Hostname: page-component-7b7567c8c-7pwfn Total loading time: 0 Render date: 2026-11-15T10:03:19.234Z Has data issue: false hasContentIssue false
  • English
  • Français

Family history of chronic disease and participation in healthy behaviours

Published online by Cambridge University Press:  24 August 2007

GODFREY A. GIBBISON and
C. DOUGLAS JOHNSON
GODFREY A. GIBBISON*
Affiliation:
School of Economic Development, Georgia Southern University, USA
C. DOUGLAS JOHNSON
Affiliation:
College of Business, Georgia Gwinnett College, USA
*
*Correspondence to: Godfrey A. Gibbison. School of Economic Development, Georgia Southern University, P.O. Box 8152, Statesboro, GA 30460, USA. Email: ggibbiso@georgiasouthern.edu
Rights & Permissions [Opens in a new window]

Abstract:

This article presents an investigation into whether individuals incorporate information on their family genetic (health) endowment into decision making regarding participation in behaviours related to the production of health capital. Competing theoretical perspectives exist as to how genetic endowment should affect participation in health-related behaviours. There is also a growing consensus that the availability of genetic information holds important consequences for the individual in terms of investment in human capital, the pricing and availability of health insurance, and labour market opportunity. Therefore, the question of how individuals use the greater availability of genetic data is important. Our empirical results indicate that poor genetic endowment tends to be associated with a lower probability of participation in cigarette smoking, but also a lower probability of participation in regular exercise.

Information

Type
Articles
Information
Health Economics, Policy and Law , Volume 2 , Issue 3 , July 2007, pp. 297 - 315
Copyright
Cambridge University Press

Introduction

Genetic endowment sometimes plays a significant role in the production of health. The link between genes and susceptibility to a number of diseases (including breast and ovarian cancer, early onset Alzheimer’s diseaseFootnote 1) is well established (Osann, Reference Osann1991; Schwartz et al., 1995; Tabarrok, Reference Tabarrok1994; Macdonald, Reference Macdonald2003). Therefore, an individual’s ability to produce a certain stock of health capital (or a particular health outcome) can be significantly influenced by the genes she inherited. Several dozen studies have documented the immediate and lifetime costs associated with morbidity due to these common diseases (e.g., Manning et al., Reference Manning, Keller, Newhouse, Sloss and Wasserman1991; Chodick et al., Reference Chodick, Heymann, Wood and Kokia2005; Payne et al., Reference Payne, Kuybrechts, Craig Green and Klittich2002, Ory et al., Reference Ory, Vanderplas, Dezii and Chang2005; Lee et al., 2004).

However, as Macdonald (Reference Macdonald2003) points out, at the current state of knowledge there are relatively few ‘simple’ genetic disorders, by which he means disorders that can be traced to a single gene or to several genes, each of which can still cause the disorder. To date, no single gene or set of genes has been assigned a causal role in the majority of cases of chronic diseases, such as heart disease and stroke. However, it is generally accepted that these tend to aggregate within families, suggesting possible genetic links (Macdonald et al., Reference Macdonald, Waters and Wekwete2005). It is far more likely that the connection between genes and health outcome occurs in a complicated manner, with many genes contributing to the same disease. Also important is the interaction between genes and behavioural and environmental factors (Tabarrok, Reference Tabarrok1994; Macdonald, Reference Macdonald2003). According to Macdonald et al. (Reference Macdonald, Waters and Wekwete2005) various gene locations have been linked to increased risk of developing hypertension, hypercholesterolemia, and diabetes, all of which are associated risk factors for heart disease and stroke.

The risk of developing many common diseases, including heart disease, diabetes, and stroke, can be modified by participation in certain healthy behaviours (for example, maintaining a certain diet and regular physical activity) or by avoiding certain unhealthy behaviours (for example, smoking), even if these diseases exhibit a tendency to occur within a family (Macdonald et al., Reference Macdonald, Waters and Wekwete2005). A comprehensive study by Nicholl et al. (Reference Nicholl, Coleman and Brazier1994) shows a significant reduction in the risk of developing coronary heart disease by physically active men to the order of 40%, reduction in the risk of stroke by up to 41%, and lower risk of diabetes by up to 29%.

