Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Knowledge and use of antibiotics in six ethnic groups: the HELIUS study

Knowledge and use of antibiotics in six ethnic groups: the HELIUS study Background: The increase of antimicrobial resistance, mainly due to increased antibiotic use, is worrying. Preliminary evidence suggests that antibiotic use differs across ethnic groups in the Netherlands, with higher use in people of non-Dutch origin. We aimed to determine whether appropriate knowledge and use of antibiotics differ by ethnicity and whether knowledge on antibiotics is associated with antibiotic use. Methods: We performed a cross-sectional study analyzing baseline data (2011–2015) from a population-based cohort (HELIUS study), which were linked to data from a health insurance register. We included 21,617 HELIUS participants of South-Asian Surinamese, African-Surinamese, Turkish, Moroccan, Ghanaian, and Dutch origin. Fifteen thousand seven participants had available prescription data from the Achmea Health Data-base (AHD) in the year prior to their HELIUS study visit. Participants were asked five questions on antibiotic treatment during influenza-like illness, pneumonia, fever, sore throat and bronchitis, from which higher versus lower antibiotic knowledge level was determined. Number of antibiotic prescriptions in the year prior to the HELIUS study visit was used to determine antibiotic use. Results: The percentage of individuals with a higher level of antibiotic knowledge was lower among all ethnic minority groups (range 57 to 70%) compared to Dutch (80%). After correcting for baseline characteristics, including medical conditions, first-generation African Surinamese and Turkish migrants received a significantly lower number of antibiotic prescriptions compared to individuals of Dutch origin. Only second-generation Ghanaian participants received more prescriptions compared to Dutch participants (aIRR 2.09, 95%CI 1.06 to 4.12). Higher level of antibiotic knowledge was not significantly associated with the number of prescriptions (IRR 0.92, 95%CI 0.85 to 1.00). Conclusions: Levels of antibiotic knowledge varied between ethnic groups, but a lower level of antibiotic knowledge did not correspond with a higher number of antibiotic prescriptions. Keywords: Antibiotics, Antibiotic knowledge, Antibiotic use, Ethnic groups Background A recent meta-analysis showed a higher prevalence of The emergence of antimicrobial resistance, along with antimicrobial resistance among migrants in Europe [2]. the steady decline in antibiotic development, has been There is preliminary evidence in the Netherlands that identified as a major health threat for the coming decade the use of antibiotics also differs across ethnic groups, by the World Health Organization (WHO). Increase in with a higher use of antibiotics among people of non- antibiotic use is the main reason for this development Dutch origin [3]. The reason for this difference, however, [1] and as such, antibiotics should only be prescribed is unclear. It could be explained by increased incidence when there is a clear indication for use. of bacterial infections, but, to the best of our knowledge, there is no evidence to support this hypothesis. Alterna- tively, knowledge about antibiotic use might vary across * Correspondence: evdulm@ggd.amsterdam.nl ethnic groups. As expectations and knowledge of the Emelie C. Schuts and Eline van Dulm contributed equally to this work. patient could potentially drive a physician’s decision to Department of Infectious Diseases, Public Health Service Amsterdam, prescribe antibiotics, receiving prescriptions could also Nieuwe Achtergracht 100, 1018, WT, Amsterdam, The Netherlands Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 2 of 17 differ between ethnic groups [4–6]. There are also Ethnicity was defined according to the country of birth cultural-specific approaches to dealing with authority, of the participant as well as that of their parents [12]. being the physician in this setting, which have explained Specifically, a participant is considered to be of non- differences in antibiotic use between countries [7]. Dutch ethnic origin if they fulfill either of the following The HELIUS (Healthy life in an Urban Setting) study criteria (i): they were born abroad and had at least one is a large-scale, population-based cohort study among parent born abroad (first generation) or (ii) they were different ethnic groups, which was established with the born in Netherlands but both their parents were born aim to investigate mechanisms underlying the impact of abroad (second generation). Dutch participants were ethnicity on communicable and non-communicable dis- born in the Netherlands and had both parents who eases [8, 9]. In 2018, approximately 13% of the popula- were born in the Netherlands. After HELIUS data tion of the Netherlands was of non-Western origin [10]. collection, the Surinamese group were further classi- The largest non-Western population groups were fied according to self-reported ethnic origin (obtained individuals of Turkish (2.4%), Moroccan (2.3%) and by questionnaire), into ‘African Surinamese’, ‘South- Surinamese (2.0%) descent [10]. In Amsterdam, approxi- Asian Surinamese’, ‘Javanese Surinamese’ and ‘other/ mately 36% of the population in 2018 was of non- unknown Surinamese’. Western descent [11]. The ethnic groups included in the HELIUS study are the largest ethnic minority groups of Data linkage Amsterdam [9]. Amongst other data, data on antibiotic Permission to link participants’ individual data to outside knowledge were collected. We were able to link these health registries was asked in the written informed con- data at the individual level to data from a health insur- sent form [8]. Of the 22,165 HELIUS participants, 19, ance register on recent antibiotic use. 895 agreed. HELIUS data of these individuals were This study then provides a unique opportunity to de- linked to reimbursement data from the Achmea insur- termine whether knowledge about and use of antibiotics ance company (Achmea Health Database, AHD) from vary between ethnic groups, and if so, whether differ- 2010 until 2015. The AHD, obtained from the largest ences in antibiotic use can be attributed to differences in health insurance company in Amsterdam, contains all knowledge about antibiotics. We hypothesized that healthcare expenditures of every insured participant, antibiotic use differs among ethnic groups as a result of including medications. A trusted third party linked differences in knowledge. data on reimbursed antibiotic prescriptions using an encrypted social security number and returned data without any identifying information. Procedures were Methods in accordance with the General Data Protection Study population and design Regulation [13]. The HEalthy LIfe in an Urban Setting (HELIUS) study is a multiethnic cohort study conducted in Amsterdam, Inclusion and exclusion criteria for present study which focuses on cardiovascular disease (e.g. diabetes), Of the 22,165 participants, we excluded those of Javan- mental health (e.g. depressive disorders), and infectious ese Surinamese or other/unknown Surinamese origin diseases [8, 9]. In brief, baseline data collection took and those with another/unknown ethnic origin because place in 2011–2015 and included people aged 18 to 70 of small participant numbers. For analyses on antibiotic years of Dutch, Surinamese, Ghanaian, Moroccan, and use, we included those who gave permission for data Turkish origin. A random sample of participants, strati- linkage and could be linked to the AHD. To reduce bias fied by ethnic origin, was taken from the municipality for individuals with short-term insurance, we excluded register of Amsterdam. Participants filled in an extensive those who were insured with Achmea for less than 365 self-administered questionnaire (variables included in the days in the year preceding their HELIUS study visit. questionnaire are described elsewhere) [9] and underwent a physical examination during which biological samples Outcome variables were obtained [9]. No information was provided regarding The primary outcomes were level of antibiotic know- appropriate antibiotic use. Between 2011 and 2015, 24,789 ledge and antibiotic use during the year prior to the persons were included. Data collection procedures have HELIUS visit. Level of antibiotic knowledge was based been previously described in detail [9]. Both questionnaire on five questions, used in other studies [4, 6, 14], which data and physical examination data were available for 22, asked the perceived necessity (yes/no) for antibiotic 165 participants. The HELIUS study was conducted in treatment during influenza-like illness, pneumonia, fever, accordance with the Declaration of Helsinki and was ap- sore throat and bronchitis. Using these questions, we proved by the AMC Ethical Review Board. All participants created an overall knowledge score of antibiotic use by provided written informed consent. summing the total number of correct responses, Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 3 of 17 resulting in a score ranging from 0 to 5. A two- Analysis on antibiotic use in the year prior to HELIUS parameter logistic regression model was fitted to the five study visit included all HELIUS participants who were binary items based on the assumptions of item response linked to the AHD and were insured for at least 365 days theory (see Additional file 1). From this model, “higher” with Achmea in the year prior to their HELIUS study and “lower” levels of antibiotic knowledge were defined visit. Determinants for having received ≥1 antibiotic pre- by a knowledge score of ≥4 and < 4, respectively. scription were assessed using logistic regression. The Antibiotic use was obtained from linked AHD data same multivariable approach as above was used for this and was based on the total number of reimbursed antibi- outcome. We also compared antibiotic use during the otics (classified by ATC code J01; anti-infectives for sys- entire period insured at Achmea versus the year prior to temic use) dispensed by community pharmacies from HELIUS study visit to assess differences when consider- 2010 until 2015. We evaluated antibiotic use (yes/no) in ing longer time periods. the year prior to the HELIUS study visit, as well as the Determinants for the total number of antibiotic pre- number of antibiotic prescriptions over the past year scriptions were then evaluated. As this outcome con- and during the entire insured period. tained a high proportion of zero values and was over- dispersed, we used a zero-inflated negative binomial re- Other variables gression model. This model contains two parts: one ac- Independent variables were obtained from the HELIUS counting for zero values in the count distribution (zero- study questionnaire (migration generation; sex; age; level inflated) and another accounting for the over-dispersed of education; marital status; self-reported medical condi- count distribution (negative binomial). Covariates for the tions; smoking; alcohol consumption; difficulty with the zero-inflated part were determined a priori from the Dutch language and perceived health) and physical risk-factor analysis on ≥1 antibiotic prescription. Covari- examination (body mass index (BMI, kg/m )). Variables ates for the negative binomial part were selected from on antibiotic-related behavior were: not having finished covariates with a p-value < 0.2 in univariable analyses antibiotic treatment; having saved antibiotics for later; and variables above this p-value were removed in and ever having asked the general practitioner (GP) for backwards-stepwise fashion. Incidence risk ratios (IRR) antibiotics. Definitions and grouping of variables are comparing the number of antibiotics prescribed over the extensively described elsewhere [8]. past year across levels of determinants were estimated from this model. Statistical analyses Multicollinearity was verified using variance inflation Sociodemographics, health status, antibiotic knowledge factors, while any variable with an inflation factor of ≥4 level and questions on antibiotic use were presented by was considered multicollinear and excluded from the ethnicity. To assess selection bias resulting from AHD model. To understand whether the association between data linkage, the same variables were compared between ethnicity and outcome was modified by demographic participants who were successfully versus unsuccessfully variables, interaction between ethnicity and other demo- linked. Comparisons between ethnic groups were made graphic variables was also assessed in all multivariable using a Pearson’s χ or Fisher exact test for categorical models. data and Kruskal-Wallis rank test for continuous The three variables involving antibiotic-related behav- variables. ior were not initially considered in the final multivariable Analysis on level of antibiotic knowledge included all models. To assess whether ethnic differences in anti- HELIUS participants with available data. Odds ratios biotic use could be explained by patterns of antibiotic- (OR) comparing levels of antibiotic knowledge across related behavior, additional multivariable models includ- determinants and their 95% confidence intervals (CI) ing these variables were constructed for the endpoints (i) were estimated using logistic regression. All variables having received ≥1 antibiotic prescription and (ii) total with an associated p-value < 0.2 in univariable analyses number of antibiotic prescriptions. were included in a full multivariable model and variables Figure 1 provides an overview of all descriptive ana- with a p-value above this level were removed in lysis and modeling used in the study. Significance was backwards-stepwise fashion. Given that the research aim determined using a p-value < 0.05. All analyses were was to determine differences between ethnicity, ethnic conducted with Stata 13.1 (StataCorp., College Station, groups were forced in all models. This multivariable ap- Texas, USA). proach was chosen to not only assess other variables as- sociated with antibiotic knowledge, but also to Results understand the extent of confounding bias when asses- Participants sing the relationship between ethnicity and outcome Of the 22,165 HELIUS participants with available data, variables. 21,617 were eligible after applying exclusion criteria. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 4 of 17 Fig. 1 Overview of descriptive analysis and models used in the study. Abbreviations: HELIUS – Healthy Life in an Urban Setting; AHD – Achmea Health Database Their baseline characteristics, stratified by ethnicity, are their GP for antibiotics ranged from 0.6% in African shown in Table 1. Median age of participants was 46 Surinamese participants to 1.9% in Turkish and Moroc- years (IQR 34 to 55) and 58% were women. The propor- can participants. tion of several medical conditions predisposing individ- As shown in Table 2, there was a significantly lower uals to antibiotic treatment differed by ethnicity. Of odds of individuals with higher level of antibiotic know- these conditions, South-Asian Surinamese participants ledge among all non-Dutch ethnic groups compared to had the highest prevalence of self-reported diabetes mel- Dutch individuals (overall p < 0.001) (Table 2). Across all litus (17%) and cerebrovascular accident (CVA) (6%) non-Dutch groups, second-generation participants had a over the last 12 months. Turkish individuals had more higher level of antibiotic knowledge than first-generation prevalent artery stenosis (10%), severe or chronic fatigue participants; however, results remained significantly (45%) and respiratory diseases (15%), whereas Ghanaians lower compared to the Dutch group. more frequently reported high blood pressure (33%). In multivariable analysis, all ethnic minority groups Excellent perceived health was reported in 12% of Dutch had lower odds for higher level of antibiotic knowledge participants in contrast to 3.3% of Turkish participants. compared to Dutch (overall p < 0.001), although the ef- fect for second-generation Ghanaian participants was not statistically significant. The odds for higher level of Ethnic differences in antibiotic knowledge antibiotic knowledge were higher in all age groups > 25 In several ethnic groups, there were substantial propor- years of age (except for those ≥65) when compared to tions of individuals reporting the need to be treated with ≤25 years of age. Furthermore, women had a significantly antibiotics for illnesses without indication, as shown in higher odds of having a higher level of antibiotic know- Table 1. The number of people reporting to have been ledge compared to males. Lower odds for a higher level treated with antibiotics and not having regularly com- of antibiotic knowledge were found for the following pleted their antibiotic treatment was low across all eth- medical conditions: myocardial Infarction (MI), severe nic groups, ranging from 0.1% in Dutch participants to or chronic fatigue, respiratory diseases and having a BMI 2.1% in Ghanaian participants. Few individuals regularly ≥25. Lower odds for higher level of antibiotic knowledge saved their antibiotics for later use, ranging from < 0.1% were also seen among individuals who regularly or occa- in Dutch participants to 0.3% in Turkish participants. sionally requested antibiotics from their GP or who The percentage of participants having regularly asked regularly or occasionally did not finish treatment. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 5 of 17 Table 1 Characteristics of the HELIUS study population (N = 21,617) by ethnicity Variables Ethnicity Dutch South-Asian African Ghanaian Turkish Moroccan (N = 4564) Surinamese Surinamese (N = 2339) (N = 3614) (N = 3906) (N = 3043) (N = 4151) Sociodemographics Female sex 2475 54% 1672 55% 2535 61% 1434 61% 1980 55% 2392 61% Age in years, median (IQR) 47 (34–58) 48 (35–56) 50 (40–57) 47 (38–53) 42 (31–50) 40 (30–50) Migration generation 1st generation N.A. N.A. 2328 77% 3468 84% 2231 95% 2544 70% 2680 69% 2nd generation N.A. N.A. 715 23% 683 16% 108 4.6% 1080 30% 1226 31% Educational level Unknown 25 0.6% 16 0.5% 36 0.9% 42 1.8% 38 1.1% 38 1.0% No school/elementary school 150 3.3% 437 14% 231 6% 660 28% 1135 31% 1205 31% Lower vocational/lower secondary 646 14% 1010 33% 1477 36% 917 39% 889 25% 694 18% school Intermediate vocational/ intermediate 994 22% 885 29% 1464 35% 578 25% 1020 28% 1294 33% secondary school Higher vocational/university 2749 60% 695 23% 943 23% 142 6% 532 15% 675 17% Marital status Married/registered partnership 1724 38% 1043 34% 766 19% 420 18% 2208 61% 2285 59% Cohabiting 914 20% 311 10% 441 11% 427 19% 132 3.7% 110 2.8% Unmarried/never married 1474 32% 1001 33% 2231 54% 779 34% 761 21% 1010 26% Divorced/separated 356 8% 580 19% 617 15% 656 28% 407 11% 414 11% Widow/widower 87 1.9% 92 3.0% 65 1.6% 23 1.0% 90 2.5% 69 1.8% Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 102 2.2% 521 17% 419 10% 185 8% 336 9% 389 10% CVA/one-sided loss of bodily 160 3.5% 212 7% 261 6% 95 4.1% 196 5% 195 5% function ≤1 day MI incl. ≥half hour chest pain or 233 5% 491 16% 440 11% 225 10% 591 16% 476 12% dotter/bypass operation Severe heart condition 67 1.5% 120 4.0% 105 2.5% 75 3.2% 153 4.3% 58 1.5% Malignant disorder 103 2.3% 70 2.3% 85 2.1% 33 1.4% 73 2.0% 46 1.2% Severe or chronic fatigue 633 14% 1032 34% 956 23% 186 8% 1602 45% 1465 38% High blood pressure 534 12% 720 24% 1230 30% 770 33% 610 17% 546 14% Artery stenosis 85 1.9% 193 6% 181 4.4% 115 5% 348 10% 200 5% Respiratory diseases 345 8% 433 14% 354 9% 117 5% 556 15% 446 11% Serious/persistent intestinal disorders 249 5% 248 8% 308 7% 70 3.0% 433 12% 391 10% Psoriasis 136 3.0% 168 6% 128 3.1% 71 3.1% 154 4.3% 121 3.1% (Chronic) eczema 420 9% 406 13% 370 9% 71 3.1% 471 13% 423 11% Incontinence 309 7% 326 11% 342 8% 108 4.7% 464 13% 300 8% Body Mass Index (kg/m ), median (IQR) 24.1 (21.9– 25.7 (23.2– 27.0 (23.9– 27.9 (25.0– 27.9 (24.6– 27.0 (23.9– 26.7) 28.8) 30.8) 31.2) 31.7) 30.7) Smoking Yes 1129 25% 861 28% 1309 32% 104 4.5% 1240 35% 525 13% No, never 1689 37% 1758 58% 2016 49% 2027 87% 1700 47% 2874 74% No, but ever 1737 38% 413 14% 805 19% 191 8% 648 18% 492 13% Alcohol consumption Never 297 7% 1072 35% 1002 24% 806 35% 2414 67% 3265 84% Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 6 of 17 Table 1 Characteristics of the HELIUS study population (N = 21,617) by ethnicity (Continued) Variables Ethnicity Dutch South-Asian African Ghanaian Turkish Moroccan (N = 4564) Surinamese Surinamese (N = 2339) (N = 3614) (N = 3906) (N = 3043) (N = 4151) Not in previous 12 months 110 2.4% 251 8% 292 7% 408 18% 358 10% 338 9% Monthly or less 436 10% 758 25% 1242 30% 508 22% 367 10% 127 3.3% 2–4 times per month 894 20% 541 18% 873 21% 291 13% 257 7% 87 2.2% 2–3 times per week 1413 31% 262 9% 439 11% 193 8% 132 3.7% 56 1.4% ≥ 4 times per week 1408 31% 147 5% 272 7% 109 4.7% 57 1.6% 16 0.4% Difficulty with Dutch language Yes N.A. N.A. 711 23% 520 13% 1926 83% 2136 60% 1774 46% Perceived health Excellent 541 12% 162 5% 303 7% 226 10% 117 3.3% 166 4.3% Very good 1381 30% 310 10% 571 14% 458 20% 383 11% 384 10% Good 2205 48% 1623 53% 2335 56% 1180 51% 1871 52% 1871 48% Mediocre 402 9% 811 27% 834 20% 383 16% 921 26% 1223 31% Bad 28 0.6% 131 4.3% 101 2.4% 88 3.8% 307 9% 241 6% Antibiotics Knowledge concerning antibiotics Antibiotics effective for influenza 324 7% 554 19% 592 15% 658 29% 744 21% 648 18% Antibiotics effective for pneumonia 4166 92% 2304 77% 3114 77% 1312 58% 2587 73% 2741 73% Antibiotics effective for fever 689 15% 552 19% 679 17% 586 26% 898 26% 614 17% Antibiotics effective for sore throat 672 15% 760 26% 1089 27% 720 32% 1203 34% 978 26% Antibiotics effective for bronchitis 2246 50% 1385 50% 1862 46% 919 41% 1235 35% 1485 40% Higher level of antibiotic knowledge 3638 80% 1996 68% 2737 69% 1248 57% 2128 62% 2528 70% Did not finish antibiotic treatment Yes, regularly 5 0.1% 49 1.6% 44 1.1% 48 2.1% 41 1.2% 44 1.1% Yes, occasionally 332 7% 312 10% 527 13% 174 8% 424 12% 445 12% Always finished or no antibiotics 4203 93% 2646 88% 3524 86% 2053 90% 3104 87% 3361 87% Saved antibiotics for later Yes, regularly 2 0.0% 7 0.2% 9 0.2% 5 0.2% 10 0.3% 6 0.2% Yes, occasionally 37 0.8% 45 1.5% 68 1.7% 62 2.7% 60 1.7% 46 1.2% No, never 297 7% 304 10% 492 12% 146 6% 387 11% 430 11% Not applicable (no antibiotics) 4203 93% 2646 88% 3524 86% 2053 91% 3104 87% 3361 87% Ever asked GP for antibiotics Yes, regularly 38 0.8% 34 1.1% 26 0.6% 36 1.6% 67 1.9% 71 1.9% Yes, occasionally 824 18% 491 16% 607 15% 401 18% 734 21% 634 17% No, never 3674 81% 2482 83% 3441 84% 1835 81% 2744 77% 3074 81% Missing data, n: marital status 128; diabetes 78; stroke 55; myocardial infarction 33; heart condition 83; malignant disorders 137; fatigue 145; high blood pressure 101; artery stenosis 140; respiratory diseases 115; bowel diseases 115; psoriasis 98; eczema 117; incontinence 126; BMI 23; smoking 107; alcohol 127; perceived health 61; AB effective for influenza 614; AB effective for pneumonia 477; AB effective for fever 685; AB effective for sore throat 625; AB effective for bronchitis 663; asked GP for AB 414; did not finish treatment 292; saved AB 325 Abbreviations: IQR Inter Quartile Range, CVA Cerebro Vascular Accident, MI Myocardial infarction, N.A. Not Applicable, GP General Practitioner N.A. Not applicable (categories not applicable due to Dutch ethnicity) All variables are reported as n (%), unless otherwise indicated Answered “yes” to the statements below The Dutch General Practitioners guidelines (and those of other European countries) advise against the use of antibiotics for fever in general or sore throat, as they usually constitute viral infections, with only a few exceptions in both cases. Therefore, antibiotics are in general not appropriate for these conditions Based on a summed score with cutoff determined by an Item Response Theory model (≥4 out of 5 antibiotic knowledge questions correctly answered was considered as having a higher level of knowledge) Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 7 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Sociodemographics Ethnicity <.001 <.001 Dutch Ref Ref South-Asian Surinamese 1st generation 0.49 0.44–0.55 0.53 0.47–0.60 2nd generation 0.56 0.47–0.67 0.60 0.50–0.73 African Surinamese 1st generation 0.51 0.46–0.57 0.53 0.47–0.59 2nd generation 0.75 0.62–0.91 0.79 0.64–0.96 Ghanaian 1st generation 0.31 0.27–0.34 0.31 0.27–0.35 2nd generation 0.64 0.41–0.98 0.74 0.47–1.18 Turkish 1st generation 0.35 0.31–0.39 0.40 0.36–0.45 2nd generation 0.56 0.48–0.65 0.62 0.53–0.74 Moroccan 1st generation 0.51 0.45–0.57 0.56 0.50–0.63 2nd generation 0.71 0.61–0.83 0.75 0.63–0.89 Female sex 1.18 1.11–1.25 <.001 1.32 1.23–1.40 <.001 Age <.001 <.001 < 25 years Ref Ref 25–34 years 1.24 1.01–1.34 1.32 1.16–1.50 35–44 years 0.99 0.81–1.05 1.30 1.14–1.49 45–54 years 0.85 0.71–0.91 1.19 1.04–1.37 55–64 years 0.95 0.78–1.02 1.26 1.08–1.45 ≥ 65 years 1.06 0.84–1.22 1.15 0.95–1.39 Educational level <.001 Unknown Ref No school/elementary school 1.10 0.79–1.55 Lower vocational/lower secondary school 1.27 0.91–1.77 Intermediate vocational/ intermediate secondary school 1.54 1.11–2.15 Higher vocational/university 2.06 1.47–2.87 Marital status <.001 Married/registered partnership Ref Cohabiting 1.15 1.04–1.27 Unmarried/never married 1.08 1.01–1.16 Divorced/separated 0.81 0.74–0.89 Widow/widower 1.10 0.88–1.36 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 0.70 0.63–0.77 <.001 CVA/one-sided loss of bodily function ≤1 day 0.85 0.75–0.97 .015 MI incl. ≥half hour chest pain or dotter/bypass operation 0.71 0.65–0.77 <.001 0.89 0.81–0.98 .017 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 8 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) (Continued) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Severe heart condition 0.61 0.52–0.73 <.001 Malignant disorder 0.97 0.78–1.20 .752 Severe or chronic fatigue 0.80 0.75–0.85 <.001 0.89 0.82–0.95 .001 High blood pressure 0.78 0.73–0.84 <.001 Artery stenosis 0.69 0.61–0.78 <.001 Respiratory diseases 0.68 0.62–0.75 <.001 0.80 0.72–0.88 <.001 Serious/persistent intestinal disorders 0.88 0.79–0.98 .024 Psoriasis 0.76 0.66–0.89 .001 (Chronic) eczema 0.93 0.84–1.02 .121 Incontinence 0.84 0.76–0.93 .001 Body Mass Index <.001 .001 < 18.5 Ref Ref 18.5–25 1.03 0.81–1.31 1.02 0.80–1.31 25–30 0.80 0.63–1.01 0.96 0.75–1.24 30–40 0.67 0.52–0.85 0.86 0.66–1.11 ≥ 40 0.58 0.43–0.78 0.78 0.57–1.08 Smoking <.001 Yes Ref No, never 1.03 0.96–1.11 No, but ever 1.19 1.09–1.30 Alcohol usage <.001 Never Ref Not in previous 12 months 0.89 0.80–1.00 Monthly or less 1.10 1.01–1.20 2–4 times per month 1.27 1.16–1.40 2–3 times per week 1.35 1.22–1.49 ≥ 4 times per week 1.59 1.42–1.78 Difficulty with Dutch language <.001 No Ref Yes 0.65 0.61–0.69 Not applicable 1.70 1.56–1.85 Perceived health <.001 Excellent Ref Very good 0.97 0.85–1.12 Good 0.83 0.73–0.93 Mediocre 0.63 0.56–0.72 Bad 0.48 0.40–0.57 Antibiotics Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 0.51 0.40–0.65 0.60 0.46–0.77 Yes, occasionally 0.57 0.53–0.61 0.59 0.55–0.64 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 9 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) (Continued) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Did not finish treatment <.001 <.001 Always finished or no antibiotics Ref Ref Yes, regularly 0.51 0.39–0.67 0.71 0.54–0.94 Yes, occasionally 0.73 0.66–0.80 0.80 0.73–0.88 Abbreviations: OR Odds Ratio, aOR adjusted Odds Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner Fewer observations in the multivariable model than in the total study population were due to missing observations on certain covariates #We found significant interactions between ethnicity and sex (p = 0.007) and ethnicity and age (p = 0.047) Ethnic difference in antibiotic use (p < 0.001) and multivariable analysis (p < 0.001). In mul- Of the 19,895 HELIUS participants consenting to link tivariable analysis, compared to Dutch individuals, first their data to other health registries, 15,461 were linked and second generation Ghanaian individuals and first- to the AHD (77.7%). Of these 15,461 participants, 15,007 generation Moroccan individuals had significantly higher (97%) were insured for ≥365 days in the year prior to odds of receiving ≥1 antibiotic prescription. Adding vari- their HELIUS study visit. Additional file 2: Table S1 ables on antibiotic-related behavior and level of anti- shows the characteristics of the study participants linked biotic use knowledge to the multivariable model did not versus not linked to the AHD. Participants present in change these associations. the AHD register had a lower level of education, higher Table 5 shows the results from the analysis on the as- prevalence of medical conditions, and less often had sociation between ethnicity and total number of anti- higher levels of antibiotic knowledge. biotic prescriptions received in the year prior to the Table 3 describes antibiotic use according to ethnicity HELIUS study visit. Differences across ethnic groups for participants registered in the AHD. In total, 31,530 were observed overall for the number of antibiotic pre- antibiotic prescriptions were recorded over the study scriptions in both univariable and multivariable analysis period. The proportion of participants receiving ≥1 anti- (both p = 0.004). First-generation African Surinamese biotic prescription in the year prior to their HELIUS and Turkish migrants had a significantly lower number study visit was highest among first-generation Turkish of antibiotic prescriptions compared to individuals of participants (25%) and was comparably high among Dutch origin. Only second-generation Ghanaian partici- second-generation Ghanaian and first-generation pants has more prescriptions compared to Dutch partici- Moroccan participants (both 25%). The proportion of pants. Furthermore, female sex, diabetes mellitus, MI, participants receiving ≥1 antibiotic prescription in the malignant disorder, respiratory disease, eczema and year prior to the HELIUS study visit was lowest in Dutch worse perceived health were significantly associated with and second generation South-Asian Surinamese partici- a higher number of antibiotic prescriptions. pants (both 16%). Having a higher level of antibiotic knowledge was not When considering the entire period during which par- significantly associated with the number of prescriptions ticipants were insured at Achmea prior to the HELIUS when included in multivariable analysis (p = 0.446). No study visit (median 6.0 years, IQR 5.0 to 6.0), the propor- significant interactions between ethnicity and sex or tion of participants receiving ≥1 antibiotic prescription education were observed. Finally, adjusting the associ- was highest among first generation Turkish participants ation between ethnicity and antibiotic use for antibiotic- (69%) and lowest in second-generation Ghanaian partici- related behaviors did not change these associations. pants (49%). The mean number of prescriptions during the entire insured period was comparable to the mean Discussion number of prescriptions in the year prior to HELIUS Our study shows that knowledge on the need to use an- study visit for all ethnic groups (Table 2). tibiotics for treatment is lower among all ethnic minority groups compared to Dutch, with second generation eth- Determinants of antibiotic use and number of nic minorities showing higher levels of knowledge com- prescriptions pared to first generation migrants. We also observed Table 4 shows the results from the analysis on the asso- ethnic differences in the use of antibiotics, with a higher ciation between ethnicity and having received ≥1 anti- proportion having received at least one prescription, but biotic prescription in the year prior to the HELIUS study a lower mean number of antibiotic prescriptions among visit. Differences across ethnic groups were observed some ethnic minority groups compared to Dutch. The overall for any antibiotic prescription in both univariable only ethnic group with a significantly higher number of Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 10 of 17 Table 3 Antibiotic use in participants linked to AHD (N = 15,007) stratified by ethnicity Ethnicity Dutch South-Asian Surinamese African Surinamese Ghanaian Turkish Moroccan (N = 2071) 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen (N = 1645) (N = 452) (N = 2334) (N = 432) (N = 1789) (N = 84) (N = 2102) (N = 776) (N = 2119) (N = 857) Duration of insurance at 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 Achmea (in years) (4.0–6.0) (6.0–6.0) (4.0–6.0) (6.0–6.0) (5.0–6.0) (6.0–6.0) (4.0–6.0) (96.0–6.0) (5.0–6.0) (6.0–6.0) (4.0–6.0) between 2010 and 2015, median (IQR) Within year prior to HELIUS study visit Participants with ≥1 ABP 16% 22% 16% 17% 17% 22% 25% 25% 19% 25% 17% Number of ABP among all participants included in the AHD Mean 0.26 0.39 0.26 0.28 0.28 0.33 0.55 0.40 0.34 0.41 0.28 Median (IQR) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.5) (0.0–1.0) (0.0–0.0) (0.0–1.0) (0.0–0.0) Number of ABP among participants with ≥1 ABP Mean 1.66 1.75 1.59 1.58 1.64 1.51 2.19 1.57 1.78 1.64 1.63 Median (IQR) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 2.00 (1.0–3.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) During entire insured period Participants with ≥1 ABP 51% 63% 53% 56% 51% 62% 49% 69% 59% 67% 54% Number of ABP per year among all participants included in the AHD Mean 0.31 0.46 0.30 0.31 0.30 0.35 0.34 0.44 0.38 0.43 0.33 Median (IQR) 0.17 (0.0–0.3) 0.17 (0. 0–0.6) 0.17 (0. 0–0.3) 0.17 (0.0–0.3) 0.17 (0.0–0.3) 0.17 (0.0–0.5) 0.00 (0.0–0.7) 0.25 (0.0–0.7) 0.17 (0.0–0.5) 0.17 (0.0–0.6) 0.17 (0.0–0.5) Abbreviations: ABP Antibiotic Prescription, IQR Inter Quartile Range Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 11 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values Sociodemographics Ethnicity <.001 <.001 .004 Dutch Ref Ref Ref South-Asian Surinamese 1st generation 1.57 1.33–1.85 1.05 0.86–1.27 1.04 0.85–1.26 2nd generation 1.05 0.79–1.38 0.95 0.71–1.28 0.92 0.68–1.24 African Surinamese 1st generation 1.15 0.98–1.35 0.89 0.75–1.07 0.88 0.73–1.05 2nd generation 1.14 0.87–1.50 1.02 0.76–1.36 0.96 0.71–1.29 Ghanaian 1st generation 1.53 1.30–1.81 1.38 1.14–1.68 1.28 1.05–1.56 2nd generation 1.81 1.09–3.01 1.92 1.12–3.27 1.64 0.94–2.87 Turkish 1st generation 1.84 1.58–2.15 1.07 0.88–1.31 1.00 0.82–1.22 2nd generation 1.27 1.02–1.57 1.02 0.80–1.30 1.01 0.79–1.29 Moroccan 1st generation 1.81 1.55–2.11 1.22 1.00–1.49 1.15 0.94–1.41 2nd generation 1.14 0.92–1.41 0.93 0.73–1.19 0.89 0.69–1.14 Female sex 1.91 1.75–2.08 <.001 1.77 1.60–1.95 <.001 1.70 1.54–1.88 <.001 Age <.001 < 25 years Ref 25–34 years 1.04 0.87–1.24 35–44 years 1.28 1.09–1.50 45–54 years 1.34 1.15–1.56 55–64 years 1.42 1.21–1.66 ≥ 65 years 1.59 1.29–1.96 Educational level <.001 .005 .001 Unknown Ref Ref Ref No school/elementary school 1.41 0.95–2.09 1.55 0.93–2.58 1.55 0.87–2.75 Lower vocational/lower secondary school 1.05 0.71–1.55 1.50 0.90–2.50 1.47 0.83–2.60 Intermediate vocational/ intermediate secondary school 0.93 0.63–1.39 1.43 0.86–2.39 1.38 0.78–2.44 Higher vocational/university 0.66 0.44–0.99 1.18 0.71–1.99 1.12 0.63–2..00 Marital status <.001 Married/registered partnership Ref Cohabiting 0.66 0.56–0.77 Unmarried/never married 0.78 0.71–0.86 Divorced/separated 1.14 1.02–1.27 Widow/widower 1.26 0.98–1.64 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 1.75 1.56–1.96 <.001 1.29 1.13–1.47 <.001 1.30 1.13–1.48 <.001 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 12 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values CVA/one-sided loss of bodily function ≤1 day 1.38 1.18–1.62 <.001 MI incl. ≥half hour chest pain or dotter/bypass operation 1.78 1.60–1.98 <.001 1.28 1.13–1.45 <.001 1.24 1.09–1.40 .001 Severe heart condition 1.45 1.18–1.79 .001 Malignant disorder 1.93 1.52–2.47 <.001 1.33 1.02–1.74 .037 Severe or chronic fatigue 1.86 1.71–2.02 <.001 1.20 1.08–1.33 .001 1.16 1.04–1.29 .008 High blood pressure 1.36 1.25–1.49 <.001 Artery stenosis 1.46 1.25–1.70 <.001 0.80 0.67–0.96 .014 0.77 0.64–0.92 .004 Respiratory diseases 2.19 1.96–2.44 <.001 1.66 1.47–1.87 <.001 1.59 1.41–1.81 <.001 Serious/persistent intestinal disorders 1.87 1.64–2.12 <.001 1.24 1.07–1.43 .004 1.22 1.05–1.41 .009 Psoriasis 1.28 1.05–1.56 .015 (Chronic) eczema 1.30 1.15–1.47 <.001 Incontinence 2.08 1.85–2.35 <.001 1.32 1.15–1.52 <.001 1.32 1.15–1.52 <.001 Body Mass Index <.001 < 18.5 Ref 18.5–25 1.03 0.72–1.46 25–30 1.24 0.87–1.76 30–40 1.64 1.15–2.33 ≥ 40 1.97 1.31–2.97 Smoking .184 <.001 .003 Yes Ref Ref Ref No, never 0.99 0.90–1.09 0.78 0.69–0.87 0.82 0.72–0.92 No, but ever 0.90 0.80–1.02 0.91 0.79–1.04 0.93 0.81–1.07 Alcohol usage <.001 .017 .012 Never Ref Ref Ref Not in previous 12 months 0.88 0.77–1.02 0.96 0.82–1.12 0.95 0.81–1.12 Monthly or less 0.68 0.61–0.77 0.83 0.72–0.95 0.82 0.71–0.94 2–4 times per month 0.73 0.64–0.83 0.93 0.79–1.09 0.92 0.78–1.08 2–3 times per week 0.64 0.55–0.75 0.92 0.76–1.10 0.89 0.74–1.08 ≥ 4 times per week 0.48 0.39–0.58 0.70 0.55–0.88 0.68 0.54–0.86 Difficulty with Dutch language <.001 No Ref Yes 1.32 1.21–1.43 Not applicable 0.80 0.71–0.91 Perceived health <.001 <.001 .002 Excellent Ref Ref Ref Very good 1.09 0.86–1.37 1.05 0.83–1.34 1.03 0.80–1.31 Good 1.55 1.27–1.90 1.22 0.99–1.51 1.20 0.97–1.49 Mediocre 2.47 2.01–3.04 1.37 1.09–1.72 1.32 1.05–1.67 Bad 3.85 3.02–4.91 1.69 1.28–2.23 1.59 1.20–2.11 Antibiotic-related behavior Higher antibiotic knowledge <.001 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 13 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values No Ref Yes 0.77 0.71–0.84 Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 4.72 3.59–6.21 3.07 2.28–4.14 Yes, occasionally 2.42 2.20–2.66 2.11 1.91–2.34 Did not finish treatment <.001 <.001 Always finished or no antibiotics Ref Ref Yes, regularly 2.70 2.02–3.62 1.80 1.29–2.50 Yes, occasionally 1.62 1.45–1.83 1.32 1.16–1.50 Abbreviations: OR Odds Ratio, aOR adjusted Odds Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner antibiotic prescriptions was second generation Ghanaian ethnic groups received less antibiotics for viral infections participants. Furthermore, we showed that a lower than non-Hispanic white children. level of antibiotic knowledge was not associated with Lower odds for higher level of antibiotic use know- receiving antibiotics or average number of antibiotic ledge were also seen among individuals who regularly or prescriptions, and that ethnic differences in antibiotic occasionally requested antibiotics from their GP or who use therefore cannot be explained by level of know- regularly or occasionally did not finish treatment. These ledge on antibiotics. findings suggest that improving antibiotic knowledge A previous study in Dutch primary care centres dem- might decrease the number of requests for antibiotics in onstrated higher use of antibiotics among first- primary care and improve appropriate use. generation migrants from Turkey, Morocco, Surinam or Our study has several strengths. First, the HELIUS the Antilles compared to Dutch, after adjustment for study consists of a large number of participants from age, sex, education, presence of chronic diseases, and major ethnic groups living in the same city, with repre- smoking [3]. We found that the odds of having ≥1 anti- sentation from all socioeconomic levels. Second, all out- biotic prescription was higher in some ethnic groups in comes and determinants were measured using the same unadjusted analysis, but after adjusting for several vari- methodology across all ethnic groups and HELIUS used ables including medical conditions, the odds were sig- translated questionnaires and had ethnically-matched in- nificantly higher among Ghanaian and first-generation terviewers and research assistants to provide assistance Moroccan participants only. In contrast, in our analyses during data collection. These procedures enhance the on the number of antibiotic prescriptions as an outcome, comparability between ethnic groups. Another major only second-generation Ghanaian migrants were at strength of the current study is that HELIUS data could higher risk of receiving a higher number of prescriptions be linked to data from a health insurance register cover- compared to Dutch participants. For all other ethnic ing the majority (77.7%) of the study population. groups, no evidence of a higher risk for more frequent Our study has also limitations. First, although HELIUS prescriptions was found, while even a lower number was participants were recruited via an ethnicity-stratified present for first-generation African Suriname and Turk- random selection of the municipal registry of ish participants. To the best of our knowledge, no other Amsterdam, the response rate for HELIUS study was studies have evaluated the variation in level of antibiotic 28% and there may be selection bias [8]. However, ana- knowledge and antibiotic use between ethnic groups and lysis from a previous HELIUS study have shown that thus our findings need to be confirmed. Notably, our participants are not exceedingly different from non- findings on antibiotic prescriptions and ethnicity are in respondents regarding sociodemographic variables [8]. line with a large retrospective cohort study performed in Second, we did not take into account the use of antibi- pediatric emergency departments in the United States otics purchased over the counter in the home country of [15]. This study also looked at the association between participants [6, 16–18], and we might therefore have ethnicity and antibiotic prescribing, showing that other underestimated antibiotic use in non-Dutch ethnic Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 14 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values Sociodemographics Ethnicity .004 .004 .001 Dutch Ref Ref Ref South-Asian Surinamese 1st generation 1.06 0.85–1.31 0.86 0.68–1.10 1.02 0.82–1.28 2nd generation 0.82 0.55–1.21 0.94 0.61–1.44 0.94 0.64–1.36 African Surinamese 1st generation 0.79 0.64–0.99 0.73 0.58–0.93 0.81 0.65–1.00 2nd generation 0.83 0.57–1.22 0.90 0.59–1.36 0.93 0.65–1.33 Ghanaian 1st generation 0.75 0.60–0.94 0.77 0.60–1.00 0.81 0.65–1.02 2nd generation 1.66 0.89–3.11 2.09 1.06–4.12 2.70 1.47–4.94 Turkish 1st generation 0.85 0.69–1.04 0.74 0.59–0.92 0.79 0.64–0.97 2nd generation 1.14 0.87–1.51 1.14 0.84–1.53 1.10 0.84–1.44 Moroccan 1st generation 0.99 0.81–1.21 0.89 0.70–1.11 0.89 0.72–1.10 2nd generation 0.84 0.63–1.13 0.92 0.67–1.27 1.02 0.76–1.37 Female sex 1.36 1.20–1.54 <.001 1.35 1.18–1.54 <.001 1.29 1.15–1.46 <.001 Age .023 < 25 years Ref 25–34 years 1.09 0.91–1.30 35–44 years 1.16 0.98–1.38 45–54 years 1.09 0.92–1.29 55–64 years 1.14 0.95–1.36 ≥ 65 years 1.44 1.16–1.80 Educational level .096 Unknown Ref No school/elementary school 1.74 0.86–3.52 Lower vocational/lower secondary school 1.74 0.86–3.52 Intermediate vocational/ intermediate secondary school 1.57 0.77–3.17 Higher vocational/university 1.45 0.71–2.96 Marital status .212 Married/registered partnership Ref Cohabiting 0.85 0.72–1.00 Unmarried/never married 0.99 0.89–1.10 Divorced/separated 1.01 0.90–1.13 Widow/widower 1.14 0.90–1.45 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 1.34 1.17–1.54 <.001 1.21 1.04–1.41 .015 1.22 1.06–1.41 .005 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 15 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values CVA/one-sided loss of bodily function ≤1 day 1.10 0.95–1.29 .208 MI incl. ≥half hour chest pain or dotter/bypass operation 1.35 1.19–1.54 <.001 1.22 1.06–1.41 .005 Severe heart condition 1.26 1.04–1.52 .019 Malignant disorder 1.91 1.47–2.48 <.001 1.60 1.21–2.12 .001 1.60 1.28–2.00 Severe or chronic fatigue 1.29 1.15–1.44 <.001 High blood pressure 1.14 1.04–1.25 .006 Artery stenosis 1.30 1.08–1.57 .005 Respiratory diseases 1.50 1.32–1.71 <.001 1.34 1.16–1.54 <.001 1.29 1.13–1.47 <.001 Serious/persistent intestinal disorders 1.24 1.07–1.44 .004 Psoriasis 1.08 0.90–1.31 .406 (Chronic) eczema 1.21 1.07–1.37 .002 1.14 1.00–1.29 .042 Incontinence 1.31 1.13–1.50 <.001 Body Mass Index .736 < 18.5 Ref 18.5–25 0.98 0.70–1.37 25–30 0.98 0.70–1.37 30–40 1.04 0.74–1.46 ≥ 40 1.06 0.71–1.57 Smoking .099 Yes Ref No, never 1.13 1.00–1.29 No, but ever 1.02 0.87–1.21 Alcohol usage .075 Never Ref Not in previous 12 months 0.93 0.77–1.13 Monthly or less 0.94 0.80–1.11 2–4 times per month 1.01 0.85–1.20 2–3 times per week 0.73 0.58–0.91 ≥ 4 times per week 0.81 0.61–1.08 Difficulty with Dutch language .320 No Ref Yes 0.94 0.86–1.04 Not applicable 1.05 0.88–1.25 Perceived health <.001 .001 <.001 Excellent Ref Ref Ref Very good 0.98 0.68–1.42 1.03 0.71–1.49 0.99 0.70–1.39 Good 1.29 0.94–1.77 1.19 0.85–1.65 1.17 0.86–1.59 Mediocre 1.73 1.25–2.39 1.40 1.00–1.97 1.44 1.05–1.97 Bad 2.30 1.62–3.26 1.71 1.17–2.49 1.72 1.22–2.43 Antibiotic-related behavior Higher antibiotic knowledge .054 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 16 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values No Ref Yes 0.92 0.85–1.00 Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 2.96 2.34–3.73 1.87 1.45–2.41 Yes, occasionally 1.75 1.59–1.92 1.15 1.06–1.30 Did not finish treatment <.001 Always finished or no antibiotics Ref Yes, regularly 1.50 1.12–2.00 Yes, occasionally 1.32 1.18–1.48 Abbreviations: IRR Incidence Risk Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner Accounts for zero-inflated distribution groups. As a recent HELIUS study found that Dutch Conclusions people of Turkish or Moroccan origin were more likely To our knowledge, this study is the first to examine eth- to use healthcare in the Netherlands as well as their nic disparities in level of antibiotic knowledge and use in country of origin [19], underestimation of antibiotic use a large population-based sample among adults with dif- in non-Dutch ethnic groups seems unlikely. Third, since ferent ethnic backgrounds. Health policy makers and several characteristics, such as education level and med- healthcare professionals are increasingly developing in- ical conditions, of HELIUS participants insured at Ach- terventions to improve the quality of antibiotic use, mea differed from those insured elsewhere, selection which is needed to help contain antimicrobial resistance. bias could have been introduced in analysis on antibiotic Targeted campaigns can be considered, for instance, use. This difference could be due to the fact that the during the annual European Antibiotic Awareness Day, City of Amsterdam provided health insurance discounts since this event addresses improvement in the quality of with Achmea for low-income individuals. These differ- antibiotic use to the general public [20]. Still, this study ences were corrected for during multivariable analyses shows that a lower level of antibiotic knowledge is not to the most possible extent. Fourth, the variable ‘ever necessarily linked to higher antibiotic usage, indicating asked GP for antibiotics’ does not discriminate between that interventions aimed at improving knowledge alone appropriate or inappropriate requests for antibiotics and might be insufficient to reduce antibiotic use. Neverthe- misclassification might have occurred. However, this less, the underlying reasons for these findings need variable gives some information on participants’ atti- further evaluation. tudes towards antibiotic use. Furthermore, due to pri- vacy restrictions, we were unable to include indication Supplementary information Supplementary information accompanies this paper at https://doi.org/10. for antibiotic therapy and duration of antibiotic use as 1186/s13756-019-0636-x. additional indices for antibiotic use (apart from the number of antibiotics prescribed). Moreover, since this Additional file 1. Supplementary Methods. was a cross-sectional study, we were unable to model Additional file 2: Table S1. Characteristics of participants not linked antibiotic knowledge with future antibiotic prescriptions. versus linked to the Achmea Health Database. Further research should examine the association of anti- biotic knowledge with future antibiotic prescriptions. Fi- Abbreviations AHD: Achmea Health Database; aIRR: adjusted Incidence Risk Ratio; nally, we are unable to determine if individuals were aOR: adjusted Odds Ratio; BMI: Body Mass Index; CVA: Cerebrovascular more demanding towards their GP or if their GPs Accident; GP: General Practitioner; HELIUS: Healthy life in an Urban Setting; were more lenient in prescribing antibiotics during IQR: Interquartile Range; IRR: Incidence Risk Ratio; MI: Myocardial Infarction; OR: Odds Ratio; WHO: World Health Organization illness [4, 5]. Neither completing antibiotic therapy, assessed by pill count, nor duration of antibiotic use Acknowledgements could be taken into account as these data were not The authors would like to acknowledge the HELIUS participants for their available. contribution; the HELIUS team for data collection and management. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 17 of 17 Authors’ contributions 7. Deschepper R, Grigoryan L, Lundborg CS, Hofstede G, Cohen J, Kelen GV, ES, MS, SG, MP and JP were involved in the conception and design of the et al. Are cultural dimensions relevant for explaining cross-national study. ES, EvD and MS were involved in the acquisition of data. EvD and AB differences in antibiotic use in Europe? BMC Health Serv Res. 2008;8:123. analyzed the data. ES and EvD wrote the manuscript. ES, EvD, AB, MS, SG, MP 8. Snijder MB, Galenkamp H, Prins M, Derks EM, Peters RJG, Zwinderman AH, and JP were major contributors in revising the manuscript. All authors read et al. Cohort profile: the Healthy life in an urban setting (HELIUS) study in and approved the final manuscript. Amsterdam, The Netherlands. BMJ Open. 2017;7(12):e017873. 9. Stronks K, Snijder MB, Peters RJ, Prins M, Schene AH, Zwinderman AH. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. Funding BMC Public Health. 2013;13:402. The HELIUS study is conducted by the Academic Medical Center Amsterdam 10. StatLine. Population key figures 2019. https://opendata.cbs.nl/statline/#/CBS/ and the Public Health Service of Amsterdam. Both organizations provided nl/dataset/37296ned/table?ts=156888214600. core support for HELIUS. The HELIUS study is also funded by the Dutch Heart 11. StatLine. Population data; age, migration background, gender and region. Foundation (2010 T084), the Netherlands Organization for Health Research 2019. https://opendata.cbs.nl/statline/#/CBS/nl/dataset/37713/table?ts=156 and Development (ZonMw) (200500003), the European Union (FP-7) (278901), and the European Fund for the Integration of non-EU immigrants 12. Stronks K, Kulu-Glasgow I, Agyemang C. The utility of 'country of birth' for (EIF) (2013EIF013). The authors were independent from funders and had full the classification of ethnic groups in health research: the Dutch experience. access to all data. The decision to submit for publication was uninfluenced Ethn Health. 2009;14(3):255–69. by the funders. The authors of this study take full responsibility for the integ- 13. Data Protection Authority. Ministry of Health, Welfare and Sport. https:// rity of the data and the accuracy of the data analysis. autoriteitpersoonsgegevens.nl/en. 14. Special Eurobarometer 338: Antimicrobial Resistance (2010). http://ec. Availability of data and materials europa.eu/public_opinion/archives/eb_special_339_320_en.htm#338. The datasets used in the current study are available from the corresponding 15. Goyal MK, Johnson TJ, Chamberlain JM, Casper TC, Simmons T, Alessandrini author on request, contingent on approval from the HELIUS scientific EA, et al. Racial and Ethnic Differences in Antibiotic Use for Viral Illness in committee. Emergency Departments. Pediatrics. 2017;140(2):e20170203 (4). 16. Lindenmeyer A, Redwood S, Griffith L, Ahmed S, Phillimore J. Recent Ethics approval and consent to participate migrants’ perspectives on antibiotic use and prescribing in primary care: a The HELIUS study is conducted in accordance with the Declaration of qualitative study. Br J Gen Pract. 2016;66(652):e802–e9. Helsinki and has been approved by the AMC Ethical Review Board. All 17. Hu J, Wang Z. Non-prescribed antibiotic use and general practitioner service participants provided written informed consent. utilisation among Chinese migrants in Australia. Aust J Prim Health. 2016; 22(5):434–9. Consent for publication 18. Hu J, Wang Z. In-home antibiotic storage among Australian Chinese Not applicable. migrants. Int J Infect Dis. 2014;26:103–6. 19. Sekercan A, Snijder MB, Peters RJG, Stronks K. Is healthcare consumption in Competing interests the country of origin among Moroccan and Turkish migrants of older age The authors declare that they have no competing interests. (55+) associated with less use of care in the Netherlands? Tijdschr Gerontol Geriatr. 2018;49(6):253–62. Author details 20. European Centre for Disease Prevention and Control. Antimicrobial Department of Internal Medicine, Division of Infectious Diseases, Amsterdam consumption. In: ECDC, editor. Annual epidemiological report for 2016. UMC, University of Amsterdam, Meibergdreef 9, 1105, AZ, Amsterdam, The Stockholm: ECDC; 2018. https://ecdc.europa.eu/en/publications-data/ Netherlands. Department of Infectious Diseases, Public Health Service antimicrobial-consumption-annual-epidemiological-report-2016. Amsterdam, Nieuwe Achtergracht 100, 1018, WT, Amsterdam, The Netherlands. INSERM, Sorbonne Université, Institut Pierre Louis d’Épidémiologie et de Santé Publique, Paris, France. Department of Public Publisher’sNote Health, Amsterdam UMC, University of Amsterdam, Amsterdam Public Health Springer Nature remains neutral with regard to jurisdictional claims in Research Institute, Amsterdam, The Netherlands. Department of Clinical published maps and institutional affiliations. Epidemiology, Biostatistics and Bioinformatics, Amsterdam UMC, University of Amsterdam, Amsterdam Public Health Research Institute, Amsterdam, The Netherlands. Received: 31 July 2019 Accepted: 24 October 2019 References 1. WHO. Antimicrobial resistance: global report on surveillance 2014; 2014. p. 257. 2. Nellums LB, Thompson H, Holmes A, Castro-Sanchez E, Otter JA, Norredam M, et al. Antimicrobial resistance among migrants in Europe: a systematic review and meta-analysis. Lancet Infect Dis. 2018;18(7):796–811. 3. Hogenhuis CC, Grigoryan L, Numans MM, Verheij TJ. Differences in antibiotic treatment and utilization of diagnostic tests in Dutch primary care between natives and non-western immigrants. Eur J Gen Pract. 2010;16(3):143–7. 4. Grigoryan L, Burgerhof JG, Degener JE, Deschepper R, Lundborg CS, Monnet DL, et al. Determinants of self-medication with antibiotics in Europe: the impact of beliefs, country wealth and the healthcare system. J Antimicrob Chemother. 2008;61(5):1172–9. 5. Grigoryan L, Burgerhof JG, Degener JE, Deschepper R, Lundborg CS, Monnet DL, et al. Attitudes, beliefs and knowledge concerning antibiotic use and self-medication: a comparative European study. Pharmacoepidemiol Drug Saf. 2007;16(11):1234–43. 6. Norris P, Ng LF, Kershaw V, Hanna F, Wong A, Talekar M, et al. Knowledge and reported use of antibiotics amongst immigrant ethnic groups in New Zealand. J Immigr Minor Health. 2010;12(1):107–12. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Antimicrobial Resistance & Infection Control Springer Journals

Knowledge and use of antibiotics in six ethnic groups: the HELIUS study

Loading next page...
 
/lp/springer-journals/knowledge-and-use-of-antibiotics-in-six-ethnic-groups-the-helius-study-H0O4F03w2N

References (24)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s).
Subject
Biomedicine; Medical Microbiology; Drug Resistance; Infectious Diseases
eISSN
2047-2994
DOI
10.1186/s13756-019-0636-x
Publisher site
See Article on Publisher Site

Abstract

Background: The increase of antimicrobial resistance, mainly due to increased antibiotic use, is worrying. Preliminary evidence suggests that antibiotic use differs across ethnic groups in the Netherlands, with higher use in people of non-Dutch origin. We aimed to determine whether appropriate knowledge and use of antibiotics differ by ethnicity and whether knowledge on antibiotics is associated with antibiotic use. Methods: We performed a cross-sectional study analyzing baseline data (2011–2015) from a population-based cohort (HELIUS study), which were linked to data from a health insurance register. We included 21,617 HELIUS participants of South-Asian Surinamese, African-Surinamese, Turkish, Moroccan, Ghanaian, and Dutch origin. Fifteen thousand seven participants had available prescription data from the Achmea Health Data-base (AHD) in the year prior to their HELIUS study visit. Participants were asked five questions on antibiotic treatment during influenza-like illness, pneumonia, fever, sore throat and bronchitis, from which higher versus lower antibiotic knowledge level was determined. Number of antibiotic prescriptions in the year prior to the HELIUS study visit was used to determine antibiotic use. Results: The percentage of individuals with a higher level of antibiotic knowledge was lower among all ethnic minority groups (range 57 to 70%) compared to Dutch (80%). After correcting for baseline characteristics, including medical conditions, first-generation African Surinamese and Turkish migrants received a significantly lower number of antibiotic prescriptions compared to individuals of Dutch origin. Only second-generation Ghanaian participants received more prescriptions compared to Dutch participants (aIRR 2.09, 95%CI 1.06 to 4.12). Higher level of antibiotic knowledge was not significantly associated with the number of prescriptions (IRR 0.92, 95%CI 0.85 to 1.00). Conclusions: Levels of antibiotic knowledge varied between ethnic groups, but a lower level of antibiotic knowledge did not correspond with a higher number of antibiotic prescriptions. Keywords: Antibiotics, Antibiotic knowledge, Antibiotic use, Ethnic groups Background A recent meta-analysis showed a higher prevalence of The emergence of antimicrobial resistance, along with antimicrobial resistance among migrants in Europe [2]. the steady decline in antibiotic development, has been There is preliminary evidence in the Netherlands that identified as a major health threat for the coming decade the use of antibiotics also differs across ethnic groups, by the World Health Organization (WHO). Increase in with a higher use of antibiotics among people of non- antibiotic use is the main reason for this development Dutch origin [3]. The reason for this difference, however, [1] and as such, antibiotics should only be prescribed is unclear. It could be explained by increased incidence when there is a clear indication for use. of bacterial infections, but, to the best of our knowledge, there is no evidence to support this hypothesis. Alterna- tively, knowledge about antibiotic use might vary across * Correspondence: evdulm@ggd.amsterdam.nl ethnic groups. As expectations and knowledge of the Emelie C. Schuts and Eline van Dulm contributed equally to this work. patient could potentially drive a physician’s decision to Department of Infectious Diseases, Public Health Service Amsterdam, prescribe antibiotics, receiving prescriptions could also Nieuwe Achtergracht 100, 1018, WT, Amsterdam, The Netherlands Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 2 of 17 differ between ethnic groups [4–6]. There are also Ethnicity was defined according to the country of birth cultural-specific approaches to dealing with authority, of the participant as well as that of their parents [12]. being the physician in this setting, which have explained Specifically, a participant is considered to be of non- differences in antibiotic use between countries [7]. Dutch ethnic origin if they fulfill either of the following The HELIUS (Healthy life in an Urban Setting) study criteria (i): they were born abroad and had at least one is a large-scale, population-based cohort study among parent born abroad (first generation) or (ii) they were different ethnic groups, which was established with the born in Netherlands but both their parents were born aim to investigate mechanisms underlying the impact of abroad (second generation). Dutch participants were ethnicity on communicable and non-communicable dis- born in the Netherlands and had both parents who eases [8, 9]. In 2018, approximately 13% of the popula- were born in the Netherlands. After HELIUS data tion of the Netherlands was of non-Western origin [10]. collection, the Surinamese group were further classi- The largest non-Western population groups were fied according to self-reported ethnic origin (obtained individuals of Turkish (2.4%), Moroccan (2.3%) and by questionnaire), into ‘African Surinamese’, ‘South- Surinamese (2.0%) descent [10]. In Amsterdam, approxi- Asian Surinamese’, ‘Javanese Surinamese’ and ‘other/ mately 36% of the population in 2018 was of non- unknown Surinamese’. Western descent [11]. The ethnic groups included in the HELIUS study are the largest ethnic minority groups of Data linkage Amsterdam [9]. Amongst other data, data on antibiotic Permission to link participants’ individual data to outside knowledge were collected. We were able to link these health registries was asked in the written informed con- data at the individual level to data from a health insur- sent form [8]. Of the 22,165 HELIUS participants, 19, ance register on recent antibiotic use. 895 agreed. HELIUS data of these individuals were This study then provides a unique opportunity to de- linked to reimbursement data from the Achmea insur- termine whether knowledge about and use of antibiotics ance company (Achmea Health Database, AHD) from vary between ethnic groups, and if so, whether differ- 2010 until 2015. The AHD, obtained from the largest ences in antibiotic use can be attributed to differences in health insurance company in Amsterdam, contains all knowledge about antibiotics. We hypothesized that healthcare expenditures of every insured participant, antibiotic use differs among ethnic groups as a result of including medications. A trusted third party linked differences in knowledge. data on reimbursed antibiotic prescriptions using an encrypted social security number and returned data without any identifying information. Procedures were Methods in accordance with the General Data Protection Study population and design Regulation [13]. The HEalthy LIfe in an Urban Setting (HELIUS) study is a multiethnic cohort study conducted in Amsterdam, Inclusion and exclusion criteria for present study which focuses on cardiovascular disease (e.g. diabetes), Of the 22,165 participants, we excluded those of Javan- mental health (e.g. depressive disorders), and infectious ese Surinamese or other/unknown Surinamese origin diseases [8, 9]. In brief, baseline data collection took and those with another/unknown ethnic origin because place in 2011–2015 and included people aged 18 to 70 of small participant numbers. For analyses on antibiotic years of Dutch, Surinamese, Ghanaian, Moroccan, and use, we included those who gave permission for data Turkish origin. A random sample of participants, strati- linkage and could be linked to the AHD. To reduce bias fied by ethnic origin, was taken from the municipality for individuals with short-term insurance, we excluded register of Amsterdam. Participants filled in an extensive those who were insured with Achmea for less than 365 self-administered questionnaire (variables included in the days in the year preceding their HELIUS study visit. questionnaire are described elsewhere) [9] and underwent a physical examination during which biological samples Outcome variables were obtained [9]. No information was provided regarding The primary outcomes were level of antibiotic know- appropriate antibiotic use. Between 2011 and 2015, 24,789 ledge and antibiotic use during the year prior to the persons were included. Data collection procedures have HELIUS visit. Level of antibiotic knowledge was based been previously described in detail [9]. Both questionnaire on five questions, used in other studies [4, 6, 14], which data and physical examination data were available for 22, asked the perceived necessity (yes/no) for antibiotic 165 participants. The HELIUS study was conducted in treatment during influenza-like illness, pneumonia, fever, accordance with the Declaration of Helsinki and was ap- sore throat and bronchitis. Using these questions, we proved by the AMC Ethical Review Board. All participants created an overall knowledge score of antibiotic use by provided written informed consent. summing the total number of correct responses, Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 3 of 17 resulting in a score ranging from 0 to 5. A two- Analysis on antibiotic use in the year prior to HELIUS parameter logistic regression model was fitted to the five study visit included all HELIUS participants who were binary items based on the assumptions of item response linked to the AHD and were insured for at least 365 days theory (see Additional file 1). From this model, “higher” with Achmea in the year prior to their HELIUS study and “lower” levels of antibiotic knowledge were defined visit. Determinants for having received ≥1 antibiotic pre- by a knowledge score of ≥4 and < 4, respectively. scription were assessed using logistic regression. The Antibiotic use was obtained from linked AHD data same multivariable approach as above was used for this and was based on the total number of reimbursed antibi- outcome. We also compared antibiotic use during the otics (classified by ATC code J01; anti-infectives for sys- entire period insured at Achmea versus the year prior to temic use) dispensed by community pharmacies from HELIUS study visit to assess differences when consider- 2010 until 2015. We evaluated antibiotic use (yes/no) in ing longer time periods. the year prior to the HELIUS study visit, as well as the Determinants for the total number of antibiotic pre- number of antibiotic prescriptions over the past year scriptions were then evaluated. As this outcome con- and during the entire insured period. tained a high proportion of zero values and was over- dispersed, we used a zero-inflated negative binomial re- Other variables gression model. This model contains two parts: one ac- Independent variables were obtained from the HELIUS counting for zero values in the count distribution (zero- study questionnaire (migration generation; sex; age; level inflated) and another accounting for the over-dispersed of education; marital status; self-reported medical condi- count distribution (negative binomial). Covariates for the tions; smoking; alcohol consumption; difficulty with the zero-inflated part were determined a priori from the Dutch language and perceived health) and physical risk-factor analysis on ≥1 antibiotic prescription. Covari- examination (body mass index (BMI, kg/m )). Variables ates for the negative binomial part were selected from on antibiotic-related behavior were: not having finished covariates with a p-value < 0.2 in univariable analyses antibiotic treatment; having saved antibiotics for later; and variables above this p-value were removed in and ever having asked the general practitioner (GP) for backwards-stepwise fashion. Incidence risk ratios (IRR) antibiotics. Definitions and grouping of variables are comparing the number of antibiotics prescribed over the extensively described elsewhere [8]. past year across levels of determinants were estimated from this model. Statistical analyses Multicollinearity was verified using variance inflation Sociodemographics, health status, antibiotic knowledge factors, while any variable with an inflation factor of ≥4 level and questions on antibiotic use were presented by was considered multicollinear and excluded from the ethnicity. To assess selection bias resulting from AHD model. To understand whether the association between data linkage, the same variables were compared between ethnicity and outcome was modified by demographic participants who were successfully versus unsuccessfully variables, interaction between ethnicity and other demo- linked. Comparisons between ethnic groups were made graphic variables was also assessed in all multivariable using a Pearson’s χ or Fisher exact test for categorical models. data and Kruskal-Wallis rank test for continuous The three variables involving antibiotic-related behav- variables. ior were not initially considered in the final multivariable Analysis on level of antibiotic knowledge included all models. To assess whether ethnic differences in anti- HELIUS participants with available data. Odds ratios biotic use could be explained by patterns of antibiotic- (OR) comparing levels of antibiotic knowledge across related behavior, additional multivariable models includ- determinants and their 95% confidence intervals (CI) ing these variables were constructed for the endpoints (i) were estimated using logistic regression. All variables having received ≥1 antibiotic prescription and (ii) total with an associated p-value < 0.2 in univariable analyses number of antibiotic prescriptions. were included in a full multivariable model and variables Figure 1 provides an overview of all descriptive ana- with a p-value above this level were removed in lysis and modeling used in the study. Significance was backwards-stepwise fashion. Given that the research aim determined using a p-value < 0.05. All analyses were was to determine differences between ethnicity, ethnic conducted with Stata 13.1 (StataCorp., College Station, groups were forced in all models. This multivariable ap- Texas, USA). proach was chosen to not only assess other variables as- sociated with antibiotic knowledge, but also to Results understand the extent of confounding bias when asses- Participants sing the relationship between ethnicity and outcome Of the 22,165 HELIUS participants with available data, variables. 21,617 were eligible after applying exclusion criteria. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 4 of 17 Fig. 1 Overview of descriptive analysis and models used in the study. Abbreviations: HELIUS – Healthy Life in an Urban Setting; AHD – Achmea Health Database Their baseline characteristics, stratified by ethnicity, are their GP for antibiotics ranged from 0.6% in African shown in Table 1. Median age of participants was 46 Surinamese participants to 1.9% in Turkish and Moroc- years (IQR 34 to 55) and 58% were women. The propor- can participants. tion of several medical conditions predisposing individ- As shown in Table 2, there was a significantly lower uals to antibiotic treatment differed by ethnicity. Of odds of individuals with higher level of antibiotic know- these conditions, South-Asian Surinamese participants ledge among all non-Dutch ethnic groups compared to had the highest prevalence of self-reported diabetes mel- Dutch individuals (overall p < 0.001) (Table 2). Across all litus (17%) and cerebrovascular accident (CVA) (6%) non-Dutch groups, second-generation participants had a over the last 12 months. Turkish individuals had more higher level of antibiotic knowledge than first-generation prevalent artery stenosis (10%), severe or chronic fatigue participants; however, results remained significantly (45%) and respiratory diseases (15%), whereas Ghanaians lower compared to the Dutch group. more frequently reported high blood pressure (33%). In multivariable analysis, all ethnic minority groups Excellent perceived health was reported in 12% of Dutch had lower odds for higher level of antibiotic knowledge participants in contrast to 3.3% of Turkish participants. compared to Dutch (overall p < 0.001), although the ef- fect for second-generation Ghanaian participants was not statistically significant. The odds for higher level of Ethnic differences in antibiotic knowledge antibiotic knowledge were higher in all age groups > 25 In several ethnic groups, there were substantial propor- years of age (except for those ≥65) when compared to tions of individuals reporting the need to be treated with ≤25 years of age. Furthermore, women had a significantly antibiotics for illnesses without indication, as shown in higher odds of having a higher level of antibiotic know- Table 1. The number of people reporting to have been ledge compared to males. Lower odds for a higher level treated with antibiotics and not having regularly com- of antibiotic knowledge were found for the following pleted their antibiotic treatment was low across all eth- medical conditions: myocardial Infarction (MI), severe nic groups, ranging from 0.1% in Dutch participants to or chronic fatigue, respiratory diseases and having a BMI 2.1% in Ghanaian participants. Few individuals regularly ≥25. Lower odds for higher level of antibiotic knowledge saved their antibiotics for later use, ranging from < 0.1% were also seen among individuals who regularly or occa- in Dutch participants to 0.3% in Turkish participants. sionally requested antibiotics from their GP or who The percentage of participants having regularly asked regularly or occasionally did not finish treatment. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 5 of 17 Table 1 Characteristics of the HELIUS study population (N = 21,617) by ethnicity Variables Ethnicity Dutch South-Asian African Ghanaian Turkish Moroccan (N = 4564) Surinamese Surinamese (N = 2339) (N = 3614) (N = 3906) (N = 3043) (N = 4151) Sociodemographics Female sex 2475 54% 1672 55% 2535 61% 1434 61% 1980 55% 2392 61% Age in years, median (IQR) 47 (34–58) 48 (35–56) 50 (40–57) 47 (38–53) 42 (31–50) 40 (30–50) Migration generation 1st generation N.A. N.A. 2328 77% 3468 84% 2231 95% 2544 70% 2680 69% 2nd generation N.A. N.A. 715 23% 683 16% 108 4.6% 1080 30% 1226 31% Educational level Unknown 25 0.6% 16 0.5% 36 0.9% 42 1.8% 38 1.1% 38 1.0% No school/elementary school 150 3.3% 437 14% 231 6% 660 28% 1135 31% 1205 31% Lower vocational/lower secondary 646 14% 1010 33% 1477 36% 917 39% 889 25% 694 18% school Intermediate vocational/ intermediate 994 22% 885 29% 1464 35% 578 25% 1020 28% 1294 33% secondary school Higher vocational/university 2749 60% 695 23% 943 23% 142 6% 532 15% 675 17% Marital status Married/registered partnership 1724 38% 1043 34% 766 19% 420 18% 2208 61% 2285 59% Cohabiting 914 20% 311 10% 441 11% 427 19% 132 3.7% 110 2.8% Unmarried/never married 1474 32% 1001 33% 2231 54% 779 34% 761 21% 1010 26% Divorced/separated 356 8% 580 19% 617 15% 656 28% 407 11% 414 11% Widow/widower 87 1.9% 92 3.0% 65 1.6% 23 1.0% 90 2.5% 69 1.8% Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 102 2.2% 521 17% 419 10% 185 8% 336 9% 389 10% CVA/one-sided loss of bodily 160 3.5% 212 7% 261 6% 95 4.1% 196 5% 195 5% function ≤1 day MI incl. ≥half hour chest pain or 233 5% 491 16% 440 11% 225 10% 591 16% 476 12% dotter/bypass operation Severe heart condition 67 1.5% 120 4.0% 105 2.5% 75 3.2% 153 4.3% 58 1.5% Malignant disorder 103 2.3% 70 2.3% 85 2.1% 33 1.4% 73 2.0% 46 1.2% Severe or chronic fatigue 633 14% 1032 34% 956 23% 186 8% 1602 45% 1465 38% High blood pressure 534 12% 720 24% 1230 30% 770 33% 610 17% 546 14% Artery stenosis 85 1.9% 193 6% 181 4.4% 115 5% 348 10% 200 5% Respiratory diseases 345 8% 433 14% 354 9% 117 5% 556 15% 446 11% Serious/persistent intestinal disorders 249 5% 248 8% 308 7% 70 3.0% 433 12% 391 10% Psoriasis 136 3.0% 168 6% 128 3.1% 71 3.1% 154 4.3% 121 3.1% (Chronic) eczema 420 9% 406 13% 370 9% 71 3.1% 471 13% 423 11% Incontinence 309 7% 326 11% 342 8% 108 4.7% 464 13% 300 8% Body Mass Index (kg/m ), median (IQR) 24.1 (21.9– 25.7 (23.2– 27.0 (23.9– 27.9 (25.0– 27.9 (24.6– 27.0 (23.9– 26.7) 28.8) 30.8) 31.2) 31.7) 30.7) Smoking Yes 1129 25% 861 28% 1309 32% 104 4.5% 1240 35% 525 13% No, never 1689 37% 1758 58% 2016 49% 2027 87% 1700 47% 2874 74% No, but ever 1737 38% 413 14% 805 19% 191 8% 648 18% 492 13% Alcohol consumption Never 297 7% 1072 35% 1002 24% 806 35% 2414 67% 3265 84% Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 6 of 17 Table 1 Characteristics of the HELIUS study population (N = 21,617) by ethnicity (Continued) Variables Ethnicity Dutch South-Asian African Ghanaian Turkish Moroccan (N = 4564) Surinamese Surinamese (N = 2339) (N = 3614) (N = 3906) (N = 3043) (N = 4151) Not in previous 12 months 110 2.4% 251 8% 292 7% 408 18% 358 10% 338 9% Monthly or less 436 10% 758 25% 1242 30% 508 22% 367 10% 127 3.3% 2–4 times per month 894 20% 541 18% 873 21% 291 13% 257 7% 87 2.2% 2–3 times per week 1413 31% 262 9% 439 11% 193 8% 132 3.7% 56 1.4% ≥ 4 times per week 1408 31% 147 5% 272 7% 109 4.7% 57 1.6% 16 0.4% Difficulty with Dutch language Yes N.A. N.A. 711 23% 520 13% 1926 83% 2136 60% 1774 46% Perceived health Excellent 541 12% 162 5% 303 7% 226 10% 117 3.3% 166 4.3% Very good 1381 30% 310 10% 571 14% 458 20% 383 11% 384 10% Good 2205 48% 1623 53% 2335 56% 1180 51% 1871 52% 1871 48% Mediocre 402 9% 811 27% 834 20% 383 16% 921 26% 1223 31% Bad 28 0.6% 131 4.3% 101 2.4% 88 3.8% 307 9% 241 6% Antibiotics Knowledge concerning antibiotics Antibiotics effective for influenza 324 7% 554 19% 592 15% 658 29% 744 21% 648 18% Antibiotics effective for pneumonia 4166 92% 2304 77% 3114 77% 1312 58% 2587 73% 2741 73% Antibiotics effective for fever 689 15% 552 19% 679 17% 586 26% 898 26% 614 17% Antibiotics effective for sore throat 672 15% 760 26% 1089 27% 720 32% 1203 34% 978 26% Antibiotics effective for bronchitis 2246 50% 1385 50% 1862 46% 919 41% 1235 35% 1485 40% Higher level of antibiotic knowledge 3638 80% 1996 68% 2737 69% 1248 57% 2128 62% 2528 70% Did not finish antibiotic treatment Yes, regularly 5 0.1% 49 1.6% 44 1.1% 48 2.1% 41 1.2% 44 1.1% Yes, occasionally 332 7% 312 10% 527 13% 174 8% 424 12% 445 12% Always finished or no antibiotics 4203 93% 2646 88% 3524 86% 2053 90% 3104 87% 3361 87% Saved antibiotics for later Yes, regularly 2 0.0% 7 0.2% 9 0.2% 5 0.2% 10 0.3% 6 0.2% Yes, occasionally 37 0.8% 45 1.5% 68 1.7% 62 2.7% 60 1.7% 46 1.2% No, never 297 7% 304 10% 492 12% 146 6% 387 11% 430 11% Not applicable (no antibiotics) 4203 93% 2646 88% 3524 86% 2053 91% 3104 87% 3361 87% Ever asked GP for antibiotics Yes, regularly 38 0.8% 34 1.1% 26 0.6% 36 1.6% 67 1.9% 71 1.9% Yes, occasionally 824 18% 491 16% 607 15% 401 18% 734 21% 634 17% No, never 3674 81% 2482 83% 3441 84% 1835 81% 2744 77% 3074 81% Missing data, n: marital status 128; diabetes 78; stroke 55; myocardial infarction 33; heart condition 83; malignant disorders 137; fatigue 145; high blood pressure 101; artery stenosis 140; respiratory diseases 115; bowel diseases 115; psoriasis 98; eczema 117; incontinence 126; BMI 23; smoking 107; alcohol 127; perceived health 61; AB effective for influenza 614; AB effective for pneumonia 477; AB effective for fever 685; AB effective for sore throat 625; AB effective for bronchitis 663; asked GP for AB 414; did not finish treatment 292; saved AB 325 Abbreviations: IQR Inter Quartile Range, CVA Cerebro Vascular Accident, MI Myocardial infarction, N.A. Not Applicable, GP General Practitioner N.A. Not applicable (categories not applicable due to Dutch ethnicity) All variables are reported as n (%), unless otherwise indicated Answered “yes” to the statements below The Dutch General Practitioners guidelines (and those of other European countries) advise against the use of antibiotics for fever in general or sore throat, as they usually constitute viral infections, with only a few exceptions in both cases. Therefore, antibiotics are in general not appropriate for these conditions Based on a summed score with cutoff determined by an Item Response Theory model (≥4 out of 5 antibiotic knowledge questions correctly answered was considered as having a higher level of knowledge) Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 7 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Sociodemographics Ethnicity <.001 <.001 Dutch Ref Ref South-Asian Surinamese 1st generation 0.49 0.44–0.55 0.53 0.47–0.60 2nd generation 0.56 0.47–0.67 0.60 0.50–0.73 African Surinamese 1st generation 0.51 0.46–0.57 0.53 0.47–0.59 2nd generation 0.75 0.62–0.91 0.79 0.64–0.96 Ghanaian 1st generation 0.31 0.27–0.34 0.31 0.27–0.35 2nd generation 0.64 0.41–0.98 0.74 0.47–1.18 Turkish 1st generation 0.35 0.31–0.39 0.40 0.36–0.45 2nd generation 0.56 0.48–0.65 0.62 0.53–0.74 Moroccan 1st generation 0.51 0.45–0.57 0.56 0.50–0.63 2nd generation 0.71 0.61–0.83 0.75 0.63–0.89 Female sex 1.18 1.11–1.25 <.001 1.32 1.23–1.40 <.001 Age <.001 <.001 < 25 years Ref Ref 25–34 years 1.24 1.01–1.34 1.32 1.16–1.50 35–44 years 0.99 0.81–1.05 1.30 1.14–1.49 45–54 years 0.85 0.71–0.91 1.19 1.04–1.37 55–64 years 0.95 0.78–1.02 1.26 1.08–1.45 ≥ 65 years 1.06 0.84–1.22 1.15 0.95–1.39 Educational level <.001 Unknown Ref No school/elementary school 1.10 0.79–1.55 Lower vocational/lower secondary school 1.27 0.91–1.77 Intermediate vocational/ intermediate secondary school 1.54 1.11–2.15 Higher vocational/university 2.06 1.47–2.87 Marital status <.001 Married/registered partnership Ref Cohabiting 1.15 1.04–1.27 Unmarried/never married 1.08 1.01–1.16 Divorced/separated 0.81 0.74–0.89 Widow/widower 1.10 0.88–1.36 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 0.70 0.63–0.77 <.001 CVA/one-sided loss of bodily function ≤1 day 0.85 0.75–0.97 .015 MI incl. ≥half hour chest pain or dotter/bypass operation 0.71 0.65–0.77 <.001 0.89 0.81–0.98 .017 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 8 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) (Continued) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Severe heart condition 0.61 0.52–0.73 <.001 Malignant disorder 0.97 0.78–1.20 .752 Severe or chronic fatigue 0.80 0.75–0.85 <.001 0.89 0.82–0.95 .001 High blood pressure 0.78 0.73–0.84 <.001 Artery stenosis 0.69 0.61–0.78 <.001 Respiratory diseases 0.68 0.62–0.75 <.001 0.80 0.72–0.88 <.001 Serious/persistent intestinal disorders 0.88 0.79–0.98 .024 Psoriasis 0.76 0.66–0.89 .001 (Chronic) eczema 0.93 0.84–1.02 .121 Incontinence 0.84 0.76–0.93 .001 Body Mass Index <.001 .001 < 18.5 Ref Ref 18.5–25 1.03 0.81–1.31 1.02 0.80–1.31 25–30 0.80 0.63–1.01 0.96 0.75–1.24 30–40 0.67 0.52–0.85 0.86 0.66–1.11 ≥ 40 0.58 0.43–0.78 0.78 0.57–1.08 Smoking <.001 Yes Ref No, never 1.03 0.96–1.11 No, but ever 1.19 1.09–1.30 Alcohol usage <.001 Never Ref Not in previous 12 months 0.89 0.80–1.00 Monthly or less 1.10 1.01–1.20 2–4 times per month 1.27 1.16–1.40 2–3 times per week 1.35 1.22–1.49 ≥ 4 times per week 1.59 1.42–1.78 Difficulty with Dutch language <.001 No Ref Yes 0.65 0.61–0.69 Not applicable 1.70 1.56–1.85 Perceived health <.001 Excellent Ref Very good 0.97 0.85–1.12 Good 0.83 0.73–0.93 Mediocre 0.63 0.56–0.72 Bad 0.48 0.40–0.57 Antibiotics Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 0.51 0.40–0.65 0.60 0.46–0.77 Yes, occasionally 0.57 0.53–0.61 0.59 0.55–0.64 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 9 of 17 Table 2 Variables associated with higher antibiotic knowledge in HELIUS study population (N = 21,617) (logistic regression analysis) (Continued) Univariable Multivariable (N = 20,081 )# OR (95% CI) P-values aOR (95% CI) P-values Did not finish treatment <.001 <.001 Always finished or no antibiotics Ref Ref Yes, regularly 0.51 0.39–0.67 0.71 0.54–0.94 Yes, occasionally 0.73 0.66–0.80 0.80 0.73–0.88 Abbreviations: OR Odds Ratio, aOR adjusted Odds Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner Fewer observations in the multivariable model than in the total study population were due to missing observations on certain covariates #We found significant interactions between ethnicity and sex (p = 0.007) and ethnicity and age (p = 0.047) Ethnic difference in antibiotic use (p < 0.001) and multivariable analysis (p < 0.001). In mul- Of the 19,895 HELIUS participants consenting to link tivariable analysis, compared to Dutch individuals, first their data to other health registries, 15,461 were linked and second generation Ghanaian individuals and first- to the AHD (77.7%). Of these 15,461 participants, 15,007 generation Moroccan individuals had significantly higher (97%) were insured for ≥365 days in the year prior to odds of receiving ≥1 antibiotic prescription. Adding vari- their HELIUS study visit. Additional file 2: Table S1 ables on antibiotic-related behavior and level of anti- shows the characteristics of the study participants linked biotic use knowledge to the multivariable model did not versus not linked to the AHD. Participants present in change these associations. the AHD register had a lower level of education, higher Table 5 shows the results from the analysis on the as- prevalence of medical conditions, and less often had sociation between ethnicity and total number of anti- higher levels of antibiotic knowledge. biotic prescriptions received in the year prior to the Table 3 describes antibiotic use according to ethnicity HELIUS study visit. Differences across ethnic groups for participants registered in the AHD. In total, 31,530 were observed overall for the number of antibiotic pre- antibiotic prescriptions were recorded over the study scriptions in both univariable and multivariable analysis period. The proportion of participants receiving ≥1 anti- (both p = 0.004). First-generation African Surinamese biotic prescription in the year prior to their HELIUS and Turkish migrants had a significantly lower number study visit was highest among first-generation Turkish of antibiotic prescriptions compared to individuals of participants (25%) and was comparably high among Dutch origin. Only second-generation Ghanaian partici- second-generation Ghanaian and first-generation pants has more prescriptions compared to Dutch partici- Moroccan participants (both 25%). The proportion of pants. Furthermore, female sex, diabetes mellitus, MI, participants receiving ≥1 antibiotic prescription in the malignant disorder, respiratory disease, eczema and year prior to the HELIUS study visit was lowest in Dutch worse perceived health were significantly associated with and second generation South-Asian Surinamese partici- a higher number of antibiotic prescriptions. pants (both 16%). Having a higher level of antibiotic knowledge was not When considering the entire period during which par- significantly associated with the number of prescriptions ticipants were insured at Achmea prior to the HELIUS when included in multivariable analysis (p = 0.446). No study visit (median 6.0 years, IQR 5.0 to 6.0), the propor- significant interactions between ethnicity and sex or tion of participants receiving ≥1 antibiotic prescription education were observed. Finally, adjusting the associ- was highest among first generation Turkish participants ation between ethnicity and antibiotic use for antibiotic- (69%) and lowest in second-generation Ghanaian partici- related behaviors did not change these associations. pants (49%). The mean number of prescriptions during the entire insured period was comparable to the mean Discussion number of prescriptions in the year prior to HELIUS Our study shows that knowledge on the need to use an- study visit for all ethnic groups (Table 2). tibiotics for treatment is lower among all ethnic minority groups compared to Dutch, with second generation eth- Determinants of antibiotic use and number of nic minorities showing higher levels of knowledge com- prescriptions pared to first generation migrants. We also observed Table 4 shows the results from the analysis on the asso- ethnic differences in the use of antibiotics, with a higher ciation between ethnicity and having received ≥1 anti- proportion having received at least one prescription, but biotic prescription in the year prior to the HELIUS study a lower mean number of antibiotic prescriptions among visit. Differences across ethnic groups were observed some ethnic minority groups compared to Dutch. The overall for any antibiotic prescription in both univariable only ethnic group with a significantly higher number of Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 10 of 17 Table 3 Antibiotic use in participants linked to AHD (N = 15,007) stratified by ethnicity Ethnicity Dutch South-Asian Surinamese African Surinamese Ghanaian Turkish Moroccan (N = 2071) 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen 1st gen 2nd gen (N = 1645) (N = 452) (N = 2334) (N = 432) (N = 1789) (N = 84) (N = 2102) (N = 776) (N = 2119) (N = 857) Duration of insurance at 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 Achmea (in years) (4.0–6.0) (6.0–6.0) (4.0–6.0) (6.0–6.0) (5.0–6.0) (6.0–6.0) (4.0–6.0) (96.0–6.0) (5.0–6.0) (6.0–6.0) (4.0–6.0) between 2010 and 2015, median (IQR) Within year prior to HELIUS study visit Participants with ≥1 ABP 16% 22% 16% 17% 17% 22% 25% 25% 19% 25% 17% Number of ABP among all participants included in the AHD Mean 0.26 0.39 0.26 0.28 0.28 0.33 0.55 0.40 0.34 0.41 0.28 Median (IQR) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.0) (0.0–0.5) (0.0–1.0) (0.0–0.0) (0.0–1.0) (0.0–0.0) Number of ABP among participants with ≥1 ABP Mean 1.66 1.75 1.59 1.58 1.64 1.51 2.19 1.57 1.78 1.64 1.63 Median (IQR) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 2.00 (1.0–3.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) 1.00 (1.0–2.0) During entire insured period Participants with ≥1 ABP 51% 63% 53% 56% 51% 62% 49% 69% 59% 67% 54% Number of ABP per year among all participants included in the AHD Mean 0.31 0.46 0.30 0.31 0.30 0.35 0.34 0.44 0.38 0.43 0.33 Median (IQR) 0.17 (0.0–0.3) 0.17 (0. 0–0.6) 0.17 (0. 0–0.3) 0.17 (0.0–0.3) 0.17 (0.0–0.3) 0.17 (0.0–0.5) 0.00 (0.0–0.7) 0.25 (0.0–0.7) 0.17 (0.0–0.5) 0.17 (0.0–0.6) 0.17 (0.0–0.5) Abbreviations: ABP Antibiotic Prescription, IQR Inter Quartile Range Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 11 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values Sociodemographics Ethnicity <.001 <.001 .004 Dutch Ref Ref Ref South-Asian Surinamese 1st generation 1.57 1.33–1.85 1.05 0.86–1.27 1.04 0.85–1.26 2nd generation 1.05 0.79–1.38 0.95 0.71–1.28 0.92 0.68–1.24 African Surinamese 1st generation 1.15 0.98–1.35 0.89 0.75–1.07 0.88 0.73–1.05 2nd generation 1.14 0.87–1.50 1.02 0.76–1.36 0.96 0.71–1.29 Ghanaian 1st generation 1.53 1.30–1.81 1.38 1.14–1.68 1.28 1.05–1.56 2nd generation 1.81 1.09–3.01 1.92 1.12–3.27 1.64 0.94–2.87 Turkish 1st generation 1.84 1.58–2.15 1.07 0.88–1.31 1.00 0.82–1.22 2nd generation 1.27 1.02–1.57 1.02 0.80–1.30 1.01 0.79–1.29 Moroccan 1st generation 1.81 1.55–2.11 1.22 1.00–1.49 1.15 0.94–1.41 2nd generation 1.14 0.92–1.41 0.93 0.73–1.19 0.89 0.69–1.14 Female sex 1.91 1.75–2.08 <.001 1.77 1.60–1.95 <.001 1.70 1.54–1.88 <.001 Age <.001 < 25 years Ref 25–34 years 1.04 0.87–1.24 35–44 years 1.28 1.09–1.50 45–54 years 1.34 1.15–1.56 55–64 years 1.42 1.21–1.66 ≥ 65 years 1.59 1.29–1.96 Educational level <.001 .005 .001 Unknown Ref Ref Ref No school/elementary school 1.41 0.95–2.09 1.55 0.93–2.58 1.55 0.87–2.75 Lower vocational/lower secondary school 1.05 0.71–1.55 1.50 0.90–2.50 1.47 0.83–2.60 Intermediate vocational/ intermediate secondary school 0.93 0.63–1.39 1.43 0.86–2.39 1.38 0.78–2.44 Higher vocational/university 0.66 0.44–0.99 1.18 0.71–1.99 1.12 0.63–2..00 Marital status <.001 Married/registered partnership Ref Cohabiting 0.66 0.56–0.77 Unmarried/never married 0.78 0.71–0.86 Divorced/separated 1.14 1.02–1.27 Widow/widower 1.26 0.98–1.64 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 1.75 1.56–1.96 <.001 1.29 1.13–1.47 <.001 1.30 1.13–1.48 <.001 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 12 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values CVA/one-sided loss of bodily function ≤1 day 1.38 1.18–1.62 <.001 MI incl. ≥half hour chest pain or dotter/bypass operation 1.78 1.60–1.98 <.001 1.28 1.13–1.45 <.001 1.24 1.09–1.40 .001 Severe heart condition 1.45 1.18–1.79 .001 Malignant disorder 1.93 1.52–2.47 <.001 1.33 1.02–1.74 .037 Severe or chronic fatigue 1.86 1.71–2.02 <.001 1.20 1.08–1.33 .001 1.16 1.04–1.29 .008 High blood pressure 1.36 1.25–1.49 <.001 Artery stenosis 1.46 1.25–1.70 <.001 0.80 0.67–0.96 .014 0.77 0.64–0.92 .004 Respiratory diseases 2.19 1.96–2.44 <.001 1.66 1.47–1.87 <.001 1.59 1.41–1.81 <.001 Serious/persistent intestinal disorders 1.87 1.64–2.12 <.001 1.24 1.07–1.43 .004 1.22 1.05–1.41 .009 Psoriasis 1.28 1.05–1.56 .015 (Chronic) eczema 1.30 1.15–1.47 <.001 Incontinence 2.08 1.85–2.35 <.001 1.32 1.15–1.52 <.001 1.32 1.15–1.52 <.001 Body Mass Index <.001 < 18.5 Ref 18.5–25 1.03 0.72–1.46 25–30 1.24 0.87–1.76 30–40 1.64 1.15–2.33 ≥ 40 1.97 1.31–2.97 Smoking .184 <.001 .003 Yes Ref Ref Ref No, never 0.99 0.90–1.09 0.78 0.69–0.87 0.82 0.72–0.92 No, but ever 0.90 0.80–1.02 0.91 0.79–1.04 0.93 0.81–1.07 Alcohol usage <.001 .017 .012 Never Ref Ref Ref Not in previous 12 months 0.88 0.77–1.02 0.96 0.82–1.12 0.95 0.81–1.12 Monthly or less 0.68 0.61–0.77 0.83 0.72–0.95 0.82 0.71–0.94 2–4 times per month 0.73 0.64–0.83 0.93 0.79–1.09 0.92 0.78–1.08 2–3 times per week 0.64 0.55–0.75 0.92 0.76–1.10 0.89 0.74–1.08 ≥ 4 times per week 0.48 0.39–0.58 0.70 0.55–0.88 0.68 0.54–0.86 Difficulty with Dutch language <.001 No Ref Yes 1.32 1.21–1.43 Not applicable 0.80 0.71–0.91 Perceived health <.001 <.001 .002 Excellent Ref Ref Ref Very good 1.09 0.86–1.37 1.05 0.83–1.34 1.03 0.80–1.31 Good 1.55 1.27–1.90 1.22 0.99–1.51 1.20 0.97–1.49 Mediocre 2.47 2.01–3.04 1.37 1.09–1.72 1.32 1.05–1.67 Bad 3.85 3.02–4.91 1.69 1.28–2.23 1.59 1.20–2.11 Antibiotic-related behavior Higher antibiotic knowledge <.001 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 13 of 17 Table 4 Variables associated with having received ≥1 antibiotic prescription in the year prior to HELIUS visit in participants linked to AHD (N = 15,007) (logistic regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior OR (95% CI) P-values aOR (95% CI) P-values aOR (95% CI) P-values No Ref Yes 0.77 0.71–0.84 Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 4.72 3.59–6.21 3.07 2.28–4.14 Yes, occasionally 2.42 2.20–2.66 2.11 1.91–2.34 Did not finish treatment <.001 <.001 Always finished or no antibiotics Ref Ref Yes, regularly 2.70 2.02–3.62 1.80 1.29–2.50 Yes, occasionally 1.62 1.45–1.83 1.32 1.16–1.50 Abbreviations: OR Odds Ratio, aOR adjusted Odds Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner antibiotic prescriptions was second generation Ghanaian ethnic groups received less antibiotics for viral infections participants. Furthermore, we showed that a lower than non-Hispanic white children. level of antibiotic knowledge was not associated with Lower odds for higher level of antibiotic use know- receiving antibiotics or average number of antibiotic ledge were also seen among individuals who regularly or prescriptions, and that ethnic differences in antibiotic occasionally requested antibiotics from their GP or who use therefore cannot be explained by level of know- regularly or occasionally did not finish treatment. These ledge on antibiotics. findings suggest that improving antibiotic knowledge A previous study in Dutch primary care centres dem- might decrease the number of requests for antibiotics in onstrated higher use of antibiotics among first- primary care and improve appropriate use. generation migrants from Turkey, Morocco, Surinam or Our study has several strengths. First, the HELIUS the Antilles compared to Dutch, after adjustment for study consists of a large number of participants from age, sex, education, presence of chronic diseases, and major ethnic groups living in the same city, with repre- smoking [3]. We found that the odds of having ≥1 anti- sentation from all socioeconomic levels. Second, all out- biotic prescription was higher in some ethnic groups in comes and determinants were measured using the same unadjusted analysis, but after adjusting for several vari- methodology across all ethnic groups and HELIUS used ables including medical conditions, the odds were sig- translated questionnaires and had ethnically-matched in- nificantly higher among Ghanaian and first-generation terviewers and research assistants to provide assistance Moroccan participants only. In contrast, in our analyses during data collection. These procedures enhance the on the number of antibiotic prescriptions as an outcome, comparability between ethnic groups. Another major only second-generation Ghanaian migrants were at strength of the current study is that HELIUS data could higher risk of receiving a higher number of prescriptions be linked to data from a health insurance register cover- compared to Dutch participants. For all other ethnic ing the majority (77.7%) of the study population. groups, no evidence of a higher risk for more frequent Our study has also limitations. First, although HELIUS prescriptions was found, while even a lower number was participants were recruited via an ethnicity-stratified present for first-generation African Suriname and Turk- random selection of the municipal registry of ish participants. To the best of our knowledge, no other Amsterdam, the response rate for HELIUS study was studies have evaluated the variation in level of antibiotic 28% and there may be selection bias [8]. However, ana- knowledge and antibiotic use between ethnic groups and lysis from a previous HELIUS study have shown that thus our findings need to be confirmed. Notably, our participants are not exceedingly different from non- findings on antibiotic prescriptions and ethnicity are in respondents regarding sociodemographic variables [8]. line with a large retrospective cohort study performed in Second, we did not take into account the use of antibi- pediatric emergency departments in the United States otics purchased over the counter in the home country of [15]. This study also looked at the association between participants [6, 16–18], and we might therefore have ethnicity and antibiotic prescribing, showing that other underestimated antibiotic use in non-Dutch ethnic Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 14 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values Sociodemographics Ethnicity .004 .004 .001 Dutch Ref Ref Ref South-Asian Surinamese 1st generation 1.06 0.85–1.31 0.86 0.68–1.10 1.02 0.82–1.28 2nd generation 0.82 0.55–1.21 0.94 0.61–1.44 0.94 0.64–1.36 African Surinamese 1st generation 0.79 0.64–0.99 0.73 0.58–0.93 0.81 0.65–1.00 2nd generation 0.83 0.57–1.22 0.90 0.59–1.36 0.93 0.65–1.33 Ghanaian 1st generation 0.75 0.60–0.94 0.77 0.60–1.00 0.81 0.65–1.02 2nd generation 1.66 0.89–3.11 2.09 1.06–4.12 2.70 1.47–4.94 Turkish 1st generation 0.85 0.69–1.04 0.74 0.59–0.92 0.79 0.64–0.97 2nd generation 1.14 0.87–1.51 1.14 0.84–1.53 1.10 0.84–1.44 Moroccan 1st generation 0.99 0.81–1.21 0.89 0.70–1.11 0.89 0.72–1.10 2nd generation 0.84 0.63–1.13 0.92 0.67–1.27 1.02 0.76–1.37 Female sex 1.36 1.20–1.54 <.001 1.35 1.18–1.54 <.001 1.29 1.15–1.46 <.001 Age .023 < 25 years Ref 25–34 years 1.09 0.91–1.30 35–44 years 1.16 0.98–1.38 45–54 years 1.09 0.92–1.29 55–64 years 1.14 0.95–1.36 ≥ 65 years 1.44 1.16–1.80 Educational level .096 Unknown Ref No school/elementary school 1.74 0.86–3.52 Lower vocational/lower secondary school 1.74 0.86–3.52 Intermediate vocational/ intermediate secondary school 1.57 0.77–3.17 Higher vocational/university 1.45 0.71–2.96 Marital status .212 Married/registered partnership Ref Cohabiting 0.85 0.72–1.00 Unmarried/never married 0.99 0.89–1.10 Divorced/separated 1.01 0.90–1.13 Widow/widower 1.14 0.90–1.45 Health status Self-reported medical conditions (previous 12 months) Diabetes mellitus 1.34 1.17–1.54 <.001 1.21 1.04–1.41 .015 1.22 1.06–1.41 .005 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 15 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values CVA/one-sided loss of bodily function ≤1 day 1.10 0.95–1.29 .208 MI incl. ≥half hour chest pain or dotter/bypass operation 1.35 1.19–1.54 <.001 1.22 1.06–1.41 .005 Severe heart condition 1.26 1.04–1.52 .019 Malignant disorder 1.91 1.47–2.48 <.001 1.60 1.21–2.12 .001 1.60 1.28–2.00 Severe or chronic fatigue 1.29 1.15–1.44 <.001 High blood pressure 1.14 1.04–1.25 .006 Artery stenosis 1.30 1.08–1.57 .005 Respiratory diseases 1.50 1.32–1.71 <.001 1.34 1.16–1.54 <.001 1.29 1.13–1.47 <.001 Serious/persistent intestinal disorders 1.24 1.07–1.44 .004 Psoriasis 1.08 0.90–1.31 .406 (Chronic) eczema 1.21 1.07–1.37 .002 1.14 1.00–1.29 .042 Incontinence 1.31 1.13–1.50 <.001 Body Mass Index .736 < 18.5 Ref 18.5–25 0.98 0.70–1.37 25–30 0.98 0.70–1.37 30–40 1.04 0.74–1.46 ≥ 40 1.06 0.71–1.57 Smoking .099 Yes Ref No, never 1.13 1.00–1.29 No, but ever 1.02 0.87–1.21 Alcohol usage .075 Never Ref Not in previous 12 months 0.93 0.77–1.13 Monthly or less 0.94 0.80–1.11 2–4 times per month 1.01 0.85–1.20 2–3 times per week 0.73 0.58–0.91 ≥ 4 times per week 0.81 0.61–1.08 Difficulty with Dutch language .320 No Ref Yes 0.94 0.86–1.04 Not applicable 1.05 0.88–1.25 Perceived health <.001 .001 <.001 Excellent Ref Ref Ref Very good 0.98 0.68–1.42 1.03 0.71–1.49 0.99 0.70–1.39 Good 1.29 0.94–1.77 1.19 0.85–1.65 1.17 0.86–1.59 Mediocre 1.73 1.25–2.39 1.40 1.00–1.97 1.44 1.05–1.97 Bad 2.30 1.62–3.26 1.71 1.17–2.49 1.72 1.22–2.43 Antibiotic-related behavior Higher antibiotic knowledge .054 Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 16 of 17 Table 5 Variables associated with number of antibiotic prescriptions in participants linked to ADH (N = 15,007) (zero-inflated negative binomial regression analysis) (Continued) Univariable Multivariable excluding Multivariable including variables on antibiotic- variables on antibiotic- related behavior related behavior IRR (95% CI) P-values IRR (95% CI) P-values IRR (95% CI) P-values No Ref Yes 0.92 0.85–1.00 Ever asked GP for antibiotics <.001 <.001 No, never Ref Ref Yes, regularly 2.96 2.34–3.73 1.87 1.45–2.41 Yes, occasionally 1.75 1.59–1.92 1.15 1.06–1.30 Did not finish treatment <.001 Always finished or no antibiotics Ref Yes, regularly 1.50 1.12–2.00 Yes, occasionally 1.32 1.18–1.48 Abbreviations: IRR Incidence Risk Ratio, CI Confidence Interval, CVA Cerebro Vascular Accident, MI Myocardial infarction, GP General Practitioner Accounts for zero-inflated distribution groups. As a recent HELIUS study found that Dutch Conclusions people of Turkish or Moroccan origin were more likely To our knowledge, this study is the first to examine eth- to use healthcare in the Netherlands as well as their nic disparities in level of antibiotic knowledge and use in country of origin [19], underestimation of antibiotic use a large population-based sample among adults with dif- in non-Dutch ethnic groups seems unlikely. Third, since ferent ethnic backgrounds. Health policy makers and several characteristics, such as education level and med- healthcare professionals are increasingly developing in- ical conditions, of HELIUS participants insured at Ach- terventions to improve the quality of antibiotic use, mea differed from those insured elsewhere, selection which is needed to help contain antimicrobial resistance. bias could have been introduced in analysis on antibiotic Targeted campaigns can be considered, for instance, use. This difference could be due to the fact that the during the annual European Antibiotic Awareness Day, City of Amsterdam provided health insurance discounts since this event addresses improvement in the quality of with Achmea for low-income individuals. These differ- antibiotic use to the general public [20]. Still, this study ences were corrected for during multivariable analyses shows that a lower level of antibiotic knowledge is not to the most possible extent. Fourth, the variable ‘ever necessarily linked to higher antibiotic usage, indicating asked GP for antibiotics’ does not discriminate between that interventions aimed at improving knowledge alone appropriate or inappropriate requests for antibiotics and might be insufficient to reduce antibiotic use. Neverthe- misclassification might have occurred. However, this less, the underlying reasons for these findings need variable gives some information on participants’ atti- further evaluation. tudes towards antibiotic use. Furthermore, due to pri- vacy restrictions, we were unable to include indication Supplementary information Supplementary information accompanies this paper at https://doi.org/10. for antibiotic therapy and duration of antibiotic use as 1186/s13756-019-0636-x. additional indices for antibiotic use (apart from the number of antibiotics prescribed). Moreover, since this Additional file 1. Supplementary Methods. was a cross-sectional study, we were unable to model Additional file 2: Table S1. Characteristics of participants not linked antibiotic knowledge with future antibiotic prescriptions. versus linked to the Achmea Health Database. Further research should examine the association of anti- biotic knowledge with future antibiotic prescriptions. Fi- Abbreviations AHD: Achmea Health Database; aIRR: adjusted Incidence Risk Ratio; nally, we are unable to determine if individuals were aOR: adjusted Odds Ratio; BMI: Body Mass Index; CVA: Cerebrovascular more demanding towards their GP or if their GPs Accident; GP: General Practitioner; HELIUS: Healthy life in an Urban Setting; were more lenient in prescribing antibiotics during IQR: Interquartile Range; IRR: Incidence Risk Ratio; MI: Myocardial Infarction; OR: Odds Ratio; WHO: World Health Organization illness [4, 5]. Neither completing antibiotic therapy, assessed by pill count, nor duration of antibiotic use Acknowledgements could be taken into account as these data were not The authors would like to acknowledge the HELIUS participants for their available. contribution; the HELIUS team for data collection and management. Schuts et al. Antimicrobial Resistance and Infection Control (2019) 8:200 Page 17 of 17 Authors’ contributions 7. Deschepper R, Grigoryan L, Lundborg CS, Hofstede G, Cohen J, Kelen GV, ES, MS, SG, MP and JP were involved in the conception and design of the et al. Are cultural dimensions relevant for explaining cross-national study. ES, EvD and MS were involved in the acquisition of data. EvD and AB differences in antibiotic use in Europe? BMC Health Serv Res. 2008;8:123. analyzed the data. ES and EvD wrote the manuscript. ES, EvD, AB, MS, SG, MP 8. Snijder MB, Galenkamp H, Prins M, Derks EM, Peters RJG, Zwinderman AH, and JP were major contributors in revising the manuscript. All authors read et al. Cohort profile: the Healthy life in an urban setting (HELIUS) study in and approved the final manuscript. Amsterdam, The Netherlands. BMJ Open. 2017;7(12):e017873. 9. Stronks K, Snijder MB, Peters RJ, Prins M, Schene AH, Zwinderman AH. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. Funding BMC Public Health. 2013;13:402. The HELIUS study is conducted by the Academic Medical Center Amsterdam 10. StatLine. Population key figures 2019. https://opendata.cbs.nl/statline/#/CBS/ and the Public Health Service of Amsterdam. Both organizations provided nl/dataset/37296ned/table?ts=156888214600. core support for HELIUS. The HELIUS study is also funded by the Dutch Heart 11. StatLine. Population data; age, migration background, gender and region. Foundation (2010 T084), the Netherlands Organization for Health Research 2019. https://opendata.cbs.nl/statline/#/CBS/nl/dataset/37713/table?ts=156 and Development (ZonMw) (200500003), the European Union (FP-7) (278901), and the European Fund for the Integration of non-EU immigrants 12. Stronks K, Kulu-Glasgow I, Agyemang C. The utility of 'country of birth' for (EIF) (2013EIF013). The authors were independent from funders and had full the classification of ethnic groups in health research: the Dutch experience. access to all data. The decision to submit for publication was uninfluenced Ethn Health. 2009;14(3):255–69. by the funders. The authors of this study take full responsibility for the integ- 13. Data Protection Authority. Ministry of Health, Welfare and Sport. https:// rity of the data and the accuracy of the data analysis. autoriteitpersoonsgegevens.nl/en. 14. Special Eurobarometer 338: Antimicrobial Resistance (2010). http://ec. Availability of data and materials europa.eu/public_opinion/archives/eb_special_339_320_en.htm#338. The datasets used in the current study are available from the corresponding 15. Goyal MK, Johnson TJ, Chamberlain JM, Casper TC, Simmons T, Alessandrini author on request, contingent on approval from the HELIUS scientific EA, et al. Racial and Ethnic Differences in Antibiotic Use for Viral Illness in committee. Emergency Departments. Pediatrics. 2017;140(2):e20170203 (4). 16. Lindenmeyer A, Redwood S, Griffith L, Ahmed S, Phillimore J. Recent Ethics approval and consent to participate migrants’ perspectives on antibiotic use and prescribing in primary care: a The HELIUS study is conducted in accordance with the Declaration of qualitative study. Br J Gen Pract. 2016;66(652):e802–e9. Helsinki and has been approved by the AMC Ethical Review Board. All 17. Hu J, Wang Z. Non-prescribed antibiotic use and general practitioner service participants provided written informed consent. utilisation among Chinese migrants in Australia. Aust J Prim Health. 2016; 22(5):434–9. Consent for publication 18. Hu J, Wang Z. In-home antibiotic storage among Australian Chinese Not applicable. migrants. Int J Infect Dis. 2014;26:103–6. 19. Sekercan A, Snijder MB, Peters RJG, Stronks K. Is healthcare consumption in Competing interests the country of origin among Moroccan and Turkish migrants of older age The authors declare that they have no competing interests. (55+) associated with less use of care in the Netherlands? Tijdschr Gerontol Geriatr. 2018;49(6):253–62. Author details 20. European Centre for Disease Prevention and Control. Antimicrobial Department of Internal Medicine, Division of Infectious Diseases, Amsterdam consumption. In: ECDC, editor. Annual epidemiological report for 2016. UMC, University of Amsterdam, Meibergdreef 9, 1105, AZ, Amsterdam, The Stockholm: ECDC; 2018. https://ecdc.europa.eu/en/publications-data/ Netherlands. Department of Infectious Diseases, Public Health Service antimicrobial-consumption-annual-epidemiological-report-2016. Amsterdam, Nieuwe Achtergracht 100, 1018, WT, Amsterdam, The Netherlands. INSERM, Sorbonne Université, Institut Pierre Louis d’Épidémiologie et de Santé Publique, Paris, France. Department of Public Publisher’sNote Health, Amsterdam UMC, University of Amsterdam, Amsterdam Public Health Springer Nature remains neutral with regard to jurisdictional claims in Research Institute, Amsterdam, The Netherlands. Department of Clinical published maps and institutional affiliations. Epidemiology, Biostatistics and Bioinformatics, Amsterdam UMC, University of Amsterdam, Amsterdam Public Health Research Institute, Amsterdam, The Netherlands. Received: 31 July 2019 Accepted: 24 October 2019 References 1. WHO. Antimicrobial resistance: global report on surveillance 2014; 2014. p. 257. 2. Nellums LB, Thompson H, Holmes A, Castro-Sanchez E, Otter JA, Norredam M, et al. Antimicrobial resistance among migrants in Europe: a systematic review and meta-analysis. Lancet Infect Dis. 2018;18(7):796–811. 3. Hogenhuis CC, Grigoryan L, Numans MM, Verheij TJ. Differences in antibiotic treatment and utilization of diagnostic tests in Dutch primary care between natives and non-western immigrants. Eur J Gen Pract. 2010;16(3):143–7. 4. Grigoryan L, Burgerhof JG, Degener JE, Deschepper R, Lundborg CS, Monnet DL, et al. Determinants of self-medication with antibiotics in Europe: the impact of beliefs, country wealth and the healthcare system. J Antimicrob Chemother. 2008;61(5):1172–9. 5. Grigoryan L, Burgerhof JG, Degener JE, Deschepper R, Lundborg CS, Monnet DL, et al. Attitudes, beliefs and knowledge concerning antibiotic use and self-medication: a comparative European study. Pharmacoepidemiol Drug Saf. 2007;16(11):1234–43. 6. Norris P, Ng LF, Kershaw V, Hanna F, Wong A, Talekar M, et al. Knowledge and reported use of antibiotics amongst immigrant ethnic groups in New Zealand. J Immigr Minor Health. 2010;12(1):107–12.

Journal

Antimicrobial Resistance & Infection ControlSpringer Journals

Published: Dec 6, 2019

There are no references for this article.