Structure, function and performance of Early Warning Alert and Response Network (EWARN) in emergencies in the Eastern Mediterranean Region

Highlights

• •The outbreak detection function was undermined by inclusion of non-epidemic diseases.
• •Adoption of electronic platforms expanded coverage of hard-to-reach areas.
• •Conflict emergencies limited coverage, despite progress due to electronic solutions.
• •The EWARN is being retained far longer than expected in protracted emergencies.

Abstract

Introduction

The Eastern Mediterranean Region (EMR) has experienced several protracted humanitarian crises. The affected population are served by eight Early Warning Alert and Response Network (EWARN) systems for outbreak detection and response. This study aimed to compare structure, function and performance of the systems’ adherence to current guidance, and noted emerging lessons.

Methods

This study included a review of published and unpublished literature, a structured survey and interviews.

Results

Findings showed that all systems adhered to basic EWARN structure. Four of eight systems had electronic platforms, while one was implementing one. Regarding key EWARN function of outbreak detection: of the 35 health conditions, 26 were communicable diseases and nine were non-communicable; two systems focused on epidemic-prone diseases. Half the systems achieved ≥60% population coverage, five achieved ≥80% reporting timeliness, six achieved ≥80% reporting completeness, and seven achieved verification of ≥80% of alerts of suspected outbreaks.

Conclusion

The findings showed that the systems followed the EWARN structure, while increasing adoption of electronic platforms. Performance, including timeliness and completeness of reporting, and timely verification of alerts were optimal for most of the systems. However, population coverage was low for most of the systems, and the EWARN’s primary focus of outbreak detection was undermined by the increasing number of non-epidemic diseases.

Introduction

The Eastern Mediterranean Region (EMR) has experienced several humanitarian crises of great magnitude in recent years. With a population of 680 million, the region has the highest number of people in need of humanitarian assistance (OCHA, 2019), including over 30 million who have been displaced across the region (WHO, n.d.). Twelve of 22 countries in the region are experiencing armed conflict, while Jordan and Lebanon are coping with large influxes of displaced refugees from the affected countries (Brennan et al., 2020). In Syria alone, 17 million people have been internally displaced and a further 5.9 million have fled to neighboring countries as refugees (UN, 2018).

Internally displaced persons (IDPs) or refugees from affected countries often settle in densely populated temporary shelters and camps, with inadequate food, unsafe water, poor sanitation, and exposure to the elements and disease-causing vectors. These circumstances increase the risk of communicable disease outbreaks among the displaced population (Gayer et al., 2007, Ismail et al., 2016, Toole, 1995). In addition, host communities of displaced populations tend to be disproportionately affected due to increased demand for local health services and excessive strain on fragile water, sanitation and other civil infrastructure (Al-Lami et al., 2013, Lam et al., 2016).

The need for rapid detection and prompt response to outbreaks of epidemic-prone diseases among victims of a humanitarian crisis is critical in ensuring that excess morbidity and preventable loss of life among the vulnerable population is avoided (WHO, 2012). However, during humanitarian emergencies, health systems are often affected and quickly become overwhelmed to meet the needs of the emergency. To address this gap, the Early Warning Alert and Response Network (EWARN) was developed by the World Health Organization (WHO) and its partners as a temporary emergency response intervention (WHO, 2012). It is a surveillance system focusing on a few priority, epidemic-prone conditions, which can be rolled out quickly. Its primary objective is to quickly detect and verify alerts or unusual health events (WHO, 2012). The first EWARN was implemented in the EMR in 1999, in the present day South Sudan, formerly Sudan, following protracted civil war and an outbreak of relapsing fever (WHO, 2002). Subsequently, the EWARN has been implemented in a number of humanitarian emergencies (Husain and Mala, 2018).

The EMR currently has eight active EWARN systems in seven countries experiencing protracted humanitarian emergencies: Afghanistan, Iraq, Libya, Somalia, Sudan, Syria, and Yemen. When humanitarian emergencies exceed local capacities to respond, as has been witnessed in these countries, local authorities usually request external support. When countries make such requests, the WHO leads the health sector response through its health emergencies program and the EWARN is one of the contributions. The Infectious Hazards Management (IHM) unit of the WHO Emergencies program at the regional office has lead implementation of the systems and monitored their operations. According to the guidelines, the EWARN should have an immediate reporting component for alerts of suspected disease outbreaks and weekly reporting components of aggregated data for monitoring patterns of disease occurrence. The primary function of the EWARN should be outbreak detection; therefore, it should focus on priority epidemic-prone diseases among the affected population, about 9–12 in number, based on risk assessment. In terms of performance, the system is expected to achieve at least 60% coverage of the target population. At least 80% of alerts received should be verified within 24−48 h; and timeliness and completion of weekly reporting should be at least 80%. Other key EWARN features should include written procedures to guide operations, epidemiological bulletins for feedback to stakeholders, and evaluations and exit strategies, as the EWARN is a temporary solution in humanitarian crises.

