1. Background
Monitoring health programs is necessary to determine whether investments from governments and development partners are producing the expected results. Even when an intervention works well in a controlled setting, it may not perform the same way in routine health services. Everyday challenges like staff shortages, limited training, and weak infrastructure can reduce a program’s effectiveness. Therefore, it is important to track whether services are reaching the right people and improving treatment outcomes.
Traditionally, this is done by measuring key indicators related to processes, outputs, outcomes, and impact. The World Health Organization recognizes routine health facility data as an important source for monitoring these indicators, tracking trends over time, and comparing results across different locations (1). However, the value of health facility-based data depends heavily on its quality, completeness, and how well it is analyzed.
Population-based surveys and health facility studies provide strong evidence, but they are often costly and time-consuming. They require extensive planning, tool development, staff training, and separate data collection. In many low- and middle-income countries, these heavy resource demands limit how often such studies can be conducted.
Routine health information systems like DHIS2 offer a practical alternative. By collecting data during regular service delivery, these systems are already widely used to monitor program performance. When data quality is strong, individual-level systems like the DHIS2 Tracker make it possible to follow individuals over time and monitor long-term outcomes effectively.
In this article, we share lessons learned from Nepal on how the existing DHIS2 Tracker system supported cost-effective research for the national HIV program. Drawing on examples from studies that analyzed nearly 10,000 individuals in total, we demonstrate how routine program data was used to monitor viral load, track drug resistance, estimate loss to follow-up, and evaluate the elimination of vertical transmission of HIV (2-4).
2. Developing a Sampling Frame
A sampling frame is simply the list of people in a target population who can be selected for a study. In our nationally representative drug resistance survey, we needed to assess viral load status and acquired drug resistance among people receiving HIV treatment in Nepal. To do this, we needed an accurate list of people currently on treatment across health facilities enrolled in antiretroviral therapy (ART) so that participants could be selected randomly.
DHIS2 Tracker made this process much easier. It provided a real-time list of people living with HIV who were on treatment by health facility and using secure client codes. This allowed the study team to understand the client load at each facility and prepare the sampling frame without having to travel to the health facilities or study sites.
Using this information, the team applied a two-stage cluster sampling method. First, health facilities were selected based on their size. Then, people living with HIV were chosen from each selected facility using systematic random sampling. Because all the necessary baseline information was already available in the DHIS2 Tracker, the study team saved significant time and money. Without it, we would have needed to send trained field staff to health facilities just to count and list individuals, or rely on aggregated data that lacked the detail needed to apply our inclusion criteria.
3. Supporting Sample Size Estimation
DHIS2 Tracker also supports sample size planning. Because the system already holds exact figures on the number of eligible individuals at each health facility, study teams can generate the exact parameters needed for sample size calculations.
In Nepal’s acquired drug resistance study, DHIS2 Tracker data helped us instantly identify the number of eligible individuals at various health facilities before any sampling began. This allowed the team to calculate the required sample size efficiently and plan the survey based on concrete routine data, rather than relying on outdated estimates or starting from scratch. This capability helps teams plan national studies more efficiently and with fewer resources.
4. Streamlining Fieldwork Planning
Once health facilities and participants are selected, field teams must be mobilized to collect biological samples or behavioral data. DHIS2 Tracker helps optimize this logistical hurdle.
Because selected participants are identified in the system through de-identified client codes, the field team can plan their exact workload in advance. In the Nepal drug resistance survey, the central research team shared the randomly selected client codes with the respective health facilities once the sampling was complete. Health workers at the facilities then used their local records to link the codes to actual people living with HIV and arranged appointments for biological data collection.
This coordination meant that field researchers knew exactly how many assessments to expect each day. Field visits became highly organized, saving time and keeping data quality high. Importantly, this approach protects individuals’ privacy. Only authorized personnel at the health facility can link the de-identified study code back to the people living with HIV actual medical record, fully complying with national guidelines.
