The aim, design and setting of the study
The specific objectives of this project comprise three phases, namely:
- To determine cumulative incidence (i.e., seroprevalence) of SARS-CoV-2 infection in the Burkinabe population (phase 1: sero-epidemiological study);
- To identify the predictors of the occurrence of severe complications, and to identify the prognostic factors for mortality (phase 2: clinical epidemiology);
- To analyze the healthcare pathways and patient experiences of management of COVID-19 in healthcare centers dedicated to SARS-CoV-2 infected patients, and to explore the perceptions, acceptability, and application of preventive strategies among the population (phase 3: socio-anthropological study).
This study will be conducted in Ouagadougou and Bobo Dioulasso, the two main cities of Burkina Faso a landlocked country in West Africa.
This will be a multidisciplinary study using mixed methods (i.e., combining quantitative and qualitative methods), in three distinct parts, detailed in Figure 1. Each part of the study is presented separately.
Phase 1: Sero-epidemiological study
For the investigation of seroprevalence, the protocol recommended by the WHO for sero-epidemiological studies in the general population will be followed .
A longitudinal cohort will be constituted from among the general population in the cities of Ouagadougou and Bobo-Dioulasso (Burkina Faso), which are the hardest hit by the pandemic. Households and persons be randomly selected, with stratification by age group (<60 years, 60 years and over) and gender (male, female) and included in the study. Random selection of the households will be performed based on census units of the national statistics institute of Burkina Faso. A household is considered as at least persons living in the same dwelling, to consider so-called “celibatoriums”, which are grouped premises comprising several individual lodgings. In each household, all the members of the household will be invited to participate in the study, to collect data at days 0, 21, 42 and 63, to enable analysis by subgroup (age and gender).
Published statistics to date regarding COVID-19 in Burkina Faso indicate that there are very few cases in persons under 18 years of age.
The inclusion criteria for the study are as follows:
- Age 18 years or older.
- Resident in a census unit of the cities of Ouagadougou or Bobo-Dioulasso that is selected for participation in the study, based on data from the national statistics institute of Burkina Faso.
- Provide written informed consent.
- Persons aged 18 years or over from within the selected zones in Ouagadougou and Bobo-Dioulasso who refuse to participate in the study;
- Persons under legal guardianship or tutorship or other form of legal protection.
The primary endpoint will be the proportion of the population with specific antibodies against SARS-CoV-2.
Secondary endpoints will be as follows:
- The proportion of the population with specific antibodies against SARS-CoV-2 in each subgroup of age (<60 years, 60 years and over) and gender (males, females).
- The proportion of asymptomatic infections (presence of antibodies in patients with no clinical signs of SARS-CoV-2 infection) in the population, in each subgroup of age (<60 years, 60 years and over) and gender (males, females).
Sample size calculation
Using the sample size calculator available on Openepi  (http://www.openepi.com/SampleSize/SSPropor.htm) the minimum final sample size needed at the final round (fourth timepoint) should be 768 for each stratum. This number is obtained using the following parameters:
- Expected frequency: 10% (Usually set at 50% if unknown but most countries are reporting seroprevalence of 10% according to correspondence with WHO);
- Acceptable margin of error: 3% (This can vary depending on need);
- Confidence interval: 95% (This should be set at 95%);
- Design Effect: 2 (A design effect of 2 is a minimum as multiple individuals are being selected from the same household and in the same area);
For the purposes of this study, we also considered the following parameters:
- Only 30% of persons in the households contacted would accept to participate in the study, given that there is a need for a blood draw (to extract serum);
- We estimate that 30% of participants will be lost to follow-up or will withdraw their participation at each collection timepoint (especially if the epidemic wanes over the course of the study period).
- We plan to perform subgroup analysis (by age group and gender)
- The average size of households in Ouagadougou is 4.8 persons; 2.35 (48.97%) men aged <60 years; 2.29 (47.71%) women aged <60 years; and 0.151 (3.15%) persons aged >60 years.
According to these calculations, if 29705 persons are contacted:
- At day 0: 30% accept to participate in seroprevalence study =6717;
- At day 21: 70% attend follow-up = 4702;
- At day 42: 70% attend follow-up (3150) =5486;
- At day 63: 70% attend follow-up (2205) =3840.
In total, 20,745 serological tests need to be performed to achieve a total of 3840 anticipated participants at day 63, for a distribution across the subgroups of approximately 768 persons per subgroup (and on the assumption that the rate of attrition is not influenced by age or gender).
