HMIS
INTRODUCTION
A general guide for interpreting data from the Health Management Information System (HMIS)
The Ministry uses a comprehensive and integrated Health Management Information System (HMIS) to collect and report on routine health services and disease data, in facilities and in communities. Data is recorded in specially designed registers as health workers are providing services. At the end of each month, data from the registers are compiled, aggregated, and reported on a monthly basis using both program-specific reports (e.g. Maternity, ANC, etc.) And composite reports (HMIS 15 for health centers and hospitals.
Challenges with using HMIS-based indicators to estimate population prevalence or incidence
All HMIS-based indicators depend on the quality and completeness of reporting. Using HMIS-based indicators to measure prevalence and/or incidence in the population will likely lead to underestimation, limited by data capture rates, reporting rates, healthcare-seeking behaviors, and healthcare access.
Population-based estimates for HMIS-based indicators
Many of the HMIS-based indicators currently rely upon population estimates for denominators. The accuracy of these indicators depends on the accuracy of the population estimates. These estimates are most likely to be accurate soon after a census but decrease in accuracy over time. They are also less accurate for small geographic areas. Inaccuracies in estimating the population can lead to over or underestimates. For example, coverage rates of over 100% are possible if estimates of the target population are too low. These errors should be explored and corrected when possible.
Impact of under-reporting from both private and public health facilities
While private health facilities are supposed to report into the HMIS system, the degree to which this happens is inconsistent; the same is sometimes true for public facilities. Central hospital reporting, through HMIS 17, is also under development. When an HMIS-based indicator aims to assess disease occurrence in the general population (e.g. malaria incidence) or coverage of service in the general population (e.g. immunization), under-reporting from facilities will likely lead to lower estimates. The denominator will be based on population projections for the entire population, but the numerator will only include what is captured in HMIS reports. Reporting rates give an indication of the degree of under-estimating. For example, if the indicator looks at the quality of care among those who attend facilities (e.g. IPTp >3 times during ANC), the indicator will be representative only of those facilities reporting and not necessarily all women who have had an ANC visit. Similarly, if road traffic deaths are presented per 100,000 in the population, but reporting rates are low, then the indicator likely represents a proportionately low estimate. As reporting from both private and public facilities improves, this will no longer be a limitation.
District Health Information Software (DHIS2)
Ministry of Health uses DHIS as an open-source software platform for reporting, analysis, and dissemination of data for all health programs, developed by the Health Information Systems Programme (HISP). The core development activities of the DHIS 2 platform (see note on releases and versions further down) are coordinated by the Department of Informatics at the University of Oslo and supported by NORAD, PEPFAR, The Global Fund to Fight AIDS, Tuberculosis and Malaria, UNICEF and the University of Oslo.
In summary, several biases may lead to underestimates, overestimates, or may have little effect. Also, several factors may influence estimates simultaneously, with sometimes differing effects. These potential biases, and others, should be taken into consideration when interpreting each indicator for which they apply.
