The Use of Data for Planning and Services Improvement in Tanzanian Primary Healthcare Facilities: Experience from Star Rating Assessment ()
1. Introduction
Quality data availability is essential for making decisions in health systems in a way that can help to address public health issues, as well as, in planning and delivery of health services through World Health Organization (WHO) health systems building blocks—service delivery, health workforce, access to essential medicines and technologies, financing, governance and leadership, and information systems ( AbouZahr & Boerma, 2005 ; Nutley & Li, 2018 ). Data use has been defined to include the following issues “data sharing, visualization, dissemination, and review” ( Nutley & Li, 2018 ); and measures of data used to include the following dimensions: “transparency, timeliness, visibility, accessibility, dissemination of information, calculation of key indicators, preparation of information products (reports), and presentation of the achievement of targets” ( Abajebel et al., 2011 ; Mwencha et al., 2017 ; Nutley & Li, 2018 ). Having a culture of data use at all levels of health services delivery is an essential element for health system strengthening ( Nutley & Reynolds, 2013 ). The recent report of The Lancet Global Health Commission on high-quality health systems in the sustainable development goals era has noted that “health systems need to develop the capacity to measure and use data to learn” ( Kruk et al., 2018 ).
However, data use in the health sector in Low- and Middle-Income Countries (LMICs) is sub-optimal. Issues affecting data use have been grouped into three areas—organizational, behavioural, and technical issues ( Hoxha et al., 2022 ; MEASURE Evaluation, 2018 ). In a study that documented a seven-year experience of interventions in Mozambique, Rwanda, and Zambia, it was recommended that to improve data use for quality improvement; it is important to conduct data quality assessments and to improve the skills of service providers in data analysis and visualization ( Wagenaar et al., 2017 ). Also, the implementation of a combination of interventions that target behavioural and technical factors has been shown to improve data quality and use ( Lemma et al., 2020 ). Involving local stakeholders in designing data collection, has been reported to be successful as measured by the completion and accuracy of data collection tools ( Mayston et al., 2020 ).
In a systematic review that looked at challenges in the use of routine health information data in LMICs, Hoxha and colleagues identified organizational or environmental challenges as the most commonly reported barriers to data use ( Hoxha et al., 2022 ). Some interventions have been shown to have the potential for helping to establish a culture of data use including decision-makers/supervisors being examples to others (role models) through advocating for data use at the council level; using incentives such as the implementation of performance-based financing; and use of accountability system through the use of open performance review and appraisal system ( Lippeveld, 2017 ). In Uganda, challenges found in health centres in implementing data quality assurance of routine information systems included: “laborious and tedious manual system; the difficulty to archive and retrieve records; insufficient health management information system forms; and difficulty in delivering hard copies of reports to relevant stakeholders” ( Kagoya & Kibuule, 2018 ).
In 2015, Tanzania started to implement the Star Rating Assessment (SRA) process for Primary Health Care (PHC) facilities to improve the quality of services provided ( Yahya & Mohamed, 2018 ). The assessment used a set of tools separate for dispensaries, health centres, and level one hospitals (hospitals at the council), with a total of 12 services areas in which service area three is on “use of facility data for planning and services improvement” ( Yahya & Mohamed, 2018 ; Kinyenje et al., 2020 ). This study aimed to assess the extent of data use in PHC facilities in terms of: functionality of the Health Management Information System (HMIS); information use and dissemination; and medical records status (recording and retrieval, as well as the confidentiality of patient’s records).
This paper adds novelty to the body of knowledge on the performance of PHC system in Tanzania by providing a countrywide baseline picture of data use in PHC facilities. The vision for PHC in the 21st century, emphasized on the need for a research-oriented PHC that can help to inform policies, strategies, and operational plans ( WHO & UNICEF, 2018 ). The operational framework for PHC further emphasized on the need for health facilities to use data for improving performance and also improving data quality ( WHO & UNICEF, 2020 ); and the measurement framework and indicators have provided a link between measurement and improvement in PHC performance ( WHO & UNICEF, 2022 ). Therefore, the results of this paper will help to inform the national and subnational level stakeholders on the situation of data use in PHC facilities, so that they can design strategies to improve data use.
2. Methods
2.1. Conceptual Framework
The design of the SRA Tool (SRT) in a section related to the use of facility data for planning and services improvement (Area 3) adapted the “Performance of Routine Information System Management (PRISM)” framework developed by Aqil and colleagues ( Aqil et al., 2009 ). The elements assessed matched with the components of the PRISM framework are shown in Table 1. The outcome measure which is improvement in data use for planning and services improvement is related to the outcome component of the PRISM framework which is “improved Routine Health Information System (RHIS)”. Therefore, based on the core components of the PRISM framework (technical, organizational and behavioural factors) matched with the assessed components of the tool (Table 1), we conceptualized that improved data use for planning and improvement of services in PHC facilities is a function of the functionality of HMIS, information use and dissemination, and medical records in a particular facility, and demographics of health facilities. The extent of the components of the PRISM framework covered in the tool has the potential to provide comprehensive information that will inform the improvement of data used for planning and improvement of service delivery in PHC facilities in Tanzania ( Nutley & Reynolds, 2013 ).