The EWARN systems in the EMR have been in place for several years. Even though these EWARN systems are built upon the same principles following WHO guidelines, they have been implemented in settings that vary considerably in terms of humanitarian settings, population size, geographical coverage, and resources. While individual country-level evaluations have been conducted for most of these systems, a landscape analysis to describe and compare the various EWARN systems in the region has not been attempted. The aim of this analysis was to examine structural, functional and compare performance among the various EWARN systems, while gauging the extent to which current recommendations are respected, as well as developing strategies for future practice.

Methods

This analysis of EWARN systems was conducted from April to November 2018. The approaches included: (1) review of published and unpublished literature on EWARN, in-country documentation of the systems, evaluation reports, operational manuals, and epidemiological reports; (2) an online survey using a structured questionnaire; and (3) follow-up interviews and emails with key stakeholders and EWARN focal points.

A self-reporting questionnaire was designed based on the EWARN evaluation protocol (WHO, 2018a). The eight categories of questions included priority diseases; human resources; system coverage; alert and epidemic thresholds; tools and guidelines; system design; reporting flow; information products from EWARN data; and performance data. The questionnaire was emailed to the EWARN focal points at the Ministries of Health and WHO in the seven countries with active EWARN systems in the EMR–Afghanistan, Iraq, Libya, Somalia, Syria, Sudan, and Yemen–in April 2018.

Follow-up telephone calls were made to seek clarity from the respondents regarding their responses, to prompt those who delayed in responding, and to seek further information where partial information was identified in response to the questionnaire. During a regional workshop for EWARN focal points in the EMR in August 2018 in Beirut, Lebanon, submitted data were presented and outstanding information gaps were addressed. The data were compiled and a descriptive analysis of key characteristics were performed, including structure, function and performance of the EWARN systems.

Results

According to the findings, the eight active EWARN systems in the EMR were established between 2004 and 2016 (Table 1). Chronologically: Sudan was the first country in the region to implement EWARN in Darfur states in May 2004; EWARN in Afghanistan, referred to as the Disease Early Warning System, was implemented in 2006; EWARN in Somalia, initially referred to as the Communicable Disease Surveillance and Response System, was established in 2011; EWARN in the government-controlled part of Syria, referred to as the Early Warning Alert and Response System, was established in 2012, while EWARN in opposition-controlled Syria was established in 2013; EWARN in Iraq was implemented in 2013, mainly in the refugee camps in Northern Iraq, and was upgraded to an electronic platform in 2014; EWARN in Yemen, referred to as the Disease Early Warning System, was implemented in 2013, mainly in the southern parts of the country, and expanded to all parts of the country following escalation of the civil war since 2015; and the EWARN in Libya, referred to as the Early Warning Alert and Response System, was established in 2016 as an electronic system (Tables 2 and 3 ).

Table 1.

Characteristics of EWARN systems in the EMR.

Characteristics	Afghanistan	Iraq	Libya	Somalia	Sudan	Syria (O)	Syria (G)	Yemen
Type of emergency	Conflict	Conflict	Conflict	Conflict	Conflict	Conflict	Conflict	Conflict
Year of implementation	2006	2013	2016	2011	2004	2013	2012	2013
Total population served (millions)	30	6	4	17	2	–	19	28
Number of reporting sites [no. at inception]	1095 [7]	226 [2]	146 [26]	798 [195]	160 [54]	450 (25)	1183 [104]	1982 [100]
Number of staff	52	18	37	31	72	–	220	35
Electronic platform implemented	No	Yes	Yes	Yes	No	No	No	Yes
Year of last evaluation	2018	2016	–	2014	2017	2019	2016	2014

O = Opposition; G = Government.

Table 2.

Diseases under EWARN in the EMR.

Diseases and health events	Yemen	Libya	Iraq	Afghanistan	Somalia	Sudan	Syria (O)	Syria (G)
Priority epidemic-prone and diseases targeted for elimination	12	11	12	11	12	10	9	9
Other infectious diseases	14	8	6	4	2	1	1	1
Non-communicable diseases	0	0	0	1	0	2	0	0
Total number	26	19	18	16	14	13	10	10

O = Opposition; G = Government.

