Inappropriate antibiotic use is a significant contributor to infections with antimicrobial resistant bacteria, which cause around 1.27 million deaths annually (1). Such inappropriate use arises from a complex interplay of structural and individual determinants, including inaccurate prescriptions in the face of clinical uncertainty, system-level shortcomings, patient expectations and practices, and mistaken beliefs about antibiotic benefits. In this context, individuals’ behaviors related to infection prevention in general and antibiotic use in particular may be decisively shaped by their health literacy (2). The meaning of the term health literacy (HL) has evolved significantly since its earliest use in the context of school-based health education (3, 4), but it can be broadly defined as the ability to find health-relevant information, to understand and evaluate this information, relate it to one’s life situation, and to use it to maintain and promote and maintain health (5–7). Limited HL may influence (in)appropriate antibiotic use in several ways. For example, a patient’s belief that antibiotics are effective for flu, together with a paucity of appropriately communicated information to correct this belief, could increase antibiotic use. Similarly, a patient’s misunderstanding that “antibiotic resistance” refers to the patient’s body becoming tolerant could result in a belief that it is better to take fewer or lower doses (8–10).

Large-scale surveys across Europe have revealed that a substantial proportion of adults have limited HL. For example, a survey in eight European countries found that around 47% of people have limited health literacy (11). In Germany, around half of the population report difficulties in finding, understanding, appraising and applying information on disease prevention (7, 12). In particular, representative German data using a brief measure of self-reported HL, the HLS-EU-16, indicate that approximately 56% of adults have adequate HL; 32% have problematic HL and 12% have inadequate HL. However, HL varies considerably between different population subgroups, with individuals with lower socioeconomic status and immigrants having, on average, lower HL (12).

Refugees and asylum seekers constitute one such subgroup that experiences a confluence of structural disadvantages and limited health literacy, placing them at heightened risk for inappropriate antibiotic use. There are at least three reasons for this. First, refugees and asylum seekers often encounter substantial challenges, including disrupted education, traumatic experiences, and language barriers, that can severely limit their HL. Once in the host country, they must navigate an unfamiliar and often complex health system, access new forms of health information, and manage their own health as well as the health of their families, frequently without adequate social support. Existing evidence suggests that refugee populations in high-income countries experience lower HL (13–15), and that this is associated with poor health outcomes (16). Second, refugees and asylum seekers are at heightened risk for infections, for example, due to missed vaccinations or overcrowded and unsanitary living conditions (17). Moreover, these infections are more likely to be caused by bacteria with antimicrobial resistance compared to the general population (18, 19). Third, these structural pressures intersect with heightened risk of inappropriate antibiotic use, due to, for example, differences in healthcare practices and medication access in their countries of origin (20, 21), lack of reliable and accessible information, and communication barriers in the host health system. As a result, limited HL, elevated risk of infections including those associated with antimicrobial resistance and inappropriate antibiotic use may converge in refugee and asylum seeker populations, posing challenges for individual and family health and for broader AMR mitigation efforts (22).

To date, there has been a paucity of research examining relationships between HL and antibiotic-related knowledge, attitudes and practices amongst newly-arrived refugees and asylum seekers in Germany. In this pilot study, we aimed to close this gap by assessing health literacy and antibiotic knowledge, attitudes and practices amongst refugees and asylum seekers in our local setting. Specifically, we aimed to:

Methods

Study Design and Setting

This cross-sectional study used a mixed-methods approach to assess health literacy and antibiotic knowledge among refugees and asylum seekers in Freiburg im Breisgau, a city of approximately 233,000 inhabitants in southwest Germany. Data collection took place from November 2024 to May 2025 at the Freiburg initial reception center for refugees and asylum seekers (Landeserstaufnahmeeinrichtung, LEA) in Freiburg. During the recruitment period for this study, the LEA had between 367 and 512 residents, mostly of whom were male (78% over the whole recruitment period). The most common reported countries of origin amongst LEA residents during the recruitment period were Türkiye (30% of LEA residents), Syria (26%), Afghanistan (14%) and Algeria (11%). Quantitative data are reported here; qualitative interview results will be reported separately.

Participants and Recruitment

Eligible participants were LEA residents aged 16–65, who were literate in one of the study languages (Arabic, Persian, Pashto, Turkish, or English), and had been in Germany for less than 12 months. The study languages were chosen due to available language skills within the study team. However, the languages also reflected the countries of origin of LEA residents during the recruitment period: of LEA residents with a reported country of origin, more than 90% were from a country where one of the study languages is a primary and/or official language.

