Following our publication, we have engaged further methods of mixed-method cluster analysis and wish to present these to the reader. Analyses of the raw data (as in Fig. 2 of the paper) with baseline as a fixed factor reveal the following mean differences (95% CI) for H1N1 A-strain, 263.6 (−1.62 to 525.59) p = .05, H3N2 A-strain, 3.35 (−99.10 to 92.41) p = .95, and B-strain, 9.39 (−20.23 to 1.44) p = .09. Further, using log transformed data, the analogous statistics are: log mean difference (95% CI) for H1N1 A-strain, 0.53 (−1.00 to −0.07) p = .04, H3N2 A-strain, 0.20 (−0.08 to 0.48) p = .16, and B-strain, 0.23 (−0.49 to 0.03) p = .08. These reanalyses yield the same message as the original paper, but the effect for the B-strain now becomes non-significant/a trend.
The authors would like to apologise for any inconvenience caused.