History

Not long after the diagnoses I started looking into developing my own pump. Much to my surprise it wasn't long before I found the micropump developed and sold by a German company called Bartels Mikrotechnik (www.bartels-mikrotechnik.de). So I started experimenting with this and immediately found the first problem. Although the pump was supposed to work at low voltages, it was not effective at battery voltages with water or insulin. It would only work at 240V! Oh dear, no one's going to want to walk around attached to the mains. I nearly threw in the towel at this point.

I quickly realised I needed to generate a high voltage from a battery, but only for a short time. The first thing that sprang to mind was a camera photoflash. The Xenon lamps used in photoflash units require around 330V and these things fit into mobile phones that are pretty small. It didn't take me long to find tiny chips on Farnell's (www.farnell.co.uk) that are designed exactly for this purpose. This seemed simple - unfortunately once again it wasn't.

The circuit looked simple but one of the components was a nightmare to find. Even though the information supplied with the chip I choose specifically identified a tiny transformer I just could not find a supplier. So I had to find an equivalent. Even when I found an equivalent I could not easily source that either. Nearly gave up again, however after quite a few emails and phone calls I managed to get some free samples, so it looked like I was back in business. Oh dear, wrong again.

I managed to get a demo board from the company with the charger circuit all ready to go. But could I get it to work, nope! This became a real struggle, and I nearly gave up for the third time. Took quite a while and contact with various engineers in the UK and USA before it was cleared up. I was using the wrong type of battery. It would only work with a lithium-polymer rechargable battery (the type you get in mobile phones).

So I'd broken the back of the problem. I now had a circuit that could power the micropump from a battery. Now I had to work out how I was going to control the pump. This was one of the easier areas to sort out, and I had some luck. I new I wanted to control the pump via a mobile phone using bluetooth for communications, and already had quite a lot of experience with bluetooth so I started looking around for a bluetooth module. There are quite a few around and I discovered that most are programmable and have various inputs and outputs. The one I settled on is the AIRCable SMD (www.aircable.net), it's a fantastic device, it's tiny, it has lots of inputs and outputs, its easily programmable and I can link to it via bluetooth.

There are two components left. The infusion set was relatively easy. It makes sense to just use presently available sets; I wasn't going to invent this. Getting hold of them was a little more difficult. First I rang around asking for samples but only managed to get two samples of the Inset 30 distributed by Animas (www.animascorp.com). The Inset 30 is actually designed and manufactured by a Danish company called Unomedical (www.infusion-set.com), so later I decided to contact them directly to see if I could buy the Inset 30 directly from them. It was a long shot, but it paid off. They were very helpful and interested in the project. They couldn't supply the Inset 30 because Animas have the rights but they were happy to supply the Comfort - Short, what's more they supplied them for free, which I did not expect.

There's one last component, the insulin reservoir. This is the trickiest component and trying to come up with a solution for this nearly resulted in me giving up. I just can't find an off-the-shelf product so I've come up with something based on an idea in a science paper I found, but it's still not satisfactory. The details of what I've come up with and the problems I need to resolve are set out in "The Insulin Reservoir" webpage. This is an area where I'm probably going to have to get some outside help. If you can help then please do get in touch.

That's the story so far.