2.2. Study Objectives
The objectives of our study were to determine the proportion of PHC facilities: that use data for planning and services improvement; that use HMIS tools correctly; that disseminate information; with proper medical records; and determine predictors associated with data use in PHC facilities.
2.3. Study Design
This was a retrospective cross-sectional study that used secondary data related to the use of facility data for planning and services improvement that are found in a national data warehouse known as District Health Information System 2 (DHIS2). This data was collected during the Star Rating assessment that was conducted in between July 2017 and December 2018.
![]()
Table 1. Data use areas assessed matched with the PRISM framework.
2.4. Target Population
All operating PHC Facilities in Tanzania (Dispensaries, Health Centers and Hospitals at Council level) regardless of their ownership, i.e. public (Local Government Authority (LGA), Military, Police, Prisons, Parastatals and other Ministries, Departments and Agencies (MDAs)) or private (Faith-Based Organization (FBO), Non-Governmental Organization (NGO), Private for Profit).
2.5. Study Population
A nation-wide assessment was conducted to all 7289 PHC facilities in 2017/18. This study targets all of these facilities which constitute the majority of healthcare facilities (more than 97%) in Tanzania.
2.6. Inclusion Criteria
All facilities whose performance and characteristics were able to be identified from the DHIS2 after data cleaning were included in this study.
2.7. Exclusion Criteria
Facilities with incomplete data related to relevant questions under study were excluded during the analysis.
2.8. Data Sources and Collection
2.8.1. Data Collection
This section explains how data that make up the SRA dataset in the DHIS2 were collected. Data collection from each facility was done by at least four trained healthcare personnel representing all four tiers of Tanzanian healthcare administrative levels, i.e. national, regional, council and facility-level to ensure transparency and fairness. Before enrollment for data collections, the team received a five-days training mainly focusing in the use of the assessment tool, data cleaning, compilation and reporting. Data collectors captured the score per question at the facility level using the SRT and subsequently entered the values in Microsoft Excel Sheets to produce the score.
The data collection tool tracks three components (variables for this study) that measure the use of Facility’s data for planning (additional file Appendix 1). The components, i.e. functionality of the HMIS, information use and dissemination and medical records are further divided into indicators, questions, assessment criteria and pre written responses. For each question, one response was chosen among the following possible scoring options; “Yes” (score = 1), Partial (score = 0.5) or “No” (score = 0).
Data from the sheets were cleaned and compiled at Council level immediately to ensure correctness, completeness, accuracy and timeliness of the data. The SRA at national level was coordinated by Health Services Inspectorate and Quality Assurance Section of the then Health Quality Assurance Division, which is currently Health Quality Assurance Unit (HQAU) of the Ministry of Health ( President’s Office—Public Service Management and Good Governance, 2018 ).
2.8.2. Data Sources
The targeted data were extracted from the DHIS2 database in form of Microsoft Excel Sheets. The sheets were checked for data quality and cleaned before use.
2.9. Study Variables and Measurements
The main dependent variable of interest for this study was “use of facility data for planning and services improvement” that was used to determine its association with facility-related independent variables, i.e. facility ownership status (public or private owned), location of the facility (rural or urban-based), gender of the facility in-charge (male or female) and health facility service level (dispensary, health centre or hospital level one).
The facility was considered using data for planning and services improvement if scored at least 80% of the assessment criteria. This cut-off point is provided in the National Guidelines for Recognition of Implementation Status of Quality Improvement Initiatives in Health Facilities ( Ministry of Health and Social Welfare, 2014 ). Therefore, since there were 12 questions weighing a total of 12 points; a facility needed 9.6 points from 12 available points to meet the cut point and be qualified as using data for planning and services improvement.
Other dependent variables were the three components that contribute to data use that are stated in Table 1, i.e. functionality of HMIS, data dissemination and medical records. Likewise, a facility needed 4.6 points out of 6 points to have a functional HMIS. Also, since both variables: data dissemination and medical records were made of three questions each; a facility needed 2.4 points to qualify in each of the category.
2.10. Data Management and Analysis
The facilities were characterized based on ownership status, location, gender of the facility in-charge and health service level. Data cleaning and analysis was done using STATA 15. The data was cleaned and checked for completeness and outliers before analysis. We determined an association between the binary variable and independent variables to determine the predictors of data use for planning and services improvement at the PHC facilities. Multivariate logistic regression analysis was conducted to measure the association that was presented in odds ratio with a 95% confidence interval. A p-value of <0.05 was considered as statistically significant.
