1. Patients launch $1.27 million crowdfunding campaign for ME/CFS gut microbiome study.
    Check out the website, Facebook and Twitter. Join in donate and spread the word!
Can You Come for a Visit? My ME/CFS Says No
My daughter and son-in-law just had a baby last week. We are thrilled. But we won't be able to see the baby or hold her any time soon. We won't be able to take over little gifts or help out with housework or babysitting.
Discuss the article on the Forums.

Collecting data from users

Discussion in 'General ME/CFS Discussion' started by optimist, Jul 11, 2014.

  1. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    @JT1024 I skimmed over it, and i must say - WOW! I will try to read up a bit when brain energy is back :)
  2. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    Make and I began discussing this issue again about a month ago. The most important things:
    1. we need to establish at this point is what do people want to get out of this?
    2. what high level aims do we want to support.
    The following items occurred to me:
    • To store illness severity, symptom table, treatment table for each member
    • To allow mapping of symptom improvement for particular treatments
    • To allow searching of symptoms and treatments for members with similar illness severity
    • To store any test results for each member
    Two uses of this information occurred to me:
    • Providing members with a more reliable and quick way to identify what treatments might cause benefit or harm
    • Providing researchers and medics with a reliable source of information (with members consent).
    What are your combined thoughts on this as a starting point? Does anything need to be added or removed?

    Leo
  3. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    @Leopardtail Those are good points. By saying "To store any test results for each member" - do you meant things like blood values, genetics tests, etc...

    Parts that I've thought about that could also be included is:
    • Pre-sickness experiences. With myself as an example: stressful life situation, long term antibiotic treatment, lyme, ...
    • Short "biography" - explaining in words the situation now and then.
    • Smart data comparison system - allowing the user to see how effective Vitamin B12 was for everyone, OR, only for those who struggles with this, this, that and this, but not this, that and this.
    • Among those who had success with "B12" - 40% also did this, 63% took this...
    • Those who struggle with this, has bad experience with this, this and this...
  4. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    I guess I should introduce myself: Mark (who is our CEO) and I discussed the project linked above a little while ago, and I have agreed to lead the project (for the analysis and design stage at least). Due to the cost involved we have both agreed we need to be very clear where we are going before we start.

    What will this do?

    I see you have given some thought to the information you would like stored.

    I have developed a few web base systems like this in that past and the thing most people do not write down, but the most important thing by far is: what are we going to do with the information? In other words who is going to use the information stored, and how are they going to use it. At this stage we are trying to get a crystal clear idea of what our members want from the system - by that I don't mean what information is stored, but how that information will be used. What are your thoughts on that?

    What's more important early on is 'how is the information going to be used'? Answering that question first allows me to make better decisions about how to do things, but also allows all of us to work out what information we need to store.

    What information it will store

    By results, yes I mean tests results whether for vitamin levels, hormones, or whatever else one might test for.

    Most of that sounds like it's in the right direction. Comparing a specific symptom with a treatment was something I expected might be needed. Complex comparisons allowing 'thus but not that' are likely to be a phase two issues, bearing in mind we will need to deliver this over the web and simply to start with.
    • Among those who had success with "B12" - 40% also did this, 63% took this...
    I had thought about querying combinations of say B12 & B9 (substitute any combination). I had not thought about ranking the accompanying treatments in that way - good thinking..
    • Those who struggle with this, has bad experience with this, this and this..
    Would it be enough simply to have members rate the treatment when they give details of trying it? Given that people often try treatments in packages, I had assumed we would need some way of 'grouping them'.

    Experience
    You mention a couple of things in your question. Did you mean all pre-existing medical conditions, and all life experience of just those the member thing triggered their condition?

    Quite a lot of the things (triggers) are 'standard items' example infections with XXXX, overwork due to family or career. Do you think we might be able to get this boiled down to some check-boxes plus additional info? How do you see that information being made use of?

    Look forward to hearing form you,

    Leo
  5. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    I fully agree that good planning is very wise. I am my self a sort of hobby programmer. I have some knowledge of the most common web technologies, but nothing too advanced. My mind is a bit spaghetti like now, so sorry for this not coming out too organized:

    I see two main purposes for the data collected:
    1. Letting fellow sufferers match their profile with cases similar to themselves, and learning what experiences they had, both good and bad, in order to hopefully find the quickest path to the best treatment plan for their special situation. This is done mainly by different statistics based on the data from the matched profiles.
    2. It will also be a hopefully great tool for researchers and doctors to generate data reports that can help them in their research; say they need data about people with digestion issues - then they could just filter out those.

      I believe for this purpose it would perhaps be wise to somehow be able to grade the quality of data provided by each users - perhaps making filters like "Have a diagnose" or "Self diagnosed". So if the researchers only want data from people who has been properly diagnosed by medical staff, they can filter out those...

      To secure the quality of data, it should perhaps be required that the user somehow must provide certain details that can help establish trust between PR and the patient.
    "Would it be enough simply to have members rate the treatment when they give details of trying it?"

    Yes, my though is that the user rates it - positively or negatively, but the system can at the same time look at the "struggles" the patient has registered on his profile and let the user know that most of the patients who reported this product/treatment as a negative experience also struggled with "these" issues. So then, if I myself do not struggle with the same issues, perhaps the product/treatment could be working for me.


    "You mention a couple of things in your question. Did you mean all pre-existing medical conditions, and all life experience of just those the member thing triggered their condition?

