Introduction

This paper discusses the rationale for using the Andersen behavioural model of health service utilisation (Andersen, 1968, 1995; Andersen & Newman, 1973) in a study of informal support in later life.  The aim of the study is to ascertain the extent to which ethnic group is a determining factor for informal support transfer in later life, as described in an earlier article (Willis, 2008b).  Informal support transfer indicates unpaid help either received by an individual or given by them to family members, friends or neighbours.  In this study, it includes instrumental tasks such as personal care, cooking and transport.  Social inequalities in ill-health and socio-economic status which influence the need for support and the ability to access support (Broese van Groenou, et al., 2006) are present among Britain’s minority ethnic groups (Modood, et al., 1997).  This study sought to investigate whether ethnicity influences informal support when controlling for socio-economic, demographic and health factors.  This paper is part of a study which uses the Andersen model as a framework for a secondary analysis of the Home Office Citizenship Survey (Home Office, 2006).  Results from this analysis have been reported elsewhere (Willis, 2008a, 2009). 

The Andersen model is useful because of its flexibility in allowing researchers to choose independent variables related to their specific hypotheses.  Further, the adaptations of the model to study different outcome variables in later life, spanning the utilisation of health care (Wolinsky & Johnson, 1991), formal social care (Bass & Noelker, 1987) and informal care (Gaugler & Kane, 2001), demonstrate its usefulness for gerontological research.  This model has become one of the most widely used frameworks to predict healthcare use since its inception more than 40 years ago (Phillips, et al., 1998).

Review of the Andersen behavioural model

Andersen's behavioural model (Andersen, 1968) was created to empirically test hypotheses about inequality of access to health services in the USA.  It addresses the concern that some sectors of society - in particular people from ethnic minority groups, people who live in inner cities and people who live in rural areas - receive less health care provision than the rest of the population (Andersen & Newman, 1973).  Andersen’s model views access to services as a result of decisions made by an individual, which are constrained by their position in society and the availability of health care services.  This model, therefore, allows us to explore hypotheses regarding social inequalities.

The model contains three sets of predictive factors: predisposing, enabling and need factors (see Figure 1).  It assumes that a sequence of factors determines the utilisation of health services: the predisposition to use services, the ability to use services and the need to use services.  Andersen’s first study focused on the family as the unit of analysis, and hence several family-level variables were used.  Later versions of the model focus on the individual as the unit of analysis (Andersen & Newman, 1973).

Figure 1 : Andersen's 1968 model

The predisposing factors are based on the argument that a family’s propensity to use health services can be predicted from a set of personal characteristics which predate the illness.  These characteristics can be divided into three sets: family composition, social structure and health beliefs (Andersen, 1968).  Specific variables include age, sex, family size, ethnicity and social class.  These indicate the position of the family in society which could influence their lifestyle and their physical and social environments.

The enabling factors are based on the argument that even if a family has a predisposition to use health services, certain characteristics must be in place to enable them to access services.  Such enabling factors include material resources, having health insurance and the availability of health services.  Without the ability to access services a predisposition will not necessarily translate into utilisation. 

Finally, in order for a health service to be used, there must first be a need to use that service.  Therefore, need factors are included in the model.  There are two types of need factor: illness variables and response variables (Andersen, 1968).  Not only must the family recognise that there is an illness, but they must also respond appropriately in order to access services.

According to Andersen, access to services is considered equitable if it can be predicted by immutable demographic characteristics such as age and sex (Andersen & Aday, 1978), or solely by the need factors, such as illness.  However, access is considered inequitable if it can be predicted wholly or partly by variables such as ethnicity or enabling factors (Andersen & Newman, 1973).  This provides the theoretical basis for using the Andersen model to study social inequalities in health service utilisation.

Andersen’s model does not specify which variables mustbe used to operationalise the need, predisposing and enabling factors (Andersen & Newman, 1973).  Instead, the decision of how to operationalise them should derive from the theoretical relationship between the independent and dependent variables.  As such, the choice of variables, within the framework of need, predisposing and enabling factors, is up to each researcher.

