Effect of comprehensive smoke-free legislation on neonatal mortality and infant mortality across 106 middle-income countries: a synthetic control study.

Background: There are few quantitative studies into the effect of comprehensive smoke-free legislation on neonatal and infant mortality in middle-income countries. We aimed to estimate the effects of implementing comprehensive smoke-free legislation on neonatal mortality and infant mortality across all middle-income countries.

Methods: We applied the synthetic control method using 1990-2018 country-level panel data for 106 middle-income countries from the WHO, World Bank, and Penn World datasets. Outcome variables were neonatal (age 0-28 days) mortality and infant (age 0-12 months) mortality rates per 1000 livebirths per year. For each middle-income country with comprehensive smoke-free legislation, a synthetic control country was constructed from middle-income countries without comprehensive smoke-free legislation, but with similar prelegislation trends in the outcome and predictor variables. Overall legislation effect was the mean average of country-specific effects weighted by the number of livebirths. We compared the distribution of the legislation effects with that of the placebo effects to assess the likelihood that the observed effect was related to the implementation of smoke-free legislation and not merely influenced by other processes.

Findings: 31 (29%) of 106 middle-income countries introduced comprehensive smoke-free legislation and had outcome data for at least 3 years after the intervention. We were able to construct a synthetic control country for 18 countries for neonatal mortality and for 15 countries for infant mortality. Comprehensive smoke-free legislation was followed by a mean yearly decrease of 1·63% in neonatal mortality and a mean yearly decrease of 1·33% in infant mortality. An estimated 12 392 neonatal deaths in 18 countries and 8932 infant deaths in 15 countries were avoided over 3 years following the implementation of comprehensive smoke-free legislation. We estimated that an additional 104 063 infant deaths (including 95 850 neonatal deaths) could have been avoided over 3 years if the 72 control middle-income countries had introduced this legislation in 2015. 220 (43%) of 514 placebo effects for neonatal mortality and 112 (39%) of 289 for infant mortality were larger than the estimated aggregated legislation effect, indicating a degree of uncertainty around our estimates. Sensitivity analyses showed results that were consistent with the main analysis and suggested a dose-response association related to comprehensiveness of the legislation.

Interpretation: Implementing comprehensive smoke-free legislation in middle-income countries could substantially reduce preventable deaths in neonates and infants.

Funding: Dutch Heart Foundation, Lung Foundation Netherlands, Dutch Cancer Society, Dutch Diabetes Research Foundation, Netherlands Thrombosis Foundation, Health Data Research UK.