Skip to main content

Table 10 Main results for clinical file

From: De-identifying a public use microdata file from the Canadian national discharge abstract database

  TOTAL_LOS_DAYS TOTAL_LOS_WEEKS
  0.2 0.05 0.04 0.2 0.05 0.04
  Comb. Complete Comb. Complete Comb. Complete Comb. Complete Comb. Complete Comb. Complete
GENDER_CODE 0.002 0.003 0.005 0.005 0.006 0.006 0.002 0.004 0.004 0.01 0.004 0.012
AGE_GROUP 2.5 6.2 7.6 16.7 8.86 19.2 1.32 3.7 4.2 9.8 4.8 11.25
TOTAL_LOS_XXX 1.2 6.8 1.8 12.3 1.9 13.3 1.2 4.7 1.8 7.08 1.9 7.4
MRDx 15.8 16.4 30.4 29 33 31.1 10.2 10.8 19.8 20 21.6 21.7
CMG_CODE 15.8 16.4 30.4 29 33 31.1 10.2 10.8 19.8 20 21.6 21.7
DIAG_BLOCK 11.4 5.4 24.2 10.2 27 11.1 6.96 3.5 15.1 6.8 16.5 7.5
DIAG_CHAPTER 2.22 0.14 5 0.28 5.6 0.34 1.4 0.066 3.3 0.16 3.7 0.19
CCI_CODE 7.4 9.16 12.3 14 13.2 14.8 4.9 6.7 8.9 11.22 9.6 12
SHORT_CCI 7.4 9.16 12.3 14 13.2 14.8 4.9 6.7 8.9 11.22 9.6 12
Total % Cells Suppressed 7.08 7.74 13.77 13.94 15.1 15.1 4.6 5.2 9.1 9.6 9.92 10.4
Entropy (%) 100 123 201 237 219 259 50 68 114 143 126 158
  1. Missingness (as a percentage of all records) for different probability threshold levels and levels of LOS detail for the quasi-identifiers in the second PUMF for combinations ("comb") and complete algorithms. The last two rows show the percentage of cells in the quasi-identifiers that are suppressed and the entropy change from the baseline as a percentage.