Table 10 Adjusted ORs and 95% CIs from the LR model based on Colombian DENV data, where binary symptoms are coded as 1 (present) versus 0 (absent). Categorical predictors use treatment (indicator) coding; with the omitted level serving as the reference (baseline) category.

From: Integrating machine learning and time-to-event models to explain and predict risk of hospitalization due to dengue in Colombia

Variable

OR

95% CI lower limit

95% CI upper limit

Subregion: ORI

0.4875

0.4086

0.5811

Subregion: other

1.3506

1.0044

1.8190

Age group: other

1.2672

1.0352

1.5523

Age group: early childhood

1.5426

1.2209

1.9510

Abdominal pain

8.4232

7.1791

9.9096

Diarrhea

1.7371

1.3865

2.1856

Drowsiness

0.2312

0.1611

0.3328

Headache

0.5985

0.5104

0.7012

High hematocrit

1.6429

1.1025

2.5125

Hypotension

2.1205

0.9654

5.1912

Low platelet count

9.3092

7.9653

10.9110

Mucosal bleeding

2.4036

1.5660

3.7765

Rash

0.6710

0.5834

0.7716

Retro-orbital pain

0.5702

0.4926

0.6598

Vomiting

2.9380

2.3819

3.3879

  1. For subregion, the rows ORI and other represent contrasts to the baseline subregion (with other denoting the pooled non-baseline, non-ORI subregions). For age group, early childhood and other are contrasts to the baseline age category (with other pooling the remaining age groups). The intercept is omitted, as it is not interpretable as an OR
Back to article page