Skip to main content

Table 2 Univariable and multivariable logistic regression analysis in the derivation dataset (N = 2838)

From: Risk scores for predicting early antiretroviral therapy mortality in sub-Saharan Africa to inform who needs intensification of care: a derivation and external validation cohort study

 

Alive/TF by 6 months of ART (N = 2755)

Died by 6 months of ART (N = 83)

Unadjusted

Model A—adjusted model excluding CD4

Model B—adjusted model including CD4

n

N

Median (IQR)/%

n

N

Median (IQR)/%

OR

95% CI

p

AOR

95% CI

p

AOR

95% CI

p

Demographics

 Age, years (for every 10-year increase)

 

2755

34 (29–41)

 

83

39 (31–49)

1.44

(1.19–1.73)

< 0.001

      

 Sex and pregnancy status

Pregnant

517

520

99%

3

520

1%

1.00

1.00

1.00

Female non-pregnant

1379

1418

97%

39

1418

3%

4.87

(1.66–14.3)

0.004

2.45

(0.76–7.88)

0.133

2.04

(0.68–6.09)

0.201

Male

859

900

95%

41

900

5%

8.23

(2.72–24.91)

< 0.001

5.47

(1.49–20.17)

0.011

4.35

(1.27–14.88)

0.019

 Marital status

Married/civil union

292

300

97%

8

300

3%

1.00

      

Single

2367

2441

97%

74

2441

3%

1.14

(0.65–2)

0.646

      

Widowed/divorced

96

97

99%

1

97

1%

0.38

(0.05–3.14)

0.369

      

 Smoking history

Never

2262

2321

97%

59

2321

3%

1.00

      

Current/ex-smoker

493

517

95%

24

517

5%

1.87

(1.11–3.15)

0.019

      

 Employed

Employed

1233

1270

97%

37

1270

3%

1.00

      

Unemployed

1522

1568

97%

46

1568

3%

1.01

(0.68–1.48)

0.971

      

 Education

None

188

196

96%

8

196

4%

1.00

      

Primary

664

687

97%

23

687

3%

0.81

(0.36–1.85)

0.610

      

Secondary

1687

1734

97%

47

1734

3%

0.65

(0.31–1.39)

0.265

      

Higher

216

221

98%

5

221

2%

0.54

(0.23–1.26)

0.300

      

HIV/TB history

 Previous TB treatment

No

2489

2561

97%

72

2561

3%

1.00

      

Yes

266

277

96%

11

277

4%

1.43

(0.74–2.77)

0.290

      

 Number of WHO TB symptoms

0

1955

1975

99%

20

1975

1%

1.00

1.00

1.00

1

407

427

95%

20

427

5%

4.80

(2.8–8.23)

< 0.001

3.39

(1.88–6.09)

< 0.001

3.16

(1.84–5.43)

< 0.001

2

188

202

93%

14

202

7%

7.28

(3.58–14.8)

< 0.001

4.03

(1.82–8.92)

0.001

3.64

(1.63–8.12)

0.002

3 or 4

205

234

88%

29

234

12%

13.83

(9.04–21.13)

< 0.001

5.05

(3.31–7.7)

< 0.001

4.68

(2.93–7.48)

< 0.001

Clinical characteristics

 WHO stage

I/II

2441

2484

98%

43

2484

2%

1.00

1.00

1.00

III/IV

314

354

89%

40

354

11%

7.23

(3.87–13.52)

< 0.001

2.57

(1.32–4.99)

0.005

2.47

(1.24–4.89)

0.010

 CD4 (per 10-cell increase)a

 

2755

249 (149–317)

 

83

98 (41–218)

0.94

(0.9–0.98)

0.002

   

0.98

(0.95–1.01)

0.211

 Weight (per 1-kg increase)b

 

2755

59 (52–68)

 

83

51 (45–60)

0.96

(0.93–0.98)

0.001

      

 BMI (per 1-unit increase)b

 

2755

21.6 (19.0–25.0)

 

83

19.0 (17.0–21.8)

0.86

(0.79–0.94)

0.001

      

 Hemoglobin (per 1-g/dL increase)

 

2755

11.9 (10.5–13.3)

 

83

9.9 (8.5–11.7)

0.69

(0.61–0.79)

< 0.001

0.73

(0.65–0.81)

< 0.001

0.74

(0.67–0.81)

< 0.001

 Temperature (per 1 °C increase)c

 

2755

36.2 (35.8–36.5)

 

83

36.5 (36.0–37.0)

2.09

(1.47–2.96)

< 0.001

1.26

(0.96–1.65)

0.092

1.25

(0.94–1.67)

0.127

 Heart rate (per 1 beat/min increase)d

 

2755

84 (75–94)

 

83

102 (85–121)

1.05

(1.04–1.06)

< 0.001

      

 Respiratory rate (per 1 breath/min increase)

 

2755

18 (18–20)

 

83

20 (18–22)

1.02

(1.01–1.04)

0.009

      
  1. Abbreviations: ART antiretroviral therapy, TF transfer-out, CI confidence interval, WHO World Health Organization, BMI body mass index, OR odds ratio, AOR adjusted odds ratio, IQR interquartile range
  2. aDue to non-linearity in the association between CD4 and log odds of death, CD4 was modeled as two terms (term 1 = X − .2432641563 and term 2 = X*ln(X) + .3438800025, where X = CD4/1000). Output shown is for the linear term. The p value associated with each CD4 term was < 0.001
  3. bBMI and weight were correlated, and BMI was retained as the preferred candidate variable. However, BMI was eliminated from the backward stepwise regression at p > 0.01
  4. cDue to non-linearity in the association between temperature and log odds of death, temperature was modeled as two terms (term 1 = temperature^3 − 47,148.67774 and term 2 = temperature^3*ln (temperature) − 169,123.2696). Output shown is for the linear term. The p value associated with each squared term for temperature was < 0.001
  5. dHeart rate was excluded as a candidate variable due to correlation with temperature