Note: To protect individual privacy and confidentiality, only the authorized focal person should share de-identified client codes with the field team in line with country-specific guidelines. Because only de-identified codes are shared, a lost code cannot be used to identify an individual. When needed, the authorized focal person can work with the responsible health facility to link the de-identified study code to the actual individual record.
5. Reducing the Data Collection Burden
DHIS2 Tracker already contains much of the baseline information required for public health studies. This drastically reduces the amount of data that field researchers need to collect from scratch.
For cross-sectional studies, socio-demographic details and clinical histories can be extracted directly from the system. In our drug resistance study, individual-level details like age, sex, date of ART initiation, and current treatment regimen were pulled straight from the Tracker. The field team only needed to collect biological samples for viral load and genotype testing. This eliminated the need for lengthy interviews, reducing both fieldwork costs and the potential for recall errors.
Beyond cross-sectional surveys, routine program data is highly valuable for cohort studies. In another national study, we used routine HIV program data for 8,192 individuals to estimate loss to follow-up and identify associated risk factors. Similarly, to monitor the elimination of vertical transmission, the team analyzed records for 322 pregnant women to understand treatment retention and infant outcomes.
For both of these cohort studies, the research team did not need to conduct any new primary data collection. By using de-identified national data already recorded in the system, we completely eliminated the costs associated with hiring, training, and mobilizing field researchers, while still generating robust evidence to strengthen maternal and child health programs including HIV treatment outcomes.
6. Supporting Data Validation and Quality Assurance
Finally, DHIS2 Tracker acts as a powerful tool for data validation during and after fieldwork. Because the authorized focal person can review the electronic tracker records alongside the newly collected field data, it is easy to spot missing values or inconsistencies.
If a discrepancy is found, the authorized focal person can easily follow up with the specific health facility to verify the information against paper-based registers or other local records. This ongoing quality assurance improves data completeness and prevents analyses based on flawed records. This level of verification is often impossible in traditional, standalone research models where dedicated field teams are only hired for a brief data collection period and leave the site shortly after.
7. Conclusion
Published studies from Nepal demonstrate that the DHIS2 Tracker and routine HIV program data can support a wide variety of rigorous research. These range from nationally representative drug resistance surveillance to longitudinal cohort studies on loss to follow-up and the elimination of vertical transmission.
These experiences prove that routine individual-level data is not just for daily program monitoring; it is a rich resource for cost-effective public health research. The DHIS2 Tracker streamlines every phase of the research cycle, including sample size estimation, developing sampling frames, participant selection, fieldwork logistics, and data validation.
The value of this approach extends far beyond HIV programs. Many other national health initiatives, such as tuberculosis and immunization programs, also rely on routine health facility data to monitor progress. Our experience with the HIV program highlights a broad, untapped potential. Using systems like the DHIS2 Tracker offers a practical and highly scalable pathway for generating timely, efficient, and cost-effective evidence across all areas of public health.
References
- World Health Organization. Analysis and Use of Health Facility Data: General Principles. Geneva: World Health Organization; 2018.
- Deuba K, Panta G, Rajbhandari RM, et al. Prevalence of viral load suppression and acquired drug resistance among people living with HIV in Nepal: a nationally representative surveillance study. Journal of Global Antimicrobial Resistance. 2023;35:122-127. doi:10.1016/j.jgar.2023.08.016.
- Shrestha A, Poudel L, Adhikari B, Deuba K et al. Determinants of loss to follow-up among people living with HIV receiving antiretroviral therapy in a universal test and treat setting: A retrospective cohort study in Nepal. Public Health in Practice. 2025;10:100634. doi:10.1016/j.phip.2024.100634.
- Shrestha U, Pandey LR, KC MB, Mirzazadeh A, Deuba K. Progress Toward the Elimination of Vertical HIV Transmission in Nepal: A Retrospective Cohort Study. Journal of Community Health. 2025;50:1105-1114. doi:10.1007/s10900-025-01474-6.