The SETA program investigated (1) the burden and severity of invasive Salmonella infections (prospective surveillance with active screening at selected healthcare facilities); (2) host immunity and acute and chronic carriage associated with natural S. Typhi/S. Paratyphi A, B, and C (hereafter S. Paratyphi)/iNTS infections over a 1-year follow-up period (prospective case-controlled and cohort study design method); (3) prevalences of S. Typhi/S. Paratyphi/NTS carriage in immediate household members of confirmed TF, PF, and iNTS disease cases (prospective active surveillance and cohort study); (4) public and private expenditures for treatment and productivity loss (cost of illness) associated with TF/PF/iNTS disease (cohort) (5) effects of invasive salmonellosis on the quality of life of patients and long-term socioeconomic study (cohort); and (6) validation of a new reverse-transcription polymerase chain reaction (rt-PCR) assay for the diagnosis of invasive Salmonella infections (Table 1). In addition to TF/PF/iNTS disease, other etiologies of bacteremia, Plasmodium infections, and viral hepatitides were also sought, where feasible, through testing at the study healthcare facilities and laboratories.

Building on the Typhoid Fever Surveillance in Africa Program (TSAP) network [6], the SETA program utilized and expanded on previously established fever surveillance infrastructure in sub-Saharan Africa. Six countries were selected exhibiting high disease endemicity (Burkina Faso, Ghana, Madagascar), further need for in-country investigations on TF (Ethiopia), and value of extending to additional study sites to countries with large population numbers (Democratic Republic of Congo and Nigeria); study sites in these countries have been integrated into the SETA program to enable harmonized multicountry surveillance and data comparability (Figure 1). Each SETA site has two distinct study areas (Table 2): a medically served area where surveillance and subsequent studies (case follow-up, enrollment and follow-up of neighborhood controls [NCs] and household contacts [HCs], healthcare utilization survey, and cost of illness and long-term socioeconomic studies) were performed, and a medically underserved area where the frequency of mortality due to suspected severe TF was assessed through postmortem questionnaires. The overall study period was between 2016 and 2019, but varies per site (Table 2).

Surveillance sentinel sites were established including primary, secondary, and tertiary healthcare facilities (Table 2). Selection of healthcare facilities was based on previous documentation of the occurrence of TF/PF/iNTS disease; case reports or series of patients with severe TF; capacity to conduct blood culture and epidemiological research; accessibility of local population to the healthcare facilities; and preference of healthcare facilities by the local population in the event of febrile illness. Each study site was characterized by having access to at least one tertiary hospital and at least one secondary or primary healthcare facility where patients were enrolled into the SETA program. Surveillance catchment areas were predefined based on the geographic or administrative coverage of healthcare services provided by respective healthcare facilities (hospital records or information materials). Healthcare-seeking behavior data collected during TSAP and other existing national census and/or demographic data, generated through the Health and Demographic Surveillance System (HDSS) under the International Network for the Demographic Evaluation of Populations and Their Health (INDEPTH), were also considered [21, 22].

Patients visiting SETA sentinel healthcare facilities were actively screened at selected entry points of each healthcare facility, including triages where applicable, outpatient wards, inpatient medical and surgical wards, and emergency wards. Laboratories of the selected healthcare facilities/collaborating institutes were involved for laboratory-based detection of any invasive Salmonella from clinical samples such as blood and specimens from perforation cases, which may be collected outside of the routine SETA patient screening procedure. The generic SETA study forms are available in the Supplementary Data.

Patient screening for eligibility was based on the following inclusion criteria (Table 3): history of fever (≥3 consecutive days in the last seven days prior to visiting the healthcare facility), or current objectively assessed fever (≥38°C tympanic/rectal and/or ≥37.5°C axillary), or clinically suspected TF or blood culture positive for invasive salmonellosis, or gastrointestinal perforations. Residency within the predefined catchment area was also an inclusion criterion, except for patients with clinically suspected severe TF complications such as gastrointestinal perforations who were eligible for enrollment regardless of catchment residence (special cases; Table 4). Blood and stool samples were collected for culture from eligible patients at enrollment to investigate causative pathogens, antimicrobial resistance, and acute carriage (shedding), and urine samples were examined for antimicrobial pretreatment status (Table 5). Oropharyngeal swab samples were collected additionally for a supplemental investigation of group A Streptococcus carriage.

