Anyone interested in alternative software?

I have to admit that I’m not very satisfied with the DM3 software from Dexcom. It seems overly complicated, not very intuitive, and cumbersome to use. Fortunately it has an “export to XML” option which is all I need to create my own alternative - at least the analysis part. For the time being I’ll have to live with the extraordinarily slow download process.


For me the DM3 software mainly suffers from "too much information". I really don't need to go back and see specific values from 3 weeks ago or day by day pie charts. I'm more interested in the following:

  • What's my average glucose for all data and what is the equivalent estimated HbA1c ?
  • Has this average been moving up or down over time?
  • How close are the Dexcom values to manually entered meter values?
I've already made a first attempt at the above and would certainly be interested in hear what others feel is important or perhaps missing from DM3. I'm also happy to provide the software (windows-only for now) to anyone interested - once I make some more progress with it.

I would certainly be interested. I think the DexCom software does have some redeeming features, but overall, I agree with you. I really liked the original Bayer WinGlucofacts software - granted it was for a BG meter, and not a CGM, but the analyses, interpetation and commentary available were both very patient friendly, and provided useful informatin to the Health Cre Professional as well.

Hi, Mike.

I’ve had similar concerns with the DM3 software, and have been thinking about a web-based alternative interface - one that you could use as a platform to share your data with your doctor or peers, as well as being able to at least view your reading history on a mobile device (or a Mac).

I had also considered a web interface but unfortunately that’s not my area of expertise. A web application would require someone to host it and people might have privacy concerns uploading their data. I do like the idea of mobile access (I’m writing this at chuck e cheese using my android device).

Ideally I would love it if dexcom got rid of the receiver and switched to a low power bluetooth transmitter. As much as I want to replace the dexcom desktop software I would love it even more to write my own receiver software for an iPhone or Android. Picture a much better display, predictive alerts, custom alarms by time, inetgration with an online food database, etc.

Saw something about a third party writing something called CHAOS TRACKER. Might want to check it out.

I found the old thread mentioning that but it doesn’t appear to have gone anywhere. The person developing it, Val, has only posted here twice (both Nov of last year) and the website listed, http://www.chaostracker.com, is now for sale by godaddy.

i agree. i would be interested in what you purpose.most of it has no day to day value for me.
thank you.

There is already a web based alternative solution developed by Holger Schmeken who is a member of Tudiabetes. It is glucosurfer.org and there are multiple import options including Dexcom via xml.

Thanks. I’d already taken a look at that but didn’t quite understand it. I’m really just looking to create something to quickly analyze my own data on my machine. I intend to create something useful for myself and will happily share it with anyone interested. I’ll try to get an initial version uploaded somewhere tomorrow for others to have a look.

thanks mike.
i looked at it to.i might as well write it down on paper thats a lot of input imho.and really for me not what im looking for either.

Ok, I’m done with the initial version and have uploaded the windows executable The application is a single exe file - no installation is needed. Just copy it to any directory and run it from there. You’ll need to export your data from the DM3 software as XML in order to use this.

Since I only have 35 days worth of data I’ll be interested to know how this works for people who have been the Dexcom longer.

Here’s a screen for those who may still be interested.

For some reason the upload URL went missing from the post. Let’s try it here: http://drop.io/dexanalyst

thanks i will give this a look.

HI. Thanks so much for the informative input. My Mom has just begun living alone as my Dad recently passed away suddenly of a heart attack. She is 65 with Adult Onset Type 1. What concerns myself and my sister is how we would be able to safely allow her to continue to live alone. She recently began with the Dexcom but what is missing is the ability to push alert data to a smart phone (SMS or call) in the event of the activation of the low blood sugar alarm. That would potentially be a lifesaver for her as there are times when there is no monitoring taking place such as the two hour calibration time, there is the potential to ignore the low blood glucose alarm, and of course there is the danger during sleep when the alarm/vibration is not heard. We would be able to call for help or come ourselves through the addition of such a feature as we (2nd party) are now notified that she is low and can take action to help her before it is too late. Such a feature could be added through a Bluetooth function I imagine. I just don’t know the technical details. Can you provide me some assistance. Can someone write the code for such a feature that could be added to the hardware currently? Greatly appreciated. Thanks Mike and anyone else who can help.

Avi

While that’s a great suggestion it simply isn’t possible with the current hardware. The idea of using Bluetooth in the future is simply something that I’d like to see. I believe that Dexcom mentioned in a financial report that they’re looking into smart phone integration so Bluetooth seems like the obvious path to me.



