Estimating the health effects of COVID-19-related immunisation disruptions in 112 countries during 2020-30: a modelling study

Abstract

Background: There have been declines in global immunisation coverage due to the COVID-19 pandemic. Recovery has begun but is geographically variable. This disruption has led to under-immunised cohorts and interrupted progress in reducing vaccine-preventable disease burden. There have, so far, been few studies of the effects of coverage disruption on vaccine effects. We aimed to quantify the effects of vaccine-coverage disruption on routine and campaign immunisation services, identify cohorts and regions that could particularly benefit from catch-up activities, and establish if losses in effect could be recovered.

Methods: For this modelling study, we used modelling groups from the Vaccine Impact Modelling Consortium from 112 low-income and middle-income countries to estimate vaccine effect for 14 pathogens. One set of modelling estimates used vaccine-coverage data from 1937 to 2021 for a subset of vaccine-preventable, outbreak-prone or priority diseases (ie, measles, rubella, hepatitis B, human papillomavirus [HPV], meningitis A, and yellow fever) to examine mitigation measures, hereafter referred to as recovery runs. The second set of estimates were conducted with vaccine-coverage data from 1937 to 2020, used to calculate effect ratios (ie, the burden averted per dose) for all 14 included vaccines and diseases, hereafter referred to as full runs. Both runs were modelled from Jan 1, 2000, to Dec 31, 2100. Countries were included if they were in the Gavi, the Vaccine Alliance portfolio; had notable burden; or had notable strategic vaccination activities. These countries represented the majority of global vaccine-preventable disease burden. Vaccine coverage was informed by historical estimates from WHO-UNICEF Estimates of National Immunization Coverage and the immunisation repository of WHO for data up to and including 2021. From 2022 onwards, we estimated coverage on the basis of guidance about campaign frequency, non-linear assumptions about the recovery of routine immunisation to pre-disruption magnitude, and 2030 endpoints informed by the WHO Immunization Agenda 2030 aims and expert consultation. We examined three main scenarios: no disruption, baseline recovery, and baseline recovery and catch-up.

Findings: We estimated that disruption to measles, rubella, HPV, hepatitis B, meningitis A, and yellow fever vaccination could lead to 49 119 additional deaths (95% credible interval [CrI] 17 248-134 941) during calendar years 2020-30, largely due to measles. For years of vaccination 2020-30 for all 14 pathogens, disruption could lead to a 2·66% (95% CrI 2·52-2·81) reduction in long-term effect from 37 378 194 deaths averted (34 450 249-40 241 202) to 36 410 559 deaths averted (33 515 397-39 241 799). We estimated that catch-up activities could avert 78·9% (40·4-151·4) of excess deaths between calendar years 2023 and 2030 (ie, 18 900 [7037-60 223] of 25 356 [9859-75 073]).

Interpretation: Our results highlight the importance of the timing of catch-up activities, considering estimated burden to improve vaccine coverage in affected cohorts. We estimated that mitigation measures for measles and yellow fever were particularly effective at reducing excess burden in the short term. Additionally, the high long-term effect of HPV vaccine as an important cervical-cancer prevention tool warrants continued immunisation efforts after disruption.

Funding: The Vaccine Impact Modelling Consortium, funded by Gavi, the Vaccine Alliance and the Bill & Melinda Gates Foundation.

Translations: For the Arabic, Chinese, French, Portguese and Spanish translations of the abstract see Supplementary Materials section.

Figure 1

Summary of recovery vaccine-coverage scenarios NA=not available. WUENIC=WHO–UNICEF Estimates of National Immunization Coverage.

Figure 2

Mean estimated additional deaths due to vaccine-coverage disruption by region (years of vaccination 2020–30) Calculated by comparing the baseline recovery and no-disruption scenarios (appendix 6 pp 2–8). HIB=Haemophilus influenzae type b. HPV=human papillomavirus. PCV=pneumococcal conjugate vaccine. VIMC=Vaccine Impact Modelling Consortium.

Figure 3

Excess deaths due to vaccine-coverage disruption and proportion of those deaths mitigated by catch-up activities (calendar years 2020–30) HPV=human papillomavirus.

Figure 4

Additional deaths due to vaccine-coverage disruption and deaths averted by catch-up activities (calendar years 2020–30) (A) Hepatitis B. (B) Measles. (C) Yellow fever. 50%, 85%, and 95% CIs are shown by ribbons; median estimates are shown by solid lines. Diseases for which the median excess burden did not exceed 50 in any one year are not shown. Scales are variable between diseases.