Figure 5

Differences of distribution of ICD categories within the primary clustering (\(k=6\)). For every ICD category, we measure the frequency within the entire population (denoted in parentheses after the category name) and then plot the difference in frequency in different clusters compared to the population frequency. Categories are sorted by the magnitude of the largest difference found within any clusters for the category. Note that since patients may have many diagnoses, clusters can have a higher frequency than the population for many categories (e.g. Cluster 5).