While Dexcom apparently has the ability to update/upgrade the current receiver software that’s clearly not something that end users like me has access to and I’m sure there’s a whole FDA approval process since the software is so critical to this system. Even with updated receiver software if the hardware support for something like Bluetooth is not there then there’s simply no way to implement this with new software. In your example even if there was Bluetooth hardware built in and the ability to update the receiver to support it it still wouldn’t be able to make a phone call or send an SMS message without hardware and software support for those technologies. At best it could transmit to a nearby Bluetooth listener like a cell phone that itself would send the emergency message. This is yet another reason why using a smart phone as the receiver makes perfect sense.



Even with custom hardware listing on the Dexcom radio frequency you wouldn’t be able to detect a low blood sugar unless the custom hardware supported the same calibration idea. My guess is that the only thing transmitted is a value (their documentation says 76 bits of information) that in turn is converted to a BG reading based on the calibration data in the receiver.

Thanks for the quick reply Mike. The example you gave is precisely what I was contemplating. Having the Bluetooth listening device (smart cell phone) that is on the person anyway act as the device that sends out the emergency message through text or phone call. As you said, it makes perfect sense to have the cell phone act as the receiving device.



If what you said is true, then even attempting to develop an Iphone app for instance would do no good as the Dexcom proprietary calibration software is needed to turn that transmitted value into a BG reading. Has anyone attempted to try and take that transmitted value and analyze it to your knowledge? Perhaps there is a way of calibrating it w/o the Dexcom software which again I assume is proprietary.



Do you know if Dexcom is working on such a concept? Thanks so much.

I certainly have no knowledge of Dexcom’s future plans so I have no idea what they’re working on. One thought I had about a possible solution to the calibration issue would be to have a future version of the transmitter, assuming the use of Bluetooth (which is obviously bi-directional), contain the calibration logic. My assumption is that currently the transmitter only transmits and simply sends a voltage value from the sensor that the receiver then converts to a BG value using calibration data. If instead the receiver simply sent the calibration BG value to the transmitter the transmitter could then in turn apply special logic and then send out the actual BG value instead of just a voltage. This would enable third party Bluetooth devices such as your emergency device or an iPhone or a wristwatch to be able to simply receive and display the values without needing to know anything about the Dexcom or in any way compromise the data integrity. The ability to send calibration data to the transmitter could remain off limits to third parties. The receiver software may well be sophisticated enough to not only deal with calibration but also deal with the blood versus interstitial fluid difference in terms of lag, i.e., it’s rising so the current BG is likely higher than what the sensor currently indicates so there’s likely a lot that a third party wouldn’t be able to do reliably. I certainly wouldn’t want the Dexcom receiver to show on value but my fancy wristwatch or phone to show a different value.

Hi Mike,
I wrote some rather clunky software that takes DexCom and Animas export files (which also contain the One Touch meter readings), dumps them into a MySQL database and then graphs the output.

To answer question #3, attached are two output pages showing my readings from 9/1 - 11/16 and from 3/1 - 9/1. The Y-axis is the dexCom, while the X-Axis is from the OneTouch. You can see why DexCom says that you shouldn’t make BG corrections without poking your finger.

Also attached is one of my daily pages showing carbs, bolus and basal readings along with the actual BG readings. There are two points where you can see the Dexcom correcting it’s calibration factors: 10:15am and 22:00pm. The line between approx 7:30am and 10:15am is where the transmitter was out of range of the receiver. When it reacquired the signal, it was way out of calibration. By 11:30. it was tracking the BG readings again. The dip at 4:30am was due to sleeping on transmitter. I got up, tested, swore a few times and went back to sleep on my other side.
5979-31to91.pdf (46.1 KB)
5980-91to1115.pdf (40.3 KB)
5981-82210.pdf (43 KB)

When I enter carbs and insulin in Dexcom, it shows on the print of a Glucose Trend Chart.

I can only see the values for these if I hover my mouse over the sign.

Does anyone know if it is possible to have the values printed on the chart?

I can’t speak for the DM3 software itself but I can sat that the carb and insulin values are available in the exported data file so my alternative software could display them in any manner needed. It currently does nothing with that data since I don’t take advantage of it myself. I occasionally enter insulin values but only when trying to bring down a high. Since I can’t specify Humalog vs Lantus it doesn’t have as much use for me.

If I added the ability to display these values would you take advantage of it? I’ve had almost no feedback from anyone trying it so I’m not sure that there’s much interest in my effort.