Table 4

Multivariate linear regression models of HIV prevalence (/100,000 adults) among developing countries.1

Sub-Saharan African countries with heterosexual contact as primary mode of HIV transmission

Non-Sub-Saharan African countries with heterosexual contact as primary mode of HIV transmission

Non-Sub-Saharan African countrieswith homosexual contact or injection-drug use as primary mode of HIV transmission




No. of countries

Regression coefficient

p-value

R2

No. of countries

Regression coefficient

p-value

R2

No. of countries

Regression coefficient

p-value

R2


Unadjusted regression model

35

0.68

24

0.70

32

0.49

Percent of population Muslim

-0.013

0.023

0.031

0.02

0.021

0.51

Percent of population Christian

0.0015

0.84

0.021

0.04

0.017

0.003

Male circumcision prevalence2

-0.61

<0.001

-2.19

<0.001

-1.26

0.35

Adjusted regression model3

27

0.85

22

0.72

23

0.73

Percent of population Muslim

-0.011

0.026

0.014

0.19

-0.036

0.26

Percent of population Christian

-0.0044

0.48

0.0011

0.94

0.015

0.15

Male circumcision prevalence2

-0.82

0.001

-1.60

0.001

0.47

0.70


1 Analyses conducted with natural log of HIV seroprevalence.

2 Male circumcision prevalence was coded as 1 = low (<20%), 2 = intermediate (20–80%), and 3 = high (>80%).

3 Model adjusted by country-specific measures including years since HIV was first reported, major geographical region, percent of population younger than age 25, percent of female adult illiteracy rate, percent of children fully immunized for diphtheria, tetanus, and pertussis, and number of doctors per 100,000 people.

Drain et al. BMC Infectious Diseases 2006 6:172   doi:10.1186/1471-2334-6-172

Open Data