Infection-linked cancers hit 2.3 million cases as wearable research reframes health data

That figure comes from the International Agency for Research on Cancer. It anchors a week in which digital health research kept complicating the simple story that more data equals better outcomes. The same day the cancer estimate landed, Northwestern University published an iScience paper showing that something as ordinary as the breath cycle changes how fast a person reacts to a visual cue.

Northwestern's study, dated 29 September, found that people responded faster on an exhalation than on an inhalation, by an average of 41 milliseconds. That is roughly one twenty-fifth of a second. Responses during the pauses between breaths were also faster than inhalation, by 21 milliseconds. It is the first experiment to measure response speed across all phases of respiration, including those pauses.

What the cancer study actually counted

The IARC team, part of the World Health Organization, estimated that infections were responsible for 2.3 million new cancer cases in 2024. That is 12 per cent of the global total, according to CBC News. Researchers first identified which infections are known to cause cancer. They then calculated what share of those cancers could be tied to specific agents, and applied those fractions to global databases of new cases.

Helicobacter pylori, a bacterium that raises the risk of stomach cancer, accounted for 760,000 cases, mostly in eastern Asia. Human papillomavirus followed at nearly 750,000 cases worldwide, with high rates in sub-Saharan Africa. HPV causes 100 per cent of cervical cancer cases, along with anal, vulvar, vaginal, penile and some head and neck cancers.

Hepatitis B and hepatitis C were linked to a significant number of liver cancers. For the first time, the researchers expanded the list to include 16 additional cancers, adding HIV as a cause of cervical cancer and Merkel cell polyomavirus, a rare virus tied to an aggressive skin cancer.

"There's been new infectious agents identified in this time as being cancer-causing, and even the infectious agents that we knew about in the past, the evidence has increased about the number of different cancers that they can cause," said Gary Clifford, a cancer epidemiologist and study co-author, according to CBC News.

Epstein-Barr virus, which was not counted in past estimates, was linked to about 260,000 cases in 2024, including nasopharyngeal cancer, some gastric cancers and Hodgkin's lymphoma. The same virus is a key risk factor for multiple sclerosis. Unlike HPV or hepatitis, there is no vaccine or antibiotic treatment for it.

Geography dominates the picture. High-income countries that built screening and vaccination programmes over recent decades are doing better than lower- and medium-resourced countries where access to prevention and treatment is limited, the study found. In North America the burden is much lower, but H. pylori and HPV remain the most common sources. In Canada, a 2019 study put the share of cancers attributable to infections at four per cent, according to CBC News.

Breath, reaction time and the limits of a lab finding

The Northwestern experiment sits at the other end of the scale from global cancer estimates: 35 adults aged 18 to 33, a computer task and a device under the nostrils to measure airflow. Participants pressed a space bar as fast as possible when a red square turned yellow. They completed the task twice, once before and once after a nap or an eight-hour sleep in the lab.

Ken Paller, the study's senior author and a psychology professor at Northwestern, said the mechanism is not yet known. "We don't yet know exactly how they are linked, but we suspect neurophysiological efficiency, because other studies have shown that various brain oscillations are synchronized with the rhythms of one's breathing," he said in the university's release.

Lead author Erika M. Yamazaki, a former member of Paller's lab, framed the result as a bridge between two fields that rarely share a table. "Study of the brain and body connection is still a new field of research, which makes the study findings all the more exciting," she said.

The lab is now running related work funded by the National Institutes of Health and led by Yamazaki, focused on sleep apnea, a condition the release describes as often under-diagnosed and inadequately treated. The researchers hope noninvasive strategies for improving sleep, sometimes called sleep engineering, can counter some consequences of poor sleep.

For anyone reading this as a wearable story, the caveats matter. Forty-one milliseconds is an average across 35 young adults in a controlled task, not a promise that a consumer device can time your breath and sharpen your reflexes. The study measured airflow with a lab device, not a wristband.

AI shame and the guidance gap

Guidance, or the lack of it, is the thread running through a third study published this week. Nieman Journalism Lab reported on 28 September on research in Information, Communication and Society titled "Guidance over guidelines? Unpacking the uses and concerns of generative AI in communication science."

Researchers from National Yang Ming Chiao Tung University in Taiwan, the University of Amsterdam and Utrecht University surveyed 1,138 communication scholars across 77 countries and held four focus groups. Nearly 70 per cent of respondents had already used generative AI for at least one research-related task, yet most considered AI use inappropriate for most research tasks. Just over half said using AI for writing texts, 55 per cent, and generating code, 50.4 per cent, were appropriate.

The authors argue that the absence of institutional guidance forces researchers to invent their own standards, producing stigma. Disclosing AI use can invite doubts about credibility; staying silent can slow a career. The study calls this tension a transparency paradox and names the resulting feeling "AI shame."

"Simply forbidding the use of genAI, or pretending that it is not being used, is neither sustainable nor constructive," the authors write.

Their recommendation is a disclosure-positive culture in which individuals, institutions and journals require researchers to state the stages of work where generative AI was used. That is a governance argument, not a technology one. It lands in the same week as a marketplace listing for an AI study tool called Studlark, priced at 2,999 US dollars on SideProjectors. The tool offers private study rooms, group chat and an on-demand AI assistant with monthly credit plans at 5 and 11 dollars.

The listing is not research and should not be read as evidence of anything. It is useful as a marker of how fast AI features are being packaged for students while the scholarly debate about disclosure is still unresolved. The buyer receives source code, database migrations and setup documentation; the seller keeps the Supabase project, environment files and API credentials.

What connects these three papers

Each study, in its own way, is about measurement and the assumptions baked into it. IARC counted 2.3 million cancer cases by deciding which infections qualify as causes and what fraction of each cancer type they explain. That method is why Epstein-Barr virus appears in the 2024 estimate when it did not before, and why the authors can say the burden is larger without claiming the world got sicker.

Northwestern counted milliseconds in a lab task and found a difference that is real but small, then said plainly that the link between breathing and response is not understood. The communication scholars counted self-reported AI use and found a gap between what researchers do and what they say is acceptable, then argued that the gap is a policy failure rather than a personal one.

Read together, they suggest that the digital health conversation is shifting away from sensors and toward the rules that govern what the numbers mean. A wearable that tracks breathing, a database that tracks diagnoses and a survey that tracks AI use all depend on someone deciding what counts, what gets disclosed and what stays in the footnote.

The cancer paper puts a number on prevention: 2.3 million cases in 2024 tied to agents that include vaccine-preventable viruses and a treatable bacterium. Clifford's point, as quoted by CBC News, is that the tools exist and the implementation lags. "We really still have a lot of work to do implementing things that we know work and we've known for a long time but haven't really made it into all populations in the world," he said.

That is the least glamorous finding of the week and probably the most consequential. It does not need a wearable, a model or a dashboard. It needs vaccination, screening and antibiotics delivered where the 760,000 stomach cancer cases and 750,000 HPV cases actually are.