Article Text

Download PDFPDF

Examining cultural, psychosocial, community and behavioural factors in relationship to socioeconomic inequalities in limiting longstanding illness among the Arab minority in Israel
  1. N Daoud1,2,
  2. V Soskolne3,
  3. O Manor1
  1. 1
    The Braun School of Public Health, Hadassah and the Hebrew University, Jerusalem, Israel
  2. 2
    Department of Epidemiology and Health Systems Evaluation, Ben-Gurion University of the Negev, Beer Sheva, Israel
  3. 3
    School of Social Work, Bar-Ilan University, Ramat-Gan, Israel
  1. Dr N Daoud, Braun School of Public Health and Community Medicine, Hebrew University-Hadassah, P.O.B. 12272, Jerusalem 91120, Israel; nihaya{at}hadassah.org.il

Abstract

Background: Few studies have examined the explanatory pathways to social inequalities in health within ethnic minorities. The current study examined the relative contributions of specific pathways explaining the associations between socioeconomic status (SES) and limiting longstanding illness (LLI) among the Arab minority in Israel.

Methods: A cross-sectional study of a random sample of 902 individuals aged 30–70 selected in a multistage sampling procedure. SES was measured by education, land ownership and relative family income. Five-stage logistic regressions assessed the attenuations in the odds of LLI among those with lower SES compared to higher SES after including relevant groups of explanatory factors: psychosocial, behavioural and community, and their integration.

Results: Rates of LLI were significantly higher in participants with lower SES. Inclusion of groups of explanatory variables attenuated all SES–LLI associations in a similar pattern: psychosocial factors played a main explanatory role, yielding 15–40% attenuation in odds ratios (OR). The contribution of community indicators was modest (10–21%); health behaviours had a marginal contribution (6–7%). Cultural factors were not associated with SES or LLI. The integrative model contributed up to 49% reduction in the OR.

Conclusions: The significant associations between SES and LLI suggest that formative policy to reduce SES–LLI disparities should emphasise creating opportunities for economic development to improve SES, which was the main predictor of inequalities. Combining strategies of community capacity building and reinforcement of individual inner resources might be complementary. Such conclusions might apply to other minorities in a similar context, for which future studies are required.

View Full Text

Statistics from Altmetric.com

Numerous studies have demonstrated the association between lower socioeconomic status (SES) and a greater prevalence of limiting longstanding illness (LLI), which refers to an illness that limits daily activities.13 Increasing cross-sectional and longitudinal data that have focused on the pathways that explain this social disparity have shown the contribution of psychosocial, community and behavioural factors in explaining this association.47 While most research has focused on general populations, few studies have examined social inequalities in health within ethnic minorities, and even fewer have examined the explanatory pathways for these disparities.810 Given the emergence of chronic diseases as a major public health problem among ethnic minorities in recent decades,11 explaining how SES is linked with LLI is of particular importance for informing equitable public health policies.

Previous evidence indicates that psychosocial, cultural, community and behavioural factors might contribute to explaining the association between SES and LLI among minorities. For example, exposure to high levels of stress and smaller social networks,1214 cultural attributes that are linked to differences in management of LLI,15 low levels of acculturation9 or community indicators (eg social participation),16 and unhealthy behaviours.17 Community variables, such as poor living conditions in disadvantaged neighbourhoods,18 or inaccessibility to health and social services19 are additional factors.

Arab citizens in Israel comprise approximately one-fifth of the population. Their current low SES20 reflects historical and political events since 1948. Many (56.6%) families live in poverty21 with indications of growing disparities.22 The steep increase in the incidence and prevalence of LLI23 may be partially explained by imposed rapid social and economic transitions,24 modifications of class hierarchy in the wake of confiscations of lands – traditionally the main source of income,25 and shifts from an agricultural to semi-industrial society.26 These have contributed to sedentary lifestyles,26 and were accompanied by changes in the family structure – from extended to nuclear – affecting the social ties and collective mixture of the society.24 Despite enactment of Israel’s National Health Insurance Law, this population is under-served with respect to health services.27 To date, very little research has explored health disparities among Arabs in Israel and their potential explanations.

