Table 3.
Effect of overall vaccine misclassification rate (p1 + p2) on vaccine effectiveness estimation by test-negative design specimen-based analysis
| Cases vaccinated by Vi-TT | Controls vaccinated by Vi-TT | VE against typhoid*(95% CI) | ||
|---|---|---|---|---|
| 0% vaccine misclassification—gold standard | 17/101 (16·8%) | 4092/8060 (50·8%) | 80·4% (66·9 to 88·4) | |
| 5% vaccine misclassification | ||||
| Misclassifying vaccinated as unvaccinated, both groups† | 12/101 (11·9%) | 3689/8060 (45·8%) | 84·0% (70·8 to 91·3) | |
| Differential misclassification, lowest possible VE‡ | 22/101 (21·8%) | 3689/8060 (45·8%) | 67·0% (47·0 to 79·5) | |
| Differential misclassification, highest possible VE§ | 12/101 (11·9%) | 4495/8060 (55·8%) | 89·3% (80·4 to 94·2) | |
| 10% vaccine misclassification | ||||
| Misclassifying vaccinated as unvaccinated, both groups† | 7/101 (6·9%) | 3286/8060 (40·8%) | 89·2% (76·7 to 95·0) | |
| Differential misclassification, lowest possible VE‡ | 27/101 (26·7%) | 3286/8060 (40·8%) | 47·0% (17·5 to 66·0) | |
| Differential misclassification, highest possible VE§ | 7/101 (6·9%) | 4898/8060 (60·8%) | 95·2% (89·6 to 97·8) | |
| 15% vaccine misclassification | ||||
| Misclassifying vaccinated as unvaccinated, both groups† | 2/101 (2·0%) | 2883/8060 (35·8%) | 96·4% (85·3 to 99·1) | |
| Differential misclassification, lowest possible VE‡ | 32/101 (31·7%) | 2883/8060 (35·8%) | 16·9% (−27·0 to 45·4) | |
| Differential misclassification, highest possible VE§ | 2/101 (2·0%) | 5301/8060 (65·8%) | 98·9% (95·7 to 99·7) | |
| 20% vaccine misclassification | ||||
| Misclassifying vaccinated as unvaccinated, both groups† | 0 | 2480/8060 (30·8%) | 100% (91·4 to 100) | |
| Differential misclassification, lowest possible VE‡ | 37/101 (36·6%) | 2480/8060 (30·8%) | −30·1% (−95·5 to 13·5) | |
| Differential misclassification, highest possible VE§ | 0 | 5704/8060 (70·8%) | 100% (98·4 to 100%) | |
Data are n/N (%), unless otherwise indicated. See the appendix (pp 3–4) for further details. p1=probability of misclassifying vaccinated as unvaccinated. p2=probability of misclassifying unvaccinated as vaccinated. VE=vaccine effectiveness. Vi-TT=typhoid Vi polysaccharide-tetanus toxoid-conjugate vaccine.
VE=(1 – odds ratio) × 100%.
Only misclassifying vaccinated as unvaccinated for both cases and controls due to the loss of vaccination cards, that is, p1 + p2=p1, hence p2=0 among both groups.
p1=0 among cases (misclassifying unvaccinated as vaccinated among cases) and p2=0 among controls (misclassifying vaccinated as unvaccinated among controls), resulting in the lowest possible VE.
p2=0 among cases (misclassifying vaccinated as unvaccinated among cases) and p1=0 among controls (misclassifying unvaccinated as vaccinated among controls), resulting in the highest possible VE.