Prevalence of Helicobacter pylori infection among young adults aged 18–25 years and development of machine learning-based screening models
This study reveals a 29.1% prevalence of *Helicobacter pylori* infection among young adults aged 18–25 and demonstrates that a support vector machine (SVM) model, utilizing factors such as family size and testing history, effectively predicts infection risk for large-scale screening.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Inside the human stomach, a tiny bacterium called Helicobacter pylori often lives quietly, but it holds a powerful influence over long-term health. This microbe is the primary driver behind many stomach and intestinal diseases, ranging from common inflammation to serious cancers. While it is a widespread issue globally, scientists have long recognized that the timing of intervention matters deeply. Removing this bacterium in young adulthood offers a unique opportunity: it is a moment when the stomach lining is still resilient, and the cost of prevention is far lower than the cost of treating advanced disease later in life. Despite this clear window for action, there has been a lack of specific data on how common the infection is among people in their late teens and early twenties, and a shortage of simple, effective tools to identify who needs testing before symptoms appear.
A team of researchers in China set out to fill this gap by focusing on a specific group: university students aged 18 to 25. They gathered a large group of volunteers from Shihezi University and asked them to undergo three different non-invasive tests to check for the presence of the bacteria. These tests included a blood test for antibodies, a test that looks for the bacteria in stool, and a breath test that detects a specific chemical reaction caused by the microbe. To ensure accuracy, the researchers classified a person as infected only if at least two of these three tests came back positive. Alongside these medical checks, the students answered a detailed questionnaire about their lives, covering everything from their family size and household income to their history of stomach problems and whether they had ever been tested for the bacteria before.
The study involved 832 young adults, and the results revealed that nearly 30 percent of them were carrying the infection. While this rate is lower than the average seen in the general Chinese population, it is high enough to warrant attention, especially given that this age group is considered the ideal time for prevention. The researchers then turned to computer algorithms, a branch of artificial intelligence known as machine learning, to see if they could predict who was infected just by looking at the answers in the questionnaire. They tested three different mathematical approaches to build a screening tool, training them on most of the data and then checking their accuracy on the remaining portion.
The results showed that one specific type of algorithm, known as a support vector machine, performed the best at distinguishing between infected and uninfected individuals. This model did not rely on complex medical history or invasive procedures; instead, it identified five key factors that strongly influenced the risk of infection. The most significant predictor was the size of the family a person grew up in, with larger families showing a higher risk. Other important factors included whether the person had been tested for the bacteria before, their annual household income, whether they had ever undergone a stomach camera examination, and whether they owned a pet at home. The model found that people from larger families, those with lower incomes, and those who had never been tested before were at the highest risk.
The researchers found that this computer-based screening tool could successfully identify the majority of infected individuals while keeping false alarms relatively low. It offered a way to sort through a large population quickly, flagging those who would benefit most from further medical testing and treatment. The study suggests that by using simple questions about family life and economic status, health officials could create a practical, low-cost method to screen young adults for this infection. While the model was built using data from a single university and may need further testing in other settings, it provides a promising blueprint for catching the infection early, potentially preventing stomach diseases before they have a chance to develop. The work highlights that even in a young, educated population, the infection remains common, and targeted screening based on everyday life factors could be a vital step toward better public health.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.