One in Three Adults Had an Infection Within a Year in Denmark’s Massive New Health Cohort
A new nationwide resource built from Denmark’s health surveys and medical registries has revealed the scale of infection across adult life with unusual precision: nearly one in three adults experienced a medically documented infection within a year of completing a health survey. The Danish Infection Cohort contains information on 609,224 adults and records more than 315,000 infections, giving researchers an unprecedented opportunity to examine how infectious disease intersects with smoking, body weight, exercise, diet, alcohol consumption, mental health and socioeconomic circumstances. Unlike many infection databases, which focus on a single pathogen or disease, the new cohort spans a broad spectrum of infections treated in hospitals and the community. Its creators say the database could help answer questions that have remained difficult to study, including why some people develop severe infections, how lifestyle alters infection risk, and whether apparently ordinary infections contribute to long-term disease.
The cohort was assembled by linking respondents to the Danish National Health Survey conducted in 2010, 2013, 2017 and 2021 with several national registries. Denmark’s health system assigns every resident a unique personal identifier, allowing information from separate databases to be connected at the individual level while tracking migration and death over time. The researchers used the Civil Registration System to establish each person’s vital and residency status, the National Patient Registry to identify hospital-based diagnoses, and the National Prescription Registry to identify anti-infective medicines dispensed by community pharmacies. The survey supplied information that is usually missing from routine medical records, including self-reported height and weight, smoking status, physical activity, diet, alcohol use, education, employment, social support, general health and mental health. This combination transforms disconnected administrative records into a longitudinal epidemiological map of infection and its potential risk factors.
The analysis began with 706,940 people who completed one of the four surveys. After restricting the study to adults aged 18 or older, requiring people born outside Denmark to have lived in the country for at least five years before the survey, and including only the first survey for people who participated more than once, 609,224 individuals remained. The researchers then followed each participant for 365 days after an assumed index date of May 1 in the relevant survey year. Because exact questionnaire completion dates were unavailable, May 1 served as a common reference point. Follow-up ended at the first recorded infection, emigration, death or completion of the one-year period. Importantly, the cohort did not require infections to be new: a diagnosis or treatment could represent either an incident infection or an ongoing condition, depending on the research question.
The team divided infections into two broad categories reflecting where care was received. Hospital-diagnosed infections were identified through primary or secondary International Classification of Diseases codes recorded during inpatient, emergency or hospital outpatient encounters. Community-treated infections were identified through prescriptions for systemic anti-infective medicines redeemed at pharmacies. In Denmark, systemic anti-infectives are prescription-only, making pharmacy records a useful way to capture infections managed outside hospitals. The researchers excluded medicines prescribed for prophylaxis, such as malaria prevention or surgical protection, as well as drugs used for noninfectious conditions such as acne or rosacea. To avoid counting repeated records for the same episode, they applied rules that linked a hospital diagnosis to a community prescription within 30 days, retained the primary diagnosis when several appeared during one encounter, and counted only one prescription when multiple anti-infectives were dispensed on the same day.
Overall, 196,980 people—32.3 percent of the eligible cohort—had at least one infection during the year after their survey response. Of these, 11,850 had a hospital-diagnosed infection and 185,130 received treatment in the community. Since individuals could experience more than one infection, the total number of recorded episodes rose to 315,689: 25,385 hospital-diagnosed infections and 290,304 community-treated infections. Respiratory infections, including pneumonia, dominated the hospital records, accounting for 8,858 episodes, or 34.9 percent of all hospital-diagnosed infections. Urinary tract infections were the next most frequent hospital category, with 4,225 episodes, representing 16.6 percent. Among community treatments, antibiotics accounted for 233,994 episodes, or 80.6 percent of the total. Beta-lactam antibiotics made up 68.0 percent of antibiotic-treated infections, while macrolides accounted for 14.3 percent.
The results also exposed striking differences between people whose infections were treated in hospital and those treated in the community. People with hospital-diagnosed infections tended to be older and had higher rates of conditions such as hypertension, cancer and short- or long-term mental illness. They also used prescription medicines more frequently before their infection, suggesting a greater burden of underlying disease. Mortality was highest in the hospital-diagnosed group, consistent with the fact that hospital records preferentially capture more severe infections. Community-treated infections were more common among women. Across the infection groups, smoking and body-mass-index data were relatively complete, with only 3.7 to 5.2 percent of smoking values and 5.5 to 7.5 percent of body-mass-index values missing. Current smokers represented roughly 20 percent of participants with infections, while approximately 16 to 18 percent were classified as obese, although these descriptive figures do not establish that smoking or obesity caused the infections.
