# Descriptive Epidemiology of COVID-19 in Hokkaido, Japan: Regional Burden and the Role of Railway-Driven Population Mobility

## TL;DR

This study examines how railway mobility and regional differences affect the spread of COVID-19 in Hokkaido, Japan.

## Contribution

The study introduces a novel approach to analyzing regional disparities in COVID-19 incidence by incorporating railway transport density as a mobility metric.

## Key findings

- Sapporo had the highest age-standardized incidence ratios (SIR) during multiple pandemic waves.
- Railway transport density was significantly associated with higher SIR in both univariable and multivariable models.
- The effective reproduction number (Rt) showed a negative association with SIR in the multivariable model.

## Abstract

Background

Understanding regional disparities in COVID-19 incidence and the factors influencing them is essential for effective public health responses. In Japan, particularly in Hokkaido, significant differences in case burden have been observed across municipalities. This study aimed to describe the epidemiological patterns of COVID-19 in Hokkaido by adjusting for age structure and to examine the association between railway-based population mobility and regional case burden.

Method

We categorized all municipalities in Hokkaido into five public health center jurisdictions: Sapporo, Otaru, Asahikawa, Hakodate, and Other Areas. COVID-19 incidence was assessed across the third to sixth waves of the pandemic (November 2020 to June 2022). To account for demographic differences, age-standardized incidence ratios (SIR) were calculated using national age-specific incidence rates and the indirect standardization method. Data on confirmed cases were obtained from government open datasets, and population mobility was measured using railway transport density from the Hokkaido Railway Company. Statistical analyses included chi-square tests and multiple regression analysis, including both full and simplified models to assess statistical associations, with the latter addressing multicollinearity among demographic variables.

Results

A total of 378,281 cases were reported during the study period, with the highest burden in Sapporo. SIR exceeded 1.0 in Sapporo, Otaru, and Asahikawa during multiple waves, indicating higher-than-expected incidence after age adjustment. Notably, Sapporo exhibited elevated SIR, particularly among older adults, while Otaru and Asahikawa showed higher rates among those aged 20-50 years. In regression analyses, railway transport density showed a statistically significant positive association with SIR in both univariable (p = 0.002) and simplified multivariable models that included the effective reproduction number (Rt). In the multivariable model, railway transport density remained positively associated (p < 0.001), while Rt showed a negative association (p = 0.032), with an adjusted R² of 0.524. These findings suggest that railway-based mobility and real-time transmission dynamics jointly influence regional disparities in COVID-19 burden.

Conclusion

This study highlights the significant impact of railway-based population mobility on COVID-19 transmission within Hokkaido, Japan. Age-standardized comparisons revealed regional variations in risk, likely influenced by urban proximity and transport infrastructure. These findings support the incorporation of geographic and mobility-related metrics into geographically tailored public health policies. Further research using larger datasets and multimodal transport indicators is recommended to validate and expand upon these insights.