3. Results
A total of 6663 PHC facilities; which is equivalent to 91.4% of all 7289 PHC facilities that were included in the SRA assessment of 2017/18 ( Yahya & Mohamed, 2018 ; Kinyenje et al., 2020 ) met inclusion criteria and were included in the final analysis under this study. A separate analysis found no significant difference between facilities that met inclusion criteria and those that did not (Table 2).
![]()
Table 2. A separate analysis output for 626 (8.6%) PHC facilities that were excluded from final analysis.
*p-value was calculated from paired t-test.
Among the facilities that met inclusion criteria; 1198 (18.0%) had used data for planning and services improvement. These are the facilities that gained at least 80% in all assessment questions. Regarding the three components of data use for planning and services improvement; 3792 (56.9%) of the facilities had functional HMIS; 1752 (26.3%) had disseminated data; and 631 (9.5%) had proper medical records.
As presented in Table 3, all facility-related variables that were included in our study were significantly associated to “data use for planning and services improvement at PHC facilities” during both bivariate and multivariate analysis except the location of the facility. The facilities whose in-charges were females were likely to use data compared to those whose in-charges were males (AOR 1.19; 95% CI: 1.04 - 1.36). Data use was likely to increase as the level of facility increases-with significant association at health centre level but not at hospital level. Public-owned PHC facilities had a 23% increased likelihood of using data for planning and services improvement compared to private-owned facilities (AOR 1.23; 95% CI: 1.02 - 1.47).
A deeper multivariate analysis focusing at association of the healthcare facility variables and the three components that form outcome variable “data use for planning and services improvement” was done and presented in Table 4.
The significance of the association for ownership variable remained very strong for data dissemination and functional HMIS but very weak for medical records. The variables “facility type” and “gender of the facility in-charge” did not change significance of the associations except for component “medical records”. The location of the facility remained not significantly associated to the three components of data use.
![]()
Table 3. Predictors associated with data use in primary healthcare facilities.
p-values are calculated using chi-square test. *Predictors whose association were found significant in the final logistic regression model. COR = Crude/unadjusted Odds Ratio, AOR = Adjusted Odds Ratio.
![]()
![]()
Table 4. A deeper analysis showing an association of healthcare facility’s characteristics with individual components that contributes to data use.
p-values are calculated using chi-square test. *Variable whose significance of association with data uses (the main outcome of the study) would differ based on the components. AOR = Adjusted Odds Ratio
4. Discussion
The findings have shown health facility data were used for planning and services improvement in only one-fifth of the PHC facilities in Tanzania. Among three components that determine the use of data in the PHC facilities; most of facilities failed in having proper medical records whereby only one-tenth of the facilities had capability to record, retrieve and ensure patient records are handled in a way that assures confidentiality. More than half of the facilities had HMIS system functional and therefore this was relatively the most performed component among the components that determine data use in the health facilities. The national target is at least 80% and therefore effort is still needed to improve the performance. The facility was considered having HMIS functional if at least one staff is trained in HMIS and HMIS tools are used and filled correctly. Moreover, about a quarter of the facilities had analysed data and displayed it for public view.
A study in five countries (Cambodia, Ghana, Mozambique, Nigeria and Tanzania) has documented the burden of recording and reporting of health data in PHC facilities. The findings show that the functionality of HMIS and medical records in PHC facilities in Tanzania may be affected by the burden of recording and reporting due to the volume of registers in which for completing monthly reporting forms it takes up to 65-hours, with 29 reporting forms ( Siyam et al., 2021 ). Besides the above challenges, data usage in Tanzania has been further emphasized in publicly-owned PHC facilities through Direct Health Facility Financing (DHFF) initiatives that were targeting public facilities since 2017 ( Kapologwe et al., 2019 ). Among others, DHFF requires public facilities to collect quality data that will be used for planning interventions targeting specific local needs. This may be the reason why public facilities have performed better in this study compared to private ones.
In a study which was conducted in 12 health facilities in three Tanzanian regions (Arusha, Lindi, and Geita) to assess the capacity of health workers to analyze and use data for family planning services, they found that: health workers have inadequate skills for data analysis and computer use; the facilities have a weak culture of data analysis and use; lower-level health facilities, lacked internet access, hence affecting their access to DHIS2; and lack of data ownership among health workers thinking in which they believed that data generated at health facilities belong to the Council Health Management Team (CHMT) and not the facilities; and lack of training on collecting, analyzing, presenting, and using data ( Anasel et al., 2019 ). Also, the findings from 115 healthcare facilities from 11 Councils in Tanzania found that poor data use was common, due to inadequate data management skills, lack of training, inadequate supportive supervision visits and feedback from CHMT ( Mboera et al., 2021 ). Training and education have been positively linked to data use (especially electronic data) ( Nwankwo & Sambo, 2018 ; Asemahagn, 2017 ); nevertheless, organizational culture and technical and managerial support are very key to sustainability of the processes ( Pesec et al, 2021 ).