This study examines the associations between SES and LLI, and assesses the relative contribution of individual and community factors in explaining the social disparities in LLI among Arabs in Israel. The adapted study framework28 (see figure 1) draws upon integrative approaches to explaining health inequalities,29 30 and postulates that SES may be associated with LLI via: (1) an individual pathway that includes psychosocial factors (exposure to chronic stress and stressful life events, lower levels of psychosocial resources of social support, social networks, coping efficacy and mastery) and cultural factors (adherence to cultural norms, consanguineous marriages); (2) a community pathway that includes social factors (lower levels of social capital, social involvement and civic participation) and structural constructs (neighbourhood problems and lower access to primary and specialist health services; and (3) a health behaviours pathway. In addition, it was assumed that the combination of all factors will have a greater contribution to explanation of the SES–LLI associations.

Figure 1

Conceptual model of individual and community pathways for explaining socio-economic inequalities in limiting longstanding illness (Adapted from Manor and Soskolne, 2004).

METHODS

Sample and procedure

A national random sample of 902 individuals aged 30–70 was selected from the Arab population in Israel in a multi-stage sampling procedure. First, 35 villages and towns were randomly selected from a total list of 155 localities; from them, statistical geographic areas were selected according to stratum based on population size and area-level socioeconomic index. Within each geographic area, a field sampling procedure was conducted to overcome difficulties in locating addresses of single households. Based on maps, the area boundaries and starting point were defined. The interviewer progressed along the right side of the streets selecting every fifth household, until 25–30 residential households were contacted. A total of 1158 households with residents eligible for inclusion were approached; 213 refused to be interviewed, and in 43 households no-one was present after three visits. The final sample consisted of 902 (78%) respondents. After signing an informed consent form, participants were interviewed at home in Arabic. The study was approved by the Institutional Ethics Committee at Hadassah Hebrew University Medical Center.

In the final sample, more men (57%) than women were interviewed; mean age was 45 years (SD = 11.3). The distribution of selected demographic variables (eg age, education and religion) in the sample were representative of the total Arab population in Israel (20); for example, 22.5% and 23.5%, respectively, were 30–34 years old; 39.6% and 40%, respectively, had 0–8 years of schooling.

Measures

Most of the measures had been previously used and validated or specifically developed for the Arab population – all at the individual level. Control demographic variables were age, gender and religious affiliation (Muslim, Christian and Druze).

The outcome variable, LLI, was measured by a yes/no question asking if the participant suffers from an illness that limits his/her daily functioning compared to others in the same age group.1

The independent variables were selected to include measures that represent different dimensions of SES, objective and subjective, that may not have the same relationships to health:31 (a) Education – number of years of formal schooling (0–8, 9–11, 12 and 13+ years; (b) Land ownership – a unique, important measure in this population that indicates inherited social prestige and current material gains by the participant or his/her family32 yes/no response categories); and (c) Relative family income – a subjective measure of SES: is average family income lower than, similar to or higher than other Arab families?

Explanatory variables

  1. Cultural attributes included factors that reflect cultural values and practices that are potentially associated with health in this population: 1. Consanguineous marriage − 40% in this population and might increase the probability of genetic-related diseases.33 Measured by the question: Is the spouse a relative (cousin, extended family)? 2. Adherence to cultural traditions – a five-statement scale, specifically developed for measuring attitudes towards traditions (eg feel obligated to help other extended family members when needed, believe that parents have the right to intervene with their children’s decisions on marriage).34 Higher scores represent greater adherence (Cronbach’s Alpha = 0.66).

  2. Psychosocial factors: 1. Exposure to stressors35 included chronic stress: ten yes/no questions representing exposure to stressful situations (eg financial, social, family, or work problems), and stressful life events: a list that measures the exposure of the respondent or any of his close family or friends to nine events during the last year. The final score for each variable was calculated as the sum of positive responses. 2. Psychosocial resources included: Social support – a six-item scale adapted from previous research.36 Higher scores represent greater support (Cronbach’s Alpha = 0.87). Social networks37 – one question regarding the number of close friends and relatives that the participant feels comfortable confide in. Mastery – a well-established seven-item scale to determine personal ability to control life situations.38 Higher scores represent higher mastery (Cronbach’s Alpha = 0.81). 3. Coping efficacy – the sum of two questions that reflect the participant’s ability to cope with most recent stressful events, ranging from 1 (not at all) to 5 (very much).39

