Challenges of IDSR in SSA
This review indicates that in most countries, data generated through the routine HMIS, which is the key source of IDSR, are rarely assessed for their quality, analysed and used to support decision-making [29]. Several studies in SSA have revealed weaknesses in case identification and recording at the primary healthcare facilities [14,24,37,57,72,82]. The quality of the data management system remains a challenge, with incomplete and inconsistent data frequently being reported at different levels of the surveillance system. Moreover, HMIS data are considered to be biased because they reflect only the population seeking care from health care facilities.
In Ethiopia, Liberia and Tanzania, assessments of the information systems have identified some data quality issues and lack of use of the generated data [45,56,72,83,84]. In a study in Ethiopia, though the surveillance system was found to be simple, useful, flexible, acceptable and representative, it lacked regular data analysis and feedback dissemination [35]. Moreover, studies in Kenya and Nigeria have indicated that there are gaps between knowledge and practice of disease surveillance among health care workers [85,86]. Incomplete data filing and inadequate organisation have been reported as an inbuilt shortcoming at all levels of IDSR in SSA countries [38,39,58,87, 88]. Routine data analysis is still insufficient at facility and district levels in the majority of the countries mainly due to lack of clear guidelines on how and when to analyse data [57,58,63,72]. Reasons identified for limited data analysis included a shortage of skilled personnel, poor understanding of the use of surveillance data in planning, as well as inadequate infrastructure [72].
There are reports of a few countries (Burkina Faso, Ghana, Liberia, Uganda) that analyse and use routine HMIS data at sub-national levels in Africa [29,45]. In terms of data utilisation, in both Tanzania [56] and Liberia [45], it was found that analysis and data-use have not been given adequate attention. The studies reported under-utilisation of IDSR data at all levels of the health system as a result of poor data management and analysis skills. The culture of data analysis was lacking, and the relevance of surveillance data for decision making at sub-national levels was grossly underestimated. The use of paper-based reporting was likely to lead to severe limitations in the transmission of the data from the point of generation to a higher level [38,58]. Despite significant investment in early outbreak detection in SSA, there is very little evidence that even high utilisation of HMIS data will influence earlier detection [89].
For the integrated system to be efficient, it requires strong coordination and communication, clear organisation structure, adequate resources [90,91] and reliable sources of data. Integration may range from interconnectivity which requires simple transferring of files with basic applications to complex convergent integration which involves merging of technology with processes, knowledge, and human performance [92]. IDSR strategy strives for the convergent integration route, but the majority of countries have never achieved total integration. Implementation of the strategy is partially done [23,48], and there is more focus on technical aspects than organisational, and workforce issues hence impair the performance of the systems [61,93]. Nevertheless, some countries such as Uganda have taken the actions to rectify those systemic challenges and reported improvement in the implementation [62].
Opportunities for improving IDSR
Health information systems
In SSA, several government ministries, agencies as well as academic and research institutions are involved in managing different aspects of the health information systems. The Ministries of Health run the routine HMIS as the major source of information for decision making and planning data that are generated via healthcare facilities. National Statistical Offices are responsible for most of the nation-wide household demographic and health surveys as well as census [29]. Other key health information systems include civil registration, demographic surveillance sites and research outputs [94]. Demographic surveillance sites function in several countries, but the data generated are not integrated into the national health information system because of concerns, about representativeness [29]. Besides, health research and academic institutions are increasingly generating evidence on human and animal health that could be used for disease surveillance purposes. However, most of the findings are only used for estimating national disease distribution rather than for planning national control programmes [95]. A warning of an impending epidemic can help relevant authorities and communities to prepare and take immediate actions to reduce morbidities and mortalities. However, such information is not available for planning, disease surveillance and outbreak management. It is recommended that the governments in SSA to consider establishing a national platform for infectious disease epidemics early warning system. Many of the epidemic diseases are known to be highly sensitive to long-term changes in climate and short-term fluctuations in weather. Meteorological data are made available daily by the National Meteorological Agencies, yet they are rarely used in the monitoring of the occurrence of diseases. Meteorological data can be combined with geospatially referenced data, population densities or road networks to generate estimates of environmental indicators that are relevant to infectious diseases [74]
It is critical for a good and efficient surveillance system to incorporate other sources such as mortality data from demographic surveys, environmental data, vital statistics and civil registration, antimicrobial resistance, systematic survey, meteorological data and research data. In most countries, despite an enormous amount of data generated by these systems, they run in parallel and independently, not well-coordinated, and sharing of information between them is minimal. Each of the existing systems operates its data collection and utilisation framework. Moreover, much of the information is generated outside the health sectors – making it not readily available for disease surveillance purposes. It is a fact that the innovations, including the use of Big Data and artificial intelligence, could transform infectious disease surveillance and response and complement the existing traditional disease surveillance systems and improve detection and response to epidemics [74].
