Introduction

Colombia is located in the north of South America, is the third-largest in the Latin America region, with 48 million in habitants¹, of which 77.1% live in the capital cities. As of June 28, 2019, the country’s political and administrative division is: 32 departments, 1,101 municipalities, and island (San Andrés, Providence and Santa Catalina) and 20 non-municipal areas. Additionally, departments are grouped into six geographical regions: Caribbean, Central, Pacific, Eastern, and Bogotá C.D. defined by the National Administrative Department of Statis tics (DANE, by its acronym in Spanish)².

According to the latest data related with the 90-90-90 targets in Latin America the 80% (62 - >95%) of people living with HIV and AIDS (PLWHA) knew their status, 62% of PLWHA accessed to antiretroviral therapy (ART), and 55% were virally suppressed², which may translate that the region has additional challenges.

By 2018, in Latin America there were an estimated of 100,000 (79,000 - 130,000) people with acquired HIV, a 7% increa se compared with 2010. The highest increases were observed in Chile (34%) and Bolivia (22%) and the lowest in El Salvador (-48%). In Colombia there are 160,000 (130,000 - 180,000) PLWHA and ranks the third among countries with the lowest rates of HIV infection (incidence per 1,000 population: 0.14 (95% CI 0.10-0.19)). The region’s incidence-prevalence ratio continues to decrease, reaching 5.4% (95% CI 4.1-6.8%), but many efforts are needed to reach the 3.0% epidemic transition benchmark³.

In Colombia, the health system is funding by public and pri vate sources, and its coverage is closer to 96% of the total population, the remaining 4% is grouped under individual in surance. There are two insurance regimes, and both of them include the same services, procedures, medicines, and inter ventions. However, different health payers are responsible for managing resources and paying to healthcare providers⁴.

Since the implementation of the national registry of HIV/AIDS managed by the High-Cost Diseases Fund (CAC, by its acronym in Spanish) in 2011-2012 and actualized within the framework of the national resolution 0273 (2019), 109,056 prevalent cases and 10,930 new cases of HIV have been reported until 2019⁵. We aim to assess the epidemiological situation (incidence, pre valence, and mortality) of PLWHA in the municipalities and geo graphical regions of Colombia during 2018.

Materials and methods

We performed a cross-sectional study with secondary data sources, including the information of PLWHA reported to the CAC from February 1st, 2018 to January 31st, 2019, by their health payers and providers.

Health providers must collect annual data from their affiliates to update the information on a web platform, which was de signed to validate the structure, coherence and consistency of the data. Once the information is reported, a cross-checking with other official sources of vital statistics and the affiliation of PLWHA to the health system with the “Unique Affiliate Da tabase” (BDUA, by its acronym in Spanish) is performed.

A data monitoring process is used to validated that the infor mation uploaded to the platform corresponds to the electro nic/printed medical records. This ensures that the informa tion reported corresponds to described in the supports. We included all confirmed cases with HIV, classified according to the clinical practice guidelines for the care of HIV of the Mi nistry of Health and Social Protection (MSPS, by its acronym in Spanish) ^(6,7.

If the information reported is different from the observed, the data available in the supports were captured and co rrected in the database; otherwise, when the data cannot be found in medical records was considered as missing. To ensure the anonymization of the PLWHA, a unique number was assigned, which also made it possible to include unique registries to estimate the epidemiological indicators. People who died before of the beginning of the period and those without a confirmed diagnosis of HIV/AIDS were excluded from the analysis (Figure 1).

Figure 1 Flow chart of cases included in the analysis 

We included sociodemographic variables, e.g., age and sex, and related to location: geographical region (classification given by DANE) according to the gross domestic product of the departments (Figure 2), as well as department and mu nicipality of residence; variables related to the clinical con dition were the current clinical status of the disease, CD4 T cells counts- and HIV viral load. Variables were analyzed with measures of central tendency according to their marginal distribution. Categorical data were presented as absolute va lues and proportions.

Figure 2 Distribution of geographical regions of Colombia according to DANE, 2018. 

We calculated the three main epidemiological indicators: in cidence, prevalence and mortality. HIV incidence was defined as PLWHA, whose infection diagnosis date occurred within the reporting period, and prevalence was defined as people diagnosed with HIV and reported during the period. Finally, to estimate the mortality, deaths were identified by using the administrative information reported by the providers, and verified in the single register of affiliates from the MSPS. All of these indicators were estimated according to the Colom bian geographical regions, department and municipality of residence, highlighting the places with the highest rates.

