This study aimed to examine the organization and functioning of the HIS in Burkina Faso in terms of expected and actual activities by comparing official documents to primary information obtained from key informants and direct observations. Based on the WHO HMN frameworks, we focused specifically on inputs, processes and outputs.
Limitations of the study
The study has some limitations, including potential bias due to the purposive selection of the two Districts and participants, as well as the risk of not being representative of health districts in Burkina Faso. Moreover, private health facilities, religious-owned health services and actors, and animal health services were not included in this study. However the number of such services are limited and they are mainly concentrated in urban areas. They could have provided additional information and given a more complete view of epidemic surveillance as part of the HIS. Nevertheless, the results obtained offer inputs for reflection on improvements strategies in Burkina Faso and similar settings.
Main gaps between the planned and existing system
Input category
Regarding the activities relating to community-level surveillance in Burkina Faso, the official forms for reporting epidemic diseases and events in the community are in French, which poses a challenge for some CBHWs in terms of reading and completion. Our study also revealed, that some forms contain certain biomedical terms that do not align with local symptoms and disease perceptions. This may result in CBHWs’ failure to detect and report disease cases as required. In their evaluation of the surveillance systems in Burkina Faso, Geers and colleagues (2018) reported that some CBHWs suggested the use of community definitions of meningitis and measles for better comprehension [24]. Moreover, in community settings, the registration of deaths by the civil registry office in municipal councils is often insufficient and verbal autopsies on causes of death are not carried out systematically [25]. This may partly be explained by the lack of collaboration between the Ministry of health and the Ministry in charge of civil registration. That may also question the reliability of the detection process and hamper the functioning of early warning systems to contain epidemic diseases.
We noted that most of the CBHWs are primarily trained to provide services to the community for reproductive health, malaria, vaccines and nutrition. Furthermore, the community-based surveillance of epidemic diseases is not well developed in terms of training on case identification and reporting [24]. With the appearance of new epidemic diseases such as dengue and COVID-19, a stronger focus on emerging diseases and CBHWs participation is needed, including improved data sharing, and feedback between the different HIS levels and the community.
At community level, it may be an advantage for example to involve traditional healers and school teachers in a direct collaboration with the CBHWs for epidemiological surveillance. As a matter of fact, the limited engagement of traditional healers and school teachers was documented in the stakeholder analysis of Dandé’s 2018 and 2019 Health District action plan [26, 27]. In a similar setting in Niger, Ndiaye and colleagues (2000) reported that due to cultural beliefs, patients would consult traditional healers before going to health centers, which causes delays in early detection, reporting and investigation of disease cases [28]. According to the World Bank (2012), “Information on the types and volumes of services offered by private facilities is almost non-existent in Burkina Faso. As a result, the role played by the private health care sector is difficult to assess, and the quality of services is difficult to appreciate as well” [29]. As some people prefer to attend private facilities for various reasons, such as faster treatment, better comfort and the availability of specific services, the insufficient involvement of private actors in the surveillance of epidemics may constitute a source of vulnerability for the existing health systems in terms of early detection and notification of epidemics.
The clustering of the actors involved in disease surveillance, as noted in this study, is common in African health systems [30]. Information and communication technologies (ICTs) could improve health system management including the relations between actors [31, 32]. This solution may already be in a process toward implementation with the new WHO e-IDSR strategy [33]. In some countries, an SMS-based mobile phone network has been set up to enable CBHWs to exchange short text messages [34]. Using mobile applications could allow CBHWs to feel part of the system and change their expectations and roles in positive ways and increase their personal commitment [35].
