Our analysis sought to answer two questions: first, whether there is sufficient health and economic value in making the vaccines available in India and Indonesia, given what is currently known about the vaccine; secondly, how soon the vaccine should be made available, to yield its highest health and economic benefits. The answer to the first question is promising. The economic value of the vaccine can be substantial: in the best case, where we assumed that the efficacies demonstrated for populations aged 18–49 years are transferable to populations of all ages: 30.3% and 23.1% of economic losses can be averted in India and Indonesia (respectively), if initiated in 2025, and 22.6% and 16.0% if initiated in 2030. Our assumption regarding the transferability of efficacies is reasonable as the 3-year final analysis of the M72 phase 2b trial did not show differential efficacy by age16.
The answer to the second question is less clear, however. The health and economic losses due to not initiating vaccination in 2020 could be fully recovered even if vaccination is initiated in 2025, provided all who are latently infected are vaccinated. Indeed, in Indonesia, vaccinating all who are latently infected in 2025 is a better option than vaccinating the 18–49 year age group in 2020 (23.05% of the projected economic losses averted compared to 18.64%, respectively), as a large proportion of the latently infected are ≥ 50 years of age. This conclusion, however, is contingent on the vaccine providing a lifetime of protection. Whether the vaccine can provide such protection is still unknown. Knowing the duration of protection is important as our results demonstrate that in both countries, 10 years of protection at the lower efficacy bound, results in a greater number of TB deaths averted if the individuals of all ages are vaccinated (Figs. 1D and 2D).
We also find that there are specific windows within each country that can yield a demographic dividend, in addition to an economic dividend, given when vaccination is initiated. In India, the total dependency ratio declines from 2020 to 2040, then begins a steady upward trajectory13. Initiating vaccination early will allow capturing this window by reducing the burden of TB on the productive population. If vaccination is initiated in 2020 or 2025, then the highest number of deaths are averted in the 30–35 age group. Incidence is more sensitive to early initiation; if initiated in 2020, then the highest incidence is averted in the 30–35 age group. Waiting until 2030, moves the age group likely to experience the greatest benefit to 40–45 years. In Indonesia, unlike in India, the total dependency ratio steadily increases. Therefore, in Indonesia there is no optimal window during which a demographic dividend can be gained. The earlier vaccination is initiated, the greater the burden on the productive population that can be averted, especially as the contribution towards life expectancy at birth is greater in Indonesia when compared to India (Table 1).
This argument, however, does not account for the exogenous shock of COVID-19 in deterring economic growth. The projected upswing in the per capita income growth (7.4% and 8.2% in India and Indonesia, respectively, in 202122) is largely uncertain now. Though population age structure can fuel economic growth, effective policies are needed to capitalize on the working population. Without such policies, countries can end up with large pools of unemployed or underemployed working age populations. COVID-19 has reframed the context for social policy and it is indeed why vaccination is all the more important now. Pre-COVID projections of TB incidence likely underestimates current levels as well. Recent estimates by Cilloni and colleagues show that a two-month lockdown with two months of recovery, gave rise to 473,000 additional TB cases and 130,000 additional TB deaths in India23. Our recent estimates show that the additional incident cases due to a 3-month lockdown with 10-months of recovery can give rise to significant increases in additional government spending and additional out-of-pocket spending24. In India, where out-of-pocket spending already accounts for as much as 43.8% of total spending, the economic and social consequences of COVID-19 only stand to aggravate the burden on patients and households.
Notwithstanding the unknowns, our results provide compelling epidemiological and economic justification for making the vaccine available as soon as scientifically justifiable. In Indonesia, 5,496,793 incident cases can be averted if the vaccine is introduced in 2030 to all ages. Fuady et al. find that 36% of households in Indonesia with a case of active drug-susceptible TB, experienced catastrophic costs 25 . If generalizable, then by introducing the vaccine in 2030, 1,978,845 households can be prevented from experiencing catastrophic losses. India had a higher estimated percentage of new TB cases with rifampicin resistant TB compared to Indonesia (2.8% versus 2.4%), which means that the gains in India can be even higher. Recent estimates from Su et al. show that India’s total tuberculosis spending per incident case in 2017 is double that of Indonesia’s (US$644; 95% CI: US$489–852, versus US$322; 95% CI: US$258–403) 26 . In India, 43.8% (95% CI: 28.9%-59.6%) of the spending is borne by the patient as out of pocket spending, whereas In Indonesia, the share is 14.5% (95% CI: 6.4%-26.8%). India’s TB spending is also projected to grow at 7.9% annually whereas Indonesia’s is projected to grow at 2.6%26.
The applicability of the efficacies demonstrated in Kenya, South Africa, and Zambia, to populations in India and Indonesia is a central assumption of our analysis and thereby potentially a limitation. We assumed that latency can be established in a reliable and cost effective manner, which in countries like India with large high risk populations12, can be expensive. Vaccinating irrespective of exposure status is therefore a better option, provided the safety of the vaccine in individuals previously unexposed to Mtb is known. Our estimates of incidence and consequent mortality are contingent on assumptions inherent in our natural history model. While these assumptions reflect the most recent thinking and findings, they are nonetheless limitations. We also assumed that the efficacies remained stable over the duration of protection, which is yet to be established. Our full-income estimates are limited by our projections of per capita income, which are now all the more uncertain. Equally uncertain due to COVID-19 are the empirically established VSL values. Though the literature is scarce, it is plausible that the mean individual valuation of mortality risk reduction is influenced by the presence of a global scale pandemic. Our estimates are also limited by WPP’s projections of populations in India and Indonesia, as well as the abridged life tables through 2050 for both countries.
As much as our results point towards substantial economic benefits, that conclusion is still contingent on unknowns - primarily, the safety of the vaccine in young adolescents and older adults, and the duration of protection. Our results are all the more reason to invest in a phase 3 trial that could resolve these unknowns and thereby pave the way towards realizing the potentially substantial health and economic benefits. Excess cases due to disruption of routine TB services from COVID-19 lockdowns, will undoubtedly add to the dividends achievable with a vaccine, though possibly moderated by COVID-19 prevention measures given the common routes of transmission27. The optimal window for vaccination is fast approaching. Investment is needed now more than ever.