The marketing (e.g., television commercials) of healthy behaviours (avoidance of unhealthy behaviours) has already begun to emphasize the importance of genetic aspects of some diseases. While some health economists advocate the use of genetic data to evaluate a candidate for health insurance based on the higher insurance risk posed by individuals with certain genetic mutations (Lemaire et al., Reference Lemaire, Subramanian, Armstrong and Asch2000), considerable uncertainty exists regarding the reliability of premium rates based on genetic studies (Lu et al., Reference Lu, Macdonald and Wekwete2006; Macdonald, Reference Macdonald2003; Hedgecoe, Reference Hedgecoe1996). Additionally, there is the possibility of discrimination in access to health insurance (Armstrong et al., Reference Armstrong, Weber, FitzGerald, Hershey, Pauly, Lemaire, Subramaian and Asch2003).

As such, this paper evaluates the relationship between family genetic endowment and health-related behaviours and is organized as follows. In the next section we discuss some of the possible interactions between family health history and health behaviours, calling on both literatures in economics and health psychology. This is followed by a theoretical framework describing the relationship of health production to health behaviours and family health endowment, and includes the specific aims and research questions of interest. The data and methods used in this study are then described, followed by the results of the study. We end with a discussion of the results and concluding remarks, including policy implications.

Family health history/health behaviour interaction

It is not clear how individuals use the information on their family health history. On the one hand, poor genetic endowment might lead to behaviours that reduce the risk of developing a chronic disease. That is, the individual might substitute healthy behaviours for inherited health risks (see, for example Dickie and Gerking, Reference Dickie and Gerking1997). On the other hand, some individuals may choose unhealthy behaviours because, given their endowment, the number of (and amount of each) healthy behaviours required to modify health risks is more than the individual wishes to implement (more than the optimal amount for the individual) or because the individual finds healthy behaviours distasteful. Alternatively, some individuals prefer to substitute medication for healthy behaviours in order to achieve or maintain a certain stock of health capital (Chen et al., Reference Chen, Shogren, Orazem and Crocker2002) despite evidence suggesting the marginal impact of health care on health is low, at least in developed countries (Birch and Stoddart, Reference Birch, Stoddart and Lopez-Casasnovas1991).

Not only is it important to understand the relationship between family history of disease and health-related behaviours from a micro (or individual) perspective, it is of importance at the macro (organizational) level as well. In the past few decades, employers have experienced a dramatic increase in health care expenditures and have begun to develop and implement strategies to ensure cost containment occurs (e.g., Chiappetta, Reference Chiappetta2005; Coulter, Reference Coulter2006; Thorpe, Reference Thorpe2005). Further, some employers are beginning to assess family endowment in an effort to determine appropriate deductibles or coverage extended to employees, or potential workplace problems (e.g., Lemaire et al., Reference Lemaire, Subramanian, Armstrong and Asch2000; Brown, Reference Brown, Pagan and Bastida2005). Organizations may be reluctant to employ individuals with family endowments which could adversely affect the individual, and ultimately the organizational, performance due to increased absences/use of sick days, potential accidents leading to worker’s compensation claims, reduced morale, etc. Organizations are also initiating wellness programs in order to encourage employees and their dependents to engage in healthy behaviours. As such, the present study has significant workplace implications and should prove insightful from policy formulation and health care cost containment perspectives.

The early economic theory of health production was developed by Becker (Reference Becker1965) and Grossman (Reference Grossman1972). In this tradition individuals maximize well-being subject to constraints such as income and prices. Well-being is derived from the consumption of market goods and goods (such as health) produced by the individual. The individual solves this constrained optimization problem to determine the well-being maximizing stock of health capital, which in turn is produced by combining market inputs (such as medication), behaviours (such as regular physical activity) with the individual’s health endowment, which is predetermined by genetic inheritance. Production of health depends upon the health production function, which determines the efficiency with which inputs are converted to health and the rate of depreciation of the stock of health capital produced. An important implication of the Becker–Grossman theoretical perspective is that individuals will compensate for poor genetic endowment by selecting healthy behaviours; that is, in producing health, behaviours and endowment are substitutes.

More recent theoretical work has suggested a more complex relationship between endowment and healthy behaviours. Birch and Stoddart (Reference Birch, Stoddart and Lopez-Casasnovas1991) showed that given the resources available to the individual (for example genetic endowment), unhealthy behaviours might be an optimal choice. For example, the individual’s endowment might be such that there is little reason to believe that future benefits of healthy behaviours will be realized. In their words ‘[M]arginal benefits of health-related behaviour would also be affected if the future value of the returns to the behaviour differed across individuals. In other words, “incentives to adopt healthy lifestyles may differ”’ (pp. 182).