Table 3.

Performance of EWARN systems in the EMR.

Performance parameters	Afghanistan	Iraq	Libya	Somalia	Sudan	Syria (O)	Syria (G)	Yemen	Median (range)
Completeness (%)	90	100	80	70	75	97	78	92	85 (70−100)
Timeliness (%)	95	84	70	60	75	98	80	92	82 (60−95)
Proportion of alerts verified within 24–48 h (%)	95	82	87	2	89	92	90	–	89 (2−95)
Coverage of target population (%)	31	16	25	97	24	92	92	97	31 (16−97)

O = Opposition; G = Government.

Together, the eight EWARN systems serve 106 million people (Table 1) who have been affected by humanitarian emergencies in the region, including 90 million of the host population, 15 million IDPs and 1 million refugees. Further findings are presented by structure, function and performance of the EWAN systems.

Structure

The EWARN structure emphasizes simplicity and effectiveness. Findings showed that all eight EWARN systems had immediate reporting and weekly reporting components. Weekly reporting is a paper-based manual method in Afghanistan, Sudan, and in the two EWARN systems in Syria. Findings showed that four EWARN systems in Iraq, Libya, Somalia and Yemen have upgraded to electronic platforms. These platforms include mobile devices and applications for data transmission.

The eight EWARN systems are managed by 477 dedicated staff. The Syrian EWARN system in government-controlled parts of the country has the largest number of staff at 220 (Table 1), while Iraq has the lowest with 16 staff. These numbers do not include the thousands of health facility staff who do the actual reporting as part of their routine activities in a total of 6040 reporting sites of the eight EWARN systems. Among the eight EWARN systems in the region, the current findings showed that five of the EWARN systems were national in scope, due widespread internal conflict; two served refugees and IDPs in Darfur, Sudan and Iraq; and the EWARN systems in opposition-controlled parts of Syria served mainly northern parts of the county. The number of administrative levels of reporting varied from country to country depending on the EWARN coverage. In Afghanistan, Yemen and Syrian, where intermediate levels are more functional, there are three levels of reporting from the health facility to the district, province and central levels. In Sudan, the systems are decentralized to each of the affected Darfur states of Sudan; in Somalia, with many competing jurisdictions, the systems are decentralized to the jurisdictions with the WHO office, playing the role of consolidation of the national reports. According to the EWARN focal point in Libya, even though the electronic reporting is nationwide, the health offices at the intermediate levels, including districts and governorates, are fragmented among different combatants, lack health workers and are largely dysfunctional; hence, reporting is direct from the reporting site to the central level, enabled by an electronic platform.

All of the eight EWARN systems have operating manuals. Furthermore, while establishment of the systems has been led by the WHO, with support from other partners such as the USA CDC, the local authorities are at the center of management of the systems at all levels, with close operational and financial support from the WHO and partners in emergency response.

Functions

The primary function of the EWARN is early outbreak detection, and timely response to minimize morbidity and mortality among populations affected by humanitarian crises. Therefore, the focus should be on epidemic-prone diseases. However, findings showed that out of the 35 health conditions included in the eight EWARN systems in the seven countries, 25 (68%) were linked to epidemic-prone diseases, while 10 health conditions (32%), including three non-infectious health conditions, were not epidemic-prone. The epidemic-prone diseases included cholera, dysentery, measles, polio, hepatitis, hemorrhagic fever, meningitis, and pneumonia. The non-epidemic-prone infectious disease included scabies, leishmaniasis, schistosomiasis, tuberculosis, HIV, and dracunculiasis (Guinea worm), while non-infectious conditions included injuries, malnutrition and pregnancy-related death. Only the two EWARN systems in Syria had the recommended number of 9–12 priority health conditions, while Yemen, with 24 health conditions, had the highest number of health conditions of the EWARN systems.

Regarding immediate reporting of alerts, the four manual EWARN systems rely on emails, text messages and phone calls for data transmission, while the four EWARN systems with electronic platforms report alerts in real-time based on established alert thresholds for levels of occurrence of the diseases and syndromes, which suggest that an outbreak may be occurring.

Regarding data management, analysis is automated within the electronic EWARN platforms of Iraq, Libya, Somalia, and Yemen, with preformatted charts and tables not requiring a skilled analyst. The systems also have mapping tools for visual display of disease-occurring patterns, including geographical distribution of reported alerts and their verification status based on color codes. In the four manual systems, data analysis is performed using Excel and is heavily dependent on expertise of the focal point.

