Table 1.

Tiered options to estimate RSV disease burden in the second phase of the WHO RSV surveillance

Tier	Burden estimate	Data source/variables required (cumulative by month)	Adjustment/correction factor	Preconditions to avoid bias	Caveats/limitations
Tier 0	None	Log of RSV positive Log of patients tested	None	None	‐Can use ex‐SARI or SARI ‐Weekly aggregation ‐All‐year‐round surveillance
Tier 1.1	Proportion of respiratory or pneumonia hospital admissions due to RSV (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of admissions by respiratory or pneumonia diagnosis (for non‐enrolment)	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment	‐Systematic sampling of enrolment days in week ‐Systematic sampling of patients	‐Adjustment factor for non‐enrolment estimated during season may overestimate burden during off‐season period ‐Assumes that % positivity of RSV to be the same in those enrolled and those not enrolled ‐Rrelationship between no. of extended‐SARI cases and no. of resp. or pneumonia cases may vary by season ‐Burden estimate biased if sampling strategy is non‐random
Tier 1.2	Proportion of all‐cause hospital admissions due to RSV (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of admissions (all‐cause) diagnosis (for non‐enrolment)	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment	‐Systematic sampling of weekdays ‐Systematic sampling of patients	‐Adjustment factor for non‐enrolment estimated during season may overestimate burden during off‐season period ‐Assumes that % positivity of RSV to be the same in those enrolled and those not enrolled ‐Burden estimate biased if sampling strategy is non‐random
Tier 2.1	RSV hospitalization rate per 100,000 pop. (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of admissions by resp. or all‐cause diagnosis (for non‐enrolment) Catchment pop. Log of patients with resp. or all‐cause illness from catchment pop. that are admitted in non‐sentinel hospitals	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment ‐Adjust for patients with resp. illness from catchment pop. that seek care from other hospitals (Healthcare Utilization Survey data)	‐Systematic sampling of weekdays ‐Systematic sampling of patients ‐Secondary‐level hospital with defined catchment pop. ‐Healthcare admissions survey or healthcare utilization survey	‐Based on WHO influenza disease burden estimation method ‐Adjustment factor derived during season may overestimate burden during off‐season period ‐Burden estimate biased if sampling strategy is non‐random ‐HUS/HAS data required
Tier 2.2	Proportion of all‐cause ICU admissions due to RSV (specify age‐band (0‐1y, 0‐2y)	Log of ICU patients that are RSV positive Log of ICU patients tested Log of ICU patients screened (for enrolment) Log of ICU patients for all‐cause diagnosis (for non‐enrolment)	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment	‐Systematic sampling of weekdays ‐Systematic sampling of patients	‐Adjustment factor for non‐enrolment estimated during season may overestimate burden during off‐season period ‐Assumes that % positivity of RSV to be the same in those tested and those not tested ‐Assumes no significant bias in selection of patients for testing ‐Burden estimate biased if sampling strategy is non‐random
Tier 2.3	Case fatality ratio (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of patients by resp. or all‐cause diagnosis (for non‐enrolment) Log of RSV deaths	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment	‐Systematic sampling of weekdays ‐Systematic sampling of patients	‐Adjustment factor derived during season may overestimate burden during off‐season period ‐Assumes that % positivity of RSV to be the same in those tested and those not tested ‐Assumes no significant bias in selection of patients for testing ‐Burden estimate biased if sampling strategy is non‐random ‐Need to follow up RSV‐positive cases till discharge to determine death
Tier 2.4	Proportion of respiratory or pneumonia or all‐cause hospital deaths due to RSV (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of patients by resp. or pneumonia or all‐cause diagnosis (for non‐enrolment) Log of RSV hospital deaths Log of resp. or pneumonia or all‐cause hospital deaths	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment	‐Systematic sampling of weekdays ‐Systematic sampling of patients	‐Adjustment factor derived during season may overestimate burden during off‐season period ‐Assumes that % positivity of RSV to be the same in those tested and those not tested ‐Relationship between no. of ex‐SARI cases and no. of resp. or pneumonia cases may vary by season ‐Burden estimate biased if sampling strategy is non‐random ‐Need to follow up RSV‐positive cases till discharge to determine death
Tier 3.1	National estimate of RSV hospitalization rate per 100,000 pop. (specify age‐band (0‐1y, 0‐2y)	Log of RSV positive Log of patients tested Log of patients screened (for enrolment) Log of admissions by diagnosis (for non‐enrolment) Census data: Mid‐year pop. by specified age bands, by administrative division serving hospital Adjusted for pop. increase Adjusted for the years of surveillance DHS data: To adjust admin division estimates to pop. estimates Pneumonia or influenza rates or prevalence of risk factors (HIV, malnutrition, crowding, prematurity etc) by region HUS data (if available): To adjust for non‐medically attended resp. illness	‐Adjust for non‐enrolment ‐Adjust for weekends or days of non‐enrolment ‐Adjust for referrals from outside catchment population ‐Adjust for non‐medically attended illness (optional)	‐Systematic sampling of weekdays ‐Systematic sampling of patients ‐Secondary‐level hospital with defined catchment pop.	‐Based on method described by Murray 2015, Theo 2017 ‐Census data required ‐DHS data on prevalence of certain morbidities required ‐HUS data optionally required if adjustment for non‐medically attended illness ‐Adjustment factor derived during season may overestimate burden during off‐season period