• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Friday, November 20, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Health

Bayesian Spatiotemporal Model Predicts Scrub Typhus Risk Across South Korea

by
November 20, 2026
in Health
Reading Time: 6 mins read
0
Bayesian Spatiotemporal Model Predicts Scrub Typhus Risk Across South Korea

Bayesian Spatiotemporal Model Predicts Scrub Typhus Risk Across South Korea

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

A ten-year nationwide epidemiological study in the Republic of Korea has characterized the spatiotemporal distribution of scrub typhus and established a Bayesian predictive framework linking environmental variables to regional disease incidence. The investigation, which evaluated surveillance data across 229 municipal administrative units between January 2015 and December 2024, revealed that monthly infection risks surge dramatically during the autumn harvest season. Using an integrated nested Laplace approximation approach, the researchers quantified how temperature, relative humidity, vegetation density, and geographic elevation influence transmission rates over time, while simultaneously accounting for spatial clustering and temporal autocorrelation across the peninsula.

Scrub typhus, also known as tsutsugamushi disease, is an acute vector-borne infection caused by the bacterium Orientia tsutsugamushi, which is transmitted to humans through the bites of infected larval trombiculid mites known as chiggers. Clinically manifesting as acute fever, severe headache, chills, and characteristic eschars at bite locations, the disease represents a significant public health threat throughout the Asia-Pacific region. Within this endemic zone, commonly designated the tsutsugamushi triangle, an estimated one billion individuals reside at risk, with approximately one million clinical cases recorded annually. In the Republic of Korea, the disease was first documented in 1951 among United Nations military personnel and has since established widespread endemicity, leading public health authorities to designate it as a Group III notifiable infectious disease requiring mandatory physician reporting.

To investigate the environmental drivers governing transmission across diverse landscapes, the research team assembled a comprehensive surveillance registry containing 27,480 spatiotemporal data points derived from 229 municipal administrative units over 120 consecutive months. The primary outcome was measured as the monthly incidence of scrub typhus per 100,000 population, paired with demographic data obtained from national statistical databases. Environmental covariates were acquired and spatially aggregated using cloud-based satellite observation platforms, capturing monthly mean temperatures, relative humidity, total precipitation, and normalized difference vegetation index values, alongside static topographical metrics including mean elevation, cropland coverage proportions, and urban land ratios resampled to a standardized 5.5-kilometer spatial grid.

Initial time-series decomposition utilizing locally estimated scatterplot smoothing isolated distinct secular trends and recurring seasonal dynamics in disease occurrence nationwide. The temporal analyses revealed that disease activity remained minimal from January through August before escalating rapidly in September and reaching its annual peak in November. Mixed-effects negative binomial regression confirmed this pronounced seasonality: compared to baseline rates in February, infection risks were 91-fold higher in November, exhibiting an incidence rate ratio of 91.00 with a 95 percent confidence interval spanning 66.50 to 125.00. October also demonstrated marked elevations, with an incidence rate ratio of 39.00, while December maintained an elevated ratio of 12.50 before disease activity subsided during the winter months.

Spatial clustering assessments employing Getis-Ord statistics detected prominent geographic disparities in disease burden across the Korean peninsula. Statistically significant high-incidence hotspots were heavily concentrated within the southern provinces, notably throughout the Jeolla and Gyeongsang regions. Conversely, statistically significant cold spots of low incidence clustered predominantly across the central and northern territories, including the Seoul metropolitan area and Gangwon Province. These geographic divergences align with surveillance records showing distinct ecological distributions among the primary trombiculid vector species, with dominant chiggers such as Leptotrombidium scutellare, Leptotrombidium pallidum, and Leptotrombidium palpale displaying seasonal and regional variations that coincide with traditional autumn outdoor activities and agricultural harvesting.

To construct a predictive statistical framework while preventing parameter instability, the investigators performed variance inflation factor evaluations. Precipitation and urban land ratios were excluded from the final regression models due to strong collinearity with temperature and cropland metrics, respectively. Generalized additive models subsequently revealed non-linear, inverted U-shaped relationships between the remaining environmental variables and disease incidence. Consequently, quadratic polynomial terms were embedded within a Bayesian hierarchical negative binomial model executed via integrated nested Laplace approximations. Spatial autocorrelation was structured using the Besag-York-Mollié 2 formulation, while temporal dependencies were modeled through first-order random walks and twelve-month cyclical seasonal components.

