key: cord-0797233-9mhcw3wy
authors: Chetty-Makkan, C. M.; Thirumurthy, H.; Bair, E. F.; Bokolo, S.; Day, C.; Wapenaar, K.; Werner, J.; Long, L.; Maughan-Brown, B.; Pascoe, S. J. S.; Miot, J.; Buttenheim, A.
title: A quasi-experimental cohort study evaluating a conditional economic incentive on first-dose COVID-19 vaccination rates among older adults in South Africa
date: 2022-05-07
journal: nan
DOI: 10.1101/2022.05.06.22274712
sha: 9e24a10dec5c73ddba75b9c0d70a54c2640a39f0
doc_id: 797233
cord_uid: 9mhcw3wy

Importance: COVID-19 vaccination coverage in South Africa remains low despite increased access to vaccines. On November 1, 2021, South Africa introduced the Vooma Voucher program, which provided a small guaranteed financial incentive, a Vooma Voucher redeemable at grocery stores, for COVID-19 vaccination among older adults, a population most vulnerable to serious illness, hospitalization, and death. However, the association of financial incentives with vaccination coverage remains unclear. Objective: To evaluate the association of the conditional economic incentive program with first-dose vaccination rates among older adults (aged [≥]60 years) in South Africa. Design: A quasi-experimental cohort study using daily data on first doses administered. We ran interrupted time series (ITS) models to evaluate the Vooma Voucher program, launched on November 1, 2021, at national and provincial levels. We used data between October 1, 2021 and November 27, 2021 in models estimated at the daily level. Setting and participants: The Vooma Voucher program was a nationwide vaccination incentive program implemented for adults aged [≥]60 years from November 1, 2021 to February 28, 2022. Intervention: Individuals who received their first vaccine dose received a text message to access a ~$7 (ZAR100) voucher that was redeemable at nationwide chain of grocery stores. Main outcome: Daily first COVID-19 vaccine doses administered per 10,000 individuals aged [≥]60 years. Results: The Vooma Voucher program was associated with a of 7.15-12.01% increase in daily first-dose vaccination in November 2021 compared to late October 2021. The incentive accounted for 6,476-10,874 additional first vaccine doses from November 1-27, 2021, or 8.31-13.95% of all doses administered to those aged [≥]60 years during that period. This result is robust to the inclusion of controls for the number of active vaccine delivery sites and for the nationwide Vooma weekend initiative (November 12-14), both of which also increased vaccinations through expanded access to vaccines and demand creation activities. Conclusions/Relevance: Financial incentives for COVID-19 vaccination led to a modest increase in first dose vaccinations among older adults in South Africa. In addition to financial incentives, expanded access to vaccines may also results in higher vaccination coverage.

3 Abstract (no more than 350 words) Importance: COVID-19 vaccination coverage in South Africa remains low despite increased access to vaccines. On November 1, 2021, South Africa introduced the Vooma Voucher program, which provided a small guaranteed financial incentive, a Vooma Voucher redeemable at grocery stores, for COVID-19 vaccination among older adults, a population most vulnerable to serious illness, hospitalization, and death. However, the association of financial incentives with vaccination coverage remains unclear.

Objective: To evaluate the association of the conditional economic incentive program with firstdose vaccination rates among older adults (aged ≥60 years) in South Africa.

Design: A quasi-experimental cohort study using daily data on first doses administered. We ran interrupted time series (ITS) models to evaluate the Vooma Voucher program, launched on November 1, 2021, at national and provincial levels. We used data between October 1, 2021 and November 27, 2021 in models estimated at the daily level.

Setting and participants: The Vooma Voucher program was a nationwide vaccination incentive program implemented for adults aged ≥60 years from November 1, 2021 to February 28, 2022.

Intervention: Individuals who received their first vaccine dose received a text message to access a ~$7 (ZAR100) voucher that was redeemable at nationwide chain of grocery stores. to the inclusion of controls for the number of active vaccine delivery sites and for the nationwide Vooma weekend initiative (November 12-14), both of which also increased vaccinations through expanded access to vaccines and demand creation activities.

Conclusions/Relevance: Financial incentives for COVID-19 vaccination led to a modest increase in first dose vaccinations among older adults in South Africa. In addition to financial incentives, expanded access to vaccines may also results in higher vaccination coverage. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint

As COVID-19 vaccines became available in 2021, many countries faced substantial demand constraints sooner than expected and struggled to achieve vaccination targets. In South Africa, COVID-19 vaccines were widely available for individuals aged ≥60 years by May 2021 (1, 2) .

Despite robust efforts to promote vaccines and ensure easy access, by October 2021 63% of older adults received at least one vaccine dose, well below the South African National Department of Health (NDoH) goals to protect vulnerable populations (2) .

As in other countries, there was substantial interest for financial incentives to increase vaccine uptake. Evidence on the effectiveness of incentives for COVID-19 vaccination comes almost exclusively from high-income countries and is mixed: some studies found that guaranteed rewards can increase vaccination (3, 4) , while others showed no effect of guaranteed or lottery incentives (5-7), or indicated that incentives may backfire by decreasing vaccination intentions (8). Low-and middle -income countries (LMICs), have faced a substantial burden of COVID-19, but there is lack of evidence on whether incentives for COVID-19 are effective in LMICs.

