Evidence for significant COVID-19 community transmission in Somalia using a clinical case definition.

One reason underpinning the global magnitude of the novel coronavirus disease (COVID-19) is its transmissibility by asymptomatic or pre-symptomatic persons (Gandhi et al., 2020). Thus, early detection and isolation of infected individuals are crucial for containing the ongoing pandemic. Currently, the World Health Organization (WHO) recommends surveillance by screening for acute respiratory illness with/without fever, and/or recent contact with COVID-infected individuals or communities (World Health Organization, 2020a). We applied the WHO COVID-19 case definition in Somalia and suggest ways to improve its performance.

In March 2020, researchers at the University of Antwerp (Belgium) launched the International Citizen Project COVID-19 (ICPCovid), which uses web-based surveys to investigate the impact of COVID-19 and associated restrictions on residents of several countries (ICPCovid research group, 2020). The survey tool was translated into the Somali language and locally disseminated via social media platforms. Participants in Somalia were consecutively recruited using a snowball approach from April 23 to May 7, 2020; responses were submitted using mobile phones. The questionnaire included questions about the presence or absence of flu-like symptoms during the last 14 days and the specific symptom(s) experienced by the respondents. Participants were asked whether they had been tested for COVID-19 within the last month (tests performed by the government using nucleic acid amplification techniques), and the eventual test results. The protocol was approved by the University of Antwerp Ethics Committee and Mogadishu University’s Institutional Review Board. Informed e-consent (checkbox) was required before submitting responses.

Overall, 4124 entries were analyzed. Of the 182 participants (65.9% male, mean age: 22 ± 4) with available test results, 49 (26.9%) had tested positive for COVID-19. 16 (32.7%) of the 49 confirmed COVID-19 cases reported no flu-like symptoms. Fever and anosmia were the most frequent symptoms, being reported respectively by 45/182 (24.7%) and 23/182 (14.3%) of participants with test results (Supplementary Table 1). Applying the WHO COVID-19 diagnostic criteria with the assumption that all tested individuals were contacts of a confirmed or probable case, 25 (13.7%) were classified as suspected COVID-19 cases (presence of respiratory symptoms with/without fever) (World Health Organization, 2020a); sensitivity: 32.7% [95% CI 20–48], specificity: 88.7% [82-94]. Adding anosmia as a diagnostic criterion for COVID-19, alongside other respiratory symptoms, increased the sensitivity of the WHO definition to 38.8% [25-54] (Supplementary Table 1). Based on these findings, we propose the following clinical criteria for COVID-19 suspicion: Any individual presenting at least one major sign (fever and/or anosmia) with one or more respiratory symptoms (cough, shortness of breath, sore throat, coryza), with an epidemiological context of COVID-19 transmission. Upon applying these criteria on all 4124 respondents and making abstraction of any previous COVID-19 contacts, the participants’ self-reported symptoms predicted 334 (8.1%) suspected cases as opposed to 259 (6.3%) detected by the WHO definition (p < 0.0001).

These results strongly suggest a high level of community transmission of COVID-19 in Somalia that is most likely under-reported using current approaches. This implies that a significant portion of infected Somalis go undetected and unwittingly serve as asymptomatic spreaders. We demonstrate that the inclusion of anosmia into a case definition for COVID-19 may result in fewer false negatives, thus ensuring that a higher number of cases are quarantined until they no longer pose a public health threat. It appears that the onset of anosmia precedes the full-blown clinical disease (Giacomelli et al., 2020), thereby enabling the early detection of infected persons. Several studies have shown that hyposmia/anosmia is very prevalent in COVID-19 patients (Marchese-Ragona et al., 2020, Hopkins and Kumar, 2020). These observations are supported by two meta-analyses (Agyeman et al., 2020, Tong et al., 2020), which further highlighted that validated methods to evaluate smell and taste in COVID-19 patients are more sensitive than subjective self-reporting. Therefore, objectively assessing these symptoms could prove useful in screening for COVID-19, even in primary healthcare settings. Notably, hyposmia/anosmia is seldom reported spontaneously by the patients themselves and should be intentionally investigated during history taking, especially during an ongoing COVID-19 outbreak.

The major weaknesses of our study include:

The web-based, anonymous data collection technique meant there was no possibility of verifying the information for accuracy.

The sample size was small.

No information was available regarding the ear-nose-throat (ENT) medical history of respondents; we however acknowledge the possibility that some ENT conditions may affect their sense of smell.

Lastly, there was a possible selection bias of respondents when assessing the performance of the screening tools (those who had been tested were most likely symptomatic or were contacts of an index case, and may not be representative of the general population).

By June 16, 2020, there were 2642 COVID-19 cases and 88 deaths officially reported in Somalia (World Health Organization, 2020b). Our study suggests that these numbers may considerably underestimate the COVID-19 disease burden in Somalia. Applying the new definition via online surveys could provide valuable information about the actual disease spread. The proposed case definition could complement existing COVID-19 surveillance measures, particularly in resource-poor settings where testing and contact tracing capacities are limited. As many countries are gradually tapering off their confinement measures, this clinical definition adds to various COVID-19 detection strategies instituted in different communities to rapidly identify suspected cases, thus preventing a possible resurgence (Peto et al., 2020).

Funding

RC receives funding from the (Advanced Grant 671055).

Ethical approval

This study was approved by the Ethics Committees of the University of Antwerp (Belgium) and the Mogadishu University (Somalia)

Conflict of interest

The authors declare no conflicts of interest.