Data collection and management
Data collection in the general population will take place at four timepoints, each 21±3 days apart. Blood samples will be drawn by trained and qualified personnel. In each household, the members of the household who accepted to participate will be invited to sign the informed consent form, and complete a questionnaire (Appendix) with socio-demographic data, comorbidities, symptoms related to SARS-CoV-2 infection and exposure, namely:
- Socio-demographic characteristics: age, profession, marital status, level of education, socio-professional category;
- Prior medical history:
- History of cardiovascular and thrombo-embolic disease: arterial hypertension, rhythm disturbances, cardiomyopathy, deep venous thrombosis, pulmonary embolism;
- History of metabolic diseases : diabetes mellitus, kidney failure;
- History of pulmonary disease: tuberculosis, asthma, chronic obstructive pulmonary disease (COPD);
- Other relevant medical history.
- Lifestyle habits: tobacco and alcohol consumption;
- Clinical signs of COVID-19: fever (temperature ≥ 38°C), cough, dyspnea, etc.;
- Exposure to COVID-19: any contact with a confirmed case of COVID-19.
Persons who complete the questionnaire will be requested to give a blood sample to extract serum for the assessment of SARS-CoV-2 antibodies. Venous blood samples will be collected at the four different timepoints (Days 0, 21, 42, 63) to enable analysis in the laboratory. Blood will be collected by venous puncture performed by health workers within the participant’s home. In total, 4 ml of blood will be drawn in a 5 ml EDTA tube for each participant, at each timepoint.
The blood samples will be transported to the laboratory within 6 hours of the collection, along with the relevant follow-up sheet for the blood samples. For transport to the laboratory, blood samples will be triple packed in an isotherm container containing an icebox to preserve the cold chain.
All blood samples will be analyzed in the Laboratory of the MURAZ center in Bobo-Dioulasso. Laboratory procedures involving handling of blood samples will be performed in a class 2 biosafety cabinet. For blood samples collected in the city of Bobo-Dioulasso, samples will be sent to the laboratory daily. Samples from Ouagadougou will be sent to the laboratory of the national center for research and training on malaria, to be aliquoted and temporarily stored at -80°C before being transferred to the MURAZ center in Bobo-Dioulasso.
At the laboratory, plasma samples will be centrifugated and aliquoted within 12 hours of the blood collection. If the time to preparation (centrifugation and aliquoting) is likely to exceed 12 hours, it will be temporarily stored at +4°C for up to a maximum of 72 hours. To minimize the number of freeze-thaw cycles , two (2) aliquots of 1 ml plasma each will be prepared from each blood sample. One of these will be used for the serological tests by ELISA anti-SARS-CoV-2 for both IgM and IgG. For participants who consent to conservation of their blood samples, the second aliquot will be stored at -80°C in a secure biobank at the laboratory of the MURAZ center in Bobo-Dioulasso, in case needed for quality control, additional tests or future research.
Laboratory tests will use the WANTAI SARS-CoV-2 Ab ELISA, an Enzyme-Linked Immunosorbent Assay (ELISA) intended for qualitative detection of total antibodies (including IgG and IgM) to SARS-CoV-2 in human serum and Acid Citrate Dextrose (ACD) plasma. The WANTAI SARS-CoV-2 Ab ELISA is intended for use as an aid in identifying individuals with an adaptive immune response to SARS-CoV-2, indicating recent or prior infection. Before being used in this study, the assays will be validated for the Burkinabe context on a small series of characterized samples from Burkinabe patients (10 positive and 10 negative samples by PCR).
The results of the biological analyses and the questionnaire data will be recorded in an electronic case report form (CRF), which will be secured by personal access codes for each researcher. Data will be transferred to the data management center at the MURAZ center daily, after preliminary data checks by the supervisors in the field. Data will be cleaned and tested for coherence and completeness, and aberrant values will be identified. Queries generated in case of aberrant data will be sent to the research supervisors in the field to be resolved by the researchers before the end of data collection.
The database will be locked after the data has been cleaned and all queries have been resolved. The database will then be exported for statistical analysis.
Data statistical analysis plan
For the statistical analysis, this research project will adopt a dynamic approach and will constantly integrate latest knowledge emerging about COVID-19. A first analysis will be performed after the first collection on day 0, to enable important information that could be useful for the fight against COVID-19 to be relayed to the authorities in Burkina Faso.