The four systems with electronic platforms generate weekly epidemiological bulletins based on inbuilt data summarization tools, while the manual systems rely on manual production of the documents based on standard templates that have been provided through the EWARN guidance document. The bulletins are disseminated to relevant stakeholders manly by email, in addition to presentations made at the health cluster meetings.

Performance

Based on survey data and follow-up interviews, findings show that population coverage by the eight EWARN systems was lowest in Iraq at 16% (6 million) (Table 1) followed by Sudan (24%; 2 million), Libya (25%; 4 million), and Afghanistan (31%; 30 million). Somalia, Syria and Yemen, with high target population coverage >90%, achieved the minimum requirement of 60% population coverage. The target population included refugees, IDPS and host populations affected by the emergencies.

Results also show that the proportion of alerts verified within 24–48 h had a verification rate of 2% in all the EWARB systems except Somalia. Afghanistan had the highest verification rate at 95%, followed by EWARN in government-controlled and opposition-controlled parts of Syria at 90% and 92%, respectively.

Timeliness of weekly reporting exceeded the 80% minimum requirement in five EWARN systems. Somalia had the lowest timeliness of reporting (60%) followed by Libya (70%) and Sudan (75%). The Afghanistan EWARN system had the highest timeliness of reporting at 95%. Completeness of reporting exceeded the minimum requirement of 80% in six EWARN systems. Completeness of reporting was below threshold in the EWARN in Somalia (70%) and in the government-controlled part of Syria (78%). Iraq had the highest completeness of reporting (100%).

Other findings show that all the EWARN systems have been evaluated at least once, except for the EWARN in Libya, which has not yet been evaluated. Additionally, only the governments of Syria and Afghanistan have considered transition of EWARN to routine surveillance systems, especially in the more peaceful areas of their respective countries. Syria has gone a step further and examined the best practices of EWARN that could be retained to strengthen the national routine surveillance system after the transition. According to the findings, an exploratory workshop was conducted in this regard in Damascus, Syria, in 2018.

Discussion

This study aimed to examine characteristics – including structure, function and performance–of the eight EWARN systems that are currently active in humanitarian emergencies in the EMR, and to identify emerging lessons that could inform future practice. Key structural characteristics that were analyzed included adoption of technology; key functional characteristics analyzed included priority health conditions under surveillance, data reporting and management, and feedback to stakeholders; and performance characteristics of the systems included target population coverage, timeliness and completeness of reporting, timeliness of verification of alerts of suspected disease outbreaks, feedback to stakeholders, and considerations for transition to routine surveillance post-emergency.

One of the key requirements of the EWARN is simplicity, and all the systems adhered to the basic structure of immediate and weekly reporting components according to the EWARN guidelines. Four of the eight EWARN systems upgraded to electronic platforms, while one was in the process of being implemented. Increasing adoption of technology was in line with the emerging trend for the EWARN in emergencies (Salazar et al., 2016).

Regarding the primary EWARN functions of timely outbreak detection and optimal response in humanitarian emergencies to minimize morbidity and mortality, the findings showed that 32% of health conditions monitored by the EWARN were not epidemic-prone diseases. In two EWARN systems, the non-communicable diseases included injuries, malnutrition and pregnancy-related deaths (Afghanistan and Somalia), while the number of health conditions exceeded the recommended 9–12 priority health conditions in six systems. This undesired phenomenon in countries such as Afghanistan, Somalia and Yemen may also be explained by the protracted nature of the emergencies, associated with sustained destruction of routine surveillance systems, and the expansion of EWARN to fill the gap. Experience has shown that where EWARN has emerged as the main functioning surveillance system, demands from stakeholders, especially from local authorities, for inclusion of orphaned, non-epidemic-prone health conditions, often increase. Such pressure is difficult for the EWARN technical teams to resist, partly due to recognition of the importance of the health conditions in certain settings, and mostly due the fact that these emergency responders serve the affected population at the preference of the host government (UNHCR, 2020).

Feedback to stakeholders, one of the key functions of EWARN, was practiced mainly through weekly epidemiological bulletins by all eight EWARN systems.

In terms of performance, the findings showed that the eight active EWARN systems in the EMR currently serve an estimated 106 million people in seven countries experiencing humanitarian emergencies; these include refugees, IDPs and the host population in the affected countries. Despite the high number, coverage of the target population was sub-optimal in most of the countries (<60%). In three countries with the highest proportion of the population under EWARN–including Somalia (97%), Yemen (97%) and Syria (92%)–the high coverage may be explained by the extended presence and subsequent scaling up of EWARN to the extent that the systems have become the sole surveillance systems in these countries. In the countries with low target population coverage, such as Afghanistan, Iraq and Sudan, the approach to EWARN implementation was more geographically targeted and as the humanitarian crisis escalated to affect more people, the systems did not expand at the same pace.

