Skip to main content

Table 3 The 17 variables with highest NRIs for predicting presence of PASC among females, based on the optimal stochastic gradient boosting model with 15,377 trees, together with ORME during the year before PASC diagnosis

From: Using machine learning involving diagnoses and medications as a risk prediction tool for post-acute sequelae of COVID-19 (PASC) in primary care

Females

ICD- 10 or ATC code

Description

NRI (%)

ORME

U071 and U072

COVID- 19—in hospitalized patients

16.1

18.8

R53

Malaise and fatigue

14.5

4.6

G933

Post-viral and related fatigue syndromes

10.1

21.1

R06

Dyspnea

8.4

6.2

J01, J02, J03, J06 and J20

Upper respiratory tract infections

5.9

3.2

R43

Disturbances of smell and taste

4.4

28.9

B34

Viral infection of unspecified site

3.1

3.5

R03 A and R03B (ATC)

Adrenergics, inhalants and other drugs for obstructive airway diseases, inhalants

2.8

1.8

R00

Tachycardia

2.6

3.0

F43

Reaction to severe stress, and adjustment orders

2.3

2.1

R05

Cough

2.1

2.3

R51

Headache

2.0

2.6

Z86

Personal history of certain other diseases

1.7

5.6

R50

Fever

1.6

4.4

F41

Anxiety disorders

1.6

1.6

G03 A (ATC)

Hormonal contraceptives for systemic use

1.3

2.3

Z02

Encounter for administrative examination

1.0

2.2

  1. NRI normalized relative influence, PASC post-acute sequelae of COVID- 19, ORME odds ratio for marginal effects, ICD- 10 International Classification of Diseases Tenth Revision, ATC Anatomical Therapeutic Chemical