Evaluation of the HIV drug resistance surveillance system in Mozambique, 2017-2018

In the past ten years, the prevalence of primary Human Immunodeficiency Virus (HIV) drug resistance has ranged from zero to 25%, with higher and increasing rates in countries with access to antiretroviral therapy (ART), a specific case in Mozambique. World Health Organization (WHO) recommended that countries implement and routinely evaluate representative HIV drug resistance (HIVDR) research to monitor the emergency and transmission of HIV drug resistance mutations. This study aimed to describe the functioning of the system and also to identify gaps in the sensitivity, representativeness and quality of the data using the WHO methodology for Pre-Treatment and Acquired Approaches. We conducted a descriptive evaluation of the information system for surveillance of HIVDR in Mozambique in 2017-2018, based on updated guidelines for evaluating of public health surveillance systems from the Center for Disease Control and Prevention (CDC). The evaluation was conducted in all provinces using secondary data extracted from a cross-sectional survey database on HIVDR, with HIV positive cases at the beginning of ART aged ≥15 years. The system was described through informal conversations with HIVDR stakeholders and the simplicity, data quality and representativeness attributes were evaluated. With 322 positive cases at the beginning of ART (mean age=32.5 years, SD±11.1), about 63.0% (203/322) cases were women and 37.6% (121/322) men. The system was implemented in 25 health facilities distributed across all 11 Mozambican provinces and was considered representative. The system used two data collection instruments, the ART book and the form accompanying samples sent to the reference laboratory. The ART form, with 27 variables, was sent offline at two levels (health facility and National Institute of Health (NHI)), accompanied by dried blood spot samples for viral load testing and genotyping in the NHI virology laboratory, and was considered simple according to the standardized criteria. The system´s data quality was considered regular at 79.9%, with about 59.8% (1156/1932) of variable fields completed and 100% (1932/1932) consistency. The system used a single national laboratory to measure the prevalence of resistance to HIV drugs and was considered simple, with regular quality and representative data. We recommended public health efforts such as conducting genotyping tests be expanded to the provincial level, and periodic monitoring of system´s data collection procedures using forms.