The following analyses will be performed :
- Calculation of the cumulative incidence by age group: i.e., the proportion of individuals who are seropositive for SARS-CoV-2 infection according to age (<60 years, or 60 years and over).
- Calculation of the cumulative incidence by gender: i.e., the proportion of individuals who are seropositive for SARS-CoV-2 infection according to gender (men, women).
- Serological response to infection (if the serological tests enable us to evaluate this criterion): i.e., the course of serum levels of specific antibodies against SARS-CoV-2 (increase in titters across the different measurement timepoints).
- Describe and compare individual, socio-economic characteristics, presence of comorbidities and exposure according to serological status of the population.
Quantitative variables will be described as mean±standard deviation (SD) and median (range) and compared using the Student t or Mann-Whitney U tests as appropriate. Qualitative variables will be described as number and percentage and compared using the chi square or Fisher’s exact test, as appropriate.
Phase 2: Clinical epidemiology
The design of this clinical epidemiological phase of the study will comprise an ambispective analysis, involving first, a retrospective cohort of all patients with COVID-19 followed in one of the dedicated COVID-19 centers in the cities of Ouagadougou or Bobo-Dioulasso since the beginning of the epidemic; and second, prospective inclusion of all patients who present up until the end of the epidemic, or at the latest, 6 months after the start of the study. Retrospective data will be collected after study initiation, from the medical files of patients diagnosed since the start of the epidemic but before the study was initiated. These data will be used (if possible, depending on the number of cases collected) to construct a score to predict the occurrence of severe complications and death related to COVID-19, which could then be used to optimize patient management. The score will be validated in the prospective cohort to be included after study initiation.
The eligible population for this phase of the study will comprise all patients with a positive diagnosis of COVID-19 and followed in either of the two participating cities (Ouagadougou and Bobo-Dioulasso). In Burkina Faso, the epidemic is principally active in these two cities, the largest in the country, and the dedicated COVID-19 centers in these two cities manage almost all patients with a positive diagnosis since the start of the epidemic. The four main management sites are follows:
- In Ouagadougou: University hospital Yalgado, University hospital Bogodogo and University hospital Tengandogo;
- In Bobo Dioulasso: University hospital Sourou-Sanou.
- All patients followed for COVID-19 (whether they are or were hospitalized or not), in the reference management centers cited above, or in outpatient care, in Ouagadougou and Bobo-Dioulasso.
- Patients who refuse to provide informed consent for the study;
- Persons under legal guardianship or tutorship or other form of legal protection.
This phase of the study will have two primary endpoints, namely:
- The proportion of severe complications (defined as severe pneumonia, acute respiratory distress syndrome, sepsis or septic shock) in patients infected with SARS-CoV-2;
- The proportion of deaths in patients infected with SARS-CoV-2.
Secondary endpoints will be as follows:
- The proportion of severe complications according to age group and gender;
- The proportion of deaths according to age group and gender;
Sample size calculation
No sample size can be calculated for this phase of the study. Exhaustive inclusion of all cases of COVID-19 followed in the cities of Ouagadougou and Bobo-Dioulasso is planned. At least 1,000 cases are anticipated (as of 28 November 2020, Burkina Faso had a total of 2,886 confirmed cases of COVID-19, of whom 984 were women and 1902 were men). The statistical power of the analyses will be calculated a posteriori.
Data collection and management
Socio-demographic and clinical data (age, gender, comorbidities, vital signs, biological parameters) at the time of diagnosis of COVID-19 will be recorded by the clinical research assistants in the COVID-19 management centers from the patients’ medical files and recorded in the CRF for the study (paper format). The development of severe complications (type of complication, date, outcome) will also be recorded in the CRF. In case of death, the date of death will also be recorded. For patients who recover, the end-date for survival analysis will be the date of discharge from the COVID-19 management center.
Data will be recorded in a secure, electronic database using the in-house Ennov software from the Georges-François Leclerc Cancer Centre (Dijon, France). Data will be monitored before locking the database. Data will be analyzed at the Georges-François Leclerc Cancer Centre.
