Geodemographic clustering of social areas
In this study, data was obtained from the public resources of the Osaka City Planning and Coordination Bureau25. Based on these tabulation data of Osaka City from the 2010 National Census, social area clustering was established. Osaka City has 332 town councils. The Z-score normalization for 42 indicators reflected the familial, socioeconomic, and ethnic statuses of the residents of each town council (Supplementary Table 1). These indicators of the abovementioned three statuses have traditionally been employed in various studies on social area analysis and factorial ecology in urban sociology and geography26,27. We applied factor analysis to the 42 indicators using the maximum likelihood method and varimax rotation to extract common factors. Geodemographic clustering of the 332 town councils into social areas was performed using Ward’s method and the scores of common factors.
Geocoding of the residence of patients with TB
Data on TB patients was obtained from the Osaka City Public Health Office. The number of patients with TB who lived in each social area and were registered as new cases in Osaka City between 2012 and 2016 was counted based on their residential address. The town councils of 4,990 patients with TB were identified from their residential addresses, but those of 138 patients could not be detected. Consequently, data from 4,852 individuals were used in this study.
Standardization of TB incidence ratio in social areas
The standardized annual TB incidence ratio for each social area (Si, where i denotes the respective social area) during the study period was calculated by dividing the number of registered TB cases in each area (Pi) by the expected number of TB cases per year (Ei).
Si = Pi/Ei.
Here, Ei was calculated based on the ratio of the population of each area (Ni) to the total population of Osaka City (Nt) from the total number of registered patients with TB (4,852 persons) in the entire city. It was calculated based on the age distribution of registered patients with TB (Supplementary Table 2) to consider the age structure of each area. Hence, for a given age group j (in increments of 10 years), if the population of Osaka City as a whole is Ntj, the population of each area is Nij, and the number of patients with TB in Osaka City as a whole is Ptj, then Ei is the sum of all age groups as follows:
$$\:{E}_{i}\:=\:{\sum\:}_{j}^{}{N}_{ij}\left({P}_{tj}/{N}_{tj}\right)$$
The 95% confidence interval for Si was calculated based on a Poisson distribution28.
Characteristics of patients with TB in social areas with significantly higher standardized TB incidence ratios
Multinomial logistic regression was used to examine the characteristics of patients with TB in social areas where the standardized TB incidence ratio was significantly higher. Two social areas with high TB incidence ratios and other social areas were used as objective variables. The explanatory variables selected for the study included sex, age group, national origin, respiratory symptoms, history of medical treatment, history of TB diagnosis, and delay in the diagnosis from symptom onset (Table 4). The adjusted odds ratio, along with the 95% confidence interval and p-value, was calculated after adjusting for all explanatory variables.
Data analysis
R ver. 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria) was used for the statistical analysis. ArcGIS ver. 10.5 (ESRI, Redlands, CA) was used for mapping, unless otherwise noted in the figure legend.
Ethical considerations
This study was a retrospective observational study. Two types of data were used: one, small area aggregated results from the national census obtained from the Osaka City Planning and Coordination Bureau; two, tuberculosis patient data collected by the Osaka City Health Center through active epidemiological investigation based on the Infectious Diseases Control Law.
This study was conducted in accordance with the “Ethical Guidelines for Life Science and Medical Research Involving Human Subjects” established by the Ministry of Health, Labour, and Welfare of Japan (effective March 23, 2021). Ethical approval was obtained from the Ethical Review Committee of the first author’s institution (approval number: 1812-01-4, approval date: April 1, 2023).
As this study was a retrospective observational study using existing data, the requirement for obtaining individual informed consent was waived. However, in accordance with the aforementioned ethical guidelines, an overview of the research and opt-out procedures was posted on the institution’s website and made publicly available for a specified period. During this period, no requests to opt out of the study were received.
All information handled in this study was anonymized and did not contain personally identifiable information. Data handling strictly adhered to the policies of the Ethics Committee, and the analysis was conducted in a secure environment. Address data were converted to coordinate values in a stand-alone environment to avoid the risk of online data breach, using the basic geocoding function of ArcGIS Pro 2.4 (ESRI, Redlands, CA) with the Japan Data Contents Starter Pack (2018) (ESRI Japan, Tokyo, Japan).
AloJapan.com