D.4 In summary

For a vaccine with waning immunity, average VE across the population^iv will peak before vaccine coverage peaks – assuming the uptake of vaccination slows down toward the end of the rollout (as would normally be expected)

Exactly when peak immunity occurs will vary by the level of protection being considered (from protection against any infection through to protection against death) if either the initial VE, or the amount of waning, varies by severity.

For a future pandemic, therefore, evidence on when peak immunity is reached will be a key consideration in when to end or ease PHSMs.

There is no ‘magic’ answer as to when to lift PHSMs, but the example provided here may provide additional evidence to assist decision-making alongside a range of other criteria. For example, if a decision-maker in the future was aiming to lift restrictions in accordance with ‘peak immunity’, anticipating it might take several weeks for any uptick in infection to occur, it would make sense to aim to make significant relaxations of PHSMs a few weeks before 95 percent of peak immunity was achieved. This would mean peak immunity could occur at the same time as any resurgence in infection is happening. For the example of South Auckland in 2021 used in this appendix, peak immunity would have occurred in late September (for non-Māori, non-Pacific people) to early October (for Māori and Pacific peoples), from the perspective of protection against serious illness or hospitalisation, and early October (for non-Māori, non-Pacific people) to mid-November (for Māori) for younger adult protection against any Delta infection.

This appendix takes data on vaccine coverage one step further, combining this with evidence on the timing of vaccine waning to consider what this means for population immunity. As stated at the outset of this appendix, the actual impact of easing PHSMs on population infection and disease rates depends upon the interplay of population immunity against disease transmission and the protection against serious illness in vulnerable people (which in COVID-19 was the elderly, immunocompromised and those with co-morbidities).

Therefore, we strongly encourage full epidemiological modelling to be undertaken (with waning included) in such a circumstance in any future pandemic. The ‘average population immunity’ can be generated in real-time and forecast, both in advance of fuller simulation modelling outputs and to assist understanding such simulation modelling once it has been conducted.