Digital disease surveillance
Digital disease surveillance (DDS) is the use of data generated outside the public health system for disease surveillance [96]. It involves the aggregation and analysis of data available on the internet, such as search engines, social media and mobile phones, and not directly associated with patient illnesses or medical encounters. It has been shown that digital approaches in surveillance improve the timeliness and depth of surveillance information in high-income countries [96,97]. So far, DDS has demonstrated its potential in early detection and response to Ebola and COVID-19 epidemics [98-101]. Recently, DDS has been used in responding to COVID-19 through case detection, contact tracing and isolation, and quarantine in several countries [102]. In Taiwan, the government-linked immigration and customs data on travellers to the National Health Insurance data on health facility visits to identify COVID-19 suspected cases during travel to an affected area [103]. On the other hand, New Zealand and Thailand have used cell-phone location data to monitor the movement of a person's subject to quarantine or isolation orders [103]. In about 30 countries, algorithmic contact tracing through the use of a cell phone app or operating system has been deployed in response to COVID-19 pandemic [103, 104].
There is growing interest in using digital surveillance approaches to improve monitoring and control of infectious disease outbreaks. However, such applications are scarce in Africa, and few studies have shown a direct connection between DDS and public health actions. Currently, the Africa CDC is implementing a pilot programme in Ghana, Liberia, Madagascar, Nigeria, Sierra Leone and South Africa to develop digital surveillance indicators and online disease dashboards based on social media to inform infectious disease surveillance [105]. Moreover, there are on-going efforts to create real-time data sharing platforms for disease surveillance using mobile technologies that will allow centralised data management and use [106]. This is expected to strengthen real-time surveillance of infectious diseases in the continent, guide interventions, and build capacity in "Big Data" approaches for outbreak prediction, analysis and prevention.
With the proliferation of information technologies and increased rate of ownership of mobile phones in SSA, there are large amounts of data on social media blogs, chatrooms and local news reports that may provide governments and other stakeholders' clues about disease outbreaks in time and place daily. Such data are essential raw materials for DDS. Advancements in information technology and information sharing is giving rise to a new field known as infodemiology – defined as "the science of distribution and determinants of information in an electronic medium, specifically the internet [98]. To-date, Program for Monitoring Emerging Diseases (ProMED-mail) [107] and HealthMap [108] are among the several leading efforts in digital surveillance. The World Health Organization routinely uses HealthMap, ProMED and similar systems to monitor infectious disease outbreaks and inform public health officials and the general public [109]. The key advantages of DDS include speed and volume, which may increasingly help health officials to spot outbreaks quickly and cheaply [106]
Community event-based Surveillance
Community-based surveillance (CBS) may be defined as the systematic detection and reporting of events of public health significance within a community by community members [110-112]. The engagement of the community has long been an essential part of both human and animal health [110-115]. CBS has played a significant role in smallpox, guinea worm and polio eradication programmes [112]. Recently, CBS was reported as an important component in response to the West African Ebola virus disease outbreak of 2014-2016 where community health workers and volunteers worked together in early detection and timely reporting to the health system [116]. With CBS, public engagement is being transformed through participatory surveillance systems that enable the community to directly report on disease events via information technology and communication tools [117]. Several CBS systems have been described and have demonstrated their accuracy and sensitivity, their ability to provide more timely measures of disease activity, and their usefulness identifying risk groups, assessing the burden of illness and informing disease transmission models [118-121]. CBS can be an important component of early warning of emerging events by engaging the communities to detect potential public health events and connecting individuals to health services [3,122-125]. In a study in Ivory Coast, following the implementation of community-based surveillance, 5-fold and 8-fold increases in reporting of suspected measles and yellow fever clusters, respectively, have been reported [122]. These findings suggest that CBS strengthened detection and reporting capabilities for several suspect priority diseases and events.
