From 2008 to 2016, Colombia recorded 72,994 cases of HIV/AIDS. Most of these notifications were in men (Men: 53,432 cases, 73.20%; Women: 19,563 cases, 26.80%). HIV was the preponderant diagnostic when compared with AIDS in both men and women (Men: 42,787 HIV, 10,645 AIDS; Women: 16,322 HIV; 3,240 AIDS). During the same period of the study 21,898 of AIDS-related deaths were notified to the Health Ministry with a higher frequency in men (Men: 16,505 deaths, 75.37%; Women: 5,393 deaths, 24.63%).
Temporal trends analyses
Table 1 shows the results of the temporal trend of the HIV epidemic in Colombia, considering the incidence of HIV/AIDS and AIDS mortality rates in the nine years of the study. The incidence had an upward trend in the entire country, in all age groups except for 0-14 years of age (decreasing trend). Caribbean, Andean, Orinoquia, and Amazon regions also had an upward trend. Joinpoint regression identified two different periods in trend for the 15-44 age group with a higher APC in the second period (2008-2013: APC 6.2%, p=0.000; 2013-2016: APC 11.2%, p=0.000).
Different from the incidence, the AIDS mortality rate was stabilized in all Colombia, in both genders and all regions. Considering the age groups, the mortality rate had a decreasing trend in the 0-14 and 15-44 groups. However, in the 15-44 group the joinpoint regression distinguished two different periods with a decreasing trend in the first period and stabilization in the second one (2008-2013: APC -2.8%, p= 0.006; 2013-2016: 3.2%, p=0.41). The same also occurred in the Caribbean region where two periods were identified with different behaviors: between 2008-2011 the mortality rate was stable (APC -2.9%, p=0.079), after 2011 it increased (2011-2016: APC 2.4%, p=0.014).
Table 1: Temporal trend analysis of HIV/AIDS incidence and mortality in Colombia, 2008-2016
|
|
INCIDENCE
|
|
MORTALITY
|
Joinpoint results
|
APC (CI 95%)
|
P
|
|
APC (CI 95%)
|
P
|
Total
|
|
7.8 (6.7; 8.9)
|
0.000
|
|
-0.1 (-1.1; 1.0)
|
0.707
|
Gender
|
|
|
|
|
|
|
Male
|
|
9.8 (8.5; 11.1)
|
0.000
|
|
-0.3 (-1.5; 0.9)
|
0.464
|
Female
|
|
2.5 (1.1; 3.9)
|
0.005
|
|
0.7 (-0.3; 1.6)
|
0.158
|
Age (years)
|
|
|
|
|
|
|
0 - 14
|
|
-7.4 (-12.4; -2.0)
|
0.014
|
|
-9.7 (-16.6; -2.3)
|
0.018
|
15 - 44
|
|
8.3 (7.1; 9.5)
|
0.000
|
|
-1.7 (-2.8; -0.6)
|
0.005
|
|
2008-13
|
6.6 (4.6; 8.6)
|
0.000
|
|
-2.8 (-4.2; 1.2)
|
0.006
|
|
2013-16
|
11.2 (7.5; 14.9)
|
0.000
|
|
3.2 (-5.9; 13.1)
|
0.410
|
45 - 64
|
|
5.5 (3.7; 7.3)
|
0.000
|
|
0.6 (-0.5; 1.7)
|
0.342
|
65+
|
|
10.2 (6.9; 13.5)
|
0.000
|
|
7.5 (3.6; 11.6)
|
0.002
|
REGIONS
|
|
|
|
|
|
|
Caribbean
|
|
8.6 (6.3; 10.9)
|
0.000
|
|
0.6 (-0.6; 1.9)
|
0.317
|
|
2008-11
|
---------
|
----
|
|
-2.9 (-6.3; 0.7)
|
0.079
|
|
2011-16
|
---------
|
----
|
|
2.4 (0.8; 4.0)
|
0.014
|
Pacific
|
|
3.2 (-0.8; 7.3)
|
0.113
|
|
0.4 (-1.2; 2.0)
|
0.661
|
Andean
|
|
8.8 (7.2; 10.4)
|
0.000
|
|
-0.9 (-2.1; 0.3)
|
0.090
|
Orinoquia
|
|
10.2 (5.4; 15.3)
|
0.001
|
|
2.8 (-1.4; 7.2)
|
0.172
|
Amazon
|
|
5.1 (1.9; 8.5)
|
0.008
|
|
-0.6 (-5.0; 4.0)
|
0.736
|
Our next task was to investigate differences between gender. Table 2 shows an increasing trend of HIV/AIDS incidence in males in all regions and all age groups above 15 years old. A decreasing trend was observed only in the 0-14 age group (APC -7.5%, p=0.040). The joinpoint regression detected two periods of time when analyzing the 15-44 age group with higher APC in the second period (2008-13: APC 8.2%, p=0.000; 2013-2016: APC 15.0%, p=0.000). Similarly, the female incidence rate had a downward trend in the 0-14 age group, but upward trends were only detected after 45 years of age (45–64 years: APC 7.0%, p=0.000; 65+ years: APC 11.0%, p=0.024). Female incidence showed increases only in the Caribbean and Andean regions (APC 5.9%, p=0.006; APC 1.9%, p=0.014, respectively), as opposed to males.
Concerning the AIDS mortality rate, the only alteration in men was in the age group 65+ years old with a significantly increasing trend (APC 7.8%, p=0.003). However, although the general trend to the 15-44 age group has been described as stable, the joinpoint regression identified two different periods with a decreasing trend in the first period (2008-2012: APC -4.8%, p=0.002) and a stabilization in the second one (2012-2016: APC 0.5%, p=0.65). Among women, there was a decreasing trend in the 0-14 age group (APC -11.2%, p=0.045) and increasing trends in the 45-64 age group (APC 3.2%, p-0.033) and in the Caribbean and Pacific regions (APC 1.9%, p=0.010; APC 2.1%, p=0.007, respectively).
