Cumulative Dissolution Custom Field

I am interested in creating a custom field for our release rate procedure to stray away from the use of excel sheets in the lab. The release rate procedure runs for 8 consecutive days , and the cumulative release rate is reported for the span of 8 days. there are 6 samples run. We use the amount generated from empower to calculate the cumulative amount  in the excel sheet. is this custom field possible to create for cumulative amount? 

Answers

  • Whats the exact formula you are looking to replicate- whats the formula in Excel?
  • Are you using dissolution option in Empower ?

    Assuming you mean cumulative dissolution (%) where :

    Sum (corrected conc) /mg tablet *100 = % Cum  Diss


  • We do not currently have the dissolution option installed. We take the amount from empower to calculate cumulative amount. the amount from day 1 is added to the amount from day 2 (from each respective vessel) , the amount from day 2 is added to the amount from day 3  and so on... for 8 days . the calculation in excel is shown below 
  • Are you waiting until all samples are accumulated to run the LC analysis?  If so, you could probably create a few custom fields to address cumulative values based on sample name/label.  If they are in different sample sets, then it gets really messy and I don't think it's feasible.
  • This is run in the same sample set over an 8 day period 
  • Empower2019
    edited December 2019
    You could put the same label on all pairs of samples so Day 1 and Day 8 have the label Unk1 and Day 2/Day 8 are labelled Unk2 and so on, and create a peak, real, calculated custom field with the formula SAME.%..SUM(Amount), search order Result Set Only. This will sum the amounts from the Day1/Day8 samples which can be thrown into a peak table and reported. You would have to ensure your calibration and quantitation functions are not impeded by these labels so for Quantitate you enter U* which captures all labels starting with U...Oh and this will only work when the standard sequence is set to average through run, if you do any bracketing such as sliding with/without overlap, that custom field wont work. Average or Sum amount values do not work when multiple calibration curves are generated. 

    Make sure to put Summarize Custom Fields as the last line in the sample set. As mentioned, if there are multiple result sets it gets very awkward, with something like a search order of Result Set First being required and the formula staying the same but obviously you would need distinct labels for further runs of this kind to prevent the labels being included unnecessarily eg Unk1 over 3 or 4 result sets. Awkward!
  • gdun
    edited December 2019
    thank you! for each day , before each set of samples , 6 standards are injected upon which the samples are quantitated against (the average), will this not work? is there any way that it could work with this setup?  see below    


  • I dont think it will work like that because you are essentially creating four calibration curves in your sample set - one curve for the first 6 set of samples and the "Standard 1L" (im not sure what this achieves since you are injecting a standard and only the diluent is quantitated off this before being cleared?)and when that curve is cleared, a third curve for the next 6 samples then "Standard 1N". Im not 100% on that though so do try my formula in the previous post and see if it gives the correct value. 

    Does your test method specify to quantitate 6 samples from the preceding standard or does it just state to inject a standard every 6 samples? If the latter then you could build one calibration curve and quantitate all 12 samples off this curve, so inject standard x6 then 6 samples, then another standard x 6 but no clear calibration then the second set of samples, followed by a quantitate line at the end with a label reference of U*, using the label examples in my previous post. I do know that summaries of fields based off the calibration curve like Amount get screwed up when more than one calibration curve is generated. 
    This all depends on what your method says in relation to standards..
  • shaunwat
    edited December 2019
    Volume Corrected cumulative release:

    Pull 1 (CF named Disso1): HR1.%.%(((Disso_Amount_Released*(Vessel_Volume))/Label_Claim)*100)

    Pull 2 (CF named Disso2): HR2.%.%((((Disso_Amount_Released*(Vessel_Volume-(Sample_Aliquot)))+(((Sample_Aliquot*Disso_Spec)/(100*Vessel_Volume))*(Disso1)))/Label_Claim)*100)

    Pull N (CF named DissoN): HRN.%.%((((Disso_Amount_Released*(Vessel_Volume-(N*Sample_Aliquot)))+(((Sample_Aliquot*Disso_Spec)/(100*Vessel_Volume))*(DissoN-7+DissoN-6+DissoN-5+DissoN-4+DissoN-3+DissoN-2+DissoN-1)))/Disso_Spec)*100)

    Convention for Pull N: The DissoN-7 fields should be read as Disso(N-7) where you are not actually subtracting 7 from the disco result. In the example given, The Pull N was from the 8th sample pull, thereby making DissoN-7 = Disso1 and DissoN-1 = Disso7. 

    Getting Empower to do this calculation without the disso option all comes down to a mathematics exercise. We don't have the disso option either and I have validated the use of these custom fields against excel sheets for our sotax disso application.
  • thank you all ! one more question: Is there a way that empower can calculate the slope of the line for cumulative release over the 8 day period? we use the average of the slopes  to report the release rate. 
  • Since you aren't doing bracket calibrations and you are completing everything in a single sample set, the following approach may work or at least is something to consider...but it may all hinge on the "Standard 1L/1N" and if you then have instructions at the bottom for processing (as Empower2019 alludes to).  This approach would rely on all processing steps being set to "normal" and not jumping around.

    You'll have to specify the sample labels as the same for each sample across the brackets, so instead of smp for all, something like smp1 for sample 1, smp2 for sample 2, and so on.

    Then you could consider a boolean/ENUM field perhaps.  For a boolean approach, you'd have 6 fields resembling something like EQI(Label, "smp1")*Amount+smp1...SUM(Amount) with the smp1 reference swapped for each sample of course.  With an ENUM, possibly something where the ENUM calc is a label check like EQI(Label, "smp1")...but where the translation table calls out a field like Amount+smp1...SUM(Amount) (fc) perhaps?

    Don't utilize a summarize custom fields at the end, and I think this might work as the processing logic would look at smp1 (tp1) and say "True" for the EQI test, and return the amount for that injection and a 0 for the sum.  At the second timpeoint, it would return a true and therefore the amount for the second timepoint and then the sum portion would really just return the result from timepoint 1 as that's all that exists at that moment.

    OPTION2:
    Beyond that, I think it would get super messy.  While this alternate approach should work regardless of processing, it means you are creating 6 fields for each timepoint...
    Labels would need to be something like smp1tp1 (something to signify the sample and timepoint).  Then you could do something like the following perhaps:

    For each sample:
    Timpeoint1: EQI(Label, "smp1tp1")*Amount
    Timepoint2: EQI(Label, "smp1tp2")*Amount+smp1tp1...(Amount)
    Timepoint3: EQI(Label, "smp1tp3")*Amount+smp1tp1...(Amount)+smp1tp2...(Amount)
    and so on for each timepoint...

    I'm not sure there is any other way to simplify it as any boolean test needs the specify the criteria absolutely and can't use wildcards.