Daily progress of clinical symptoms and treatment history were recorded for all enrolled patients with clinically suspected severe TF (Table 4) until discharge, regardless of blood culture results. Patients with severe TF complications (Table 6) requiring and/or undergoing surgery were closely monitored by hospital physicians, who gauged the willingness of the patients to participate in this research on severe typhoid by providing sufficient information and accurate explanation concerning the study activities and expectations. Patient enrollment only occurred after written informed consent was obtained. In case of emergencies whereby patients with gastrointestinal perforations required an immediate surgical intervention, an existing hospital consent process/form for surgery was followed for patient treatment without delay. Such patients were approached after stabilization for informed consent for study participation. Where feasible and based on patient or/and guardian consent, treating physicians collected blood and/or surgical samples such as gastrointestinal tissue, gallbladder, bile, or peritoneal fluid for culture and/or PCR. Histopathology and cytology of surgical samples were performed at the hospital, if feasible.

Assessment of death attributed to invasive Salmonella infection, particularly TF, was also analyzed through postmortem questionnaires (see Supplementary Data) performed in the medically underserved areas in selected study sites. While the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) [23] and the verbal autopsy standards manual [24] recommended by the WHO were referenced, the SETA postmortem questionnaires adapted symptoms outlined in the case definition, which included prolonged history of fever before deterioration and death, at least 1 abdominal-related symptom, deterioration consistent with the known TF complications such as acute abdomen or gastrointestinal bleeding or encephalopathic development with gradual deterioration to death, and history of treatment or hospitalization. This questionnaire was also administered to participants enrolled in the medically served areas in case of any fatality occurring during the study period.

Enrolled patients with blood culture–confirmed TF/PF/iNTS (cases) became part of the study cohort for clinical follow-up and health economics surveys during the period of 1 year (Table 5). At each clinical follow-up visit, blood and stool samples were collected from all cases and additional oropharyngeal swab samples were obtained from iNTS cases. Immunological, biochemical, and parasitological assessments were conducted on the collected blood samples (immunoglobulin G [IgG]/immunoglobulin M [IgM] antibodies, complete blood count, malaria, creatinine, bilirubin, alanine aminotransferase, aspartate aminotransferase, and additional peripheral blood mononuclear cell and T-cell stimulations where feasible in Ghana and Burkina Faso for a substudy on the cellular immune response associated with iNTS disease cases over the period of 1 year of follow-up). Stool culture was performed to detect bacterial shedding for assessment of acute and chronic carrier status up to the 1-year time point following the case confirmation [25, 26].

For investigations of host immune response and carriage status of cases in the context of the local population, asymptomatic NCs matched to cases by age (±5 years), sex, and location of residency (1:4 case-control ratio) were enrolled (Table 3) following written informed consent. Neighborhood is defined based on the location of residency of TF/PF/iNTS cases within the surveillance catchment areas. Residency locations of cases noted for the study follow-up visits were applied to identify NCs, whereby neighbors residing within the same administrative units such as villages or communes were approached for age- and sex-matched criteria for screening and enrollment of NCs. Blood and stool samples were collected at enrollment and at days 180 and 360. Oropharyngeal swab samples were collected additionally from iNTS disease case-matched controls. Immunological assessments were conducted on the collected blood samples, and stools were cultured for Salmonella.

Upon identification of blood culture–positive TF/PF/iNTS cases, two immediate HCs were approached for study participation to investigate for carriage of S. Typhi, S. Paratyphi, and NTS. Household contacts are individuals residing in the same house as the case, sharing the common living condition and environment. Priority for enrollment was given to individuals who prepare food for the case, who is the closest in age to the case, and who spends the most time with the case (Table 3). Frequency and duration of Salmonella carriage was estimated through confirmation of bacterial shedding in stool over the period of 1 year. Blood samples were also collected from immediate HCs for the immunological characterization of Salmonella carriers (Table 4).

All samples collected and examined from study participants (cases/NCs/HCs) are described in Supplementary Table 1 (summary of sample collection for participants in SETA).

A healthcare utilization survey [27] was conducted in the surveillance catchment area to estimate the proportion of the population who seek healthcare at the SETA surveillance healthcare facilities in the event of any fever or symptoms consistent with severe enteric fever. This was used to derive age-specific adjustment factors to adjust crude incidence rates of TF, PF, and iNTS disease in the catchment population [21]. In addition, socioeconomic and water, sanitation, and hygiene (WASH)–related data collected through the survey allow further observational analyses on the surveillance catchment population and settings.