  3. Community characteristics: Social variables: 1. Social participation – a nine-item scale regarding membership in formal or voluntary organisations, participating in religious activity, etc,40 ranging from 1 (always), to 3 (never) (Cronbach’s Alpha = 0.72). 2. Civic engagement – two yes/no questions about voting in national elections.41 3. Social capital – three yes/no questions about feelings of exploitation, mutual help and trust among community members.42 Structural variables: 1. Access to healthcare services – time and mode to reach a specialist clinic (by foot, bus, taxi or private car). 2. Neighbourhood problems – a 14-item scale,43 expanded and adapted to the Arab neighbourhoods, with higher scores representing more problems (Cronbach’s Alpha = 0.84).

  4. Health behaviours: A combined44 index of four health behaviours.45 The response categories for each behaviour were dichotomised as follows: 1. Diet; healthy ‘balanced diet’ and ‘unbalanced diet’. 2. Smoking; ‘never smoked or past smoker’ and ‘current smoker’. 3. Use of sun protection; ‘always or sometimes’ and ‘seldom and never’. 4. Physical activity; ‘every day/1–2 times/week’ and ‘seldom or never’.

Statistical analysis

The associations between each SES measure and LLI were examined, followed by examining the associations between sociodemographic control variables and SES. No consistent associations were found between SES and religious affiliation, and no significant interactions were detected for either age or gender in their effect on SES–LLI associations. Therefore, all models were adjusted for age and gender. In order to determine potential explanatory variables, bivariate associations between each explanatory variable and each SES measure were examined, and the variables that were significantly associated with at least one SES measures and with LLI were included in the multivariate analysis. Logistic regressions were used for the dichotomous variables and linear regressions for the continuous variables, while adjusting for age and gender. Finally, five successive logistic regression models4 were fitted. Model 1 included the main association between each of the SES measures and LLI. Groups of variables were added to model 1 as follows: in model 2 – psychosocial variables; in model 3 – community variables; in model 4 – health behaviour; and in model 5 – all groups of variables. Correlations within each group of explanatory variables were examined to avoid multicolinearity, but no variable had to be deleted. SES disparities were evaluated by the odds of LLI in each SES category. The OR between the extreme socioeconomic categories was used to assess the relative contribution of each group of variables to changes in the SES–LLI associations. The per cent of decline in the OR was calculated in models 2 through 5 as compared to model 1.

Due to missing values for some variables of interest, the sample sizes for the logistic regression models are different across measures of SES; however, the sample size across models 1–5 within each SES measure remains constant.

RESULTS

Distribution of the study variables is presented in table 1. LLI was prevalent among one-quarter of the sample. Generally, the study population was poor: 40% had 0–8 years of schooling, close to 70% reported not owning land, and about 85% reported family income lower than or similar to other Arab families.

Table 1 Distribution of study variables

A significant and consistent social disparity in the prevalence of LLI was observed by each SES measure, although the strength of the associations varied depending on the SES measure. Lower levels of SES were associated with higher prevalence of LLI (see figure 2).

Figure 2

Associations between SES measures and LLI (OR and confidence intervals).

In the bivariate analysis (table 2A and B), most variables were significantly associated with at least one of the three SES measures. These included adherence to cultural norms, all the psychosocial variables (chronic stress, stressful life events, social support, social networks, mastery and coping efficacy), two community social variables (social participation and feelings of exploitation), the community structural variables (neighbourhood problems, time and mode of reaching specialist clinic), and the health behaviours index. One cultural variable (spouse is relative) and three community social variables (trust among community members, mutual help and civic engagement) were not associated with any SES measures and therefore were not examined further for associations with LLI.

Table 2 Associations between SES measures and explanatory variables

Of the variables that were significantly associated with at least one SES measure, several were also significantly associated with LLI (table 3). These included several psychosocial variables (chronic stress, stressful life events, mastery and coping efficacy), a community social variable (social participation) and the health behaviours index.