Crude and age-standardized rates were calculated by using the direct method, taking the Colombian population projec ted by DANE (n= 49,834,240) until July 2018 as the reference. National estimations were standardized using the world po pulation projected by the United Nations for Latin America and the Caribbean for 2020⁸. Morbidity and mortality in dicators were plotted using QGis 3.12 (Open Source Geos patial Foundation). STATA version 13.0 (STATA Corp, College Station, Texas, USA) was used to the statistical analysis. Ac cording to resolution 8430 of 1993 of the MSPS of Colombia, this investigation has no risk, and not intervention was ca rried out. However, the confidentiality and anonymization of the information was guaranteed.

Results

HIV incidence in 2018

In 2018, 10,930 new cases of PLWHA were reported. A sum mary of the characteristics of the study population is provi ded in Table 1. PLWHA frequency was higher in males, aged between 25 to 49 years. The median age was 30 years (IQR 24 - 40). Regarding the clinical status, 39.32% (n=4,298) cases were reported with AIDS (stage 3) and 35.27% (n=3,855) had undetectable HIV viral load.

Table 1 Clinical and sociodemographic characteristics of people living with HIV in Colombia, 2018 

a The regions in Colombia are made up of departments. Departments are formed by a grouping of municipalities. The grouping of regions depends on the department’s gross domestic product (GDP) and is available in anual reports from the DANE (Departamento Admistrativo Nacional de Estadística) (2). Geographical regions are: Bogotá, D.C.; Caribbean: Atlántico, Bolívar, Cesar, Córdoba, La Guajira, Magdalena and Sucre; Central: Antioquia, Caldas, Caquetá, Huila, Quindío, Risaralda and Tolima; Eastern: Boyacá, Cundinamarca, Meta, Norte de Santander and Santander; Pacific: Cauca, Chocó, Nariño and Valle del Cauca; Other departments: Amazonas, Arauca, Casanare, Guainía, Guaviare, Putumayo, San Andrés, Vaupés and Vichada.

The national age-standardized incidence rate was 22.12 per 100,000 population (95% CI 21.71-22.54). The highest inci dence rates were estimated in Bogotá, C.D. and Central re gions, which were even higher than the national (Table 2). The five municipalities in Colombia that reported the highest incidence were: Florencia (Cauca) 354.28 per 100,000 (95% CI 106.1-766.89), Barranca de Upía (Meta) 251.10 (95% CI 49.03- 852.66), Magangué (Bolívar) 228.53 (95% CI 195.24-266.02), Armenia (Antioquia) 161.41 (95% CI 63.78-340.37) and Agua da (Santander) 128.96 (95% CI 3.26-662.27). Figure 3A shows the municipalities with a higher incidence. In contrast, the lowest incidences were observed in Madrid (Cundinamarca), Cimitarra (Santander), and Potosí (Nariño). The supplemen tary table 1 describes the data for the 1,122 municipalities in Colombia.

Table 2 Epidemiological indicators of HIV/AIDS in Colombia, 2019 

*Per 100,000 population

** Per 100 population

Figure 3 Distribution of HIV incidence and prevalence in Colombia, 2018; 3a. HIV incidence in Colombia, 2018; 3b. HIV prevalence in Colombia, 2018 

HIV mortality in 2018

In 2018, 1,802 deaths were reported in PLWHA. The characte ristics of people who died during the period are provided in Table 1. Mortality was higher in males, aged between 25 to 49 years. The median age was 43 years (IQR 34 - 55). Regarding the clinical status, 84.35% (n=1,520) had AIDS. The majority of cases had no measurement of the last viral load (42.01%; n=757) and 42.51% (n=766) did not received ART.

The age-standardized mortality rate was 3.78 per 100,000 population (95% CI 3.61-3.96), with an increase of 56.4 % in the number of deaths compared to 2017. The highest mor tality rates were estimated in the Caribbean and Central re gions, which were even higher than the national (Table 2). The five municipalities in Colombia that reported the highest mortality were: Armenia (Antioquia) 283.08 (95% CI 8.07- 1,067.90), Saladoblanco (Huila) 141.13 (95% CI 3.57-542.00), Morales (Bolívar) 113.94 (95% CI 13.80-324.36), Vigia del Fuerte (Antioquia) 94.22 (95% CI 3.54-347.12) and Calamar (Guaviare) 78.97 (95% CI 9.56-260.47). Figure 4 shows the municipalities with the highest mortality. Moreover, the municipalities with the lowest mortality were Patía (Cauca), Santa Rosa del Sur (Bolívar), and Piendamó (Cauca) (supple mentary material, table 1).

