Incorrect dispersion values, dispersion upper limits and dispersion lower limits in table outcome_measurements

[Nader-B]

Hello,

There is a problem with the dispersion measurements in the table outcome_measurements. In fact, in some cases the column dispersion_value is filled with data coming from a different outcome (it should be Null). In other cases, the two columns dispersion upper limit and dispersion lower limit have this same problem.

An example trial is https://clinicaltrials.gov/ct2/show/results/NCT01597388, in the 23rd outcome entitled 'Time to AZD2014 Peak Plasma Concentration at Steady State (Tmax,ss) on Cycle 1 Day 22, Continuous Dosing, With Fulvestrant' for the first arm, we have normally a dispersion_lower_limit equaling 0.5, a dispersion_upper_limit equaling 4 and no dispersion_value.

However, for this same outcome in the table outcome_measurements we find a dispersion_value equals to 447.9 instead of Null as it should be. The value 447.9 corresponds to the 22nd outcome. Hence, it shouldn't be put in 23rd one's measurements.

I joined two figures for the example i mentioned, the first one is the data in clinicaltrials.gov and the second corresponds to what we find in the table outcome_measurements.

Thank you

[crystal-wb]

Hello,

There is a problem with the dispersion measurements in the table outcome_measurements. In fact, in some cases the column dispersion_value is filled with data coming from a different outcome (it should be Null). In other cases, the two columns dispersion upper limit and dispersion lower limit have this same problem.

An example trial is https://clinicaltrials.gov/ct2/show/results/NCT01597388, in the 23rd outcome entitled 'Time to AZD2014 Peak Plasma Concentration at Steady State (Tmax,ss) on Cycle 1 Day 22, Continuous Dosing, With Fulvestrant' for the first arm, we have normally a dispersion_lower_limit equaling 0.5, a dispersion_upper_limit equaling 4 and no dispersion_value.

However, for this same outcome in the table outcome_measurements we find a dispersion_value equals to 447.9 instead of Null as it should be. The value 447.9 corresponds to the 22nd outcome. Hence, it shouldn't be put in 23rd one's measurements.

I joined two figures for the example i mentioned, the first one is the data in clinicaltrials.gov and the second corresponds to what we find in the table outcome_measurements.

Thank you

Hello Nader -B,

I am sorry for the delay. While looking into the incorrect values being set for this particular study, we discovered that they are displayed that way on ClinicalTrials.gov. Here is the url we are using for saving the study data: https://clinicaltrials.gov/ct2/show/NCT01597388?resultsxml=true

Here is a quote from Karen Chiswell Ph.D. after looking into the issue:
"The records for the first outcome (#22) are OK. The records for the second outcome (#23) have the dispersion_value_num_numeric populated when these should be null. For this outcome (#23), the dispersion is reported by the range (min – max) and these values are given in the dispersion_lower_limit (min) and dispersion_upper_limit (max) variables.
I think the error is in the XML though. Perhaps the data submitter didn’t enter the data correctly (some data entry got incorrectly carried down from a previous outcome record), or perhaps it isn’t being mapped correctly in the XML? "

So we are taking the information the way they are delivering it, however it's delivery seems to be a wrong in this case.

[micronix]

Looks like this is resolved.