While the response to COVID-19 by the Government of India has been more or less uniform across the country, in that a lockdown was imposed throughout, its impact measured in terms of the death rate has varied across the states. This suggests state-specific characteristics are likely to be relevant to the determination of this rate. We attempt to identify these. As there is a dispersion of wealth, as reflected in per capita income, among the states we first look for a relationship between income and death from COVID-19. This actually revealed the death rate and per capita income to have a mildly positive relationship, implying that wealth is not a shield against death from the disease. It led us to investigate the possible impact of public policy interventions in the area of healthcare, notably expenditure on health and physical infrastructure in the public sector. In almost all the statistical exercises implemented here, all three measures of the death rate deployed here are strongly related to the variable health expenditure as a share of the gross domestic product but weakly so to infrastructure. We interpret this as a sign of the role of the health system as a whole - medical personnel, infrastructure and protocols - in the prevention of death and of health expenditure as a determinant of its effectiveness. A significant aspect of this study is the use of more than one measure of the death rate, modified as appropriate, in the empirical exercise. Its findings have an immediate implication for policy. At this stage of the development of health infrastructure in India, its states must prioritise spending on health in their budgets. The wide variation in health expenditure in relation to gross domestic product across states demonstrates that the laggards have the capacity to do so.