Socio-demographic and clinical characteristics (age, gender, comorbidities, vital signs, biological parameters) of all patients at the time of diagnosis will be described and compared. Quantitative data will be described as mean±SD or median (range). Qualitative variables will be described as number (percentage). Predictors of the occurrence of severe complications will be identified by logistic regression. The cohort will then be divided into two parts: the first part will be used to generate a score to predict severe complications (development set). The second part of the cohort will be used to validate the predictive score (validation set). The discriminatory capacity of the score will be assessed by the area under the receiver operating characteristic (ROC) curve (AUC), and the sensitivity, specificity, positive and negative predictive values of the score will be calculated. Prognostic factors of survival will be identified for the whole study population using Cox’s proportional hazards regression model. The outcome of interest will be death. Surviving patients will be censored at the date of discharge from the COVID-19 center. Imputation techniques and sensitivity analyses will be performed to account for missing data.
All analyses for phases 1 and 2 will be performed using STATA 14 (Copyright 1985-2015 StataCorp LLC; College Station, TX, USA) and R (www.r-project.org).
Phase 3: Socio-anthropological study
The third and final phase of the study aimed to analyze the healthcare pathway and the experiences of patients receiving care in dedicated centers for COVID-19 management in Burkina Faso, as well as the perceptions, acceptability, and application of preventive strategies in the population. An empirical study will be conducted in the COVID-19 management centers and in the community.
For phase 3, sampling will take account of the different characteristics of the study zones and the population. The study will be performed in the cities of Ouagadougou and Bobo-Dioulasso, and four types of respondents will be included, namely:
- Persons who have recovered from COVID-19 after having been hospitalized in a specialized COVID-19 center. Eligible persons will be identified with help of the healthcare services;
- Relatives and/or caregivers (spouse, children etc.) of persons who have recovered from COVID-19 after having been treated in a specialized COVID-19 center;
- Healthcare professionals directly involved in the management of patients admitted to specialized COVID-19 centers (nurses, laboratory technicians, physicians);
- All-comers from the general population without known COVID-19 infection (men, women, young or old) in the cities of Ouagadougou and Bobo-Dioulasso.
For the choice of the study population, certain diversification criteria will be respected, to maximize heterogeneity, independently of the statistical frequency (young/old, men/women, city center/suburbs, nurses/laboratory technicians/physicians, high vs low level of education, etc.). The balance of these criteria will be verified as the study advances, during the daily and weekly updates. Similarly, the choice of the areas where the study will be carried out will aim to maximize the diversity of the different areas within both cities. The four following types of city area will be included: residential (central and peripheral), and poorer neighborhoods (central and peripheral). With four zones for each city, a total of eight zones will be included, as per the Table below:
Distribution of the city zones for the socio-anthropological study
The choice of households will be at the discretion of the researcher in the field, according to the availability of persons to participate voluntarily in the study. The only restriction will be that the researchers must avoid including households who are already participating in the sero-epidemiological study, to avoid over-burdening, the household members with research participation, and to prevent any influence of each study phase on the other.
Persons aged 18 years or older living in the cities of Ouagadougou or Bobo-Dioulasso:
- who had received care for COVID-19 in one of the reference management centers, and who had recovered;
- or a relative or caregiver (spouse, children etc.) of such a person;
- or a healthcare professional directly involved in the management of patients with COVID-19 in one of the reference management centers (nurses, laboratory technicians, physicians).
Persons who refuse to provide informed consent for study participation.
The endpoint for this study will be the acceptability and application of preventive strategies in the population.
For the sample size for phase 3, the principle of data saturation will be retained, i.e. the researchers will cease recruitment when it becomes clear that additional interviews and observations yield no new information either at individual or at group level. However, for the purposes of estimation, we aim to recruit around 100 persons altogether from the two cities. Persons who have recovered from COVID-19 infection will be identified via medical records by the medical staff, who will contact them to provide information about the study. The researcher will then contact the recovered patients thus identified, and invite them to participate in the study, and obtain the contact details of any relative or caregiver who could also be invited to participate. Healthcare professionals will be contacted directly in the COVID-19 management centers. Information about the study will be provided and their consent will be obtained. For the general population, researchers will recruit volunteers directly in their homes.
Data collection and management
For this phase of the study, all persons who accept to participate (recovered patients, relatives/caregivers, medical staff) will be invited to participate in a semi-structured interview (supplements 2 to 4) to explore their experience of COVID-19, treatment, the disease itself, their professional activity. Interviews will also explore perceptions, acceptability, and application of COVID-19 prevention strategies in the population.
Interviews will be recorded and complemented by observations made by the interviewer during the interview. Data will be transcribed and analyzed with the help of QDA Miner software.