    Quite a lot of the things (triggers) are 'standard items' example infections with XXXX, overwork due to family or career. Do you think we might be able to get this boiled down to some check-boxes plus additional info? How do you see that information being made use of?
    "

    I was thinking about check-boxes yes - that way it is possible to use the data effectively. I was thinking perhaps those that was thought to contribute to developing the condition first or all, but perhaps registering all medical conditions is not a bad idea either?
    Last edited: Jul 23, 2014
  6. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    Have there been any thoughts on choice of technology to base this system on? PHP, Rails, ...?
    I've almost finished reading a book about a new Node based framework called Meteor (www.meteor.com). Have you heard abut it? In my opinion, it is really good and I believe it could serve very well for a system like this.
  7. Hip

    Hip Senior Member

    Messages:
    2,738
    Likes:
    2,481
  8. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    @Hip Thanks for the extensive list! I will have to dig into some of this :D
  9. Snowdrop

    Snowdrop Senior Member

    Messages:
    421
    Likes:
    598
    Toronto
    While reading this a thought occurred as to something that might be useful as part of the project.

    Re-reading Leopardtail's post #22 it may be that my idea is contained in the comment "To allow mapping of symptom improvement for particular treatments"

    Just to be certain, does this statement mean that a scale (or something similar that would do the job) would be developed so that when we start rating things we'll all be using the same reference point?
  10. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    It should in my opinion most definitely be a pre-defined scale. Otherwise it will be very hard to compare data.

    Mapping effect over time is a really good solution, however, I think if one takes more than one thing at a time, which is normally the case. it can be somewhat difficult to know if the effect is purely based on that particular product. To somehow remedy that possibility, I think it would be good, as I mentioned in my previous post, to inform that among those who noticed betterment from taking this product, 58% of them also was using this, 35% was using this and 93% this.

    That way it will be more easy to mimic the treatments other took and to achieve similar effect - hopefully.

    Of course, this can also be done with all sorts of data like symptoms, conditions, even climate for that matter.
  11. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    What I had at the back of my mind but did not say too explicitly (to allow other people's ideas to germiante) was that when we as a group decide to record symptom (e.g. polyyuria) we create a rating scale for that symptom (e.g. waking up at night to wee, weeing more than hourly by day, hourly, etc etc ) so that we can record symptoms before treatment, then record improvement after a supplement (or worsening).

    So yes, we would need some form of rating scale that gives fixed reference points in order to provide good quality data and avoid 'good mood' or 'wishful thinking' improving evaluations. Generating those scales would be a good bit of work for us all, but right now, I can't see a way around that issue unless existing scales are available.
    optimist likes this.
  12. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    Have not started thinking about technology as yet. I prefer to be crystal clear what we want to achieve first. This is how I usually run a project:
    1. get very clear how a system is going to improve the lives of it's end-users (our main focus now)
    2. work out how it's going to achieve that in terms of facilities provided and information stored (helps with the above)
    3. think about what's needed in terms of data and processes (the logical part of the programming stuff....)
    4. design the system
    5. select the technology.
    6. then worry about the programming etc.
    I never do more than two steps at once, that avoids muddying the water and wasting lots of effort. Bearing in mind that Phoenix is completely run by volunteers, I will also need to consider how common technical skills are in whatever we choose too - more common technology has it's advantages.

    I will check out the link, but for now lets keep our focus on what it is going to do for members though, and discuss the technical stuff with people whose interest lies there.

    Leo
    optimist likes this.
  13. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    No worries, we all suffer the same issues

    Sounds good, perhaps we could come up with some concrete examples that would 'bring it to life' for people reading the discussion?

    Two possible approaches to this spring to mind:
    1. Allow members to search for other members with similar symptoms before treatment and measure overall 'wellness' (e.g. bell scale) before and after treatment. (easier to fill in the data, easier to develop but less exact).
    2. Above plus allow members to provide concrete evaluations of particular symptoms so we can evaluate its effect for constipation vs for inflammation. Much harder to do, will require far more input from members both to develop systems and to fill forms in, but far far better data.

    Sounds good.
  14. daisybell

    daisybell Senior Member

    Messages:
    172
    Likes:
    298
    New Zealand
  15. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    daisybell likes this.
  16. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    Thanks for that hip,

    The first link raised one obvious question, should we filter results where only one or two people have given information? Very many treatments on that list were reported by only one person. That makes it very difficult to know whether it was a unique response or an interaction with another treatment, it also produces large numbers of results of which a lot are of that type. Your thoughts everybody?
  17. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    Good thinking! Much better than using a numeric scale! I guess it will be well worth the time extra for defining such scales, as the data will be of much better quality. It will not be based on "I feel I am a three on a scale of ten", but it will be "I wee once an hour".
  18. Leopardtail

    Leopardtail Senior Member

    Messages:
    999
    Likes:
    604
    England
    What did you have in mind under 'activities' here?
  19. optimist

    optimist Senior Member

    Messages:
    139
    Likes:
    102
    Norway
    Before ME I had a memory like glue, now it can be more like Teflon :) I believe I meant ways of exercising.
  20. MeSci

    MeSci ME/CFS since 1995; activity level 6

    Messages:
    3,261
    Likes:
    3,665
    Cornwall, UK
    Not time to read all messages in full yet, but generally a good idea. It is likely to suffer logistical problems, notably the phenomenon of providing more questions than answers, as is often found in surveys (e.g. the first one on Hip's list). What name/s do we use for supplements, what do we do in the case of overlaps (e.g. whole extracts vs. individual ingredients), what do we do about timescales of outcomes (look at short- and long-term outcomes? defined how?), allowing for editing/correction (like a Wiki?), how do we define outcomes (symptoms, blood results...), what about diagnostic criteria, etc. Not saying that the obstacles are insurmountable by any means, but it is likely to require a huge amount of work and discussion, unless this has already been done and some of the existing databases are good. Not time to look at them at present.

    Sorry if I have repeated anything others have already said!

See more popular forum discussions.

Share This Page