Study setting
This study was conducted in the Oromia Regional State of Ethiopia. Oromia is the largest of Ethiopia’s fourteen regions, comprising approximately 32% (363,399.8 km²) of the national land area. During the COVID-19 pandemic, the region’s healthcare response infrastructure included 25 designated COVID-19 treatment hospitals, 15 laboratories equipped for quantitative reverse transcription polymerase chain reaction (qRT-PCR) testing, and over 60 sites for point-of-care antigen rapid diagnostic tests (Ag-RDTs).
From the 12 comprehensive (referral and university teaching) hospitals in the region, this study purposively selected five Hospitals. The study selected the hospitals in four major towns of Oromia, Ethiopia, based on their high COVID-19 caseloads, ensuring data from the most active treatment and testing centers (Fig. 1). The surveyed hospitals were: Jimma University Specialised Hospital (Jimma Town), Wollega University Comprehensive Specialised Hospital, Nekemte Comprehensive Specialized Hospital (Nekemte Town), Ambo University Teaching and Referral Hospital (Ambo Town), and Adama Hospital Medical College (Adama Town)22.
Fig. 1
Map of Ethiopia showing the location of towns from which participating referral hospitals were selected in the Oromia Region. The map was created by the authors using QGIS Geographic Information System software (QGIS version 3.16.16, Hannover, 2009–2019, QGIS Development Team). The base layer of regional boundaries was sourced from publicly available administrative shapefiles (https://data.humdata.org/dataset/cod-ab-eth) for Ethiopia.
Study design and participants
A multicentre cross-sectional study was conducted across five selected hospitals from August 9 to 23, 2021. Hospitals with high patient load and those serving as COVID-19 treatment centres at the time were purposefully selected. Recruitment was conducted by dedicated research staff who approached patients across all hospital departments for SARS-CoV-2 testing, regardless of symptom status.
Sample size determination
Using a two-population formula, the sample size was calculated to compare the proportions of active (PCR-positive) and past (serology-positive) infections. Based on assumed prevalence rates of 20% (PCR+) and 30% (serology+), with a two-sided significance level (α) of 0.05, 80% power, and a minimum detectable difference of 10%, an initial sample size of 580 participants was determined. To adjust for potential clustering effects across the five participating hospitals, an intraclass correlation coefficient (ICC) of 0.02 was applied, inflating the sample size to 1,360 using a design effect adjustment. Additionally, a 10% non-response rate was incorporated to ensure robust precision and account for potential participant attrition, thereby maintaining the study’s statistical power and validity.
Data collection procedures
Trained data collectors administered a structured electronic questionnaire through face-to-face interviews to collect demographic and clinical data, with comorbidities cross-verified against medical records when available. Twelve data collectors, sixteen laboratory professionals for sample collection, and twelve for testing—all trained on protocols, tools, and software—conducted the fieldwork. Data collection instruments were translated into local languages (Afaan Oromo and Amharic) and back-translated for consistency. Daily quality checks were performed using the Open Data Kit (CSEntry) application, with supervisors making immediate corrections, and all data were transmitted to a secure central server for verification and approval.
Laboratory specimens
Eligible participants provided paired nasopharyngeal swabs (two per participant) and 5 ml venous blood samples, collected by trained laboratory personnel under strict biosafety and personal protective equipment (PPE) protocols. Following collection, samples were immediately transported under controlled cold chain conditions (2–8 °C) to designated regional laboratories (Adama Regional Laboratory and Nekemte Regional Laboratory) for processing. All laboratory procedures were conducted in Biosafety Level-2 (BSL-2) facilities, with adherence to standardized infection prevention measures, including full PPE. Specimen collection, processing, and testing were performed in compliance with the manufacturer’s protocols to ensure analytical validity23,24 and the Ethiopian Public Health Institute’s recommendations25,26,27.
