Estimating the incidence of enteric fever in children in India: a multi-site, active fever surveillance of pediatric cohorts

Active, community-based surveillance will estimate the true incidence of typhoid fever and provide a better understanding of transmission of typhoid in the population than available data from hospital based extrapolations.

We will recruit 6000 eligible children from each of the four geographically distinct study sites across India. Children are eligible to participate if they are between 6 months and 13 years + 364 days of age, likely to be in the study area for the duration of the study and are willing to receive investigations in the event of febrile illness. All eligible children living in the designated study area will be offered participation and the enrolment will continue till 6000 children have been recruited into each site’s cohort. If multiple eligible children are present within a household, all of them will be eligible to participate. Written informed consent will be obtained from the primary caregiver in addition to assent from age-eligible children.

Active weekly fever surveillance will be performed by trained study personnel either through home visits or by telephonic contacts through the course of study participation. Each participant will be followed for 24 months or until the child attains 15 years of age, whichever is earlier. The study team will ensure at least one face to face contact with the child and the primary care giver every 4 weeks. During these weekly contacts, details of any illness, visits to a health facility and treatment received since the previous visit will be recorded for each participant within the household. If contact with the parents/primary caregiver is not established for two consecutive weeks, then a supervisory visit will be made to confirm that the child is unavailable for surveillance. Parents are encouraged to report any fever episode to the study team at the earliest in addition to reporting them at weekly visits by study personnel.

In the event of fever, a daily follow up will be established until there are three consecutive fever free days. A digital thermometer and a fever diary card will be provided, and primary caregivers will be trained to use the thermometer and document the temperature three times a day. The study team will also document the temperature of the child in the fever diary card during home visits. At each daily contact during the fever follow up, the study team will document information pertaining to the previous calendar day including the highest documented temperature. If the child has any serious illness the parents will be encouraged to seek healthcare from their preferred provider. The family will be informed the need to avoid antibiotics and over the counter medicines that are not prescribed by their physician.

Any reported fever for 3 or more consecutive days will be designated as ‘Suspected Typhoid Fever’ and the child will be encouraged to visit the study clinic. At the study clinic the child will be evaluated by a study physician on the fourth day of fever and if there is history of fever in the preceding 12 h a blood culture will be requested. If the child has been afebrile for over 12 h, unless clinically indicated the blood culture will be deferred. Any subsequent fever within the next 24 h will warrant a blood culture. Prior antimicrobial therapy will not be a contraindication for blood culture. If a child has received a blood culture in the previous 14 days, blood culture will be withheld unless it is requested for by the clinician treating the participant.

A dedicated laboratory team will be present at each site for collection of blood samples and subsequent processing using automated blood culture systems using standard operating procedures across sites. For children between 6 months and 1 year, 3 mL will be inoculated into a paediatric blood culture bottle. For children above 1 year of age, 5 mL of blood will be inoculated into a paediatric blood culture bottle. Additional blood may be drawn for investigations requested by the treating physician. The total volume of blood drawn must not exceed 2 mL per kg body weight. Blood samples will be transferred to the site laboratory facility within 4 h of sample collection at ambient temperature. Where blood cultures do not provide a definite diagnosis and fever persists for 7 or more days, additional serology based tests may be performed as clinically indicated but these will not contribute to the estimates of typhoid incidence. S. Typhi isolated from other normally sterile sites will also contribute to confirmed typhoid fever estimates. All typhoid isolates will undergo antimicrobial susceptibility testing and will be archived for future genomic characterisation. All other significant culture isolates including S. Paratyphi will be documented.

The incidence of acute typhoid fever in children aged 6 months to 15 years in India was previously estimated to be between 6 and 12 per 1000 child years (CY) with the highest rates in 1–4 year olds [4, 10]. If community-level incidence has declined as hospital-based estimate suggest, the current incidence in this age-group would be between 1 and 5 per 1000 CY. Recruiting, at each site, a cohort of 5600 children between 6 months and 13 years + 364 days will provide 8400 child years of observation over 18 months that would be distributed approximately as 2400 child years in the 6-month to 4 years, and 3000 child years in the 5–10 and 11–14 age groups in an open cohort and assuming a uniform death rate. Accumulating 8400 child-years will exclude an incidence rate of 2/1000 with 95% confidence if there are 26 or more cases (IR > 3/1000) in the cohort. In turn, this sample size will be adequate to exclude at 95% confidence limits, a rate of 2/1000 in each of the 6 months-4 years, 5–10 and 11–14 years age groups in which the observed incidence is 4/1000 or greater using a Fishers exact test based on the formula (Armitage,1971; Snedecor & Cochran,1965) as described in Epidemiologic Analysis with a Programmable Calculator, 1979.

Incidence rates will be calculated as the number of discrete events of illness (for confirmed typhoid fever and acute febrile illness) over the total person-time under observation and at risk. Recurrent events will be allowed to include multiple episodes of illness over the study period. The person-time for each child will begin from the date of enrolment to the end of follow-up or date of loss to follow-up. Time with illness and when children not under surveillance will be excluded from the calculation of person-time. Rates will be expressed as per 1000 child-years. The two-year study period will be divided into three age intervals (6 m – 4 years, 5-9 year and 10-14 years) determined by the chronological age. The incidence rates for each age interval will be calculated based on the number of children at risk in the beginning of each interval, the number who had the event and the number who were lost to follow-up. The age-specific rates will be calculated as the number of new cases of illness in the specific age interval divided by total number of person-years of follow-up contributed by all children at risk in this interval. Children will be allowed to move higher up in the age intervals. Because recurrent events of illness will be considered, we will account for correlation at the level of individual child using mixed effects poisson regression model. All statistical analysis will be performed using Stata 14.0 (Stata Corp).

Annual assessment of access to safe water, sanitation and hygiene, population density, demography and socioeconomic status will be performed for risk characterisation of the population.

All children with culture-confirmed typhoid will be invited to participate in a longitudinal follow-up to estimate costs and outcomes of the illness episode. They will be contacted on the 14th, 28th and 90th days since the onset of fever. Trained study personnel will administer questionnaires to estimate costs at the end of the illness episode, and at 14 days and at 28 days from fever onset. The clinical outcomes of the illness episode will be documented at 28 and 90 days. A single blood sample of 3 mL will be collected at the 28th day visit to assess immune responses to the infection.