All eligible participants (or their responsible parents/guardians) were approached for voluntary written informed consent. The informed consent process included an explanation by study staff about the study purpose, expectations from participants, duration of participation, risks and benefits for participation, confidentiality, right to decline or withdraw from the study, and contact information of study investigators. Participants were asked to sign a statement of consent if they agreed to join the study. If the participant was an infant or child, his or her parent/guardian was asked to sign or thumbprint the statement of consent. If the participant was an adolescent, both the participant and the parent/guardian were asked to sign or thumbprint the statement of consent. If the participant was illiterate, an independent literate witness (where possible, this person should be selected by the participant and had no connection to the study team) was asked to sign or thumbprint the statement of consent. Study participants were allowed to withdraw from the study at any time without loss of clinical services or penalties of any kind. The International Vaccine Institute’s (IVI) Institutional Review Board and site-specific ethical review boards reviewed the study annually to ensure continued compliance with the ethical principles and guidelines based on the WHO (2009) [28], the Council for International Organizations of Medical Sciences (2016) [29] and the Declaration of Helsinki (World Medical Association Declaration of Helsinki, 2013).

Data from each participant required complete and adequate source documentation (hospital or medical records, laboratory reports, test results) unless the data recorded directly on our study forms were considered the source data. The surveillance study forms included the informed consent and enrollment form for all participants, case report forms for patients, follow-up forms for Salmonella-confirmed patients or/and special cases and their NCs and HCs, postmortem questionnaires, and laboratory forms (Supplementary Data). Electronic data collection was performed using the SETA Collect software (Android 5.0.1; API 23) developed by the IVI for sites with stable internet access. In parallel, paper-based data collection was available at all sites. Data collected through SETA Collect were periodically exported to the IVI server via a web-based paperless data management system (PDMS). Paper-based data were entered into the computerized data management system (CDMS) using a dual entry process, which were transferred to the IVI monthly. This system automatically backed up data at systematic intervals onto local hard disks and external media and provided an audit trail. In the event of data discrepancy or missing values, study staff referred to the original source documents or contacted respective participants to clarify as needed. All tablets and PDMS/CDMS databases were password protected. All paper-based study forms were stored in cabinets with restricted access by authorized study staff. Both systems had algorithms for checking missing data points, range, and logical errors.

SETA provides incidence rates and frequency proportions of TF. The incidence estimation is based on the formula used in TSAP [6, 30], using adjustments such as the proportion of healthcare-seeking behavior in case of fever and severe febrile illnesses in the catchment population (denominator) and recruitment rate of study-eligible patients in the screened and enrolled patients (numerator) (Table 7). Equally, adjusted incidences, as well as frequency proportions of patients with severe TF (and deaths due to confirmed or/and suspected severe TF) out of the total number of TF cases are assessed. Final adjustment factors or multipliers to be applied in the adjusted incidence estimations are harmonized with the Severe Typhoid in Tanzania (STT) project [31] and the Surveillance for Enteric Fever in Asia Project (SEAP) [30, 32, 33] sites to ensure data comparability. A descriptive analysis of frequency and prevalence of deaths due to suspected severe TF from the selected medically underserved areas will be presented. The clinical history of symptoms associated with severe TF such as fever, abdominal pain, seizure, and other symptoms (Table 6) and the cause of death recorded, if known, are being made available. Due to the absence of autopsies, any analysis on death attributable to suspected severe TF complications will remain descriptive and limited.

Analyses on severe TF will also include investigations on host risk factors that may be associated with the severity of disease (eg, age group, sex, and comorbidities) and site-specific characteristics. Various statistical methodologies will be used including the Spearman correlation coefficient and/or stratified logistic regressions for univariate and multivariable analyses to measure the odds ratios. Data from the longitudinal cohort study of the TF/PF/iNTS patients and their respective NCs and HCs will be analyzed by comparing the level of antibodies (IgG/IgM) of cases and their shedding of Salmonella species during an acute phase and up to a 1-year time point, with that of NCs and HCs (background antibody level and bacterial shedding in the respective households and communities). Salmonella strains detected in stools of cases, HCs, and NCs will be analyzed to investigate any patterns potentially related to transmissions in respective study sites, for which the population-based healthcare utilization survey data will be cross-compared for various environmental risk factors such as the household WASH conditions and domestic animals. All bacterial isolates yielded through SETA will be sequenced for further molecular epidemiological analyses such as the emergence and spread of AMR/MDR strains as performed in TSAP [34]. Further analyses will be also performed including, but not limited to, the preuse of antimicrobials and/or antimalarials, AMR/MDR of detected bacterial pathogens, geospatial analyses of healthcare-seeking behavior and cases, socioeconomic and WASH risk factors, cost of illness, and long-term socioeconomic impact to patients.