Studying ethnicity with the Andersen model

In the original behavioural model (Andersen, 1968) ethnicity was one of the predisposing variables.  The logic for this was that ethnicity was an indicator of a family’s position in society and could therefore predict their need for and use of health services.  However, this position did not allow the independent influence of ethnicity on the dependent variable to be examined; instead its influence was considered simultaneously with the other predisposing variables.  Other studies altered the position of ethnicity within the Andersen model in order to focus on its influence more directly (e.g. Bradley, et al., 2004; Davidson & Andersen, 1997; Doescher, et al., 2000; Mitchell, et al., 1997).  Some of these studies are now described.

The behavioural model has been used to predict dental health service utilisation with an emphasis on ethnicity and age (Andersen & Davidson, 1997).  Ethnic group and age group were extracted from their usual place as predisposing variables.  They became exogenous variables with hypothesised influences on each of the predisposing, enabling and need factors.  By adapting the behavioural model in this way, inequalities in dental health service utilisation by ethnicity and age cohort could be understood.

A second study explored the determinants of intended future use of informal care among two ethnic groups (Bradley, et al., 2004).  They ran a regression analysis with four models.  Model 1 contained ethnic group, Model 2 added the predisposing factors, Model 3 added the need factors and Model 4 added the enabling factors.  In this way the bivariate relationship between ethnic group and the outcome variable was measured in Model 1, and the subsequent models allowed the change in the significance of the ethnicity variable to be assessed with the addition of each set of variables.  This method of applying Andersen’s model by initially separating ethnicity as an independent variable was selected for the present study.

Conceptual framework

In the present study, the behavioural model of health service use is adapted to examine the relationship between individual-level social and demographic factors and the receipt or provision of informal support.  Ethnicity is argued to have influences at all levels of the need, predisposing and enabling factors.  Further, employing similar methods for analysis as Bradley et al (2004), ethnicity is the primary independent variable, with the need, predisposing and enabling factors added to the analysis in sequence.  Figure 2 shows the conceptual framework used in this study.

Figure 2 : Conceptual framework for the present study, derived from the Andersen behavioural model

Ethnicity is the primary variable of interest in this study of informal support in later life. There are structural inequalities by ethnic group in Britain in each of the other independent variables.  For instance, the highest rates of limiting long-term illness or disability reported in the 2001 census were among Pakistani and Bangladeshi groups (ONS, 2004), and the Fourth National Survey of Ethnic Minorities reported that Caribbean, Pakistani and Bangladeshi men were over-represented in manual types of employment (Modood, et al., 1997).  Therefore, it is argued that the need, predisposing and enabling factors will vary by ethnic group.

The need factor in this study is the presence of a chronic illness or disability, which is associated with an immediate need for informal support.  The reason for placing need second in the hierarchy is to assess the attenuating effect of illness and disability upon ethnic group independent of the other variables.  That is, health inequalities are expected to affect the relationship between ethnicity and informal support. 

The predisposing factors are next in the model; they are age, sex and education.  These factors are thought to predispose an individual to the transfer of informal support, and are also predisposing conditions for the enabling factors.

The enabling factors comprise the final set in the hierarchy.  They are income, socio-economic status, tenure, access to a vehicle, having at least one child and marital status.  The rationale behind the enabling factors is that one can have a need and a predisposition for support, but if one is not enabled to access that support it will not be achieved.  These factors can facilitate support transfer if they are present, but impede it if they are absent.

Entering the four sets of independent variables into a multivariate analysis in this order allows a test of the relationship between ethnicity and informal support when controlling for a hierarchy of theoretically important socio-demographic variables.

The outcome variable, informal support transfer, indicates support both received by the individual and given by them to others.  The rationale for this is that just as a particular factor, such as disability, signals a greater need for informal support, so it may have the opposite effect on the individual’s ability to provide informal support.

Conclusion
This article has described the Andersen behavioural model of health service use, and how it can be adapted for gerontological research.  In the present study informal support transfer is the outcome variable, rather than health service use, and ethnicity is the primary independent variable, rather than one of the predisposing variables.  The intention of this is to examine the influence of ethnicity on informal support transfer once the other factors have been taken into account.  This will show whether any previously observed ethnic group differences in informal support transfer are actually artefacts of socio-demographic inequalities.

