The evolving epidemiology of scrub typhus in Thailand (2003-2024): insights from latent process modelling of national surveillance data

By integrating 22 years of national surveillance data with seroprevalence and antibody kinetics through latent process modelling, this study reveals that scrub typhus transmission in Thailand extends significantly beyond its traditional northern focus into the northeast and south, while the at-risk demographic has progressively shifted toward older adults who face the highest mortality, necessitating regionally tailored surveillance and age-targeted clinical strategies.

The Gist

Imagine Thailand as a giant, bustling garden. For decades, gardeners have been worried about a specific, invisible pest called Scrub Typhus. This isn't a bug you can see; it's a microscopic bacteria carried by tiny mites (chiggers) that live in the grass. When these mites bite humans, they can make people very sick with high fevers.

For over 20 years, Thailand has kept a notebook (surveillance data) counting how many people get sick. But here's the problem: The notebook isn't perfect.

Sometimes, people don't go to the doctor. Sometimes, doctors are too busy or don't know what to look for. Sometimes, the notebook gets lost during a crisis (like the pandemic). So, just looking at the numbers in the notebook doesn't tell the whole story. It's like trying to guess how many fish are in a lake just by counting the ones caught in a net that sometimes breaks or gets lost.

This paper is like a team of detectives who decided to solve the mystery of the "real" number of infections. They didn't just look at the notebook; they combined it with a "blood test history" (seroprevalence data) to see how many people had been infected in the past, even if they never got sick enough to go to the hospital.

Here is what they discovered, broken down into simple stories:

1. The "Ghost" in the Data (Reporting vs. Reality)

The researchers built a special mathematical model—a kind of digital time machine—to separate the "real" infections from the "missed" ones.

• The Pandemic Dip: During 2020 and 2021, the notebook showed almost no cases. The detectives realized this wasn't because the mites disappeared. It was because everyone was staying home, hospitals were overwhelmed with COVID, and people were afraid to go out. The "net" had a hole in it.
• The Bounce Back: By 2024, the numbers jumped back up. The net was fixed, and the gardeners were counting again.

2. It's Not Just a "Northern" Problem

For a long time, people thought Scrub Typhus was a problem only in the mountainous North (like Chiang Rai and Chiang Mai), similar to how people might think "snow" only happens in the North Pole.

• The Surprise: The detectives found that the "snow" is actually falling all over the country! The Northeast and the South are also hotspots. In fact, some provinces in the South and Northeast are just as dangerous as the famous northern ones, but they've been ignored because no one was looking closely.

3. The "Aging Farmer" Shift

Think of the people getting sick as a group of hikers.

• 20 Years Ago: The hikers getting bitten were mostly young and middle-aged adults (25–45).
• Today: The group has changed. Now, the older adults (65+) are the ones getting bitten the most.
• Why? Thailand's farming workforce is getting older. Young people are moving to cities for office jobs, but the older farmers are still out in the fields, knee-deep in the grass where the mites live. They are the "veteran hikers" who keep getting bitten.
• The Danger: Not only are older people getting bitten more, but they are also the ones most likely to get very sick or die if they don't get treatment quickly. Their immune systems are a bit like old, rusty shields that don't work as well as they used to.

4. The Seasonal Rhythm (The "Dance" of the Mites)

The mites don't bite randomly; they dance to the rhythm of the weather. The researchers found three different dance styles across the country:

• The Late Dancers: In the Northeast and Central areas, the mites go wild in September and October (the end of the rainy season).
• The Early Dancers: In the North and West, the party starts earlier, around July.
• The All-Year Dancers: In the deep South, the mites are active almost all year round, like a party that never ends.
• The Good News: These dance patterns haven't changed much in 20 years. If you know the rhythm of your local area, you know when to be careful.

The Big Takeaway

This paper tells us that we need to change our strategy.

• Don't just watch the North: We need to protect the Northeast and South, too.
• Protect the Elders: Since older farmers are the most at risk, doctors and families need to be extra careful with them.
• Know the Rhythm: We need to warn people before the mites start their specific dance in their region.

In short: Scrub Typhus in Thailand is a shape-shifting problem. It's hiding in more places than we thought, it's targeting an older population, and it follows a strict seasonal schedule. By using smart math to look past the messy notebook data, the researchers have given us a clear map to protect the people who need it most.

Technical Summary

1. Problem Statement

Scrub typhus, caused by Orientia tsutsugamushi and transmitted by chigger mites, is a neglected vector-borne disease in Thailand. While the disease has been nationally notifiable for over two decades, the long-term epidemiological trends remain poorly characterized at the national level.

• The Core Challenge: Passive surveillance data (reported cases) do not directly reflect underlying transmission intensity. Reported counts are confounded by variations in healthcare-seeking behavior, diagnostic capacity, clinical awareness, and reporting practices across different regions and time periods.
• The Gap: Previous studies (e.g., Wangrangsimakul et al., 2018) analyzed data up to 2018 but were limited to a single province for trend analysis and did not disentangle true transmission changes from reporting artifacts. There is a lack of comprehensive understanding regarding how transmission intensity, disease severity, and seasonal patterns have evolved across all 77 provinces of Thailand over the last 22 years.

