Skip to main content

Table 4 Boolean features which were significant under a Bonferroni-corrected p-value, together with their respective aggregated critical threshold

From: Use of machine learning to identify patients at risk of sub-optimal adherence: study based on real-world data from 10,929 children using a connected auto-injector device

Feature in the first 3 months (with respect to a threshold value)

Bonferroni- corrected p-value

3 months prediction

6 months prediction

9 months prediction

Mean adherence < 90%

< 0.001

< 0.001

< 0.001

Adherence SD > 9%

< 0.001

< 0.001

< 0.001

Number of transmissions = 0

< 0.001

< 0.001

< 0.001

Fast needle speed

< 0.001

< 0.001

< 0.001

Age at start > 10 years old

< 0.001

< 0.001

< 0.001

No dose changes

0.002

< 0.001

Injection depth > 6 mm

0.021

< 0.001

0.002

Gender (male)

0.042

0.013

  1. SD standard deviation