A convenience sampling approach was used. Trained bilingual interviewers approached individuals in the waiting area of the LEA health clinic and explained the purpose of the study. Those interested were provided with study information and invited to participate. All questionnaires were self-completed, but bilingual staff were available to assist if required. Participants completed the questionnaires in paper form in private spaces within the LEA to ensure confidentiality. Participation typically took 20–30 minutes.

The study was approved by the Ethics Committee of the University of Freiburg (reference number 24-1300-S1). Informed consent was obtained from all participants in their preferred language. Participation was voluntary, with the option to withdraw at any time without consequences. Data were pseudonymized and stored securely in accordance with data protection standards.

Questionnaires

The questionnaires consisted of several components, designed to measure HL, antibiotic knowledge and use and demographic variables.

The HLS-EU-16 questionnaire was used to assess general HL (11, 23, 24). This tool covers the ability to access, understand, appraise, and apply health-related information in three domains (health care, disease prevention and health promotion). Participants rate how easy they find various health-related tasks, such as understanding a doctor’s instructions or understanding health-related advice from the media, on a four-point Likert scale (1 = very difficult; 2 = difficult; 3 = easy; 4 = very easy).

The omega total reliability coefficient for the HLS-EU-16 was 0.926 and Cronbach’s alpha was 0.896, indicating good reliability.

Participants provided information on age group, gender, country of origin, education, time since arrival in Germany. In addition, participants were asked to rate their general health on a five-point Likert scale (1 = very poor; 2 = poor; 3 = fair; 4 = good; 5 = very good) and to indicate whether or not they had had symptoms in the last 2 weeks suggestive of infection (fever; sneezing, runny nose, coughing etc.; diarrhea; skin rash; any other infection symptoms).

All participant materials were available in Arabic, Persian (Dari), Pashto, Turkish, or English, translated by certified translators and/or native speakers familiar with medical terminology.

106 LEA residents took part in the study; demographic characteristics of the sample are shown in Table 1.

Health literacy

We first examined HL in the whole sample. The most common response for all items was “easy”; however, there were notable differences between items (see Table S1 for responses to the individual items). For example, the items regarding getting a second opinion, finding information about mental health problems, and evaluating health information in the media were most likely to be scored as difficult or very difficult whereas over 90% of participants reported finding it easy or very easy to follow instructions from a doctor or pharmacist.

Around half of participants had ‘sufficient’ HL as measured by the HLS-EU-16 (51%); about a third had ‘problematic’ HL (30.8%) and the remaining 18.3% of participants had ‘inadequate’ HL. A one-sample proportion test showed that the prevalence of ‘sufficient’ HL was not significantly different from the prevalence of 55.8% observed in a representative German sample (Jordan and Hoebel, 2013; p = 0.37, 95% CI [0.41, 0.61]).

We next considered relationships between HL and demographic variables (education, age group and language). There was no significant evidence for an increasing trend in HL index across increasing education levels (J-T = 2382, p = 0.057) or age group (J-T = 1542, p = 0.687). Nor did HL index differ significantly between the group reporting very low, low or fair general health and the group reporting good or very good general health ((W = 1017.5, p = 0.275).

However, HL index differed significantly between language groups, χ²(2) = 6.75, p = 0.034. Post-hoc comparisons using Holm’s correction for multiple comparisons revealed that this was because Arabic speakers tended to have higher HL (median = 31.25, IQR = [27.08–34.38]) than Pashto speakers (median = 27.08, IQR = [24.76–30.21]), W = 569, p = 0.034, Cliff’s δ = 0.43. The remaining pairwise comparisons were non-significant; see Fig. 1.

Antibiotic knowledge and use

On average, participants answered 30.2% of the antibiotic knowledge questions correctly (SD = 17.0%). The range was wide (0.0% to 70.0% answered correctly), in part because many participants answered don’t know to many or all questions. The number of don’t know answers did not differ significantly by language group. Responses to the individual items are shown in Table S2. We excluded participants who only gave don’t know answers from further analyses involving antibiotic knowledge (n = 7).

There was no evidence for differences in antibiotic knowledge between different language groups (χ²(2) = 1.89, p = 0.39) or between levels of general health (W = 1091.5, p = 0.695). Nor was there evidence for increasing antibiotic knowledge with increasing education level (J-T = 2152.5, p = 0.064). However, the trend for increasing antibiotic knowledge with increasing age was significant (J-T = 1877.5, p = 0.002).

