Machine learning insights on the effectiveness of non-pharmaceutical interventions against COVID-19 in Nigeria

The lack of effective pharmacological measures during the early phase of the COVID-19 pandemic prompted the implementation of non-pharmaceutical interventions (NPIs) as initial mitigation strategies. The impact of these NPIs on COVID-19 in Nigeria is not well-documented. This study sought to assess the effectiveness of NPIs to support future epidemic responses.

Daily COVID-19 cases and deaths were analysed using smoothed variables to identify transmission trends. Regression analysis and clustering algorithms were applied to evaluate the impact of each NPI.

Multiple transmission peaks were reported, with the highest smoothed daily new cases (approximately 1790) observed around 29 December 2021 and smoothed daily new deaths (approximately 23) peaking around 8 September 2021. NPIs such as public transport (coefficient value -166.56, p=0.01) and workplace closures (coefficient value -150.06, p=0.01) strongly correlated with decreased case numbers. This finding highlights the importance of mobility control and non-essential workplace management in slowing infection transmission during an outbreak. Public transport restrictions (coefficient value -2.43, p<0.001) also had a direct effect on death reduction.

Public transport restrictions and workplace closures correlated with reductions in the number of cases and deaths. These findings can guide future pandemic responses to enhance favourable public health outcomes.