The pattern changed markedly across survey years. The proportion of participants with a hospital-diagnosed infection increased from 19.7 percent in the 2010 survey group to 31.9 percent in the 2021 group. At the same time, the share with a community-treated infection fell from 32.0 percent to 20.1 percent. The authors caution that these shifts cannot be interpreted as a simple rise in severe infections or a decline in minor infections. Changes in testing, prescribing, health-care use, coding practices, population composition and the COVID-19 pandemic could all have altered the recorded numbers. During the pandemic, respiratory-virus testing became more routine in hospitals, potentially identifying infections that would previously have gone undiagnosed. Likewise, changes in clinical guidelines, antibiotic prescribing and public behavior could have affected pharmacy records. Calendar time will therefore be a crucial variable in future studies using the cohort.
The database’s power comes from joining behavioral information to medical events, but that design also introduces limitations. Survey participation ranged from 54 to 60 percent across the relevant years, and response rates were lower among some groups, including young men, older women, unmarried people and ethnic-minority populations. If people who do not respond differ systematically from respondents, estimates may not represent the entire Danish population. Researchers can use nonresponse weights or statistical imputation to reduce some of this bias, but such methods cannot recover information that was never measured. Self-reported health conditions may also disagree with registry diagnoses; for some conditions, particularly depression, survey responses may capture illness more effectively than hospital records, while for others the registry may be more reliable.
The infection definitions themselves are deliberately broad and should be adapted to individual research questions. A single diagnostic code may identify a genuine infection, but it may also fail to capture infections that never reach hospital care or are coded nonspecifically. Conversely, a prescription indicates that an anti-infective was dispensed, not necessarily that a laboratory-confirmed infection was present. The researchers note that hospital infection codes often have high positive predictive value—meaning that most coded cases are genuine—but their sensitivity can be low, so many true infections may be missed. Previous validation studies cited by the team found that coding accuracy varies substantially by pathogen and period. Registry records may therefore be strong for identifying serious, medically attended disease but less suitable for estimating the total number of infections in the population.
That balance between scale and imperfection makes the Danish Infection Cohort less a final answer than a platform for new investigations. Researchers could examine whether physical activity changes the relationship between obesity and infection, whether smoking predicts hospitalization rather than infection itself, or how educational and economic conditions shape access to treatment and recovery. The cohort could also be linked to laboratory databases, hospital medication records, death certificates and income or employment registries. Such connections may help trace the aftermath of infection, including cardiovascular disease following respiratory illness, recurrent infections in people with chronic conditions, or long-term effects that are invisible in short clinical studies. The researchers plan to add data from the 2023 and 2025 health surveys as they become available. With nearly 200,000 adults already documented as having medically treated infections and more than 315,000 episodes recorded, Denmark has created one of the most detailed population-level laboratories yet for understanding how infections move through society—and why their consequences can be so unequal.
Subject of Research: Population-based infectious disease epidemiology using linked Danish health survey and national registry data
Article Title: The Danish Infection Cohort: a resource for population based infectious disease epidemiology
Article References: Cocoros, N. M., Oren, E., Darvalics, B. et al. “The Danish Infection Cohort: a resource for population based infectious disease epidemiology.” European Journal of Epidemiology 41, 681–691 (2026). Original research article
Image Credits: AI Generated
DOI: 10.1007/s10654-026-01410-5
Keywords: Danish Infection Cohort, infectious disease epidemiology, national health registries, health surveys, hospital-diagnosed infections, community-treated infections, antibiotics, respiratory infections, lifestyle factors
Tags: broad spectrum infectious disease researchDanish infection cohortepidemiological studies in Denmarkhealth database linkage methodologyinfection prevalence in adultsinfection risk factors and outcomeslinks between lifestyle and infection risklong-term health impact of infectionsmedical registry data analysisnationwide health surveys Denmarkpopulation-based infectious disease epidemiologysocial determinants of infectious diseases