To give a concrete example, for an individual with a family history of obesity the marginal cost of regular physical activity might exceed the present marginal benefit or any expected future benefit. Furthermore, if an individual faces multiple or severe health risks due to poor endowment, the consequences of unhealthy behaviours may seem relatively minor; that is, the incentive to invest in health production depends on the individual’s perception of the chance of survival (Dow et al., Reference Dow, Holmes, Philipson and Sala-i-Martin1999).Footnote 2. This line of reasoning suggests some individuals complement poor genetic endowment with unhealthy behaviours, such as smoking and excessive alcohol consumption, or fail to choose healthy behaviours, such as regular exercise.

There is also substantial evidence from the health psychology literature that in various cultural settings individuals sometimes take a fatalistic view to health production and health outcome (Liang et al., Reference Liang, Yuan, Mandelblatt and Pasick2004; Sharf et al., Reference Sharf, Stelljes and Gordon2005; Gorin, Reference Gorin2005). Fatalism is the sense of having no control over the onset of illness, that developments in one’s health over time is a matter of chance or fate, or in any case is determined by some external locus of control.

This paper makes a contribution to the literature on the relationship between family genetic endowment and health-related behaviours on several levels. To date we are aware of few studies (Ganz, Reference Ganz2001; Dickie and Gerking, Reference Dickie and Gerking1997) which test the relationship between family health endowment and choice of health-related behaviours by adults using directly observable exogenous measures of health endowment. Ganz, for example, found that smoking intensity falls after newly diagnosed smoking-related family cancer. While Ganz considered the effect of family history of cancer on smoking and alcohol consumption, we are able to consider family history of a number of diseases (cancer, diabetes, heart disease, high blood pressure, high cholesterol and stroke). This gives us the additional advantage of being able to have different endowment effects on different behaviours. We consider the effect of family health history on two behaviours, regular exercise and smoking.

Theoretical model

We assume individuals maximize utility subject to the standard economic constraints. Simply put, utility is a function of goods purchased in the market (M) and health produced by the individual (H).

(1)
$${\rm{Max }}\ U = U(M, \ H)$$

.

The production of health is subject to a health production technology whose inputs include the individual’s genetic endowment (E), endogenous health behaviours (V) and exogenous characteristics (X)

(1)
$$H = H(E, \ V, \ X)$$

.

Finally, the individual faces the budget constraint

(3)
$$P_V V + M = Y,$$

where PV are prices for V and Y is income. We assume H′(V)> 0 for healthy behaviours; that is, the marginal product of healthy behaviours is positive.

Our main question in this paper is whether genetic endowment and behaviours are complements or substitutes; that is, whether individuals with poor genetic endowment seek to reduce the risk their endowment pose to their health by adopting behaviours that reduce such risk or whether individuals with poor endowments increase the risk of developing the disease with unhealthy behaviours. There are several diseases that have come to be termed ‘lifestyle’ diseases because the probability of developing those diseases is modifiable by participating in a mix of healthy behaviours. So our empirical questions are whether V is a function of E and what is the sign of V′(E). Both V′(E) < 0 (Becker–Grossman) and V′(E) > 0 (Birch–Dow) are theoretical possibilities. Furthermore, since an individual has many genetic components and there are many health-related behaviours we suggest V′(E) could be negative with respect to Vj and Ei but positive with respect to Vk and El for any endowment component/health-related behaviour pair.

Data and methods

The data used in this study were collected in 2001 by the Caribbean Food and Nutrition Institute (CFNI) in Jamaica. The target population was adults aged 18 to 64 years. The survey was designed as a stratified multi-staged sample. The first stage is the selection of census Enumeration Districts (EDs). Enumeration districts are fully contained within each administrative region (parishes). The EDs are selected with probability proportional to their size (measured by the number of dwellings in each ED). The second stage is the selection of dwelling units within each ED. From each ED an equal number of dwellings were selected using systematic sampling with a random start. At the household level one person was selected for interview by trained professionals. That individual must be between 18 and 64 years of age and should be the person who last celebrated a birthday.

Interviewers told potential respondents they were from a research company in Kingston (including the name of the company), conducting a survey about dietary and exercise behaviour on behalf of CFNI. Once a respondent from the survey household was selected to be interviewed, he or she was told the interview could take up to 45 minutes and that their cooperation was appreciated. A total of 1,513 respondents participated (see Table 1 for descriptive statistics of the sample).

Table 1.

Descriptive statistics

Dependent variables

The purpose of this study is to investigate the extent to which family genetic endowment affects choice of health-related behaviours. The two behaviours of interest are cigarette smoking and physical activity.

Cigarette smoking

The survey respondents were asked if they were ever a smoker and if they are currently a smoker. We took full advantage of the presence of quitters by creating a multiple response smoking variable (never = 0, quitters = 1, current smokers = 2). Smoking is of interest because it has been identified as a risk factor for several diseases, including some types of cancer and stroke (Osann, Reference Osann1991; Oglobin and Brock, Reference Ogloblin and Brock2003).

