Evaluating the strategies to control SARS-CoV-2 Delta variant spread in New Caledonia, a zero-COVID country until September 2021

Abstract

Objectives: New Caledonia, a former zero-COVID country, was confronted with a SARS-CoV-2 Delta variant outbreak in September 2021. We evaluate the relative contribution of vaccination, lockdown, and timing of interventions on healthcare burden.

Methods: We developed an age-stratified mathematical model of SARS-CoV-2 transmission and vaccination calibrated for New Caledonia and evaluated three alternative scenarios.

Results: High virus transmission early on was estimated, with R₀ equal to 6.6 (95% confidence interval [6.4-6.7]). Lockdown reduced R₀ by 73% (95% confidence interval [70-76%]). Easing the lockdown increased transmission (39% reduction of the initial R₀); but we did not observe an epidemic rebound. This contrasts with the rebound in hospital admissions (+116% total hospital admissions) that would have been expected in the absence of an intensified vaccination campaign (76,220 people or 34% of the eligible population were first-dose vaccinated during 1 month of lockdown). A 15-day earlier lockdown would have led to a significant reduction in the magnitude of the epidemic (-53% total hospital admissions).

Conclusion: The success of the response against the Delta variant epidemic in New Caledonia was due to an effective lockdown that provided additional time for people to vaccinate. Earlier lockdown would have greatly mitigated the magnitude of the epidemic.

Keywords: COVID-19; Modelling; Non-pharmaceutical interventions; SARS-CoV-2; Vaccine; Zero-COVID.

Figure 1

Daily reported COVID-19 cases, vaccine coverage, and interventions timeline in New Caledonia from March 2021 to January 2022. Daily COVID-19 reported cases (pink crosses) and vaccine coverage in the eligible population (i.e., ≥12 years old) (green solid line).

Figure 2

Comparison of the model simulations to observations. Panel A shows simulated daily infections, panel B shows simulated daily hospitalizations, panel C shows daily ICU admissions, and panel D shows the projected ICU bed occupancy; ICU dynamics have been simulated using daily ICU admission probabilities given hospitalization. Observational data points are marked with crosses. Solid lines and shaded areas represent the mean simulation and the 2.5^th-97.5^th percentile range, respectively, from 1000 simulations using 1000 random samples from the posterior distribution of parameters estimated through Markov Chain Monte Carlo sampling. Dashed lines represent the maximum number of ICU beds under normal (orange) and COVID (red) regimes. ICU: intensive care unit.

Figure 3

Alternative scenarios compared to observations of hospital admission and ICU bed occupancy. Panel A shows simulated daily hospital admissions and panel B shows daily ICU bed occupancy (simulated using the average probability of ICU admission given hospitalization over the entire epidemic). Scenario A: purple, no lockdown is implemented; Scenario B: orange, no vaccination after lockdown; Scenario C: blue, 2 weeks earlier implementation of the lockdown and associated measures. The mean baseline simulation (no scenario) is represented by a dotted green line. Solid lines and shaded areas represent the mean simulation and the 2.5^th-97.5^th percentile range, respectively, from 1000 simulations using 1000 random samples from the posterior distribution of parameters estimated via Markov Chain Monte Carlo sampling. Dashed lines represent the maximum number of ICU beds under normal (orange) and COVID (red) regimes. ICU: intensive care unit.