Background
Data analysis and the use of information are essential components of a well-functioning health delivery system for planning and monitoring the progress of disease intervention programmes. In many low- and middle-income countries, the Health Management Information Systems (HMIS) have been established to enhance routine health facility-based data management [
1]. The HMIS is expected to measure the magnitude of disease morbidity and mortality in populations, monitor trends over time, detect and hence facilitate prompt response to any unusual trends. In Tanzania, the HMIS is the primary information system, established during the early 1990s [
2,
3]. It is composed of facility-based health records used for routine health services management, providing indicators for data on morbidity, mortality, health infrastructure, and service coverage. To strengthen the performance of HMIS in Tanzania, the Government adopted the District Health Information System (DHIS2), a web-based software package for collecting, validation, analysis, and presentation of aggregate statistical data tailored to integrated health information management activities. The adoption of DHIS2 aimed to facilitate data access and stimulate usage.
The effectiveness of a health information system depends on collecting, analysing, interpreting, and utilising the information correctly at all levels [
4‐
6]. However, most often, incapacitated systems compromise optimal functioning and system performance.
Generally, in many Sub-Saharan African countries, data utilisation at all healthcare systems is low [
4,
7‐
11]. Available literature suggests that, despite some notable successes, the impact of HMIS on the decision-making process within Africa health systems remains limited [
4,
11,
12]. Several barriers have been reported to prevent the HMIS from achieving full potential in Africa [
13,
14]. Institutional, technological, individual, and logistical capacities are perceived factors that either enable or impede the successful implementation and use of data generated from the HMIS [
15]. It has been reported that though the HMIS offers opportunities to inform health decision-making at all levels of the health systems, its usefulness is realised only when it allows for the transformation of generated data into meaningful information and knowledge for action [
7].
Routine health information is essential for both operational and strategic decision-making at all levels of the health system. Investment in health is dependent on efficient and reliable HMIS. With increased investment in disease control programmes in Tanzania, there is a critical need for a sound health information system to support decision-making at all levels and vice versa. It is envisaged that improved HMIS would enhance evidence-based decision and policy-making, leading to improved accountability and effectiveness at all health system levels. In the efforts to promote information utilisation and evidence-based decision-making, the Global Summit on Measurement and Accountability for Health has called for action for all countries to have health information flows involving utilising data locally to improve services effectiveness of disease programmes [
16]
. The objective of this study was to determine the utilisation of HMIS data and factors influencing the performance of the system at the district and primary health care facility levels in Tanzania.
Discussion
Generally, the utilisation of health data collected from health facilities in Tanzania is relatively poor. About two-thirds of the facilities reported using the HMIS data they collect, mainly comparing performance between service coverage, determining disease trends over time, and community health education and promotion. However, at the district level, the primary use of HMIS was for preparing reports and annual planning. This use suggests the underutilisation of all important data collected from health facilities. Several studies have identified weaknesses in data use and response in Tanzania while recognising that these are critical components for sound public health decision-making [
4,
12,
17‐
20]. A similar low utilisation of routine facility data has been reported from elsewhere in Sub-Saharan Africa. For example, in a study in Ethiopia, training, data analysis skills, supervision, regular feedback, and favourable attitude were factors related to routine health information system utilisation [
21]. The findings of this study and others elsewhere indicate that data is collected for reporting purposes, and there is minimal utilisation of the information to inform decisions. Usually, to most health workers in low-income countries, HMIS is associated with the filling of registers, compiling, and submitting reports to the next level without its utilisation [
22,
23].
Studies elsewhere have reported that data generated by health facilities are most often not sufficiently utilised to improve health care [
24‐
26]. Generally, in Sub-Saharan Africa, health information utilisation at primary health care and the district is poor [
22]. In Ethiopia, the level of HMIS data utilisation for different decision-making purposes has been reported at 57.9–62.7% [
27‐
29]. In a recent study in Zanzibar, it was reported that only 42% of the healthcare workers used HMIS data for monitoring and evaluation, 35% for planning, 23% for supply and drugs management, 18% for budgeting, and 10% for disease outbreak preparedness [
30]. Factors associated with good utilisation of HMIS data have been described to include staff motivation, training, supportive supervision, a good perceived culture of health information, competence, and decisions based on superior directives [
28,
29].