Table 3 Odds rations and confidence intervals in the associations between explanatory variables and LLI*

In order to examine the contribution of the explanatory variables that were associated with both SES and LLI, multivariate analyses were conducted by including different groups of variables (table 4). Cultural variables were not included in the multivariate analyses as none met the prerequisite of being associated with both SES measures and LLI. Associations between education and LLI (table 4A) were attenuated by 26% after adding psychosocial variables (model 2), about 20% after adding community variables (model 3), and only about 7% after adding health behaviours (model 4). The integrative model (model 5) showed a reduction of about 35% in the main association. Similar trends of reduction in OR were observed in the associations with land ownership (table 4B). The ORs were attenuated by 15% in model 2 after adding the psychological variables, about 10% after the inclusion of community variables (model 3), and about 20% when integrating all variables (model 5). Health behaviour index was not included here as it was not associated with land ownership. Regarding the association with relative family income (table 4C), a larger reduction (40%) in OR was observed after adding psychosocial variables (model 2), about 21% reduction after adding community variables (model 3), and 6% reduction was seen when adding health behaviour (model 4). The integrative model (model 5) contributed to a reduction of 49% in the association.

Table 4 OR and CI of LLI by education, land ownership, and relative family income and reductions in OR after including explanatory variables*

Following the larger attenuation in OR (20–49%) in the integrative models (model 5, table 4A–C), all SES–LLI associations became non-significant.

DISCUSSION

What is already known on this subject

Whereas most research has focused on examining the pathways that explain health inequalities in the general populations, few have examined the specific pathways that explain these inequalities within ethnic minorities.

What this study adds

  • Among minorities, it is vital to explore pathways that contribute to explaining the social inequalities in LLI in order to develop more exclusive health policies that can help prevent the excess of LLI in their poorer sectors.

  • Within the Arab minority in Israel, social disparities in LLI were explained mainly by psychosocial factors. The contribution of community factors was modest and that of health behaviours was marginal, suggesting that formative policy to reduce SES–LLI disparities in this minority should incorporate economic development with strategies of community capacity building and reinforcement of psychosocial resources among disadvantaged sectors.

Little research has been conducted on the pathways that explain social inequalities in health within ethnic minorities, especially Arabic-speaking populations.16 Consequently, the possibility of comparing and discussing the present results relative to populations of similar cultural context is limited.

The findings of the present study, which estimate social disparities in LLI and examine the relative contribution of specific factors in explaining these inequalities within the context of Arabs in Israel, indicate that psychosocial factors are the main contributors compared to community characteristics or health behaviour. Furthermore, integrating psychosocial factors with community characteristics and health behaviours provide even more satisfactory explanation for the health disparity under study.

The social disparities detected for LLI, by which participants with lower education, those who did not own land or reported relatively lower family income, were more likely to have LLI, compared to their better-off counterparts, accord with substantial previous evidence that SES predicts the incidence of LLI in minorities.8 10 46 The present results point to the importance of all three SES measures in estimating social inequalities in health among Arabs in Israel. In contrast with a previous study,47 the present results indicate that education is associated with health in this population; those with 0–8 years of education showed a higher risk of LLI, compared to individuals with other higher education categories. The association found with the unique SES measure of land ownership – which is used here for the first time – most likely reflecting the owners’ social prestige and material resources,32 points to the potential of this SES measure in determining health among Arabs in Israel. The fact that the subjective SES measure of relative family income tends to have slightly higher OR in the association with LLI than the two objective SES measures, might reflect the breadth of subjective SES measures as compared to objective SES ones; one might also include, in addition to their objective SES, a reflection on the situation. However, as all data were self-reported the stronger association between these two subjective measures might have been influenced by a common reporting tendency.48 49

In comparing the contribution of the different explanatory groups of variables to the SES–LLI inequality, similar patterns were observed for all SES measures; psychosocial factors provided a stronger explanation than did community indicators, whereas health behaviour had a marginal explanatory effect. Although the composition of the different groups of explanatory variables varies in each SES–LLI association, the present findings correspond with previous studies4 7 50 indicating that psychosocial factors do contribute to explaining the SES disparities in morbidity and mortality. Specifically, it is shown that lower mastery and coping efficacy, along with greater exposure to stressful life events and chronic stress, had a major contribution in explaining the SES–LLI association. Notably, these findings support previous evidence that minority groups of lower SES are exposed to higher levels of stress, and have lower psychosocial resources.12 14

