Stakeholder's perception of the health information system performance in Burkina Faso

Introduction the health information system (HIS) in Burkina Faso has improved significantly in recent years. In order to suggest further improvements, we specifically assessed the HIS performance indicators of the epidemic surveillance system from the perspectives of the stakeholders. Methods we conducted a mixed methods study to assess the performance through timeliness and completeness indicators, strengths, and weaknesses of the HIS in Burkina Faso with specific focus on epidemic surveillance in the health districts of Dandé and Tenkodogo for the period of 2016 to 2019. Results fewer than 35% of health districts were able to report at least 90% completeness of community reports since 2017. In 2018, four districts did not exceed 1% completeness of community reports. Some concerns remain related to a need of local support and inter-sectoral collaboration. The technical and organizational factors affect process and performance of the system directly or indirectly through behavioral determinants. Conclusion the ability to measure the performance of all health facilities and to share all community reports online are challenges for the health system in Burkina Faso. New technologies, training-sensitization, and the involvement of actors with influence on social or behavioral change could help to ensure dynamic performance, if perceptions of actors are taken into account.


Introduction
According to the World Health Organization (WHO), a well-functioning Health Information System (HIS) is one that ensures the production, analysis, dissemination and use of reliable and timely health information [1]. It also ensures the availability of information to guide action and assess the functioning of the health system. Indeed, a good HIS encourages the provision of dynamic and efficient health services [2].
Most countries in sub-Saharan Africa have improved their HIS considerably after the Ebola outbreak in West Africa in 2014-2016 [3,4], mainly through an increased focus on digitalization and the use of information and communications technology. However, in many cases, the HIS is designed to meet the information needs of the central level of the health system using an administrative and hierarchical logic [5,6] that may result in missed opportunities to address operational needs at the local level.
In 2010, the national HIS strategic plan of Burkina Faso identified a number of weaknesses, including poor coordination of the entire national HIS, and the inadequacy of the system to motivate the staff in charge of health information [7]. It also identified redundant data collection tools, inadequate data collection at the community level, poor data harvesting as well as low representativeness of available data, as only 40% of the population attended health facilities [7].
Since then, the HIS of Burkina Faso has improved significantly, most notably through the introduction of an electronic health management information system (HMIS) in 2013 [8]. This system includes the collection and integration of routine health data across most health programs based on the District Health Information Software 2 (DHIS2) [8]. In spite of the recent HMIS efforts, the performance of the routine health information system (HIS) remains difficult to monitor given the lack of baseline data [8]. This includes the performance of the RHIS in relation to epidemic surveillance. Moreover, few efforts have been directed at understanding possible enablers and barriers to the routine processes, as perceived by those directly involved in the surveillance of epidemic diseases in Burkina Faso.
The objective of this study was to assess the performance of the RHIS of Burkina Faso with respect to epidemic surveillance at the health district (HD) level and to explore the strengths and weaknesses of the surveillance system as perceived by relevant stakeholders within the HIS.