Several factors, including a limited number of staff and skills, low motivation, inadequate resources, lack of training and refresher courses, combined with lack of incentives and tools, have been pointed out to be responsible for the underperformance of HMIS in low- and middle-income countries [
13,
31‐
34]. This study observed inadequate human resource for data management in all health facilities. In a study in Ethiopia, the availability of a standard set of indicators, skilled human resources, well-designed reporting formats, and staff trained to fill formats increased the likelihood of achieving data quality [
35]. Organisational factors such as the culture of using information, resource availability, planning, governance, training, supportive supervision, and availability of finances have been reported to influence HMIS performance [
36]. In Tanzania, Simba and Mwangu [
19] found that trained staff on HMIS and seeking or provision of frequent feedback were significantly associated with performance.
In a recent study in Kenya, there was a significant positive relationship between the availability of adequate staffing for HMIS tasks, training of staffs, supervision of HMIS activities, availability of plans for HMIS, promotion of a culture of information, staff motivation, and the performance of routine health information system [
37]. In our current study, less than half of the health facilities reported that the district team members visited their facility for supervision. However, when made supervision visits, the team did not have a supervision checklist. Already health workers’ data analysis skills, feedback, and regular supervision have been described to affect the successful implementation and use of routine health information systems [
21,
35,
38,
39]. A study in Uganda reported that self-efficacy and the presence of RHIS staff directly influence the use of HMIS data [
40].
Moreover, it has been reported that adequate supportive supervision and health facility performance review to be significantly associated with good performance in HMIS [
41]. Lack of standard operating procedures (SOPs) was reported as a barrier to health workers’ performance in data management. SOPs for data management at the health facility level have been shown to help improve HMIS data quality in Rwanda [
42].
In this study, the majority of the HMIS focal persons were non-health information technicians. Lack of staff with core competence in data management and analysis is one of the core weaknesses identified to affect data performance in several studies [
6,
20,
32]. Similar findings have been reported in other studies elsewhere [
43]. It is high time for the Government of Tanzania to establish an appropriate carder for data analysis and define clear responsibilities at all levels. Thus, there is a need for a proper team of skilled and competent people at all levels. For the HMIS to work effectively and efficiently, there must be consistency and integrity between the human, supplies, and process aspects.
Despite the significant relevance of the findings obtained from this study, its design attempted to control various biases by ensuring representation of districts and facilities and improving the validity of findings by intertwined interviews with actual observations. It faces several limitations. First, its cross-sectional nature limits the generalisation of results into other time points. Some of the responses were time-bound, resulting in different findings if a similar study was done at different time points. Second, on the same note, the findings may only be relevant for Tanzania and countries with very similar functionality to the HMIS.
Similarly, some of the responses are based on the interviewees’ experiences. Therefore, those found at the time of this assessment contributed significantly to the nature of the results. Finally, in the context that health workers are not randomly placed, variations might be expected with a different set of respondents.
Acknowledgements
We are grateful to all the management of the districts and health facilities who participated in the study. Special thanks to the District Health Management Information Systems (HMIS) Focal persons, District Reproductive, Child and Neonatal Health Coordinators, District AIDS Control Coordinators and the entire Council Health Management Teams of the study districts for their support throughout the data collection exercise. We are grateful to our research assistants: John Ng’imba, Joyce Kaswamila, Simon Alfred, Jesca Kivinge, Leilath Mtui, Jesca Massawe, Gilbert Mwageni, Glory Lema, Nicholas Lubange, Estaban Mremi, Osyth Sylvester, Neema Lauwo, and Isolide Massawe, for their enthusiasm and dedication. Finally, the authors thanked Prof. Gasto Frumence for his critical review of the earlier manuscript version.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.