Timeliness and completeness of reporting exceeded the recommended baseline of 80% in five of the eight systems, despite insecurity in all the affected EMR countries. The high rates of timeliness and completeness of reporting may be attributed to adoption of web-based reporting electronic solutions that have made it possible to overcome distance and insecurity barriers faced by all the EWARN systems in the EMR. Electronic solutions not only reduce the burden of reporting on responsible health facility staff, but crucially, these solutions also facilitate real-time reporting of alerts of suspected disease outbreaks for timely action. However, these findings also illustrate the possible contribution of insecurity in hindering timely verification of alerts of suspected outbreaks, as this often involves field visits by rapid response teams. Despite timely reporting, the desired verification within 24–48 h of at least 80% of reported alerts was not achieved in one of the most insecure countries, namely Somalia (2%).

Further findings showed that all the systems have been evaluated at least once, except Libya. Additionally, Afghanistan has initiated transition to routine surveillance systems, while Syria has initiated preliminary exploration of the possibility of restoring routine surveillance in secured districts (WHO, 2018b, c). One of the aims of the transition is to ensure adoption of best EWARN practices and use of EWARN financial resources for a better and stronger post-EWARN routine surveillance system. This is especially important in countries with protracted emergencies where the EWARN is the main surveillance system. Transition approaches in such settings should ensure that the phasing out of EWARN does not leave a surveillance vacuum in the countries. There are currently no clear guidelines to inform transition from the EWARN to a routine surveillance system post emergency; this is a major gap that should be addressed to ensure that transition is part of the general recovery process following an emergency.

Regarding limitations of this study, it is acknowledged that this study did not address all the components of the EWARN systems in the region; for example, the laboratory component of the EWARN systems was not assessed. Additionally, data were collected from responders through a survey, which has inherent limitations such as reporting bias. Therefore, the analysis of the data needs to be interpreted with caution.

Conclusion

Overall, the findings showed that the systems followed the basic structure of EWARN and most of the EWARN performance indicators–including population coverage, timely verification of alerts, as well as timeliness and completeness of reporting EWARN systems–were within optimal levels for most of the eight active EWARN systems in the EMR. Additionally, feedback to stakeholders through epidemiological bulletins was disseminated in all the reviewed systems. Most of the EWARN systems had adopted electronic solutions to facilitate data transmission and data management.

Despite the positive observations in key parameters of the EWARN, these findings also show that even though EWARN is intended to be limited in scope and duration, and should focus on key priority epidemic-prone diseases based on risk assessment, some of these ideals have not been respected. Contributing factors include the protracted and large-scale humanitarian crisis with extensive destruction of local surveillance systems. It should also be noted that solutions to some of these challenges, including burdening of EWARN with non-epidemic prone diseases, may lie in the emerging developments of the EWARN such as adoption of electronic solutions. In settings where electronic EWARN systems have been established, the functional capacity of the IT infrastructure often exceeds the needs of EWARN. This extra capacity of the IT infrastructure could be made available to support other surveillance needs, while preserving the integrity of EWARN. For example, special reporting forms and modules for other surveillance needs could be added to the established IT infrastructure at minimal added cost through the leadership of technical units responsible for the concerned health conditions such as tuberculosis, HIV, injuries and trauma, and nutrition (Brennan and Nandy, 2001, Connolly et al., 2004, Salama and Dondero, 2001). This could go a long way toward preserving the integrity of the EWARN.

Despite the limitations of the study, the findings also raise important question for EWARN practice and research going forward: should the increasing variety and number of health conditions in the EWARN become the new norm? Given that optimal operation of EWARN is highly resource-intensive, is expansion of EWARN to function as the national surveillance system sustainable? In settings where EWARN functions as the sole surveillance system, how should transition from EWARN occur without creating a surveillance vacuum in the affected countries? Should transition from EWARN in such settings include rebuilding the routine surveillance system? Are emergency response funders willing to address the cost of rebuilding a destroyed routine surveillance system to ensure smooth transition from EWARN without creating a surveillance vacuum post-emergency? These are pertinent EWARN practice questions for national authorities, emergency response partners and researchers going forward.

Funding statement

None declared.

Ethics statement

The study did not require ethical approval.

Conflict of interest

None declared.

Appendix A. Supplementary data

The following is Supplementary data to this article: