Extent of the association between self-rated health and place of birth: a cross-sectional study among people at high risk of developing pre-diabetes and diabetes in Sweden.

OBJECTIVE: The first aim was to determine the extent of the relationship between place of birth and self-rated health (SRH) in primary healthcare patients born outside Sweden and those born in Sweden. The second aim was to investigate whether socioeconomic and lifestyle factors explained any differences.
SETTING: Two academic primary healthcare centres in Stockholm County, Sweden. PARTICIPANTS: 825 patients at high risk of developing pre-diabetes and diabetes, aged 18-74 years, attending academic healthcare centres in areas with large numbers of immigrants, 56.8% born abroad and 43.2% born in Sweden. Patients with a diagnosis of diabetes were excluded. Inclusion criteria were based on previous research showing that people born in Middle Eastern and Asian countries who live in Sweden have a high prevalence of and risk for diabetes. OUTCOME: SRH was dichotomised as optimal (very good/good) and suboptimal (fair/bad/very bad) and compared in those born outside Sweden and in Sweden.
RESULTS: There was a statistically significant difference in the SRH of the two groups (p=0.008). Logistic regression analysis showed a crude OR for reduced SRH of 1.46 (95% CI 1.10 to 1.92) in patients born outside Sweden. After controlling for education, employment and marital status, the OR increased to 1.50 (95% CI 1.11 to 2.02). After controlling for physical activity and smoking, it decreased to 1.36 (95% CI 1.00 to 1.85).
CONCLUSION: Socioeconomic and lifestyle factors influenced SRH. It could therefore be useful for clinicians to consider these factors when providing care for patients born outside Sweden and resettled in areas with large numbers of immigrants. © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

The questionnaire was available in different languages.

Study included two healthcare centres catchment areas with a large number of immigrants.

Non-random sampling limited the ability to draw conclusions.

Those born outside Sweden were grouped together regardless of place of birth, potentially masking inequalities by place of birth.

The study did not include information on cultural perceptions about health.

Background

According to the WHO, good health includes a balance between good objectively measurable and good subjectively experienced health. Whereas objective health is assessed by diagnosing diseases and measuring various clinical parameters, subjective health combines the physical, mental and social well-being of the individual.1 Self-rated health (SRH) has emerged as a unified, reliable measure of subjective health status that can provide valuable information and serve as a complement to objective health measures.2–4 SRH is widely used in the field of public health, including national and international health surveys.5 Prior studies have shown that SRH is an independent predictor of mortality.4 6 7 It is associated with higher rates of mortality from numerous diseases, including cardiovascular diseases8 and diabetes, as well as all-cause mortality.9 A meta-analysis from 2006 found that people who reported their health as fair or bad had a twofold higher risk of all-cause mortality than those who reported their health as good.7

SRH is influenced by many factors. Examples include gender,10 socioeconomic status (SES)11 and age.12 Women, people with low SES and older adults are more likely to report poor health. Lifestyle factors such as obesity, physical inactivity and smoking are also related to poor SRH.11 13 Furthermore, the way people rate their health in relation to self-reported diseases or functional ability seems to vary by country, as shown in a comparative study of people in Finland, Sweden, Spain, the Netherlands and Israel.12 However, some indicators seem to be significantly associated with SRH regardless of country of origin, including marital status, hypertension, stroke, diabetes, cancer, musculoskeletal diseases and activities of daily living.12

The factors that potentially influence how people rate their health differ from one individual to another, and even those with similar levels of objective ill health can perceive their health differently. A study of young people with type 1 diabetes mellitus found that participants varied in their SRH even after adjusting for multiple risk factors.14 Similarly, studies have found discrepancies in SRH among people with hypertension15 and advanced cancer.16

Immigrants make up a growing portion of the Swedish population, currently over 20%.17 In 2016, almost 80% of the population growth in Sweden resulted from immigration. Prior studies suggest that immigrants in Sweden generally have lower SRH than their Swedish-born counterparts.18 19 Many studies on SRH in immigrant groups in Sweden focus on Finnish immigrants,20 21 as they constitute one of the largest immigrant groups in the country. Similarly, studies have been conducted on immigrants from other European countries living in Sweden.18 22 According to such studies, people born in Finland, Eastern Europe and Southern Europe living in Sweden rate their health poorer than those born in Sweden.20 23 A similar pattern has been observed in immigrants from outside Europe, including adolescent boys,24 immigrants from the Middle East25 and Assyrians with diabetes from Turkey.26

Research also shows that immigrants are more vulnerable to diseases than the native-born populations.27–29 This raises the question of whether differences in SRH could be explained by differences in objective health status. However, in previous studies, the inclusion of diseases has not eliminated the differences in SRH between foreign-born and native-born populations. In addition, a study comparing Kurdish immigrants to people born in Sweden found that the immigrants were more likely to report lower SRH despite comparable objective health, including chronic diseases, eczema, psoriasis, urinary tracts complaints and diabetes.26

Comparison between studies of SRH in immigrants versus native populations is made difficult by differences in study populations and methodological approaches. For instance, many previous studies in Sweden focus on middle-aged and older adults which limits their representativeness. According to Statistics Sweden, people between the ages of 18 and 35 years make up around 20% of the total immigrant population.30 This group therefore needs to be included when studying immigrants’ health.

There is a gap in knowledge about the SRH of people living in Sweden who were born outside Sweden, particularly those born in the Middle East, Asia and Africa—a group with a high risk of developing pre-diabetes and diabetes.31 32 More information about SRH in this group has the potential to improve healthcare services for these people.

Aims

The first aim of the current study was to determine the extent of the relationship between place of birth and SRH in primary healthcare patients born outside Sweden and those born in Sweden. The second aim of the study was to investigate whether socioeconomic and lifestyle factors explained any differences.

Methods

The study design was cross-sectional.

Data source

Data were drawn from the ‘4D Diabetes Project: Screening and Treatment of Prediabetes and Diabetes in Primary Care – A Pilot Study’, a large study conducted by Karolinska Institutet and the Stockholm County Council. The overall aim of the 4D Diabetes Project was to develop methods and procedure for early identification and/or prevention of pre-diabetes and diabetes in people at high risk for those diseases. People born in the Middle East, Asia and Africa who live in Sweden have a higher risk of developing pre-diabetes and diabetes than native Swedes and develop diabetes on an average of 10 years earlier.31 32 Two areas with large populations of immigrants—in particular, immigrants born in the Middle East and Asia—were selected for the 4D Diabetes Project: Flemingsberg and Jakobsberg.30 Inclusion criteria for the 4D Diabetes Project were as follows: attending an academic primary healthcare centre in Flemingsberg or Jakobsberg during the period between 2013 and 2015, being between the ages of 18 and 74 years and being either a first-generation or second-generation immigrant or a person born in Sweden whose parents were born in Sweden. Patients with a diagnosis of diabetes were excluded from the study. The project aimed to consecutively recruit patients attending participating healthcare centres during the study period. All patients who fulfilled the study criteria were asked whether or not they wished to take part in the study. Information on the total number eligible patients who were asked to participate was not recorded. A total of 830 patients (389 from Flemingsberg and 441 from Jakobsberg) from 69 countries (including Sweden) agreed to participate and were included in the project. A total of 56.9% (n=469) were born outside Sweden and 42.1% (n=356) were born in Sweden. The biggest groups were from Turkey (n=86), Iran (n=52), Iraq (n=48), Finland (n=39), Bangladesh (n=24), Chile (n=23) and Syria (n=20). Five people in the 4D Diabetes Project were missing information about their place of birth. The final sample size in the current analysis was thus 825.

Data collection

The data were collected via face-to-face interviews at the academic primary healthcare centres in Flemingsberg and Jakobsberg. The interviews lasted approximately 15 to 20 min and were conducted in a separate room at the centres to maintain privacy for the participants. The questionnaire was available in Swedish, Arabic, Persian and Turkish.

Patient and public involvement

Neither patients nor the public were involved in the design of the study which was developed on the basis of previous research as well as the clinical experience of members of the 4D Diabetes Project group. Patients who met the inclusion criteria were provided with written and verbal information about the study and actively chose whether or not they wished to participate. The results of the study will be disseminated via seminars for healthcare professionals and patients at academic primary healthcare centres in the Stockholm region, including the participating centres. Centres across the Stockholm region will also receive short written summaries of the results.

Outcome variable

SRH was assessed using a single question, “How would you rate your general state of health?” There were five response alternatives as recommended by WHO, ranging from ‘very good’ to ‘very bad’.5 Responses were dichotomised in the analysis. Those who chose the alternatives ‘very good’ or ‘good’ were considered to have optimal SRH, whereas those who chose ‘fair’, ‘bad’ or ‘very bad’ were considered to have suboptimal SRH. This classification has been used in previous studies.19 33

Independent variable

Place of birth was the independent variable in this study. Place of birth was divided into two categories, ‘born outside Sweden’ and ‘born in Sweden’. The first category included immigrants from all countries and the second category included those born in Sweden whose parents were born in Sweden.

Explanatory variables

The explanatory variables included in the study were sex (male or female), age, SES and lifestyle factors. Patients were divided into two age groups, 18–44 and 45–74 years. We included three measures of SES. The first, educational level, was divided into 9 or fewer years, 10–12 years and more than 12 years. The second, employment status, was divided into employed, unemployed (including homemakers) or other (retired/student). The third, marital status, was divided into married/cohabiting or single (divorced or separated/widowed). Two lifestyle factors were included in the study. The first, physical activity, was based on self-reported time spent in any kind of physical activity per week. It included five response alternatives divided into four categories in our study: <30 min, 30–60 min, 60–90 min and >90 min. The second lifestyle factor was smoking. In the current study, patients were divided into non-smokers, former smokers and current smokers.

Statistical analysis

Descriptive statistics were used to assess the study population’s characteristics by place of birth. For the categorical variables, that is, sex, educational level, employment status, marital status, physical activity and smoking, the differences in percentages were compared using the χ2 test. A p value of <0.05 was considered statistically significant. Logistic regression was performed to determine the association between low SRH in those born outside Sweden and the explanatory variables. Those born in Sweden were used as the reference group. The results are presented as ORs with 95% CIs. The statistical analyses were performed with STATA V.14 (StataCorp).

Ethical considerations

The cross-sectional nature of this study limited the potential ethical problems, as there were no interventions that could have resulted in psychological or physical harm to the participants. Potential participants were provided with written and verbal information about the study and informed that participation was voluntary and that they had the right to withdraw from the study whenever they wished without facing any consequences. Since many potential participants were born abroad and had limited or no knowledge of Swedish, the written and verbal study information and questionnaires were available in Swedish, Arabic, Persian and Turkish. All participants provided written informed consent before inclusion in the study. Individual identities were replaced with code numbers prior to analysing the data. The link between the code number and identities is stored in a locked vault in keeping with the ethical rules of Karolinska Institutet.

The results of this study can help form the basis of future studies to deepen our understanding of SRH and could thus benefit patients in the future, including but not limited to our study participants.

Results

Demographic characteristics

Demographic characteristics of the study population by place of birth are shown in table 1. A total of 825 patients participated; 56.8% were born outside Sweden and 43.2% were born in Sweden. No significant differences were found between the two groups with regard to sex or marital status. Those born outside Sweden were significantly younger, had a higher educational level and were less physically active. On the other hand, a significantly higher percentage of those born outside Sweden than those born in Sweden were current smokers. The mean age of those born outside Sweden was 46 and those born in Sweden was 52. More than half the participants had more than 12 years of education (54.0%). More than two-thirds of the participants were married/cohabiting (70.0%). The majority of the participants reported being physically active (63.3%), and more than half were employed (53.5%). The percentage of people born outside Sweden with optimal SRH (51.7%) was lower than in those born in Sweden (57.6%).

Table 1

Characteristics of the study population† in total and by place of birth

Variables	Category	Total	Born outside Sweden	Born in Sweden	P value
Sample size, n (%)		825 (100)	469 (56.8)	356 (43.2)
Age, mean (SD)		49 (0.5)	46 (12.8)	52 (16.4)	<0.0001‡§
	18–44 years	323 (39.0)	209 (44.6)	111 (31.3)	<0.0001§
	45–74 years	506 (61.0)	260 (55.4)	244 (68.7)
Sex, n (%)	Female	475 (57.2)	269 (57.4)	203 (57.0)	0.924‡
	Male	355 (42.8)	200 (42.6)	153 (43.0)
SRH*, n (%)	Optimal	430 (52.3)	204 (51.7)	204 (57.6)	0.008‡
	Suboptimal	392 (47.7)	226 (48.3)	150 (42.4)
SES, n (%)	Educational level**				<0.0001‡
	≤9 years	148 (18.0)	97 (20.8)	51 (14.3)
	10–12 years	230 (28.0)	98 (21.0)	132 (37.1)
	>12 years	445 (54.0)	272 (58.2)	173 (48.6)
	Employment status***				<0.0001‡
	Employed	433 (53.5)	260 (56.5)	173 (49.6)
	Unemployed	102 (12.6)	73 (15.9)	29 (8.3)
	Other	274 (33.9)	127 (27.6)	147 (42.1)
	Marital status				0.066‡
	Single	248 (30.1)	129 (27.5)	119 (33.4)
	Married/cohabiting	577 (69.9)	340 (72.5)	237 (66.6)
Lifestyle factors, n (%)	Physical activity****				<0.0001‡
	<30 min	103 (12.5)	80 (17.1)	23 (6.5)
	30–60 min	105 (12.7)	76 (16.2)	29 (8.1)
	60–90 min	95 (11.5)	58 (12.4)	37 (10.4)
	>90 min	521 (63.3)	254 (54.3)	267 (75.0)
	Smoking*****
	Non-smoker	412 (50.1)	245 (52.4)	167 (47.0)
	Former smoker	248 (30.1)	115 (24.6)	133 (37.5)
	Current smoker	163 (19.8)	108 (23.0)	55 (15.5)

*3 missing, **2 missing, ***16 missing, ****1 missing, *****2 missing.

†The study population consisted of the 4D Diabetes Project population minus five individuals missing data on place of birth.

‡Bold numbers are significant.

§Χ2 test.

SES, socioeconomic status; SRH, self-rated health.

Table 2 presents the distribution of SRH by place of birth and sex. There were statistically significant differences in SRH between women and men in the study population as a whole and in people born outside Sweden. No significant sex differences were observed in people born in Sweden.

Table 2

Distribution of optimal and suboptimal self-rated health (SRH) in the study population* by place if birth and sex (n=822†)

Characteristic	Optimal SRH, n=392		Suboptimal SRH, n=430		P value
	Female	Male	Women	Men
In total, n (%)	232 (53.3)	198 (46.1)	239 (61.0)	153 (39.0)	0.042
Born outside Sweden	155 (64.1)	87 (35.9)	113 (50.0)	113 (50.0)	0.002
Born in Sweden	84 (56.0)	66 (44.0)	119 (58.3)	85 (41.7)	0.661

Bold numbers are significant.

*The study population consisted of the 4D Diabetes Project population minus five individuals missing data on place of birth.

†Data on sex were missing for three individuals.

Logistic regression

Table 3 presents the ORs of suboptimal SRH in patients born outside Sweden and patients born in Sweden in three models. Model 1 was adjusted for age and sex; model 2, for age, sex and SES (employment status, educational level and marital status); and model 3, for age, sex, SES and lifestyle factors (physical activity and smoking). The crude OR of reporting suboptimal health was 1.46 (95% CI 1.10 to 1.92) for those born outside Sweden (reference group: those born in Sweden). This relationship increased to 1.50 after controlling for educational level, employment and marital status (95% CI 1.11 to 2.02) and decreased to 1.36 but remained significant after adjustment for physical activity and smoking (95% CI, 1.00 to 1.85).

Table 3

Prevalence ratios (ORs) and their 95% CIs of suboptimal self-rated health for those born outside Sweden*

Variable	Crude OR	Model 1: +age, sex	Model 2: +educ. level, employment, marital status	Model 3: +physical activity, smoking
Born in Sweden	Reference	Reference	Reference	Reference
Born outside Sweden	1.46 1.10–1.92	1.48 1.12–1.97	1.50 1.11–2.02	1.36 1.00–1.85
Age
18–44 years		Reference	Reference	Reference
45–74 years		0.940.70–1.24	1.000.74–1.37	0.940.68–1.29
Sex
Female		Reference	Reference	Reference
Male		1.34 1.02–1.78	1.160.87–1.55	1.180.88–1.60
Educational level
≤9 years			Reference	Reference
10–12 years			1.400.90–2.17	1.300.83–2.02
>12 years			1.57 1.05–2.33	1.460.97–2.20
Employment status
Employed			Reference	Reference
Unemployed			0.470.29-0.74	0.470.29-0.75
Other			0.760.55–1.05	0.750.54–1.04
Marital status
Married/cohabiting			Reference	Reference
Single			0.61 0.45–0.83	0.62 0.45–0.86
Physical activity
<30 min				Reference
30–60 min				1.96 1.09–3.52
60–90 min				2.10 1.13–3.79
>90 min				2.80 1.74–4.50
Smoking
Non-smoker				Reference
Former smoker				0.990.70–1.41
Current smoker				1.010.67–1.50

Bold numbers are significant.

*The study population consisted of the 4D Diabetes Project population.

Discussion

In this study, we aimed to assess and compare the association between SRH and place of birth in primary healthcare patients born outside Sweden and in Sweden. Those born outside Sweden had significantly lower SRH than their Swedish-born counterparts. This result agrees with the results of previous studies that investigated SRH in patients born outside Sweden, especially those born in the Middle East.22 26

Chronic diseases are more prevalent in non-European immigrants to Europe than in native Europeans.34 35 The presence of chronic diseases can alter quality of life which could potentially negatively impact SRH. However, we were unable to explore this issue because of limited data. Additionally, psychological status can affect SRH.36 37 Psychological distress can be triggered by various situations. Migration per se is a stressful process.38 Immigrants can also be affected by past traumatic experiences; ongoing stress and worry about the situation in their home country, including family left behind; and discrimination in the host country.37 They can also face obstacles such as the need to learn a new language and to translate educational and professional competencies to those of the host country.39

The finding that SRH differs significantly between immigrants and their Swedish-born counterparts agrees with results of several previous Swedish studies.19 26 39 40 For example, our analyses confirm the results of a study of nearly 9000 older adult immigrants (age 55–74 years) in Sweden from a variety of ethnic backgrounds22 but extend the findings to a larger age range. Similarly, the current study extends the results of a study of ethnicity and low SRH in Assyrians living in the town of Södertälje, which focused solely on people with diabetes, did not include lifestyle factors as a possible confounder and collected data in a different way.26 41

Several studies carried out in Sweden have had similar results despite differences in data collection methods. The current study, carried out with face-to-face interviews and questionnaires in several languages; the study of Assyrians in Södertälje, carried out using an interpreter26 and a study of immigrants in Malmö, carried out with mailed questionnaires that were not translated into immigrants’ native languages24 all found that immigrants reported poorer health than native Swedes.

In a large cross-sectional study of immigrants from Turkey, Iran and Poland, differences between the immigrants and native Swedes in reporting significantly poorer health disappeared after accounting for SES, knowledge of the Swedish language and discrimination.39 In our study, as in the study of older immigrants in Malmö,18 the differences in SRH between groups remained statistically significant after adjustment for SES.

In contrast to our study, a study of 93 age-matched and sex-matched pairs of native born Swedes and people from Somalia and Iraq who had recent permanent residence permits found that the immigrants did not have worse subjective physical or psychological health than the Swedish controls.35 The authors of the case-control study suggested that a number of factors may have contributed to this finding, including a healthy refugee effect, the environment of the study location and the relatively high levels of education in the men from Iraq. However, more than half of those born abroad in our study (58.2%) had more than 12 years of education, and we still found a significant difference in SRH between those born abroad and native-born Swedes.

As in our study of people in two largely immigrant neighbourhoods in the Stockholm area, a study of European and non-European immigrants in a single neighbourhood in Denmark found that immigrants had lower SRH than the native-born people in the study population.42 Unlike in our study, however, the authors of the Danish study found that immigrants had fewer life resources and that this explained the results. Life resources in that study comprised a multiple factors, including ‘living with others’, ‘has children’, ‘has studied beyond primary school’, ‘has occupational income’ and more. We found fewer but similar SES factors in our study, including education, employment and marital status. They had a significant impact on the results but did not fully explain the differences between immigrants and native Swedes.

A Swedish study of patients with type 2 diabetes that had a study population similar to ours with regard to countries of birth, found a similar pattern of SRH.26 That is, the patients in the previous study who were born in the Middle East had lower SRH than the patients born in Sweden. As in our study, in the previous study, statistically significant difference in self-rated general health were not explained by socioeconomic or lifestyle factors. Future studies could focus on lifestyle, mental health, stress and comorbidity as potential explanations for SRH in people born in Middle East and Africa with a high risk of diabetes.

SRH is an easy item to collect since it is assessed through a single question. Despite that, it has an impact on a major heath outcome (mortality). Efforts to understand SRH and the factors related to can benefit the healthcare system in many ways. The current study adds support to previous research that found lower SRH in immigrant than in native populations. In our study, as in others, socioeconomic and lifestyle factors influenced SRH. It could therefore be useful for clinicians to consider these factors when providing care for patients born outside Sweden and resettled in areas with large numbers of immigrants. Future studies should further investigate what lies behind the discrepancies in SRH in immigrant and native-born primary healthcare patients.

The current study has several strengths and limitations. One strength of the study was the way data were collected. The questionnaire was available in different languages spoken by many people living in the study areas: Arabic, Persian, Turkish and Swedish. Another strength of the study was that the catchment areas of the healthcare centres in the current study included a large number of immigrants. Sampling was non-random, which limited our ability to draw conclusions. Another limitation of the study was grouping together those born outside Sweden regardless of the place of birth which might have masked potential inequalities. Furthermore, even people from the same country can come from different ethnic and religious groups, different cultures and different socioeconomic and political situations. Therefore, the final limitation of this study is the lack of information on cultural perceptions about health and lifestyle in the study population which might partially explain the low SRH in patients born outside Sweden.