Figure 4 HIV mortality in Colombia, 2018 

Discussion

In 2018, there was a slight increase of almost 7% in the num ber of newly detected HIV infections compared with 2017³. The estimated prevalent cases of HIV in 2019, reported by the MSPS through the Spectrum tool, was 157,702 from those, only 69.15% (n=109,056) were reported to the CAC. The final number of PLWHA may be higher. Regarding incident cases, around 10.000 new cases were reported to the CAC and were attended by the health system, while, the Institute of National Health-INH estimated 15,908 new HIV cases by 2019⁹.

The observed gap could be explained because while the CAC records the information provided by the health insurers, the INH performs the epidemiological surveillance of the newly diagnosed cases. This difference can be given by under-registration or the inability of the health system to linkage to care and retain the newly diagnosed HIV population. It is important to highlight that Colombia has a registry of all PLWHA treated within the framework of the health system, established by a resolution from the MSPS and all health in surers in cooperation with their healthcare providers must re port all cases to the CAC; for this reason, our approach does not incorporate methodologies such as UNAIDS, which are based on estimates or mathematical models^10,11.

The HIV infection was more frequent in municipalities located in four of six regions of Colombia. The highest age-standar dized incidence and prevalence were reported in Florencia (Cauca) (354.28 per 100,000 and 3.32 per 100 persons, res pectively). Among those municipalities, we found that almost 60% of cases had AIDS (stage 3); the above indicates delays in linkage-to-care of people with HIV for early detection as well as the continued risk of onward transmission. If the goals of 90-90-90 are considered, it is evident that Colombia has a limitation in achieving especially the first 90 related to diag nose at least 90% of the infected population. Also, the fact of detecting people when they are already in an advanced stage (34% of incident cases and 54% of people who died) should be an alert to establish early detection programs that include new test and treatment strategies.

According to mortality data, municipalities in the Central region had the highest mortality rates: Armenia (Antioquia) 283.08, Saladoblanco (Huila) 141.13, and Vigia del Fuerte (Antioquia) 94.22 per 100,000. These results may be explai ned because most cases were diagnosed in stage 3 (84.35%) and had a viral load ≥ 1000 copies/ml. It is important to note that in general, PLWHA, once they are admitted in HIV care programs, maintain a high rate of virological success (viral load <200 copies / ml) (83%).

Regarding the geographical location of PLWHA in Colombia, previous studies have described the HIV epidemic as a disea se affecting the entire country¹². In our study, the trends in all the indicators are consistent in the three major regions, and geographical differences may be explained by ethnic and cultural factors and their relationship with sexual behavior¹³.

On the other hand, there is a high proportion of people who did not receive ART (20.96%; 16.24% and 42.79% in incident, prevalent and death population, respectively). Therefore, emphasis should be placed on the initiation and retention of highly active antiretroviral therapy (HAART). Reasons to ex plain that patients have no HAART are various, from lack of adherence, administrative purposes, and not offered by prac titioners. The last reason is due to the current clinical practice guideline is outdate and requires and update including the best evidence available^6,7.

Some strengths of the study will be discussed. Taking into account that HIV/AIDS is considered a high-cost disease of public health interest in Colombia, PLWHA must be re ported by their health care insurers and providers to the administrative registry managed by the CAC and, that en sures the completeness of the data. The mandatory natu re of the reporting process also allows an epidemiological and clinical follow-up based on real-world information. In addition, there is a competitive compensation mechanism within the health system that obliges insurers to report the variables requested by regulation and allows the characte rization of the population with HIV. Among other strengths, data is verified by a data monitoring process against clinical records. Further, information reported to the CAC is one of the sources that MSPS provides to international initiatives such as GAM (Global AIDS Monitoring 2020 - Indicators for monitoring the 2016 United Nations Political Declaration on Ending AIDS) ⁽¹⁴.

Our estimations could be limited due to under-reporting effect. In fact, we only have the information reported for people enrolled with an authorized public or private insuran ce agency of the national healthcare system.

Our results allow identifying the geographical and clinical characteristics of PLWHA as well as the epidemiological si tuation (incidence, prevalence, and mortality) in Colombia. While there are 14% of the municipalities without HIV ca ses, it is necessary to keep the infection under control and to strengthen the reporting process to the CAC in order to avoid the under-reporting. Furthermore, future work is needed to integrate data from different sources and follow-up programs within the primary care¹⁵.

Finally, different strategies should be implemented to improve early detection of risk factors and prevent transmission, emphasizing on specific protection activities and monitoring other sexually transmitted infections. Also early detection strategies guarantee a timely linkage-to-care programs and avoid delays in ART initiation¹⁶.