Comparative evaluations demonstrated that a one-month lag model provided superior overall predictive fit and accuracy over a two-month lag model, yielding a lower Watanabe-Akaike information criterion of 62010.39 compared to 62533.26 and a reduced root mean square error of 7.08 versus 7.33. Within the optimal one-month lag specification, ambient temperature demonstrated a positive association with disease incidence, yielding an incidence rate ratio of 1.02 per 1 degree Celsius increase, with modeled risks peaking at 16.03 degrees Celsius. Vegetation density also displayed a significant positive correlation, with the normalized difference vegetation index showing an incidence rate ratio of 1.07 per 0.1-unit increment and reaching maximal risk at an index value of 0.65.

Elevation similarly exhibited a positive association with transmission risk, producing an incidence rate ratio of 1.32 per 100-meter rise, with estimated risks peaking at approximately 142.45 meters above sea level. This topographical association corresponds to ecological habitats where elevated relative humidity and denser vegetation growth remain less disturbed by human development, providing favorable shelter and food resources for wild rodent populations that serve as principal vertebrate hosts for chigger mites. By contrast, relative land-use ratios dedicated to cropland did not achieve statistical significance in either the one-month or two-month lag models.

The influence of atmospheric moisture revealed complex, divergent patterns across different temporal lag windows. In the one-month lag model, relative humidity demonstrated an apparent protective association, exhibiting an incidence rate ratio of 0.96 with peak risk estimated at 73.92 percent. Researchers noted that this short-term inverse pattern likely reflects human behavioral adaptations, as sustained high-humidity conditions and rainfall suppress outdoor agricultural and recreational exposures. Conversely, the two-month lag model identified a positive association between relative humidity and disease risk, generating an incidence rate ratio of 1.16 with peak risk at 78.50 percent, reflecting vector developmental cycles wherein sustained moisture enhances chigger survival, oviposition, and larval emergence over six- to eight-week intervals.

Spatial mapping of Bayesian model projections closely matched observed epidemiological patterns across southern and central administrative units, though prediction uncertainty remained noticeably elevated across northern mountainous territories, particularly within Gangwon Province. The authors attributed this geographic variation in predictive certainty to lower vector species diversity, differing vector compositions dominated by species such as Neotrombicula kwangneungensis and Leptotrombidium tectum, and regional differences in healthcare infrastructure and diagnostic reporting. The researchers acknowledged that the study’s ecological design, potential underreporting in rural settings, and uniform spatial resampling of environmental variables highlight the need for future predictive frameworks to incorporate granular entomological field surveys alongside individual-level behavioral determinants.

Epidemiological records contextualize the growing burden of scrub typhus across the Republic of Korea. Following its designation as a Group III notifiable infectious disease in 1994, when initial counts hovered near 300 annual cases, national surveillance registered 152,775 total infections between 2001 and 2024. Long-term surveillance indicates a sustained upward trajectory, with average annual incidence expanding from 4,667 cases during 2005 to 2010 up to 7,854 cases during 2011 to 2020, marking a 68.3 percent increase. Beyond this decadal expansion, the study’s seasonal-trend decomposition isolated distinct multiannual oscillations in national transmission between 2015 and 2024. The estimated underlying trend line climbed progressively from 2015 into 2018, dropped steadily toward a study-period trough spanning late 2019 and early 2020, and subsequently rebounded before leveling out into a stable baseline.

The negative binomial regression also delineated earlier seasonal risks preceding the late-autumn surge. When evaluated against February as the baseline reference, March was the only month lacking a statistically significant difference in infection risk. Significant elevations emerged by early summer, with June, July, August, and September exhibiting moderate incidence rate ratios between 2.74 and 3.80. This progressive elevation intersects with vector phenology and cultural behaviors. Specifically, the Korean Thanksgiving holiday, Chuseok, observed from late September through early October, overlaps with peak chigger questing activity. During this holiday period, family gatherings involving visits to ancestral burial grounds for traditional grave-tending rituals place substantial populations in direct contact with vegetative mite habitats.

Although the one-month lag model achieved optimal statistical fit, the two-month lag model offered corroborating insights into delayed ecological influences across municipal units. In this secondary specification, temperature maintained an estimated incidence rate ratio of 1.03 per 1 degree Celsius increase (95 percent credible interval: 1.02 to 1.05), reaching maximum disease risk at 22.72 degrees Celsius. Similarly, vegetation greenness exhibited an incidence rate ratio of 1.08 per 0.1-unit rise in normalized difference vegetation index (95 percent credible interval: 1.04 to 1.12), peaking at 0.71. Elevation displayed an incidence rate ratio of 1.29 per 100 meters (95 percent credible interval: 1.13 to 1.48), matching the 142.45-meter risk peak observed under the one-month framework.

