D3 age incorrect - ok to use D1_D2 to calculate current age?

89 views
Skip to first unread message

s.j.k...@googlemail.com

unread,
Nov 9, 2017, 7:40:32 AM11/9/17
to TADPOLE

I've noticed that AGE in D3 is baseline age, rather than current age. Current age can be approximated using this and VISCODE, but in a real situation you would have an accurate current age available. VISCODE is usually, but not always that accurate.

My question is, are we allowed to use D1_D2 to calculate current age for D3 patients?

Less important, but can VISCODE in D3 be used directly in the D3 predictions? I don't think it should be because it contains stealthy longitudinal information, i.e. an MCI with a high VISCODE is less likely to have underlying AD.

Many thanks in advance,
Steve

Razvan Marinescu

unread,
Nov 10, 2017, 5:39:05 AM11/10/17
to tadpolec...@googlegroups.com
Yes, this was indeed a mistake. You can use the corrected age from D1_D2.

We plan to provide a script for everyone that gives the correct age for every subject in D3, but until then you're welcome to do the correction yourself. We will make an announcement when the script is ready.

Raz

Neil Oxtoby

unread,
Nov 13, 2017, 7:33:41 PM11/13/17
to TADPOLE
Current age is VISITAGE = AGE + Years_bl and can be corrected in D1_D2 as follows:

MATLAB table:

dataTable_D1D2.VISITAGE = dataTable_D1D2.AGE + dataTable_D1D2.Years_bl;

% Add VISITAGE to D3 using a Left Outer Join of D3 to D1D2 on {RID,VISCODE}, keeping only VISITAGE from D1D2

dataTable_D3 = outerjoin(dataTable_D3,dataTable_D1D2,'Keys',{'RID','VISCODE'},'RightVariables','VISITAGE','Type','Left');


Python pandas DataFrame:
dataTable_D1D2['VISITAGE'] = dataTable_D1D2['AGE'] + dataTable_D1D2['Years_bl']
# Add VISITAGE to D3 using a Left Outer Join of D3 to D1D2 on {RID,VISCODE}, keeping only VISITAGE from D1D2
dataTable_D3
= dataTable_D3.merge(dataTable_D1D2[['RID','VISCODE','VISITAGE']],how='outer',on=['RID','VISCODE'])

Complete MATLAB/python code for loading the data, adjusting AGE to VISITAGE, and saving to CSV files again is available in the TADPOLE GitHub repo scripts:

Neil
Reply all
Reply to author
Forward
0 new messages