Table 1 Scenarios for evidence updates of key determinants of measles vaccination impact. Assumptions for four key determinants used in each scenario and their data sources are summarised. Abbreviations: CFR case-fatality risk, SIA supplementary immunisation activity

Base

With data sources used previously [7, 16]:

• CFRs: country-specific, time-invariant estimates for under 5 years old from Wolfson [21] and halving such estimates for older than 5 years old.

• Contact patterns: physical contact matrix from POLYMOD Great Britain [25].

• Age-dependent first-dose vaccine efficacy: step function which assumes 85% for under one year old and 95% for over 1 year old [19].

• SIA coverage in zero-dose population: equal coverage in population with or without previous vaccination (random distribution of SIA doses).

(A) CFR, Portnoy’s method

‘Base’ scenario, with updated CFRs: country-specific, time-varying, incidence-related estimates for under and over 5 years old from Portnoy et al. [9]

(B) Contact patterns, synthetic matrices

‘Base’ scenario, with update for contact patterns: country-specific synthetic matrices [10, 11]

(B’) Contact patterns, proportional mixing

‘Base’ scenario, with contact probabilities proportional to the age distribution of population

(B”) Contact pattern, uniform mixing

‘Base’ scenario, with uniform contact probabilities across age groups

(C) Age-dependent first-dose vaccine efficacy, linear trend

‘Base’ scenario, with updated first-dose vaccine efficacy: efficacy as a linear function of age with an increase of 1.49% per month of age, from 68% at birth to 98% at highest [12]

(D) SIA coverage in zero-dose population, dependency on previous vaccination

‘Base’ scenario, with update for SIA coverage in zero-dose population: informed by a weighted logistic function fitted to all available surveys [13]

(D’) SIA coverage in zero-dose population, dependency on previous vaccination, excluding the largest survey

‘Base’ scenario, with update for SIA coverage in zero-dose population: informed by a weighted logistic function fitted to available surveys except for the one with the largest sample size (Indonesia, 2002) [13]

(D”) SIA coverage in zero-dose population, 7.7% never reached

‘Base’ scenario, with random delivery of SIA doses except for an isolated 7.7% of the target population that are assumed to never receive any measles vaccine dose [26]