The Ministry of health and the health officials in Dandé and Tenkodogo should strive to improve CBHWs’ understanding of forms and produce simplified case definitions (with less technical terms) for all diseases with an epidemic potential. Therefore, information needs to be formulated according to CBHWS’ level of instruction, and that must be done under the supervision of a district, regional, or central staff [35]. As noted by the USAID, interconnecting forms and electronic platforms of various programs could allow for real-time notification, collection, analysis and use of data on diseases or events for effective public health interventions. Indeed, by training the CISSE and other data managers, such as laboratory technicians, to enter data correctly across existing programs, could improve data quality, as observed in this study. In Tanzania Nsaghurwe and colleagues (2021), showed that it is possible to integrate and share digital data between levels and programs of the health system even when people use different digital tools. Indeed, data entry errors like, for example, a discharge prior to admission dates, were resolved through interoperability filters of the health information exchange system, which spotted such errors and requested corrective actions from the point of data entry [36].
Ideally, information from the community and PHC levels are used for planning and managing of the surveillance system as well as for advocacy and policy development. According to official documents, the head nurse should be able to analyse, interpret and use health information data for local decision-making [16]. In practice, the nurse, and even the DMT, solely refer to directives from the central level in Ouagadougou. They do not have decision-making autonomy regarding the management of problems encountered in their area of responsibility, despite the fact that they carry out annual analyses in which their problems are identified. As reported by Odhiambo-Otieno in Kenya (2005), such centralized decision-making may often overrule or ignore local expectations [37].
Certain actors, particularly ICPs, report to several other actors, such as CBHWs, DMOs and the CISSE. This increases the risk of errors, widens the discrepancy between the TLOH data and the IDSR database, causes duplications during the reporting and delays in the flow of information. Ouedraogo and colleagues (2018) highlight the tedious task of entering and reporting health data, with the risks of duplicate entry in different databases [38]. A number of other factors may explain inadequate reporting, including insufficient quality control by supervisors due to lack of procedures or directives, inexperience of new staff with forms and guidelines, and individual or social factors – such as labour disputes between health workers and the government. In 2019, health workers went on strike for a continuous period of eight months, thus causing a noticeable drop in the year’s statistical reports [13, 39].
Because of the different skill levels between some community “rapporteurs” and the ICP, the interface of their collaboration gets complicated because of the other stakeholders’ inability to understand the forms provided by the ICP or the latter’s inability to explain the forms in an easy language. Schweyer and Cabe (2005) addressed this phenomenon by; for them, “procedures or tools are not those that structure the networks, but a more egalitarian approach between health professionals” [40, 41].
Official surveillance documents mention that the head nurse or any other designated person must file the CBHW monthly report. However, because of PHC understaffing, a single nurse may be linked to CBHWs from several villages and therefore be responsible for multiple reports, which causes a work overload and the risk of reporting erroneous data. A CBHW report is 38-page long, so makes it a daunting task to complete and verify. Moreover, the head nurse does not always receive the offered DMT support for CBHW supervision and investigation of rumours, including deaths in the community.
Also, late payment of financial incentives may also influence CHBWs’ motivation to carry out in-depth investigations of rumours. According to several studies, lack of in-depth investigations may lead to infodemics, which refers to false or misleading information that may in turn cause widespread public reluctance to adopt the required infection control measures promoted by health authorities – thus delaying essential interventions [42,43,44,45,46].
Several guidelines exist for the completion of standardized forms for surveillance. However, they do not specify the means by which the reported data should be verified. This creates a risk of entering erroneously aggregated data that is inconsistent with the local reality. In turn this may lead to poor resource estimation, poor planning and performance assessment and, eventually, poor decision-making at central level. Innovating through the use of online forms or electronic applications could in this case, make it possible to better supervise the completion of forms at community level. That would promote timely detection of erroneous or incomplete data [47, 48].
In the specific case of diseases with an epidemic potential, the multi-skills requirement of ICPs, including routine activities, epidemiological surveillance, administration and management of resources, community collaboration, etc., generates a critical need for training and in-service training to ensure the timely identification and correct completion of surveillance forms. The lack of training of newly assigned ICPs by DMT members on surveillance procedures may be an additional explanation of the poor completion and filling level of surveillance forms.
Strategies such as on-site training should also be devised to engage all staff at health facilities, not just ICPs, in surveillance activities.
