If both total joint counts are populated (tjc_observed) and (tjc_observed_28) and the values are different between them, should we select the largest value as a representation of the number of joints examined?

If the first (tjc_observed) is higher than the second (tjc_observed_28) then, we would advise using the first since that allows for a wider variety of joints to be counted. 

The 28 is a specific set of joints that feed into many of the disease activity measures; the broader one allows for any joints (including some not included in the TJC28) to be counted.