Methods
Study design: we conducted a convergent mixed methods study in the health districts of Dandé and Tenkodogo comparing the observed performance of the routine HIS with the strengths and weaknesses of the reporting system as perceived by those directly involved. Specifically, we assessed the performance of the HIS in a cross-sectional study evaluating the timeliness and completeness of surveillance reports submitted to and from the district level. At the same time, we used a phenomenological approach to explore the strengths and weaknesses of the HIS, as perceived by selected stakeholders engaged in epidemic surveillance at the health district level.
Study sites: situated in the West region of Burkina Faso, the health district of Dandé is on the border of Mali, whereas Tenkodogo, located in the East region, is close to Togo and Ghana. Given the continuous cross-border movement of humans, animals, and goods, the two districts were considered at increased risk of epidemic disease transmission and thus purposely selected for performance analysis of the HIS with respect to epidemic surveillance. Primary health care centers (PHCs), Soumagou (Tenkodogo) and "Vallée du Kou" (Dandé) were randomly selected among all PHCs in the two Districts.
Data collection: secondary data were retrieved from the central level database at the Health data warehouse (EnDoS/DHIS2) including the District/ Regional reports covering the period of 2016 to 2019 across all 15 HDs, including Dandé in Hauts Bassins (N=8) and Tenkodogo in Centre-East (N=7) regions. Supplementary data were collected from the national HIS data manager and corresponding statistical yearbooks [9][10][11][12]. In addition, we retrieved data from the university hospital of Bobo (CHU) and the regional hospital center (CHR) of Tenkodogo. Data included all weekly official letter telegrams (Télégramme-lettre officiel hebdomadaire (TLOH)) from 2016 to 2019, and all Monthly Activity Reports (MARs) from the health facilities as well as Community-Based Health Workers' (CBHW) reports. We referred to technical surveillance guides for the identification, notification and management of disease cases with epidemic potential [13][14][15]. Information relating to perceived strengths and weaknesses of the HIS was collected through semi-structured interviews of stakeholders associated with PHCs in Soumagou (Tenkodogo) and "Vallée du Kou" (Dandé). Informants included CBHWs as well as health workers involved in epidemic surveillance, such as the head nurse (Infirmier chef de poste (ICP)) at each PHC. Community members attending the PHC as patients were also included as were all members of the district, regional and central management teams.
Data analysis: we measured the performance of the routine HIS according to the timeliness and completeness of reported health data. Timeliness was defined as the percentage of expected reports delivered on time and completeness as the percentage of expected reports delivered during a given period [16]. The accepted national standard for both timeliness and completeness was 100% [16]. We used a document analysis guide to assess the timeliness and completeness of i) the transmission of data from Districts to the central level (TLOHs ), ii) the transmission of community reports from Districts to the central level (EnDoS/DHIS2), and iii) the transmission of MARs from Districts to the central level. We conducted a thematic analysis of all semi-structured interviews using the Performance of the Routine Information System Management (PRISM) framework, proposed by Aqil et al. Specifically, we assessed the strengths and weaknesses of the HIS as perceived by stakeholders according to the i) use of tools (technical factors), ii) coordination and organization of surveillance activities and impact of the workload on other activities (organizational factors), and iii) information and decision-making dynamics (behavioral factors) [17]. The findings from the two study components were compared for corroborations and/or disagreements.
Ethical considerations: informed consent was obtained for all study participants and anonymity was ensured during data management and analyses by assigning codes to each respondent before the transcription of interviews. The study was approved by the Ministry of Health and ethical clearance was provided by the institutional ethics committee of "Institut de la recherche en sciences de la santé" (IRSS), Burkina Faso. The study was registered under the number 32-2019/CEIRES on October 02, 2019.

Results
Descriptive statistics: this study identified a total of 34 PHCs and one medical center (MC) in Dandé, as well as 26 PHCs and one MC in Tenkodogo for which reported health data were retrieved across 15 health districts. We interviewed 60 health professionals and community members (Table 1) distributed across the community (52%), the operational or decentralized district level (31%) as well as the intermediate and central levels (17%). The mean years of professional experience at the facility or in the job were 7.6 years (variations between 0.17 and 25 years) for health workers from PHCs to the central level and 5.5 years for community members (variations between 0.08 and 18 years). Moreover, 18% (11/60) of all interviewees were female. All health professionals and CBHWs declared that they were trained at least once in integrated disease surveillance and response, integrated disease surveillance and response (IDSR), by the Ministry of Health.

Performance of the routine HIS: timeliness and completeness of data
Community reports transmitted from Districts to the central level: data concerning the timeliness of the transfer of CHBW reports from the HDs to the central level were unavailable for the study period of 2016-2019 (Table 2). In terms of completeness of reports, there were no available data for the year 2016 while for the years 2017-2019, completion percentages were recorded without absolute numbers. In 2017, 7 of 15 HDs (47%) recorded 0% completeness of reports; 4 HDs achieved less than 9% completeness, while 4 HDs reported more than 80% completeness ( Table 2). Three of the 4 'high' performance HDs maintained high completeness scores (>99%) in 2018, while 4 of the 7 'low' performance HDs remained in the lowest category (<0.1%). All other HDs recorded some or significant improvements in their completeness scores in 2018, with 1 HD changing from 0% completeness of CHBW reports in 2017 to 100% completeness in 2018. In 2019, all high-performance HDs recorded a noticeable drop, ranging between 45% and 62% completeness, while the remaining HDs recorded some improvement -though the completeness scores remained below 60% (Table 2).  (Table 3). A noticeable lower performance was observed across all HDs in 2019, with both timeliness and completeness falling below 60%.
Monthly activity reports (MARs) from health facilities were transmitted from each health district to the central level: as in the case of community reports, timeliness data for MARs were unavailable across all years. All districts, except one (Do), recorded high completeness scores (>91%) for MARs in 2016 and 2018 (Table 4). In addition to Do, two HDs, recorded relatively low completeness scores (59.3% and 83.7%) in 2017, while all HDs recorded a drop in completeness scores in 2019, with scores ranging from 49% to 97.5%.