Of the five community social characteristics considered, only social participation was found to be an explanatory variable in the SES–LLI association. This result suggests that the community as a social group has a limited contribution in reducing social disparities in LLI and may reflect the social transitions this community has undergone in recent decades – from a collective society with entrenched solidarity, extended family structure, and strong social networks, into a more individualised one.24 Interestingly, the community social characteristics were associated mostly with SES measures, but not with LLI, proposing that lower SES groups were more influenced by the individualisation processes than the higher SES groups. Neither social support nor social networks were associated with LLI, and therefore were not considered as explanatory variables in the SES–LLI association. This is in contrast with previous evidence for an association of social networks with health among older Arabs in Israel.37 However, social networks might have different meanings and composition for different age groups. In addition, the weak contribution of the community factors may be related to the fairly small variability measured for social capital (mutual help and trust among community members). Notably, in contrast with other studies,42 here social capital variables mostly were not associated with either LLI or SES. This difference might stem from different perceptions of social capital as well as the way it was measured, namely individual versus community-level measurement.51

Community structural factors (neighbourhood problems and access to specialist care) as well as cultural variables (adherence to cultural norms and spouse is relative) were not found to explain the SES–LLI association. Both groups of variables were associated with some SES measures, but they were not related to LLI. Our results are in contrast with studies proposing that cultural factors impact health among Arabs in Israel.47 Notably, the cultural variables were measured here by a scale with a moderate internal consistency (α = 0.66) and further investigation of the contribution of these variables is needed.

The marginal contribution of health behaviours as an explanatory factor in SES–LLI associations supports previous findings,52 and suggests that promoting healthy behaviours in this population cannot be an exclusive solution for reducing social disparities in LLI, even though unhealthy behaviours were major determinants of poor health among Arabs in Israel.47 This may also be attributed to the scale used in the study, in which all health behaviours were given similar weights, although they may well differ in their explanatory contribution to the SES–LLI association.

The integrative model, which included psychosocial, community and health behaviour variables, proved to be a suitable model for explaining social inequalities in LLI in this population.

While assessing the inequalities by contrasting the extreme SES categories, the present model explained up to 49% of the inequalities and the associations became non-significant. However, to reach more reduction in OR, additional factors that were not included in the current study might play a role in explaining the disparities. Such factors might include health literacy, cultural sensitivity of health services, and the Israeli–Arab conflict, which might contribute to high psychosocial distress in this population.

Limitations

A number of limitations should be noted in the present analysis. The cross-sectional design limits the ability to assess causality. The associations proposed in the study framework may well be bidirectional: LLI may lead to reduced levels of several of the explanatory variables, for example mastery, social or civic participation, and social networks and support. In addition, the current results reflect the variables selected for each group of explanatory variables, the way they were measured, and their inclusion in the different models. Although the selection was based on theoretical and empirical considerations, these choices may exert some influence on the results. Using only the OR between the extreme socioeconomic categories to assess the relative contribution of each group of variables might have de-emphasised the impact of the different pathways on social disparities in relation to the middle social categories. It should be mentioned that the modest sample size influences the precision of the OR estimates and percentage changes, thus the latter values should be interpreted as indicative of the potential relative importance of the various factors.

Conclusion and recommendation

The present results demonstrate the importance of studying the relative contribution of different factors to explain the social disparities in LLI among the Arab minority in Israel, an approach which may also be relevant to other minorities that are confronting a substantial increase in chronic diseases and in their deprived sectors. In this study, social disparities in LLI were explained mostly by an integration of psychosocial factors, community characteristics and health behaviour. This suggests that formative policy to reduce SES–LLI disparities should emphasise creating opportunities of economic development to improve SES, which was the main predictor of inequalities. A combination of community capacity building and reinforcement of individual psychosocial resources might be complementary strategies.

Acknowledgments

We would like to thank Mrs. Bella Adler for consultation in data processing and give thanks to the participants. Many thanks to the editor and reviewers for comments on a previous version of this paper.

REFERENCES

View Abstract

Footnotes

  • Funding: This work was supported by The Israel National Institute for Health Policy and Health Services Research (NIHP) Grant Number a/95/2003.

  • Competing interests: None.

  • Ethics approval: The study was approved by the Institutional Ethics Committee at Hadassah Hebrew University Medical Center.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.