2. Methodology

The authors developed a hierarchical Bayesian latent process model to jointly analyze multiple data sources, allowing them to separate true epidemiological dynamics from surveillance biases.

• Data Sources:

  • Surveillance Data: 22 years (2003–2024) of monthly reported cases and deaths from the National Disease Surveillance (Report 506) database, covering all 77 provinces and 13 health regions.
  • Seroprevalence Data: Nationwide IgG seroprevalence data from 2007–2008 and 2012 (Gonwong et al., 2022) to inform cumulative transmission.
  • Population Data: Annual population sizes by province and age group.
  • External Parameters: Antibody waning rates (fixed at 0.53 year⁻¹) and clinical fraction (fixed at 0.15) derived from literature.

• Model Structure:

  • Latent Process: The model estimates the "Force of Infection" (FOI) and the "Reporting Proportion" (probability a clinical case is reported).
  • Transmission Dynamics: FOI is modeled as a multiplicative function of:
    • Baseline FOI: Province-specific random walk to capture long-term trends.
    • Seasonality: Modeled via a Fourier series (3 harmonics) to capture monthly variation.
    • Age Susceptibility: Relative susceptibility by age group (reference: 25–35 years).
  • Observation Model:
    • Reported cases are linked to latent clinical infections via a region-specific reporting proportion (modeled as a random walk over time).
    • Deaths are modeled as a function of reported cases and age-specific Case Fatality Rates (CFR).
    • Seroprevalence is linked to the cumulative FOI over time, accounting for antibody waning.
  • Inference: Parameters were estimated using Hamiltonian Monte Carlo (HMC) sampling in Stan (via rstan).

3. Key Contributions

• Methodological Innovation: The study introduces a novel latent process model that integrates passive surveillance with seroprevalence and antibody kinetics to correct for underreporting and estimate true transmission intensity across a large geographic scale.
• Temporal Extension: Extends the epidemiological analysis of scrub typhus in Thailand from 2018 to 2024, capturing the impact of the COVID-19 pandemic on disease surveillance.
• Granular Resolution: Provides estimates for all 77 provinces and 13 health regions, moving beyond the traditional focus on the northern region.
• Demographic Shift Analysis: Quantifies the shifting susceptibility and mortality risks across age groups over two decades.

4. Key Results

A. Reporting Dynamics

• Reporting Proportion: Estimated reporting rates varied significantly by region and time. High-burden regions (North) had higher reporting proportions (peaking ~0.89 in 2013), while lower-burden regions (Central/East) had lower rates (<0.50).
• Pandemic Impact: A synchronized, sharp decline in reporting proportion occurred across all regions in 2020–2021 (e.g., dropping to ~0.18–0.26 in some regions), attributed to the COVID-19 pandemic's effect on healthcare access and surveillance.
• Rebound: Reporting rates rebounded rapidly to pre-pandemic levels by 2024.

B. Transmission Intensity (Force of Infection)

• Geographic Heterogeneity: While the North (Health Regions 1 & 2) remains the highest burden area, substantial transmission exists in the Northeast and South.
• Emerging Hotspots: Provinces like Mae Hong Son and Tak showed high FOI and increasing trends. Notably, provinces in the Northeast (e.g., Roi Et, Udon Thani) and South (e.g., Surat Thani, Chumphon) showed rising transmission trends, challenging the view of scrub typhus as solely a northern disease.
• Seasonality: Three distinct seasonal patterns were identified and remained stable over 20 years:
  1. Pattern 1 (55% of provinces): Late-year peak (September/October), high amplitude. Dominant in Northeast/Central.
  2. Pattern 2 (36%): Early peak (July), moderate amplitude. Dominant in North/West.
  3. Pattern 3 (9%): Flat, year-round transmission. Dominant in the South.

C. Age-Specific Susceptibility and Severity

• Susceptibility Shift: A dramatic reversal in age susceptibility was observed. In 2003–2007, older adults (>65) had lower susceptibility than the reference group (25–35). By 2019–2024, the >65 group showed the highest relative susceptibility (1.6x), likely due to an aging agricultural workforce.
• Case Fatality:
  • Absolute CFR for the reference group declined from 0.14% (2003–2007) to 0.06% (2019–2024), likely due to improved healthcare access.
  • However, relative CFR remained highest for adults >65 years across all periods, indicating this group faces the greatest mortality risk.

5. Significance and Implications

• Paradigm Shift: The study challenges the traditional view of scrub typhus as a purely northern disease, highlighting significant and growing burdens in the Northeast and South.
• Targeted Interventions:
  • Surveillance: Needs to be tailored to regional seasonal patterns (e.g., different peak months for North vs. South).
  • Clinical Preparedness: Public health messaging and clinical vigilance must shift focus toward older adults, who are now the most susceptible and have the highest mortality rates.
• Resilience: The rapid rebound of reporting post-pandemic suggests Thailand's surveillance infrastructure is resilient, provided resources are maintained.
• Future Research: The findings underscore the need for deeper investigation into vector ecology (chigger species distribution), rodent host dynamics, and the specific ecological drivers of the distinct seasonal patterns observed in different regions.

In summary, this paper provides the most comprehensive epidemiological reconstruction of scrub typhus in Thailand to date, revealing a disease landscape that is geographically broader, seasonally diverse, and increasingly affecting an aging population.