Even if tools are available on the ground, some concerns remain related to the case definition, production and transmission of data
Whereas standard case definitions were available at the PHCs, simplified case definitions for certain diseases were sometimes missing at the community level making it difficult for CBHWs to identify and report suspected cases: "It would be a good thing for us if they could find more effective guides where all the definitions of diseases are included. [ confirm or refute after" (D.D, member of a regional management team).
The production and transmission of CBHWs reports and monthly activity reports (MARs) experienced various challenges affecting the observed timeliness and completeness of reports from the HDs to the central level, even within the same region: "When I take our region, the transmission of reports from CBHWs to PHCs is not done in the same way. There are some, which are regular, and some, which are not. Synthesis and transmission are the same things. Data entry on EnDoS is also the same" (C.P. member of a regional management team). One explanation offered for the observed reporting irregularities across HDs included the uneven costs of data transfer: "Some PHCs are not "floated ". Therefore, we use our own credits to make calls. [...] If you take the 43 facilities here if we take out the ones that are "floated", it is no more than 23 or 24, and the rest, if we need them, we must use our own units to make calls. However, there is no unit allocated for that" (T.I, member of regional management team).

Organizational factors: coordination and organization of surveillance activities
Behind the formal coordination of data management, some coping strategies take place: decision-making took place through formal instructions by email, or in a concerted/team approach with the District Management Team (DMT), provincial, regional committee, and regional health directorate (Direction régionale de la santé (DRS)) through social networks, with additional guidance (if needed) coming from the central level.  [...]. Then, the District calls the ICPs and trains them" (B.W. member of a regional management team).
The workload impact other activities: for health workers, the workload associated with weekly TLOH reporting had a negative impact on daily activities such as consultation, care, preventive, and promotional activities, whether at the PHC, Centre medical (CM/CMA) or the Centre hospitalier regional/Centre hospitalier universitaire(CHR/CHU) level. Health workers argued that CBHW reports and MARs were difficult to produce due to their detailed reporting requirements.