Table 2 – Temporal trend analyses of HIV/AIDS incidence and mortality among men and women in Colombia, 2008-2016.
|
|
INCIDENCE
|
|
MORTALITY
|
Joinpoint results
|
APC (CI 95%)
|
P
|
APC (CI 95%)
|
P
|
MALE
|
|
|
|
|
|
Age (years)
|
|
|
|
|
|
0 a 14
|
|
-7.5 (-14.0; -0.4)
|
0.040
|
-6.0 (-16.5; 5.8)
|
0.261
|
15 a 44
|
|
10.8 (9.2; 12.5)
|
0.000
|
-2.2 (-3.5; -1.0)
|
0.003
|
|
2008-12
|
---------
|
----
|
-4.8 (-6.6; -2.9)
|
0.002
|
|
2008-13
|
8.2 (6.2; 10.3)
|
0.000
|
---------
|
----
|
|
2012-16
|
---------
|
----
|
0.5 (-1.5; 2.5)
|
0.650
|
|
2013-16
|
15.0 (11.4; 18.8)
|
0.000
|
---------
|
----
|
45 a 64
|
|
5.0 (3.1; 7.1)
|
0.000
|
0.0 (-1.3; 1.4)
|
0.869
|
65+
|
|
10.1 (7.6; 12.7)
|
0.000
|
7.8 (3.6; 12.1)
|
0.003
|
Regions
|
|
|
|
|
|
Caribbean
|
|
10.1 (8.4; 11.8)
|
0.000
|
0.3 (-1.2; 1.8)
|
0.821
|
Pacific
|
|
5.5 (1.1; 10.1)
|
0.023
|
-0.3 (-2.4; 1.7)
|
0.600
|
Andean
|
|
10.6 (8.9; 12.4)
|
0.000
|
-0.8 (-2.2; 0.5)
|
0.149
|
Orinoquia
|
|
13.7 (8.6; 19.0)
|
0.000
|
2.9 (-1.5; 7.4)
|
0.186
|
Amazon
|
|
9.2 (4.0; 14.6)
|
0.003
|
0.5 (-5.1; 6.5)
|
0.867
|
FEMALE
|
|
|
|
|
|
Age (years)
|
|
|
|
|
|
0 - 14
|
|
-7.0 (-13.0; -0.6)
|
0.035
|
-11.2 (-20.9; -0.4)
|
0.045
|
15 - 44
|
|
1.6 (0.0; 3.3)
|
0.069
|
-0.5 (-2.2; 1.1)
|
0.404
|
45 - 64
|
|
7.0 (5.0; 9.0)
|
0.000
|
3.2 (0.4; 6.0)
|
0.033
|
65+
|
|
11.1 (1.9; 21.1)
|
0.024
|
6.6 (-1.0; 14.9)
|
0.081
|
Regions
|
|
|
|
|
|
Caribbean
|
|
5.9 (2.3; 9.6)
|
0.006
|
1.9 (0.6; 3.2)
|
0.010
|
Pacific
|
|
-2.1 (-5.2; 1.1)
|
0.139
|
2.1 (0.8; 3.3)
|
0.007
|
Andean
|
|
1.9 (0.6; 3.2)
|
0.014
|
-1.0 (-3.0; 1.0)
|
0.250
|
Orinoquia
|
|
3.5 (-1.1; 8.3)
|
0.125
|
2.9 (-3.3; 9.5)
|
0.316
|
Amazon
|
|
-0.8 (-4.8; 3.3)
|
0.613
|
-2.4 (-8.8; 4.3)
|
0.400
|
Spatial analysis
Figure 2 shows the spatial distribution of HIV/AIDS incidence (Figs. 2A, C, and E) and AIDS mortality rates (Figs. 2B, D, and F). One notices a territorial expansion of HIV/AIDS incidence and a contraction for AIDS mortality. The most impacted municipalities are localized in the Caribbean, Andean, and Orinoquia regions. After 2014, there was an increase in incidence and mortality rates in municipalities bordering Venezuela, in the Caribbean and Orinoquia regions.
All univariate spatial autocorrelations of Global Moran’s I indexes were positive and statistically significant to both HIV/AIDS incidence and AIDS mortality (HIV/AIDS incidence: 2008-2010 - I=0.15 p=0.00, 2011-2013 - I=0.28 p=0.00; 2014-2016 - I=0.22 p=0.00; AIDS mortality: 2008-2010 - I=0.26 p=0.00; 2011-2013 - I=0.24 p=0.00; 2014-2016 - I=0.23 p=0.00). Figures 3A, C, and E show the LISA maps for the HIV/AIDS incidence and figures 3B, D, and F for AIDS mortality rates. One notes the high-high clusters for both rates are located on the Caribbean coast and in the ‘coffee belt’ of the Andean region. The low-low clusters were restricted to the Orinoquia and Amazon regions. Also, between 2014 and 2016, a high-high cluster of incidence and mortality appeared in the Caribbean and Orinoquia regions bordering Venezuela (Figs.3C, E, and F).
Our next question was about the direction of the HIV epidemic expansion in Colombia. To answer it, we employed the Kernel density estimator. Figure 4 shows the Kernel maps of HIV/AIDS incidence (Figs. 4A, C, and E) and AIDS mortality rates (Figs. 4A, D, and F). In these maps one can notice a higher density in municipalities of the ‘coffee belt’ for both incidence and mortality rates. While the HIV/AIDS incidence expands from the North of the Caribbean to the Northeast and Midwest of Colombia, AIDS mortality contracted in zones toward the Andean region’s ‘coffee belt’.