Would you like to discuss why they use Spearman instead of Kendall and if this negative binomial distrubition asumption is valid in this case?

It won't have any effect on the final result, which basically states that the data set is insufficient.