Computational implementation and geographic variations further illustrated regional diagnostic challenges. Statistical optimization explored 81 distinct hyperparameter combinations, testing spatial mixing parameters of 0.3, 0.5, and 0.7 alongside spatial precision parameters up to 1.5 under penalized complexity priors within the R-INLA environment. The administrative divisions varied substantially in geographic area, spanning from 3.04 square kilometers in Jung-gu, Busan, to 1,820 square kilometers in Hongcheon-gun, with a median area of 395.4 square kilometers. Although Jung-gu fell below the standardized 5.5-kilometer satellite grid resolution, the remaining 228 units retained sufficient spatial definition. Moreover, healthcare resource disparities likely influenced surveillance completeness: Gangwon Province maintained an average of only 17,630 hospital beds between 2015 and 2023, contrasting sharply with major southern endemic areas such as Jeollanam-do with 41,036 beds and Gyeongsangnam-do with 62,447 beds.

Subject of Research: Bayesian spatiotemporal prediction and environmental determinants of scrub typhus in South Korea

Article Title: Spatiotemporal prediction of scrub typhus incidence and environmental risk factors in Republic of Korea: a Bayesian hierarchical approach

Article References: Kim, Y., Kwon, D., Hasahya, E., & Lee, H. S. (2026). Spatiotemporal prediction of scrub typhus incidence and environmental risk factors in Republic of Korea: a Bayesian hierarchical approach. Infectious Diseases of Poverty, 15(1), Article 82. https://doi.org/10.1186/s40249-026-01481-2

Image Credits: AI Generated

DOI: 10.1186/s40249-026-01481-2

Keywords: scrub typhus, Orientia tsutsugamushi, Bayesian hierarchical modeling, spatiotemporal epidemiology, Republic of Korea, disease surveillance, integrated nested Laplace approximation, relative humidity, normalized difference vegetation index, vector-borne disease, chigger mites, Leptotrombidium scutellare, seasonal trend decomposition, environmental risk factors, spatial clustering, hotspot analysis, public health forecasting

News Source: Phoebe Ingram. (November 20, 2026). Bayesian Spatiotemporal Model Predicts Scrub Typhus Risk Across South Korea. Scienmag.

Tags: Bayesian hierarchical modelingchigger mitesdisease surveillanceenvironmental risk factorshotspot analysisIntegrated Nested Laplace ApproximationLeptotrombidium scutellareNormalized Difference Vegetation IndexOrientia tsutsugamushipublic health forecastingrelative humidityRepublic of Koreascrub typhusseasonal trend decompositionspatial clusteringSpatiotemporal Epidemiologyvector-borne disease
Share12Tweet7Share2ShareShareShare1

Related Posts

Oncology Staff Essays Reveal Mutual Psychological Benefits in Long-Term Cancer Care

Oncology Staff Essays Reveal Mutual Psychological Benefits in Long-Term Cancer Care

November 20, 2026
Targeting Folate Receptor Overcomes Immunotherapy Resistance in Colorectal Cancer in Mice

Targeting Folate Receptor Overcomes Immunotherapy Resistance in Colorectal Cancer in Mice

November 20, 2026

Jiangxi Cities Narrowed Gaps in Healthcare Workforce and Beds Over Ten Years

November 20, 2026

Dimethoate Metabolites Pose Greater Predicted Toxicological Risks Than Parent Insecticide

November 20, 2026

POPULAR NEWS

  • Targeting Folate Receptor Overcomes Immunotherapy Resistance in Colorectal Cancer in Mice

    Targeting Folate Receptor Overcomes Immunotherapy Resistance in Colorectal Cancer in Mice

    29 shares
    Share 12 Tweet 7
  • Jiangxi Cities Narrowed Gaps in Healthcare Workforce and Beds Over Ten Years

    29 shares
    Share 12 Tweet 7
  • Dimethoate Metabolites Pose Greater Predicted Toxicological Risks Than Parent Insecticide

    29 shares
    Share 12 Tweet 7
  • Dual-Frequency Electroacupuncture Modulates Neurotransmitters and Relieves Insomnia in Mice

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Oncology Staff Essays Reveal Mutual Psychological Benefits in Long-Term Cancer Care

Nitric Oxide Inhibition Partially Normalizes Cortical Proteins in Autism Mouse Models

Targeting Folate Receptor Overcomes Immunotherapy Resistance in Colorectal Cancer in Mice

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 89 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.