All participants reported ‘sometimes’ or ‘very often or always’ using antibiotics in ways that would be considered problematic (Fig. 2 and Table S2). There was no evidence for decreasing problematic antibiotic use with increasing education level (J-T = 1816.5, p = 0.075) nor with increasing age (J-T = 1601.5, p = 0.508). However, problematic antibiotic use differed between different language groups (H(2) = 21.54, p < 0.001). Post-hoc pairwise comparisons using Wilcoxon rank sum tests with Holm’s correction revealed significant differences in problematic antibiotic use between all three pairs (Fig. 3): Arabic vs. Dari (p = 0.002), Arabic vs. Pashto (p = 0.02), Dari vs. Pashto (p < 0.001).

Problematic antibiotic use also differed by general health (W = 1485.5, p = 0.018); participants in the very poor, poor or fair general health group had a median problematic antibiotic use score of 8 (IQR: 5.5–11.5), compared to a median of 6 (IQR: 3–10) amongst those with good or very good general health.

The correlation between health literacy index and problematic antibiotic use was significant (ρ = -0.259, p = 0.009), indicating that participants with higher health literacy had lower levels of problematic use. Moreover, the correlation between antibiotic knowledge (percentage correct) and problematic antibiotic use was also significant (ρ = -0.256, p = 0.011), indicating that participants with higher antibiotic knowledge had lower levels of problematic use.

This study assessed health literacy (HL) and antibiotic knowledge among refugees and asylum seekers in Freiburg. Approximately half (51%) reported sufficient HL – a rate remarkably similar that found in the general German population. This similarity is encouraging, and suggests that many refugees and asylum seekers have a foundation of health-information related skills and a corresponding readiness for effective self-management and integration in the German healthcare system. Conversely, 30.8% of participants reported problematic HL and 18.3% reported inadequate HL, indicating that they, in common with approximately the same proportion of the German population, face significant challenges with accessing and understanding health-related information.

These findings contrast with those of previous studies which have indicated that refugees and asylum seekers have lower HL (14–16, 38). In particular, several studies in Sweden have estimated that around 62% of refugees have inadequate or problematic HL using the same measure as we used (14, 16, 38). These discrepancies may arise from differences in sample characteristics, such as countries of origin, language, length of time in the host country, and/or dissimilarities in the healthcare system and the organizational HL of the organizations involved in providing refugee healthcare.

We found significant disparities in HL tied to language: Arabic speakers had significantly higher HL index scores than Pashto speakers. The small size of the Pashto subgroup must be noted, as this increases the uncertainty of the HL estimate in this group. Nevertheless, this result draws attention to the fact that refugees and asylum seekers are not homogeneous with regards to HL and highlights the need for tailored interventions for specific linguistic and cultural subgroups.

Recent antibiotic use was common. Approximately one-third of participants reported having used antibiotics in the preceding month and over half had used them in the preceding year. These findings align with WHO multi-country survey data, which reported that 35% of respondents had taken antibiotics in the past month (25), but suggests usage rates somewhat higher than the general German population (9). Rates of antibiotic use varies widely by country (25, 39) and use in refugees’ home countries is often considerably higher than in Germany (40). We did not ask whether our participants obtained their most recent antibiotics in Germany or elsewhere (e.g., in their home country or in a third country). However, the elevated rate of antibiotic use in our study converges with evidence of high rates of (often clinically inappropriate) antibiotic prescribing for newly-arrived refugees and migrants in Germany (41).

The majority of participants (60.8%) reported obtaining antibiotics, or a prescription for them, from a doctor or dentist. This is somewhat lower than WHO multi-country survey data (81%), though this may be in part due to the inclusion of nurses in the WHO question (25). In addition, however, some participants had obtained antibiotics from friends and family members or had them saved up from a previous occasion. This is in keeping with existing data showing that antimicrobial self-medication is common globally but varies significantly by country and region and is particularly high in many low and middle income countries (21, 42, 43). For example, one recent meta-analysis estimated that 78% of people in low and middle income countries use non-prescribed antibiotics (43).

Antibiotic knowledge was low, averaging only 30.2% correct answers. Age was positively associated with knowledge, but no significant association was found with education, health status, or language group. Problematic antibiotic practices were widespread, with all participants reporting engagement in at least one inappropriate antibiotic practice sometimes and approximately 60% reporting engagement in at least one practice very often or always. Such practices may increase the risk of antimicrobial resistant infections, which are notably common among refugee populations (19). It is important to note that, despite comparable HL to the German population, problematic antibiotic use may reflect health system differences in the refugees’ countries of origin, where antibiotics are frequently sold in open markets without prescription or sufficient regulatory oversight (20, 44). Such environmental factors may strongly shape health behavior after migration, even in the context of adequate HL.

