LAUSR

The global COVID-19 pandemic caused a sudden awakening of the global health community to the crucial importance routine health information systems (RHIS) hold in understanding the effects of the pandemic on health services in lowand middle-income countries (LMICs), offering a tipping point for a significant leap toward stronger systems. RHIS include the recording and reporting of data on service utilization and provision, morbidity and mortality, and health resources [1]. Although there have been major technological improvements in recent years, the systems remain mostly paper-based at primary health care facility level with health workers combining data compilation functions with service provision duties. Data from facilities are transferred to district health offices, where they are then typically digitized, aggregated, and transferred to the regional and national offices. Data elements are multifold, often fragmented across programs, with multiple registers and forms for recording and compilation. The mandatory monthly reporting, the lack of feedback and systematic data use mean that workers are most concerned about work time use for data compilation and transfer and usually miss the forest for the tree. Due to their continuous and real-time nature and their integration to health service provision, data quality issues in terms of data completeness, timeliness, and accuracy are pervasive. Public health professionals and many researchers recognize a critical role for RHIS as the backbone of the health system for evidence-based decision-making [2]. First, it generates near real-time data on service delivery inputs, processes and outputs for annual assessment of the performance of health plans and inform management of services. Its value in tracking service utilization trends in crisis situations was shown in West Africa during the Ebola epidemic and is now even more obvious during the COVID-19 pandemic [3]. Second, RHIS data aim to be exhaustive, covering all facilities and ideally also community service provision data, even though private sector reporting is often a challenge. The analysis can be straightforward and standardized, requiring simple comparison of numbers and proportions without a need to account for sampling errors. Third, based on their longitudinal nature and exhaustiveness, some have argued that RHIS represent a good data source for program evaluation based on quasi-experimental design [4]. Similarly, their longitudinal nature and compilation at district level make them fit for effectiveness evaluation of large-scale health programs [5]. Finally, RHIS data are owned and managed by the government, one essential requirement for sustainability and local decision-making. There is evidence that RHIS have improved in recent years for instance in terms of reporting completeness [6]. There have been substantial improvements in the technology and database organization, in the data content and quality assessment, and in methodologies to analyze these data. A major catalyst has been the transition to the web-based electronic District Health Information System (DHIS-2) by a large number of lowand middle-income countries, mostly in Africa and southern Asia [7– 10]. There have been improvements in approaches for data quality assessments, including adjustments for completeness, consistency assessments, identification of outliers and overall data curation [11–13]. Advances are also being made on the methodological front. Examples are alternative ways of estimating denominators for coverage indicators using service statistics, adjustment