Results from interrupted time-series analyses estimating the association of the Vooma voucher programme with changes in the number of first vaccine doses delivered per day per 10 000 adults ≥60 years in South Africa
| Unadjusted daily models | Adjusted daily models | |||
| Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | |
| National model | ||||
| Vooma voucher programme | 0.89 (0.52 to 1.27) | p<0.001 | 0.65 (0.33 to 0.96) | p<0.001 |
| Vooma vaccination weekend | – | – | 3.28 (1.98 to 4.58) | p<0.001 |
| Daily active facilities per 100 000 population | – | – | 0.55 (−0.60 to 1.69) | p=0.340 |
| Provincial model | ||||
| Vooma voucher programme | 0.80 (0.43 to 1.18) | p<0.001 | 0.53 (0.17 to 0.89) | p=0.004 |
| Vooma vaccination weekend | – | – | 2.10 (1.04 to 3.16) | p<0.001 |
| Daily active facilities per 100 000 population | – | – | 0.86 (0.43 to 1.29) | p<0.001 |
| KwaZulu-Natal and Gauteng | −4.15 (−7.45 to to 1.58) | p=0.003 | −3.49 (−6.94 to 0.04) | p=0.047 |
| Vooma voucher in KwaZulu- Natal and Gauteng | 1.96 (1.01 to 2.91) | p<0.001 | 1.39 (0.34 to 2.45) | p=0.010 |
Models also included day of the week and week fixed effects. In province-level models, we include an indicator for KwaZulu-Natal and Gauteng combined, compared with the rest of the country.