Discussion
The performance of health information systems (HIS) occurs within an organizational setting with actors who need motivation, knowledge and skills to perform their tasks and where specialized technical know-how is required for timely analysis and reporting [17]. In countries with limited resources, the HIS often fails to offer guidance for performance improvements at the operational level [2]. Especially as the focus is often on data production rather than on action [18].
Processes are affected by technical, organizational and behavioral factors: our study shows that data collection and the flow of data through the HIS in Burkina Faso, seem to be out of step with the overall objectives of the system. Specifically, we observed low rates of timeliness and completeness for community and monthly activity reports from most of the health districts to the central level, with negative implications on the general performance of the system -not least on fundamental surveillance activities such as case identification and notification of diseases with epidemic potential. The inadequate reporting aligned with several weaknesses of the existing HIS as perceived by those directly involved in surveillance activities [17].
The flow of information from the CISSE to the HIS encountered several difficulties such as lack of or late data entry. It remains uncertain whether the low completion of CBHWs reports (Table 2) reflects the true absence of cases or if the data generated in the communities are irrelevant to surveillance activities. Indeed, the 2020 MEASURE evaluation in the Centre-South region of Burkina Faso reported that only 67 percent of center managers would be competent to use community-level or health facilities-level data [8]. This points to the importance of strengthening data quality and training. In the context of emerging epidemic diseases and given the important role of community actors in early disease detection, alert and response [19], actions and efforts should be discussed directly with the actors, taking the point of departure in the importance of data.
Furthermore, our data suggest that more emphasis should be placed on empowering the operational level which includes communities, PHC and CM/CMA, in line with WHO recommendations, by taking into account citizens' "expectations about health and health care" and ensuring " that their voice and choice decisively influence the way in which health services are designed and operate" [20]. It is however essential that data flow from Districts is timely and complete.
In facilities, failure to use standard procedures during consultations by some health workers combined with a lack of measurement and monitoring of the performance of CHRs or CHUs may explain the current low performance in early detection of epidemics. Improved training and sensitization could alleviate this problem. The great potential of using CBHWs for social mobilization and for the implementation of community intervention could be strengthened. Notably, regular feedback from the central level to the community level must be integrated into the system, to stimulate and encourage the confidence of CHBWs in performing their tasks.
Technical and organizational factors affect HIS process and performance directly or indirectly through behavioral determinants: even if the HIS operates with standardized methods and a set of organizational rules and tools, it is essential that a sufficient number of staff is available to meet the workload. For health workers, HIS reporting represents a significant workload, and other timeconsuming HIS activities include supervision of clinical activities by the ICP and investigation of disease rumors. This is in accordance with the study by Ndongo et al. in Cameroon, which found that health professionals responsible for data collection were discouraged by the large number of forms to be filled in due to the multiplicity of vertical programs [6]. Thus, these types of technical demands may lead to unproductive behavior such as long delays in reporting and low timeliness.
A major challenge in terms of confirming suspicious case lies in providing case definitions for all epidemic diseases in communities. Also, rapid detection and more complete laboratory data and case follow-up may avoid reporting delays, as lack of feedback and acknowledgment is perceived as highly demotivating. Full computerization of the HIS including integration of patient files could facilitate and accelerate the process of monitoring, storage, and extraction of case data. It would significantly reduce the time needed to process data at district, regional and national levels, while at the same time reducing the number of errors including the risk of data being compromised when transferred via informal routes, such as WhatsApp.
With managers at the intermediate and central levels, we observed an interest in improving the quality and integrity of data through ICT. The move toward the electronic system for integrated disease surveillance and response (e-IDSR), despite the skepticism of some HIS managers, seems to be the next step in disease surveillance [21]. However, in the context of limited financial resources and low priorities in the purchase and maintenance of computer equipment by districts, this could be a challenge. In addition, as stated by Aissi et al. [22], electronic management tools, based on technological innovations to improve performance have significant side effects such as stress and feelings of invasion of privacy among health staff [22]. In any case, it would be recommendable to focus HIS strengthening efforts on regular and systematic supervision visits at all levels of the health system, alongside a consensual development of reports and plans for the tracking and implementation of recommendations [8].
Behavioral factors have a direct influence on performance: in 2004 the insufficient use of existing data or information at the central level was criticized by the WHO [18]. More than 15 years later, efforts to decentralize health information and include patients and health workers in decisionmaking continue to be a challenge. Our results indicate that decision-making remains centralized, preventing the participation of peripheral structures with their local realities. In addition, our study shows that the motivation and capacity to search for suspected cases or for investigating rumors of potential outbreaks is low. Unnecessary delays are often created by health workers when transmitting samples to laboratories, with therapeutic and social implications for the patients, and such delays impede the early detection of epidemics.
The health workers' strike in Pouytenga in 2017, which resulted in the removal of 18 ICPs in the HD, as well as the major strike in 2019, had a notable effect on the HIS performance as shown in Table  3 and Table 4 [23,24]. Moreover, the 2019 strike for financial issues blocked the transmission of surveillance data on diseases with epidemic potential, especially during the 3 rd and 4 th quarters [12]. As mentioned by Phalkey et al., financial issues at any level directly affected logistics (vehicles, transport facilities, etc.), equipment availability as well as motivation. The latter in turn affect performance [25].
Limitations: we did not analyze the relevance of the implemented HIS, including its accuracy, the use of information and data, competencies in HIS tasks, and finances due to our methodology. These indicators would have required other more complex research tools and methods for their collection, analysis and assessment of the performance. Also, we did not have the possibility to assess the quality of the HIS reporting. Finally, we did not examine the quality of HIS, as we mainly focused on the stakeholder´s perceptions of how the system worked.

Conclusion
Analysis of the performance of the HIS, based on the Routine Information System Management (PRISM) framework and from the point of view of stakeholders in Burkina Faso, makes it possible to identify and understand several weaknesses. We recommend that the strengthening of the health information management system over the next years be focused on the accuracy, updating, completeness, representativeness, relevance, complementarity of data and their appropriation by stakeholders and potential users. It is important to point out the possibility for the system to measure separate performances of the university and regional hospitals, and to share all CBHWs reports online, while applying the principle of One-health. Developing a culture characterized by regular use of available information for both operational action and strategic management and finally for a good performance is critical. New technologies, which offer opportunities for improvement should be tested in the local context, in order to explore opportunities to respond adequately to clearly defined public health objectives.

What is known about this topic
 HIS is a tool at the service of the health system, capable of improving its performance especially on data production;  Burkina Faso's 2010-2020 national HIS strategic plan identified excessive and redundant data collection tools, weaknesses in data collection at the community level.
What this study adds  The HIS should include the possibility to measure individual performance of university and regional hospitals, and a complete share online of all CBHWs reports;  The study shows that to improve the HIS performance, the capacity of CBHWs for social mobilization and for implementation of community intervention should be strengthened by associating actors in the community or the health system, with higher influence on social change and behaviour.
permission to conduct this study, Kouman Patrice, Tamini Alexis who assisted us in the field and Some Battamacka for his scientific writing advices. This paper and the research behind it would not have been possible without the financial support of "EMERGING EPIDEMICS: Improving preparedness in Burkina Faso" through the Ministry of Foreign Affairs in Denmark (project 17-06-KU).