Associations of health literacy with socioeconomic position, health risk behavior, and health statusa large national population-based survey among Danish adults

Roughly four in ten adults in one of the world's most educated and most equal countries cannot reliably find, understand, or use health information. Not in a developing nation with limited schooling infrastructure, but in Denmark. If that number holds in Denmark, the question isn't whether this is a problem elsewhere — it's whether anywhere is exempt. That is the central finding from Majbritt Svendsen and colleagues, who published the first nationally representative health literacy study ever conducted in Denmark. To their knowledge, this is also the largest single-country sample ever measured with this particular instrument. To understand what they found, you need to know what they measured. Health literacy, as defined here, isn't just reading level. It's the ability to access, understand, appraise, and apply health information across three domains: healthcare, disease prevention, and health promotion. The team used the short form of the European Health Literacy Survey, known as the HLS-EU-Q16, which contains 16 items, each asking respondents how easy or difficult a specific health task is. Easy or very easy scores as one. Difficult or very difficult scores as zero. Scores run from zero to sixteen, and they fall into three bands: inadequate, problematic, and adequate. The scale showed strong internal consistency in this sample, with a Cronbach's alpha of 0.90. The item-level results give you a feel for what the bands actually mean. Fewer than three percent of respondents had difficulty understanding how to take a prescribed medicine. However, sixty-two percent had difficulty judging whether health risk information in the media was reliable. Additionally, fifty-three percent struggled with knowing when to seek a second medical opinion. The gap between those tasks is what the three literacy levels are measuring. The study itself was a nationwide cross-sectional survey conducted between December 2016 and February 2017. A random sample of fifteen thousand seven hundred twenty-eight Danish adults aged twenty-five and older was drawn from the national civil registration system, stratified by age, gender, and postal code. Of those, nine thousand seven participated — a response rate of fifty-seven point three percent. Health literacy scores were computed for respondents who answered at least fourteen of the sixteen items, yielding eight thousand four hundred fifty-five usable scores. The design's most powerful feature was individual linkage to nationwide administrative registries. Education came from the Danish Education Registers. Income came from the Danish Income Register, averaged over three years before the survey and split at three hundred twelve thousand Danish kroner — roughly the national average. Social benefit receipt came from the DREAM database. These weren't self-reported figures. They were objective records, which matters enormously when you're trying to isolate the effect of socioeconomic position on health literacy. The analytical approach was adjusted multinomial logistic regression — a method that lets researchers ask, for example, whether lower education predicts lower health literacy even after holding income, age, and migration background constant. So, what did they find? The headline number is that eight point two percent of respondents had inadequate health literacy and thirty point nine percent had problematic health literacy. Together, that's just under thirty-nine percent — nearly four in ten. The median score across all respondents was thirteen point zero. That burden was not evenly distributed. Men had substantially higher odds of inadequate health literacy than women — an adjusted odds ratio of two point three. Younger adults were overrepresented in the inadequate group, which runs counter to the intuition that older people, navigating more complex health decisions, would struggle more. Immigrants made up ten percent of the inadequate group but only six point one percent of the adequate group. Education showed a clear stepwise pattern: twenty-one percent of the inadequate group had only basic schooling as their highest credential, compared with fifteen point five percent in the adequate group. Income followed the same gradient — thirty-nine percent of the inadequate group earned below the national average, versus twenty-six percent of the adequate group. Additionally, a third of those with inadequate health literacy were receiving social benefits, compared with nineteen percent of those with adequate literacy. These patterns held after adjustment. Migration background, education, income, and social benefit receipt all remained statistically significant predictors of low health literacy in the multivariable models. This isn't just education correlating with everything else and pulling the other variables along. Each factor carries independent weight. Now here's where the findings get interesting — and a little counterintuitive. You might expect low health literacy to map neatly onto the full suite of health risk behaviors: more smoking, heavier drinking, and less exercise. Svendsen and colleagues found a more selective pattern. In their adjusted models, smoking showed no significant association with inadequate health literacy. High alcohol consumption also did not show a significant association — the paper actually notes that never drinking above recommendations was associated with higher odds of inadequate health literacy, a pattern that sits uncomfortably with simple assumptions about literacy and behavior. What did hold up, and strongly, were two longer-term indicators. Sedentary behavior — physical inactivity — was associated with inadequate health literacy at an adjusted odds ratio of two point three one, with a confidence interval of one point eight one to two point nine five. Obesity, defined as a body mass index above thirty, was associated with an adjusted odds ratio of one point seven eight. These aren't the same as choosing to smoke a cigarette or drink a glass of wine. They're cumulative, embodied states — the kind that build over years. The authors are careful not to claim causation, but the pattern they describe suggests health literacy is more clearly linked to these chronic, accumulating risk factors than to discrete acute behavioral choices, at least once socioeconomic confounders are controlled. The health status findings reinforce this picture. Self-assessed health showed one of the strongest associations in the entire study. Compared to people with adequate health literacy, those with inadequate health literacy had an adjusted odds ratio of four point zero three for reporting poor or very poor health. That's a fourfold difference in odds. Even at the problematic level — not the worst category, just below adequate — the odds ratio for poor self-rated health was one point nine nine. Sickness-absence compensation, used in this study as a proxy for long-term health problems, was also significantly associated with lower health literacy in both the unadjusted and adjusted models. The policy framing Svendsen and colleagues apply is worth considering. Denmark has universal healthcare. Cost is not a barrier to accessing a doctor or a clinic. And yet four in ten adults still struggle to navigate health information. That means the problem isn't primarily financial — it's structural. The information and systems that people are expected to engage with are more complex than a substantial portion of the population can reliably manage. The authors call for what they term "universal health literacy precautions" — designing health services and communications to work for people across the full spectrum of literacy, rather than expecting individuals to rise to meet the system's demands. The recommendation is system-level adaptation, not individual remediation. There are real limitations here. The cross-sectional design means no causal conclusions are possible — every association described is exactly that, an association. The response rate of fifty-seven percent is solid for population surveys, but the authors acknowledge that the most vulnerable groups may be underrepresented among those who responded, which likely means the thirty-nine percent figure is a conservative estimate. Web-based survey collection also skews toward higher literacy, and the team ran sensitivity analyses accordingly. The estimates, they note, could possibly exaggerate the true effect size, though the direction of bias in prevalence is likely toward underestimating how widespread low health literacy actually is. The strengths are considerable. This is the largest single-country sample ever collected with the HLS-EU-Q16. The registry linkage gives the socioeconomic variables a quality that self-reported data simply can't match. And using a validated European instrument means these Danish findings can be directly compared to other populations measured the same way across the continent. What this study does is establish a baseline, with unusual precision. It tells us how common low health literacy is, who carries the burden, and which health outcomes are most strongly associated with it. Svendsen and colleagues call explicitly for longitudinal work to clarify causal pathways, and for systematic organizational change — not just awareness campaigns, but structural redesign. The measurement is done. The harder work is next. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.