To increase vaccination among older adults in advance of an anticipated COVID-19 fourth wave, the NDoH Vooma Voucher program was launched on November 1, 2021 to reach the vulnerable and impoverished. The voucher program was promoted through multiple communication channels (eFigures) and continued until February 28, 2022, with some changes in eligibility and voucher amounts (eMethods). We assessed the short-term association of the Vooma Voucher incentive program on first-dose vaccination rates among those aged ≥60 years.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint 6 We used national data on the number of first doses of the COVID-19 vaccine that were administered daily. The analysis was based on data provided by the NDoH for purposes of evaluating the incentive programme. Using population data from the South African Community Survey 2016 (9) we calculated the number of first COVID-19 vaccine doses administered daily per 10,000 individuals aged ≥60 years. A sensitivity analysis with weekly doses is in the eMethods. First dose vaccination rates were calculated separately at the national-and provincelevel. The Vooma Voucher program was introduced on November 1, 2021, as a nationwide incentive for COVID-19 vaccination among adults aged ≥60 years. Individuals who received their first vaccine dose received a US$7 (ZAR 100) voucher by text message that was redeemable at local grocery stores.

We estimated four separate interrupted time series (ITS) models to estimate the association between the Vooma Voucher programs and trends in vaccination rates: unadjusted and adjusted national models as well as unadjusted and adjusted provincial models. The primary outcome was the daily vaccination rate per 10,000 individuals. National models utilized linear regression with Newey-West standard errors, while provincial models utilized generalized estimating equations with robust standard errors. Daily models included day of the week indicators to control for temporal patterns within each week and week-specific fixed effects. Unadjusted models included no additional covariates; adjusted models included additional measures of vaccine supply (daily active vaccine delivery sites per 100,000 individuals and an indicator variable for the three days of the Vooma weekend). The Vooma weekend was a nationwide initiative to open more vaccination sites to boost demand (10). Province-level adjusted models included an indicator for KwaZulu-Natal and Gauteng provinces, as well as an interaction term to assess impact of the voucher program in these provinces versus the rest of the country. Our analysis included data . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

In the weeks preceding introduction of the Vooma Voucher program, the 7-day rolling average of daily first vaccine doses administered declined from 13-15 doses per 10,000 individuals aged ≥60 years in early-September to <10 doses by late-October 2021 (Figure 1) . For the first 27 days, the Vooma Voucher program was associated with an increase in daily first vaccine doses administered of 0.89 per 10,000 individuals (95% CI 0.52, 1.27; p<0.001) ( (Figure 2) . Using the last week of October 2021 as a baseline, the percentage increase in vaccinations is estimated at 7.15%-to 12.01% per 10, 000 individuals per . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint 8 day. Weekly models and implications of supply and demand creation adjustments are found in the eTable.

A nationwide incentive program that provided vouchers redeemable at grocery stores to older adults who received a first dose of a COVID-19 vaccine was associated with increased vaccinations in South Africa. In the first month of the program, before vaccine demand may have been influenced by the announcement of the Omicron variant, 7.15% to 12.01% of all first vaccine doses administered to older adults could be attributed to the program. This represents a meaningful increase in vaccinations in response to a US$7 (ZAR100) incentive.

This study is among the first to evaluate incentives for COVID-19 vaccination in LMICs.

An incentive of US$24 increased vaccination rates by 4% in a randomized trial in Sweden (3).

Moreover, several studies found that lotteries and guaranteed incentives offered by US states did not increase vaccination rates (7) . The Vooma Voucher may have had better results due to substantial advertisement of the program and amount of the incentive relative to average income levels in South Africa. It is notable that focused Vooma weekends were associated with higher vaccination rates.

A key limitation is the assumption in our ITS models there were no other factors that coincided with the introduction of the Vooma Voucher program and affected vaccine demand.

The robustness of our findings to the inclusion of supply and demand creation measures, and to provincial and weekly specifications, increases confidence in our findings. Another limitation is . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint 9 that we only study short-term effects of the incentive program, as the announcement of the Omicron variant made it challenging to study effects of incentives beyond that date.

Given the vulnerability of older adults to serious illness, hospitalization, and death as a result of COVID-19, identifying effective strategies to increase vaccine demand is crucial. More generally, as LMICs struggle to achieve sufficiently high vaccine demand despite expansions in vaccine delivery and access, our findings suggest that small financial incentives may be effective in increasing vaccination coverage. 

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint Notes: 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 to the rest of the country.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint

Here's how phase 2 of SA's COVID vaccine roll-out works

Department of Health COVID-19 Online Resources and News Portal

Monetary incentives increase COVID-19 vaccinations. Science (80-)

Guaranteed Financial Incentives for COVID-19 Vaccination: A Pilot Program in North Carolina

Lottery-Based Incentive in Ohio

Vaccination Rates

Financial Incentives and Other Nudges Do Not Increase COVID-19 Vaccinations among the Vaccine Hesitant

Association Between Statewide Financial Incentive Programs and COVID-19 Vaccination Rates

It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 7, 2022. ; https://doi.org/10.1101/2022.05.06.22274712 doi: medRxiv preprint