Andersen’s model is very useful for gerontological research as it can be adapted depending on the research question.  The present study focused on ethnicity as its primary variable of interest.  Other authors have examined age group (Andersen & Davidson, 1997) and socio-economic status (Broese van Groenou, et al., 2006) in the same way.  The outcome variable can also be adapted.  Those studied in the past have included health service utilisation, formal service utilisation and informal care receipt.  In conclusion, this model is a powerful tool for the exploration of gerontological theories.

Funding

This work was supported by the King’s College London Graduate School, Funds for Women Graduates and the Gilchrist Educational Trust.

References

Andersen, R. (1968). Behavioral model of families' use of health services. Research Series No. 25. Chicago, IL: Center for Health Administration Studies, University of Chicago.

Andersen, R. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36(1), 1-10.

Andersen, R., & Aday, L. A. (1978). Access to medical care in the US: Realized and potential. Medical Care, 16(7), 533-546.

Andersen, R., & Davidson, P. L. (1997). Ethnicity, aging, and oral health outcomes: A conceptual framework. Advances in Dental Research, 11(2), 203-209.

Andersen, R., & Newman, J. F. (1973). Societal and individual determinants of medical care utilization in the United States. The Milbank Memorial Fund Quarterly. Health and Society, 51(1), 95-124.

Bass, D. M., & Noelker, L. S. (1987). The influence of family caregivers on elder's use of in-home services: An expanded conceptual framework. Journal of Health and Social Behavior, 28(2), 184-196.

Bradley, E. H., Curry, L. A., McGraw, S. A., Webster, T. R., Kasl, S. V., & Andersen, R. (2004). Intended use of informal long-term care: The role of race and ethnicity. Ethnicity & Health, 9(1), 37-54.

Broese van Groenou, M., Glaser, K., Tomassini, C., & Jacobs, T. (2006). Socio-economic status differences in older people's use of informal and formal help: A comparison of four European countries. Ageing and Society, 26(5), 745-766.

Davidson, P. L., & Andersen, R. (1997). Determinants of dental care utilization for diverse ethnic and age groups. Advances in Dental Research, 11(2), 254-262.

Doescher, M. P., Saver, B. G., Franks, P., & Fiscella, K. (2000). Racial and ethnic disparities in perceptions of physician style and trust. Archives of Family Medicine, 9(9), 1156-1163.

Gaugler, J. E., & Kane, R. A. (2001). Informal help in the assisted living setting: A 1-year analysis. Family Relations, 50(4), 335-347.

Home Office (2006). Communities Group and National Centre for Social Research, Home Office Citizenship Survey, 2005 [computer file]. Colchester, Essex: UK Data Archive [distributor], SN: 5367.

Mitchell, J., Mathews, H. F., & Griffin, L. W. (1997). Health and community-based service use: Differences between elderly African Americans and Whites. Research on Aging, 19(2), 199-222.

Modood, T., Berthoud, R., Lakey, J., Nazroo, J., Smith, P., Virdee, S., et al. (Eds.). (1997). Ethnic Minorities in Britain: Diversity and Disadvantage. London: Policy Studies Institute.

ONS (2004). Focus on Ethnicity and Identity Retrieved 27/7/07, from http://www.statistics.gov.uk/downloads/theme_compendia/foe2004/Ethnicity.pdf

Phillips, K. A., Morrison, K. R., Andersen, R., & Aday, L. A. (1998). Understanding the context of healthcare utilization: Assessing environmental and provider-related variables in the behavioral model of utilization. Health Services Research, 33(3), 571-596.

Willis, R. (2008a). Advantageous inequality or disadvantageous equality? Ethnicity and family support among older people in Britain. Ethnicity and Inequalities in Health and Social Care, 1(2), 18-23.

Willis, R. (2008b). Older people, ethnicity and social support. Generations Review, 18(4): http://www.britishgerontology.org/08newsletter4/research1.asp.

Willis, R. (2009). More similar than different: informal support and ethnic diversity. Paper presented at the British Society of Gerontology 38th Annual Conference, Bristol.

Wolinsky, F. D., & Johnson, R. J. (1991). The use of health services by older adults. Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 46(6), S345-S357.