A study in Nigeria found that training of health workers in PHC facilities improved their data management skills including filling registers and forms, data analysis and use ( Nwankwo & Sambo, 2018 ). Findings from a study in Ethiopia have shown slightly more than half of 250 staff who participated in the study reported to use facility data routinely to develop plan; while identified factors that hindered data use included residence, data management knowledge, work load, computer skill, computer access, supportive supervision, HMIS training, and availability of HMIS guideline and formats ( Asemahagn, 2017 ). Also, research evidence in Tanzania shows that the level of education is positively associated with the use of HMIS especially if the systems are electronic ( Sukums et al., 2014 ) and therefore, we presume the use of data in high-level facilities was due to the presence of relatively more educated staff in these facilities. In Costa Rica, a strong culture of valuing data as a tool to drive improvements coupled with technical and managerial support has contributed to the use of data for improving care to patients and improving population health ( Pesec et al, 2021 ). Rendell and colleagues have identified several factors that can influence data use in PHC services which they organized them into three groups: governance (leadership, participatory monitoring, regular review of data); production of information (presentation of findings, data quality, qualitative data); and health information system resources (electronic health management information systems, organizational structure, training) ( Rendell et al., 2020 ).
The ultimate goal of SRA processes is to improve performance management of the PHC facilities that includes management of health information systems and use of data ( Yahya & Mohamed, 2018 ). Therefore, the government of Tanzania must strongly uphold and implement the digital health strategy namely “The National Digital Health Strategy 2019-2024”—which is also in line with both the Tanzania Development Vision 2025 and the Health Sector Strategic Plan 2021-2026 ( Ministry of Health, Community Development, Gender, Elderly and Children, 2019, 2021 ). The strategy aims at improving the management of health information systems in the PHC facilities particularly through the use of sustainable and interoperable electronic systems. Improving the use of data is essential in strengthening PHC in a way that will ensure its orientation towards person centred care in PHC ( Alpert et al., 2020 ). Therefore, it is recommended for Tanzania to make use of the data use partnership project to strengthen culture of data use in PHC facilities by focusing on having better data (i.e. strengthening quality of data production) and implementing interventions that will help to nudge health workers towards a behaviour of regular data use ( Arenth et al., 2017 ). By strengthening the use of data in PHC facilities, it will be an important step in building a learning health system in Tanzania ( English et al., 2016 ; Sarakbi et al., 2021 ). Also, in-charges of PHC facilities can make use of facility management meetings and quality improvement teams meetings to build capacity of staff on data use through joint data analysis and sharing to inform decision making and quality improvement interventions ( African Health Initiative Partnership Collaborative for Data Use for Decision Making, 2022 ).
The findings of this paper may have two limitations as follows: first, the SRA data used are five years old now (collected in the fiscal year 2017/2018), and some corrective actions may have taken place in this period to improve the situation of data use. However, the results help to provide a baseline of countrywide situation with regard to data use in PHC facilities; and that future analyses which will use more current data will use our findings as a baseline to monitor on improvement trends. Second, the paper analyzed the predictors of data use based on quantitative data on facility type, ownership, location, and gender of facility in-charge but did not explore some qualitative information on what support facility in-charges and other health workers need in order to improve on data use. Further research is needed to explore (both qualitatively and quantitatively) what factors are influencing data use in PHC facilities in Tanzania. This will help to increase availability of evidence on this area given the findings by Rendell and colleagues—who found that “scant evidence exists regarding factors that influence the use of health service delivery indicators to improve delivery of primary health care services in low- and middle-income countries” ( Rendell et al., 2020 ).
5. Conclusion
The use of facility data for planning and services improvement in PHC facilities is low in Tanzania, and as a result, the quality of care could be hampered. Many efforts need to be targeted at privately-owned and low-level PHC facilities (i.e. dispensaries). The facilities performed more or less equally despite locality and therefore more research is needed to explore other factors that could influence the use of data in PHC facilities in Tanzania. The very low level of PHC facilities that have proper medical records require special attention in order to ensure that there is proper recording, retrieval, and privacy of patients’ records.
Acknowledgements
The authors are passing their sincere gratitude to the Ministry of Health, especially, the Health Quality Assurance Unit for permitting us to use the SRA data.
Apart from government institutions, the authors extend appreciation to development partners such as World Bank, The United States—Centres for Disease Control and Prevention (CDC), Danish International Development Agency (DANIDA), and The World Health Organization who together made SRA possible. Others were the communities of the facilities visited, PharmAccess International, Association of Private Health Facilities in Tanzania (APHTA), Christian Social Services Commission (CSSC), and Development Partners in Health-Group (DPG-H).