I think the underestimation was in long covid, not infections