RT-PCR testing
The Abbott SARS-CoV-2 RT-PCR assay was performed on the Abbott m2000 system using nasopharyngeal samples collected using rayon or Dacron-tipped applicator swabs, which were immediately placed in sterile viral transport media (VTM) before being processed using the Abbott SARS-CoV-2 RT-PCR assay on the Abbott m2000 system. Collected samples were stored at 2–8 °C when processed within 72 h or at -70 °C for long-term preservation, with all transportation conducted following strict triple-packaging protocols to maintain sample integrity and biosafety. Samples were stored at regional laboratories and hospitals under appropriate temperatures until they were tested. The testing procedure involved RNA extraction through magnetic bead purification followed by amplification targeting the RdRp and N genes, delivering results in approximately 3 h. According to the manufacturer’s Clinical Performance Evaluation, the assay demonstrated a positive percent agreement (PPA) of 95.9% (95% CL: 86.0%-99.5%) and a negative percent agreement (NPA) of 100% (95% CL: 93.5%-100.0%), confirming its high diagnostic accuracy for SARS-CoV-2 detection28. All testing was performed in compliance with the manufacturer’s instructions to ensure standardized and reliable results.
Anti-SARS-Cov-2 serology testing
For serological testing, approximately 5 mL of venous blood was collected from each participant in standard serum tubes via venipuncture, with samples centrifuged to separate cellular components and processed daily before being aliquoted and stored at -20 °C for preservation. The samples remained stable for 8 days at 2–8 °C or for extended periods at -20 °C to -70 °C, with transport permitted at ambient temperatures for up to 5 days. Testing was performed in batches using Elecsys Anti-SARS-CoV-2 chemiluminescent immunoassay (CLIA) on the automated Cobas 6000 e601 system (Roche Diagnostics), which employs a double-antigen sandwich format to detect anti-nucleocapsid antibodies. It has a high sensitivity (99.5–100% for samples collected ≥ 14 days post-PCR confirmation) and specificity (> 99.8%), while demonstrating no cross-reactivity against other coronaviruses29. The assay provided results within 18 min (throughput: 300 tests/hour).
Data management and analysis
The data was exported from CSentry CSPro DataEntry 7.2.1 to Microsoft Excel for cleaning and then transferred to Stata version 14.2 for analysis. We have calculated the past seroprevalence and the current prevalence at the time of the study.‘Active infection prevalence was defined as the proportion of participants with a positive RT-PCR result. ‘Seroprevalence’ (past infection) was defined as the proportion with a positive Elecsys Anti-SARS-CoV-2 result. We calculated various diagnostic parameters, sensitivity, specificity, positive predictive value and negative predictive value and Cohen’s kappa coefficient (κ) between RT-PCR and Elecsys Anti-SARS-CoV-2 tests. The ROC Curve (Receiver Operating Characteristic Curve-ROC) and the area under the curve (AUC) analysis were performed using RT-PCR as the reference standard. The Elecsys Anti-SARS-CoV-2 assay, though clinically reported as binary (reactive/non-reactive), generates a continuous quantitative signal (e.g., Cut-off Index, COI). This underlying signal was used to plot sensitivity versus 1-specificity across all potential thresholds against the PCR results, allowing for the assessment of overall discriminatory performance. The diagnostic parameters were presented with a 95% Confidence Interval (CI) determined using the Clopper-Pearson method. Variables for the multivariable logistic regression models were selected based on clinical relevance and a univariable association with the outcome at p < 0.25. A multiple logistic regression model was used to identify factors associated with SARS-CoV-2 infections. A p-value < 0.05 at 95% CI was used as a level of statistical significance.
Ethical considerations
Ethical approval was received from the Jimma University Institute of Health (Ref No-JUIR/IRB /333/23) and the Oromia Health Bureau Ethical Review Board. All enrolled participants provided informed consent for the simultaneous collection of nasopharyngeal swabs and blood samples. Data collectors adhered to the standard COVID-19 prevention recommendations during the interview, including using face masks and physical distancing. The study confidentially maintained all information. Those who tested positive for PCR were transferred to treatment centres or followed at home-based and isolation centers according to the COVID-19 National Case Management and Isolation Protocol at the time25,26.All methods were performed in accordance with the relevant guidelines and regulations.