The Technical Guidelines on IDSR Guidelines [22,27] highlight the need for community-event based surveillance. This is because most of the health problems and events happen at the community level. It is through these reasons that putting a surveillance mechanism to obtain information at the community level is an added advantage to capture diseases and public health events at its early stages to allow effective preparedness and response thereby managing disease outbreaks at the source. Despite the relevance of the inclusion of community information in surveillance, by the end of 2017, of the 44 countries in the WHO Africa region, 32 (68%) had commenced CBS, and 35 (74%) had event-based surveillance [23]. However, there is only one report from Sierra Leone that data collected from the two approaches are integrated into the national IDSR system [122]. In some countries, the CBS programmes are still operating as pilot or research projects [126,127] and most cover a limited geographical area and are mainly for specific disease programmes in rural settings [122, 128].
One Health Surveillance
As part of an effective global response to diseases transmitted between animals and humans [129], there have been calls for integrating surveillance of zoonotic disease events in human and animal populations. The driving force is the fact that about three-quarters of emerging infectious diseases of humans have animal origin [130]. The concept of one health (OH) promotes the trans-sectoral collaboration between human, animal, and environmental disciplines and sectors in addressing complex health issues. The aim is to remove the traditional boundaries between disciplines and sectors and that all relevant stakeholders are involved in the definition and management of health problems [129]. Several African countries have carried out a prioritisation exercise on the zoonoses in the region. Among the diseases that were ranked high, include Anthrax, Brucellosis, Viral Haemorrhagic Fevers, Zoonotic Avian Influenza, Human African Trypanosomiasis, Rabies and Plague [131-135]. With this approach, OH surveillance is strongly encouraged at global, national and local levels to efficiently manage health events involving humans, animals and their environment [27]. With the adoption of OH surveillance, some issues need to be considered and addressed. These include the need to define the characteristics of OH surveillance and identify the appropriate mechanisms for inter-sectoral and multi-disciplinary collaboration [90, 131].
Towards multi-sectoral and multi-indicator surveillance
The emerging and re-emerging infectious diseases in Africa underlined the urgent need for the integration of public health surveillance systems [136]. As infectious disease threats increase in SSA, effective ways of predicting outbreaks and planning for outbreak responses become increasingly important. An epidemic intelligence that encompasses activities related to early warning functions for infectious diseases of humans and animals in SSA is almost non-existent. We, therefore, propose development and adoption of a national platform for public health surveillance that is multi-sectoral, multi-disease and multi-indicator epidemic intelligence system (Figure 2). Evidence-based outbreak preparedness provides ground to streamline and concentrate our efforts towards diseases that have been documented to circulate. Among other things, outbreak preparedness entails prediction of possible epidemics with regards to the possible location of involvement, the risk and vulnerability of the population, the extent of the outbreak, its spread and socioeconomic consequences. Therefore, for any effective outbreak preparedness plan, information on prior risks is crucial in setting priorities for robust outbreak management and response plan. Research findings for decades have displayed mapping of exposure patterns and the burden of infectious diseases that have the potential to cause outbreaks in the community.
Modern technologies such as artificial intelligence and machine learning are widely applied in the analysis of a significant volume of data to assess the status and forecast future dynamics of diseases [137,138]. The prediction model is not only valuable for disease prevention and saving disability-adjusted life years, but it also saves valuable financial resources due to the high costs and resource utilisation associated with poorly predicted management techniques and costs to the health system when an outbreak happens. These emerging technologies are likely to become a powerful means of helping us collect more accurate and timely information, which in turn can lead to more effective preventive measures and improved public health practice. The techniques are expected to allow decision-makers to identify areas where the model predicts with certainty a particular risk category, to effectively target limited resources to those districts most at risk for a given season.