Processes categories
In a previous study in Burkina Faso, Geers and colleagues (2018) identified the cell phone as the main tool for CBHWs to report unusual events [24]. This corroborates our findings, where this tool was used in case of transport constraints or for immediate notification of unusual events. However, the cell phone report is usually followed by a hard copy (paper) transmission of epidemiological information between the PHC and health districts. Some actors pointed out this procedure as leading to an overload of work or a double activity. That calls for a reflection on the opportunities to improve data transmission and archiving of weekly and monthly reports including digital means.
We observed a lack of simplified case definition for some disease and skills, which could explain the misidentification of suspected cases or under-detection of cases by CBHWs and health facility workers. Hence the importance of training and the use of case definitions that are understandable at local level. Issues such as understanding of forms and/or case definitions could be addressed by using a local language during training sessions of community actors or by the use of translated forms, as recently applied in the case of community management of childhood infections and illnesses in the District of Barsalogho [49]. In Nigeria, Nnebue and colleagues (2012) recommended regular training and in-service training of relevant health care workers with adequate and regular provision of IDSR forms, copies of the standard case definitions, and other necessary logistics to the PHC [50]. These additional training activities by local and state governments imply the availability of financial resources, which may require considerations as to relevant reallocation opportunities.
We observed that factors such as the poor quality of the telephone network coverage or the unavailability of financial resources affect reporting. This situation sometimes lead actors across all levels to adopt accommodative methods, in order to report information in time or to promote active feedback to the community level. These accommodative methods included the use of private telephones and/or telephone units and use of any available means of transportation such as privately owned cars or motorcycles. These choices imply a cumbersome process of reimbursement. In their study from Zanzibar, Saleh and colleagues (2021) reported similar findings in terms of availability of communication services, inadequate transportation capacities and funds, which hindered regular supervision, training, and outbreak investigation [51].
The ministry of health is planning to deploy innovative monitoring tools such as RapidSMSFootnote 4 and RECFootnote 5 [52] for pilot diseases and should hopefully replace the current accommodative measures. However, for the new tools to be efficient, they must be appropriated by all actors in order to improve the overall performance of the system.
We also observed that all data analyses were undertaken at district or regional level, as PHC workers did not perform any analysis or interpretation of collected data at local level. As Rasmussen (2018) [53] pointed out, the inability of actors to analyse and work with data at the local level may negatively affect the quality of the data reported to the district level. In this study, we noted several forms with missing or inadequate data submitted to the District level which, given local analysis, may have been noted.
Additionally, epidemic management meetings could play a catalytic role for planning and decision-making at the local level. Yet, are increasingly difficult to maintain due to lack of financial means across all levels (regional, provincial, departmental, municipal). The current strategy of integrating these meetings with other activities, such as sexual and reproductive health and malaria control programmes, faces the challenge of a limited attendance numbers by all stakeholders. As a consequence key actors are excluded from essential discussions and decision-making that align with local expectations, such as financial support.
We identified a lack of direct laboratory feedback from the District to the PHCs in the case of a negative test result. This represents a motivational challenge, also noted by Drabo and colleagues (2015), who reported that without regular communication of test results, the relevance of notifying cases and the sense of useful contribution towards the system gets questioned at community level [54]. In fact, this feedback definitely encourages health workers and the community “rapporteurs” to participate in the surveillance system and in-service training [54].
Output categories
At the national level aggregated data are translated into yearbooks, statistical yearbooks, epidemiological reports and other reference documents that are distributed at district level. The dissemination of these documents may vary between districts. These resources, in reality, are used for service and system planning and management, advocacy and policy development at district level. However, nationally aggregated data may not adequately reflect the health situation in a given district, thus posing a challenge to the effectiveness of the developed policies and plans. Whereas the HMN [2] states that local information should be used to guide local decision-making, the capacity to analyse data often lacks at the peripheral levels where data are generated and where the results should be used for planning and management. That is sometimes, due to a poor archiving system and a lack of corrective actions in local databases after data validation sessions. In their analysis of the WHO-African region, Mbondji and colleagues (2014) underlined that health management information systems generate considerable data, but the information is rarely used because of concerns about bias, quality and timeliness [55].