+ Reply to Thread
Results 1 to 5 of 5

Thread: Simulation of blood glucose

  1. #1
    Points: 4,373, Level: 42
    Level completed: 12%, Points required for next Level: 177

    Posts
    15
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Simulation of blood glucose




    Hello all,

    I would like to make a simulation model of blood glucose using a real dataset of patients' blood glucose levels.

    I have thousands of patients' glucose levels but the sampling frequencies are heterogeneous - each patient's sampling is different from the next. I am wondering if I can use these data to simulate many thousands of modelled patients with high sampling frequencies.

    Does anyone on here have any ideas, or know where to point me?

    I am more than happy to share some dummy data if that would help explain my problem.

    Thanks,
    t

  2. #2
    Omega Contributor
    Points: 38,374, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Simulation of blood glucose

    Can you better define by what you mean by heterogeneity in this setting as well as high frequency sample. What is the exact issue, some individuals are oversampled?

    Another option beyond simulations may be resampling approaches like bootstrapping.
    Stop cowardice, ban guns!

  3. The Following User Says Thank You to hlsmith For This Useful Post:

    tonytheviking (01-04-2014)

  4. #3
    Points: 4,373, Level: 42
    Level completed: 12%, Points required for next Level: 177

    Posts
    15
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Re: Simulation of blood glucose

    Thanks for the reply.

    The heterogeneity I refer to is with regard to how frequently the samples were taken from each patient. There was no controlled sampling protocol - samples were just taken as and when indicated according to clinical judgement. So some patients had 30 samples in the first 24 hours when others may have had only 2 or 3. The interval between samples is therefore highly variable.

    I want to use this dataset to simulate a cohort of patients with sampling every hour - this is quite a high frequency - for 24 hours.

    Perhaps I need bootstrapping for this; I don't know much about this technique so any pointers again would be very welcome.

    Best wishes,
    T

  5. #4
    Omega Contributor
    Points: 38,374, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Simulation of blood glucose

    Why would the time increment between measurements matter? Do you expect fluctuations in values? I am curious what type of inpatients these are? What are you going to use these data for in particular and why couldn't a linear regression model be used to predict glycemic value for a one hour unit increase when controlling for patient characterics? Is there a y systematic bias in who and when patients get level checked?
    Stop cowardice, ban guns!

  6. The Following User Says Thank You to hlsmith For This Useful Post:

    tonytheviking (01-04-2014)

  7. #5
    Points: 4,373, Level: 42
    Level completed: 12%, Points required for next Level: 177

    Posts
    15
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Re: Simulation of blood glucose


    Thanks again for the response. Yes, time does matter as the values change over time. The patients are intensive care patients.

    I am trying to simulate hourly blood glucose sampling, using a database of real glucose values from irregular sampling.

    I'm sure it's possible but I need some pointers.

    Cheers,
    T

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats