Lower knowledge about diabetes among foreign‐born compared to Swedish‐born persons with diabetes—A descriptive study

Abstract Aim To compare foreign‐ and Swedish‐born persons, diagnosed with type 2 diabetes, to study whether there are dissimilarities in knowledge about diabetes and to study determinants of knowledge. Design A cross‐sectional descriptive study was conducted. Method Data were collected between September 2014 and March 2016, using the standardized Diabetes Knowledge Test (DKT), statistically analysed. Results The results showed dissimilarities in knowledge between foreign‐ and Swedish‐born persons, supporting the hypothesis that foreign‐born persons had lower knowledge about diabetes than Swedish‐born persons. There was a relationship between poor knowledge and country of birth, marital status and employment status. Country of birth was the strongest independent determinant of knowledge about diabetes. The risk of poor knowledge was ten times higher among persons born in the Middle East or in another country outside Europe compared with Swedish‐born persons. Other influencing factors for poor knowledge about diabetes were being not gainfully employed and living alone.

. Type 2 diabetes is a progressive disease with the risk of developing micro-and macrovascular complications, influencing health negatively and possibly leading to high costs for health care and suffering for the individual (IDF, 2015, Socialstyrelsen, 2015. Although serious complications may develop, the disease can be controlled, complications can be prevented and the incidence can be delayed by practicing adequate self-care. For patients with type 2 diabetes, active participation in self-care based on knowledge about the disease is thus of great importance (IDF, 2015).
The risk of being affected with type 2 diabetes is increased in migrants, particularly those from countries outside Europe, compared with the native population of the host country and this is an increasing problem for several European countries (IDF, 2015;Shaw, Sicree, & Zimmet, 2010;Whiting, Guariguata, Weil, & Shaw, 2011;Zimmet & Alberti, 2006). This leads to personal suffering as well as public health challenges and high costs (IDF, 2015). Sweden, like many other European countries, has changed to a multicultural society due to ongoing global migration (SOS, 2015, SCB, 2016. Today, about 17% of the population in Sweden are born abroad (SCB, 2016).
The migrant population includes more than 200 different nationalities, with the largest groups coming to work voluntarily as labour migrants from Finland and former Yugoslavia and during the 1990s also as refugees forced to flee from the war (SOS, 2015, SCB, 2016.
During the last decade, migration to Sweden has been dominated by persons of non-European origin, with the largest group being refugees from the Middle East. Most non-European migrants coming to Sweden are refugees.
Health-related behaviour, including self-care, is determined by beliefs about health and illness, held by the individual and culturally determined, based on the person's knowledge and refined by experiences (Glanz, Rimer, & Lewis, 2008;Hjelm & Bard, 2013;Hjelm, Bard, Nyberg, & Apelqvist, 2003, 2005. The behaviour is also influenced by whether the person perceives the causes of health and disease as possible to influence or not, in other words having an internal vs. an external locus of control (Rotter, 1966). Furthermore, behaviour is influenced by the perceived risk, threat or susceptibility of being affected by a disease, the perceived seriousness of the disease and the likelihood of action due to perceived benefits or barriers/costs to behaviour change required by the disease (Rosenstock, Strecher, & Becker, 1988). Other influencing factors are, for example age, gender, education, ethnic origin and experiences related to migration Rosenstock et al. (1988) and Glanz et al. (2008). As regards type 2 diabetes, self-care is crucial for disease control and prevention of diabetes-related complications (IDF, 2015). A person should receive individualized information about their health and health-related behaviour, and the information should be individually tailored and culturally appropriate (Leininger & McFarland, 2006). Thus, it is important to meet the needs of people from different countries and healthcare staff must have an understanding and knowledge of how cultural background affects perceptions of health and illness (Leininger & McFarland, 2006). Culturally adapted or appropriate care or health education can be defined as education tailored to the cultural and religious beliefs and linguistic skills of the community being approached, but also taking into consideration likely literacy skills (Hawthorne, Robles, Cannings-John, & Edwards, 2008). In Sweden, persons diagnosed with type 2 diabetes are generally managed in primary healthcare centres or in specialized in-hospital-based clinics in the case of complications related to diabetes. Irrespective of origin, all patients should receive the same care (SFS, 2017). The exception is that non-Swedish-speaking persons, according to the law, have the right to get an interpreter in all contacts with authorities (SFS, 1986). However, access to interpreters varies and can be limited, and therefore, many consultations in health care are made without an interpreter or using relatives (Hadziabdic & Hjelm, 2014).
In the literature review, no previous studies comparing knowledge about diabetes between foreign and a native population have been found. A study that indicates limited knowledge about diabetes among Turkish migrants diagnosed with diabetes, living in Germany (Kofahl, Knesebeck, Hollman, & Mnich, 2013), has been found. Further, a few studies support this path, aiming to implement and evaluate intervention programmes that measure the knowledge about diabetes in different migrant groups of Korean (Choi & Rush, 2012;Song, Han, & Lee, 2010), Chinese (Chesla, Chun, & Kwan, 2013;Sun, Tsoh, Saw, Chan, & Cheng, 2012) and Bangladeshi (Islam, Wyatt, & Patel, 2013) origin living in the USA as an outcome measure.
Qualitative studies comparing foreign-and Swedish-born persons with type 2 diabetes have indicated dissimilarities in knowledge and perceived seriousness of diabetes, with limited knowledge and lower risk awareness, which might affect health-related behaviour including self-care (Hjelm & Bard, 2013;Hjelm et al., 2003Hjelm et al., , 2005.
Thus, it was hypothesized that foreign-born persons have lower knowledge about diabetes than Swedish-born persons. The aim of the study was to compare foreign-and Swedish-born persons, diagnosed with type 2 diabetes, to study whether there are dissimilarities in knowledge about diabetes mellitus and to study determinants of knowledge.

| Design and method
This was a cross-sectional descriptive study with data collected with a standardized and validated self-report instrument, Diabetes Knowledge Test (DKT; Fitzgerald et al., 1998), to investigate knowledge about diabetes mellitus. The design was chosen to give the possibility to gather information about the variable knowledge in migrants with type 2 diabetes and to enable studies of relationships with the person's background characteristics (Creswell, 2014).

| Participants
A convenience sampling procedure was used. In this study, a distinction will be made between people born in Sweden and abroad and people born in another country than Sweden are considered as foreign-born persons (SOS, 2015). Invited were all known foreign-born persons diagnosed with type 2 diabetes according to ICD E11 (WHO, 2015) managed in a diabetes clinic in a healthcare centre in an immigrant-dense area in a county in Sweden, aged ≥18 years and with duration of diabetes ≥1 year. Persons with known psychiatric diagnoses ICD F00-F29/F60-F99 (WHO, 2015) were excluded since cognitive impairment might influence the results. Inquiries, to participate in the study, were sent to persons who met the inclusion criteria (N = 379), with reminders 3 weeks apart. Fifty-two persons did not want to participate in the study, 14 inquiries came back since the person could not be reached at the current address and two persons did not understand the interview questions, despite an interpreter being present and were thus excluded; 242 persons did not answer.
Of the foreign-born persons, 69 were included in the study. The subsequent process was to recruit 69 Swedish-born persons, matched for gender, by the same procedure as the foreign-born group. The study population did not differ from the non-respondents as regards gender, age and country of birth (p = 0.49, 0.66, 0.82).

| Data collection
Data were collected between September 2014 and March 2016 by structured interviews based on questionnaires. In this study, the Diabetes Knowledge Test (DKT; Fitzgerald et al., 1998) and socio-demographic background data are reported. Permission for the study was received from the operation manager after oral and written information about the study from the principal investigator (PI; first author). Then, the diabetes specialist nurses at the clinic were informed, orally and in writing, about the implementation of the study. After this, one diabetes specialist nurse identified persons meeting the inclusion criteria, from computerized registers at the clinic. Then, a letter with information about the study and a prepaid response envelope, addressed to the clinic, was sent. Two reminders, with 3 weeks apart, were sent in the event of no answer. The inquiry was translated into Arabic and Bosnian/Croatian/Serbian, as they are the main language groups among migrants in Sweden (SCB, 2016)  A pilot test was performed in ten foreign-born persons to assess how the interview guide worked. Face and content validity were tested (Polit & Beck, 2012). The instrument worked well, and therefore, all interviews were included in the study.

The Diabetes Knowledge Test (DKT), developed by the Diabetes Research and Training Center at the University of
Michigan (Fitzgerald et al., 1998), was used as one of the several instruments, after approval was obtained from the copyright holder. DKT is a questionnaire that consists of 23 items, measuring a person's knowledge about diabetes in a general part (items 1-14), an insulin use part (items 15-23) and a total part (items 1-23). For every item, there are several alternative answers. In this study, the general part was used for all participants while the insulin part, and thereby also the total part, was answered by insulin-treated participants only (see Table 1). DKT has demonstrated adequate validity and reliability, Cronbach's alpha > 0.70 (Fitzgerald et al., 1998) and has been adapted for use in many countries throughout the world and translated by researchers into several languages (Fitzgerald et al., 2016). DKT was translated into Swedish and back-translated into English by two different authorized professional translators. Translation into Swedish was done in several steps to ensure that the essential meaning of the items was preserved (Polit & Beck, 2012). The original questionnaire in English was translated into Swedish by an authorized professional translator and then retranslated back into English by another authorized professional translator. The PI for the study then reviewed the two English versions and agreed that they were equivalent. The interviews were performed with professional interpreters in respective language. DKT has been used in several study populations with different origin and languages (Chesla 2012;Choi & Rush, 2012;Mufunda, Albin, & Hjelm, 2012;Islam et al., 2013;Kofahl et al., 2013).

| Data analysis
To describe the data, numbers and percentage, mean (SD) and median (range) were used (Altman, 1994). Comparisons between groups were made by tests of statistical significance. For continuous variables, Student's t test was used for normally distributed variables and Mann-Whitney U-test for non-normally distributed variables and chi-squared test was used to ascertain any differences between categorical variables; p < 0.05 was considered statistically significant.
To determine diabetes knowledge measured by the DKT (Fitzgerald et al., 1998), one point was awarded for a correct answer and zero for a wrong answer or no response. The total knowledge score ranged from 0 to 23 and was categorized as: <11 = poor knowledge, 11-17 = average knowledge and >17 good knowledge For analytical statistics, to identify any independent associations between knowledge and socio-demographic variables and diabetesrelated characteristics, multiple logistic regression analysis (stepwise logistic regression, backward conditional) with calculation of odds ratio (OR; 95% CI) was used (Altman, 1994). The analysis was per-  Table 5).

| Ethics
The study was approved by a Regional Ethics Committee and implemented in accordance with the Helsinki Declaration (World Medical Association, 2013). Written informed consent was obtained from the respondents. All collected data were anonymized and coded so that no participant could be identified. Collected data were stored in a locked space at the PI's workplace and inaccessible to the healthcare staff. Results were analysed and presented on a group level and in such a way that no one could be identified.

| Respondents' socio-demographic and diabetes-related data
The study population included 138 participants, both men and women, of whom 69 were foreign-born persons, aged 33-90 years and 69 were Swedish-born, aged 48-91 years. The foreign-born group included mainly persons born in the Middle East and former Yugoslavia (Table 1).
In the foreign-born group, the majority (75%) had received their diabetes diagnosis in Sweden (Table 2)

| Diabetes knowledge
The Swedish-born group had significantly higher knowledge in all three parts of the instrument, compared with the foreign-born group (

| Determinants of knowledge about diabetes
The multifactorial influence on knowledge (

| Knowledge gaps
Knowledge deficit, defined as over 50% incorrect or missing answers, was noted in the general part in eight questions in the foreign-born group, compared with four questions in the Swedishborn group (  (17) Years of residence in Sweden, mean (SD) 25 (17) (56.5%) or in carbohydrates (55.1%) and which problem is usually not associated with diabetes-related complications (53.6). Among the Swedish-born persons, knowledge deficit was noted in four questions, in descending order: Which food is highest in fat (72.5%), definition of a "free food" (60.9%), what numbness and tingling may be symptoms of (58%) and food that should not be used to treat low blood glucose (55.1%).  (Hjelm & Bard, 2013;Hjelm et al., 2003Hjelm et al., , 2005 (Choi & Rush, 2012;Song et al., 2010), China (Chesla et al., 2013;Sun et al., 2012) and Bangladesh (Islam et al., 2013), limited knowledge has been shown, as a result of interventions aiming to improve knowledge about diabetes in migrants. These studies show that the migrants' knowledge initially was limited, which needs to be considered in diabetes care and particularly when designing management strategies, including teaching activities as recommended (Creamer, Attridge, Ramsden, Cannings-John, & Hawthorne, 2015;Hawthorne et al., 2008;Socialstyrelsen, 2015).

| D ISCUSS I ON
This study showed lack of knowledge in the areas of diet, how to achieve glycaemic control and dealing with diabetes-related complications among the foreign-born group, and therefore, future patient education for this group should focus on these areas of knowledge.
In this study, the majority of foreign-born persons (75%) had received their diagnosis in Sweden and thus ought to have had the same diabetes education as Swedes (Socialstyrelsen, 2015). Studies have shown a lack of culturally adapted diabetes education for persons of foreign background living in Sweden (Socialstyrelsen, 2015) and in other countries (Creamer et al., 2015;Hawthorne et al., 2008), which needs to be considered.
However, the main results showed a difference in knowledge between the foreign-born and the Swedish-born group and to be able to improve knowledge about diabetes among foreign-born persons, possible reasons for the difference need to be identified. To start with, it is prerequisite that the person understands information given if a person's knowledge about an illness is to be improved. Hadziabdic and Hjelm (2014) highlighted the use of a professional interpreter during patient visits, to avoid communication barriers and to improve the impact on quality of health care, but it was not always obvious due to lack of authorized professional interpreters. The next step will be to understand which determinants knowledge is based on. According to Rosenstock et al. (1988) and Glanz (2008), health-related behaviour, including self-care, is determined by beliefs about health and illness held by the individual, based on the person's knowledge and refined by experiences. Therefore, it is essential to identify a person's health beliefs and attitude towards an illness to improve knowledge and to know whether a person perceives a disease as a threat or not (Hjelm et al., 2005). The feeling of not being able to influence a situation, in other words having an external locus of control, might also affect the acquisition of knowledge by leading to less interest in learning more about an illness (Rotter, 1966).
Further, limited knowledge about a disease might result in lowrisk awareness and perceived low severity of health and illness and low basic knowledge about the body, diabetes and health might affect how a person can understand information given (Hjelm et al., 2003(Hjelm et al., , 2005. Previous research has emphasized the importance of knowing a person's migratory background, which might affect the acculturation process including self-care behaviour and knowledge ingestion (Berry, 2005;Hjelm & Bard, 2013;Hjelm et al., 2003;Hull, 1979).
Today, the migrant population in Sweden has expanded to include  Johansson, 2000), which has been shown to act as a barrier to receiving information (Hjelm et al., 2005).
When considering migratory background and origin in the studied population, mainly refugees from the Middle East and former Yugoslavia, the cultural distance compared with the Swedish population was large, for example in language, social relations, family structure and cultural values (Triandis, 2000). Thus, it can be difficult to understand each other in a patient education situation and particularly as there is a lack of culturally appropriate and culturally adapted education models in diabetes care (Creamer et al., 2015;Hawthorne et al., 2008;Socialstyrelsen, 2015). Thus, it is important to consider that persons coming from countries with great cultural distance might experience difficulties in teaching situations (Triandis, 2000). Research has demonstrated that when migration was self-initiated, the acculturation process, including adaptation to a new situation, for example teaching situations, became more positive (Hull, 1979), and thus, the acculturation stress decreased (Berry, 2005). Acculturation stress which might also be another barrier against information delivered.
In addition to the migration situation that might act in different ways as a barrier to adopting new knowledge, most diabetes education is delivered when the person is newly diagnosed. This is often at the same time as the patient is in a crisis reaction, due to having been informed about having a chronic diagnosis (Cullberg, 2006).
This might also be a barrier to information, as we know that during the crisis reaction many are mentally blocked, and this may cause difficulties in assimilating new knowledge. The question is what repeated information and teaching there has been for this particular group over time. Another influencing factor is the shorter duration of diabetes in foreign-born compared with Swedish-born persons, and thus, they have had less contact with health care and exposure to diabetes-related knowledge.
Research has shown that a good social network can have a positive effect on self-care behaviour such as knowledge ingestion (Berterö & Hjelm, 2010). Being unemployed and/or living alone could lead to limited social network, which previous studies have emphasized as playing an important role in supporting positive selfcare behaviour among persons with diabetes (Lanting et al., 2008), and hence, the absence of a social network can affect the way a person manages their disease and gathers knowledge about it (Berterö & Hjelm, 2010). In this study, foreign-born compared with Swedishborn persons were to greater extent unemployed and on sick leave and had a higher risk of poor knowledge about DM than those who were gainfully employed. In addition, the study showed that participants living alone had a higher risk of poor knowledge than those living with a partner. Migrants in Sweden, particularly those of non-European origin, are more often unemployed and living in limited social networks (Socialstyrelsen, 2009).
A person's self-care ability is related to individual beliefs about health and illness, which form their attitudes to self-care and which are culturally determined and based on their knowledge, and thus, it is of great importance to assess these and base all educational activities on them to achieve a successful acculturation process and increase knowledge about an illness and improve health (Leininger & McFarland, 2006). Moreover, according to the Swedish Health and Medical Services Act, a person should be given individualized information about their health and health-related behaviour (SFS, 2014(SFS, , 2017.
The results of this study showed that foreign-born persons living in Sweden, born in the Middle East or in another country outside Europe, who were living alone or were not gainfully employed were liable to have poor knowledge about diabetes.

| Study limitations
The sample size might be seen as limited, but analysis of dropouts did not show any significant dissimilarities compared with the listed population of foreign-born persons diagnosed with type 2 diabetes concerning gender, age and country of birth. However, the study population is representative of how the migrant population in Sweden is composed, mainly with persons from the Middle East and former Yugoslavia (SCB, 2016).

| CON CLUS ION
Foreign-born persons with type 2 diabetes, particularly those from a country outside Europe, living in Sweden, had lower knowledge about diabetes compared with Swedish-born persons.
For future health policy and services, it is of great importance to assess person's own knowledge and develop appropriate individualized education including diet, glycaemic control and diabetes complications aiming to increase knowledge about diabetes among foreign-born persons and thereby improve self-care behaviour and prevent diabetes-related complications to improve or retain health. Thus, appropriate diabetes education models, for foreign-born persons, should be developed and implemented in health care.

ACK N OWLED G EM ENTS
We would like to thank the diabetes nurses at the studied diabetes clinic for good cooperation that made it possible to carry out this study. We are grateful to Associate Professor Per Nyberg for consultation concerning statistics.

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no conflict of interest.