New Insulet survey

So Insulet are sending out a survey to known users. Some of the questions might be of interest to other people on this list. I quote verbatim (question 11 of 15 or so):

Shaping the Future of Diabetes Management Technology
Question Title

    1. The next generation Omnipod System is an Automated Insulin Delivery System combining the tubeless, waterproof Pod with an embedded algorithm, the Dexcom G6® Continuous Glucose Monitor and personal Smartphone control. Through this unique wearable design, the system can help prevent highs and lows by using the sensor values sent by the CGM directly to the Pod and automatically adjusting basal insulin delivery.

The system will initially be compatible with select Samsung smartphones. For users who do not have a compatible phone or prefer to keep their phone and insulin delivery device separate, a Personal Diabetes Manager (PDM) will be provided. The PDM will display insulin delivery information as well as CGM data on the same device.

NOTE: Users can choose to continue using the Dexcom G6 mobile app on their smartphone for CGM values only. They will still experience all features and benefits of the system because the CGM communicates directly with the Pod.

Question: Assuming you have a personal smartphone that is compatible for controlling the next generation Omnipod System, please select your preferred use case below

So the next question is the biggy; and this will illustrate why these online surveys are next to useless. I give the only two possible responses here:

*12. Assuming your personal smartphone is not compatible to control the next generation Omnipod System, would you:

I would use the PDM to control the Omnipod System and my Dexcom CGM

I would use the PDM to control the Omnipod System but continue using the Dexcom mobile app on my smartphone to control the Dexcom CGM

I wouldn’t do either, I’ve already swapped dumbphones once because Insulet don’t support Android, I’m not doing so again because now they don’t support Apple… So I didn’t answer, since none of the above was not on the ballot paper.

So then things start greying out, but, next question:

*13. For each of the pairs of statements below, we’d like for you to indicate how strongly you fall on one side or the other, between the two ends. There is no mid-point, so please choose which side best describes how you feel.

Life with diabetes is completely unpredictable Life with diabetes is very manageable
Life with diabetes is completely unpredictable Life with diabetes is very manageable

I rapidly answered the stream of stupid questions and when I had done this:

! This question requires an answer.

*12. Assuming your personal smartphone is not compatible to control the next generation Omnipod System, would you:

So I answer it; there was this little “x” at the top of the browser window, I clicked on it and miraculously the idiot window disappeared.

I have to admit, it isn’t just Insulet. I can’t remember the last time I managed to make it through one of these Putin/Kafkaesk surveys without having to close the window manually.


Sounds like they’re worried about losing customers who use iOS and are now trying to get an idea If designing an app for iOS is worth the money. Getting apps pushed through the FDA is much easier with iOS over Android. It took Dexcom a year and a half to get their Android app approved after the G5 launch, while iOS was ready to go out of the box at launch.


But to get that data spidermonkey, which was running the survey, would have to record abandoned surveys. If they don’t all they have is the data from people who lied and said one or other of the two false answers, and the existing Android users.

I don’t know, do these survey companies deliberately construct annoying surveys and then monitor when the sucker who tries them gives up? I mean, I referenced Kafka, but that would be extreme.

Anyway, TidePool are working with Insulet on an iOS app. That’s an obvious third choice in the list, but I don’t see how SM can detect that people like me know about TidePool rather than being just annoyed because they don’t use a “select Samsung phone”.

It’s all weird. I suggests to me that Insulet are considering just doing the PDM and dropping the dumbphone support, which would, of course, invalidate their previous assertion that the PDM is “not required”.

1 Like

For the record, I like the Kafka reference.

1 Like

False dichotomies and suggestive or other stupid questions seem like a hallmark of online surveys. I wonder why companies keep using such poorly designed surveys. In the end they’re hurting themselves if they make decisions based on inaccurate data, right?


I also bailed out on the survey and was unable to finish it.

MDI is always an option…

1 Like

Yeah that is really poor form. honestly, a large survey like that should be created off the back of a focus group of patients and consulting with engineering to ensure they are asking about real trade-offs that could affect the development process and then only release a large survey after a small sample has been trialed and any issue (like the one you point out) addressed. Anything else is just lazy. Connecting with patients and getting non-actionable information is a sin.


Let’s distinguish between the survey companies that have the software to show questions and collect the answers, and the people/organizations that design the actual questions. Everyone can think of a bunch of questions to ask, but it’s a highly-technical task to design a scientifically sound questionnaire (one that gives valid, repeatable results from a sample of respondents to accurately predict the beliefs and behavior of a larger population.) A lot of the surveys I’ve seen on line were apparently produced by unqualified people, and they have lots of blunders such as the examples you’ve cited.

With respect your specific question, in a proper sound survey they do monitor for questions that are not answered and surveys that are abandoned part-way through. To get a valid, repeatable result almost everyone needs to be able to quickly select an answer. If the respondent needs to ponder, then their selection will be more heavily swayed by irrelevant factors (current mood, recent events, …), and if more than a tiny fraction of the respondents don’t pick an answer at all, that’s a very strong indication that the question is defective: this should have been identified in the pre-test (the trial runs on small samples to debug the survey), and well qualified survey designers are likely to know during the formulation of the question, rather than finding out the hard way when pre-testing or running the live survey.

Real survey companies, like Gallop and Opinion Research Corp. have deep knowledge about this field. One of the early books in the field is “The Art of Asking Questions” 1951 by S. Payne. It was out of print for quite a while, but it seems that it is currently available, and was recommended to me by a scientist at Opinion Research back when I was a young lad. There are lots of books with similar titles; I don’t know if they are sound science or just trying to market themselves by confusion with the original.


12 to 18 months ago they had a similar survey that a focus group was researching. The choices then were to use only a Samsung phone ( because of Knox) or use the PDM. Looks like they are still working the details.

1 Like

Is that because Knox allows them to secure the patient (user) data to prevent the patient (user) gaining access to it? I.e. the employer/employee security which, I assume, prevents employees disclosing their employers data even though they can see it on their samphone.

1 Like

That would be my assumption, kind of like AOL or Apple assuming we are all secure our devices

1 Like

That wasn’t what I meant. Knox allows an employer to secure data on an employee phone while still allowing the employee to do the stuff employees do. In effect it seems to create two phones in one. If Insulet chose Samsung beause of Knox it would surely be because they wished to prevent the owner of the phone from free access to the data that Insulet stores there.