Considering specific antibiotic knowledge items, we found both similarities and differences in our sample relative to other studies. For example, regarding knowledge about use of antibiotics to treat bacterial infections, a global meta-analysis estimated that around 34% of people give incorrect answers to such statements (45); in a German study, around 10% answered incorrectly and a further 15% did not know (26). In our sample, 18% gave the incorrect response and a further 30% chose don’t know. Regarding knowledge about use of antibiotics to treat viral infections, the global meta-analysis estimated that 54% gave incorrect answers (45); 47% of answers in the German study were incorrect (26). In our sample, 34% of participants answered incorrectly and 37% chose don’t know for this question.

Similarities and differences compared to previous studies should also be noted with regard to problematic antibiotic use. For example, do people stop taking antibiotics as soon as they start feeling better? According to the global meta-analysis, around 47% of people report doing this (45); in our sample, 42.4% reported sometimes and 21.2% reported that they very often or always do this. Do people use non-prescribed antibiotics or support others in doing this (e.g., by saving leftover antibiotics, keeping a supply of antibiotics at home, or using antibiotics prescribed for someone else for the same illness)? As noted above, estimated rates of antimicrobial self-medication vary significantly by country and region (21). In Germany, previous studies have found that approximately 20% agree with statements of this nature (9, 26). In our study, 58% of participants reported that they sometimes or very often or always keep leftover antibiotics so that they or their family and friends can use them another time.

Higher HL and better antibiotic knowledge were associated with less problematic antibiotic use. Although this does not demonstrate causality, it suggests that targeting behavior change interventions towards individuals with low HL and/or low antibiotic knowledge could have the greatest potential to improve appropriate antibiotic use in this population. Such interventions should take into account (mis)information and (false) beliefs that are relevant to the participants' country/culture of origin, since typical antibiotic prescribing and cultural practices vary by country and people's use likely made sense in that context (46, 47).

Our findings reinforce that health literacy is shaped by sociocultural, systemic, and regulatory contexts, not simply by individual knowledge or education. Refugees' HL is a strength to build upon, but ongoing education is necessary as they transition into new communities (48). The pronounced language-based disparities illustrate the nuanced impact of cultural and educational background. Refugees from specific language communities – such as Pashto – may be particularly vulnerable and should be prioritized in health promotion activities. Medical staff should be aware that patterns of antibiotic use in refugees’ and migrants’ home countries may differ from those in the host country in heterogeneous ways and address knowledge gaps and misunderstandings in a culturally sensitive way.

This study is subject to a number of limitations. The relatively small number of Pashto and Dari speakers compared to the Arabic group means that caution is warranted in interpreting the subgroup findings; moreover, the Arabic speakers represented a much larger number of countries of origin than the Pashto and Dari speakers, who came almost exclusively from Afghanistan with a small number from Iran. The small subgroup numbers precluded more complex analysis of, for example, interactions between multiple demographic factors. The fact that we only recruited one participant from Türkiye should also be borne in mind, given that this was the most common country of origin amongst LEA residents during the recruitment period. Recruitment in the waiting area for the LEA health clinic may have biased the sample towards those with health-related issues and possibly omitted those facing greater access barriers. Moreover, the use of self-report measures may result in limitations regarding accuracy or veracity of participants’ self-evaluations. Nevertheless, this study provides valuable, real-world data on HL and antibiotic use amongst recently-arrived refugees and asylum seekers in a German setting. The results offer nuanced insights into similarities and differences within this population as well as in comparison with the host population.

In conclusion, health literacy among refugees in Freiburg’s LEA is encouragingly similar to that of the German population, but patchy knowledge about and problematic use of antibiotics, possibly driven by cultural norms and unregulated medication practices in countries of origin, remain major concerns. There is an urgent need for evidence-based interventions to improve safe and effective use of antibiotics amongst refugee and asylum seeker populations, building on their existing HL. Heterogeneity in HL and antibiotic knowledge and use amongst refugees and asylum seekers underscores the need for such interventions to be targeted to the specific needs of individuals and subgroups in order to effectively address knowledge gaps, inappropriate antibiotic practices, and emerging threats such as antimicrobial resistant infections.