Physical activity

The health and healthcare benefits of exercise in reducing the risk of coronary heart disease, cerebrovascular disease (stroke) and diabetes mellitus have been shown to be substantial (Nicholl et al., Reference Nicholl, Coleman and Brazier1994). Light physical activity has also been shown to lead to modest reductions in the number of doctor’s visits by older individuals (Lee and Kobayashi, Reference Lee and Kobayashi2001). Physical exercise has also been linked to reductions in neuroendocrine stress hormones and depressive symptoms (Nabkasorn et al., Reference Nabkasorn, Miyai, Sootmongkol, Junprasert, Yamamoto, Arita and Miyashita2005).

Respondents were asked which of five statements best describe their current exercise habit. The statements ranged from ‘At present I do not exercise and I do not intend to start in the next six months’ to ‘At present I exercise regularly and have been doing so for longer than six months.’ Respondents were told that by regular exercise the interviewer means planned exercise for at least one-half hour for three or more days a week. We created two variables based on these questions. The first variable is equal to one if the respondent had been exercising regularly for at least six months and zero otherwise. The second variable is equal to one if the respondent has been exercising regularly whether for six months or less than six months. Individuals who have maintained an exercise program for more than six months were combined with individuals who have been exercising regularly for less than six months because the latter group is a relatively small group comprising only 9 percent of the sample.

Independent variables

Family health endowments

The data set contains two questions on knowledge of family health history: whether a family member has been diagnosed with one of six chronic diseases (cancer, diabetes, heart disease, high blood pressure, high cholesterol and stroke) and whether a family member has died from one of those diseases. We created six binary variables for each of the diseases with which family members have been diagnosed (yes = 1, no = 0). In addition, we created a discrete variable measuring the number of chronic diseases diagnosed in the family.

In a similar vein we created four binary variables (high blood pressure and high cholesterol excluded), one for each of the diseases from which a family member has died (yes = 1, no = 0). A count variable was also created to indicate the number of diseases from which family members had died. About 10% of respondents indicated they were not aware of any family member who had died from one of the diseases listed above (responses of ‘don’t know’). For each of the binary variables these individuals were given a zero response and a separate variable was created to identify these individuals. Hence we have a fifth binary variable which includes all the ‘don’t knows’. One benefit of having this group in the data is that if health-related behaviours are unrelated to genetic endowment then such behaviours should also be neutral with respect to not knowing one’s endowment.

It should be noted that for the purposes of the present research, family is defined as the extended family. The nuclear family carries less meaning than the extended family for the majority of the population in the Caribbean. Therefore, it is typical to define the family in the extended sense rather than the nuclear (Barrow, Reference Barrow1998; Clarke, Reference Clarke1999 revised edition). In addition, we wish to note there is a distinction between knowledge of one’s family health history and a belief that a health condition is hereditary. Knowledge of illness in the family may increase health awareness and therefore leads to participation in healthy behaviours, regardless of whether the individual believes there is familial susceptibility. A family history of chronic disease may or may not lead to a belief that an illness runs within the family.

Other independent variables

We use a set of fairly standard independent variables. Variables include education, marital status, age, race, gender, area of residence (urban versus rural) and employment status (employed versus unemployed). We created two binary education variables, one for individuals whose highest education is high school completion and another for individuals who have completed college education. For marital status we have a single variable (single versus not single). Age is entered as a continuous variable measured in years. Race is entered as a binary variable (black = 1, all others = 0).

We include two additional variables, whether the individual is satisfied with his current state of health, and whether he has been trying to improve his health. If the individual is dissatisfied with his health, then the current stock of health capital must be less than the individual’s optimal. This should encourage the individual to undertake changes (which may include physical activity) that would put one on the path towards the optimum. Similarly, if an individual is trying to improve his health, the individual must believe health outcomes are modifiable by individual action, indicating the individual does not take a fatalistic view of health production.

The major data weakness is a lack of income data. It is typical for surveys in Jamaica (except for the living standards measurement survey) to not ask questions about income because of the complexity of obtaining a full measure of income. While this is a data weakness, it is not clear that it is serious. Studies of health-related behaviour have produced at best inconsistent results of the relationship between income level and healthy behaviour.

Estimation strategy

We estimated separate equations for cigarette smoking and exercise. Participation in physical exercise is a binary variable. We used the following standard simple logistic regression model to estimate coefficients for participation in physical activity

$$\Pr (P_i = 1) = \Lambda (\beta 'x_i ){\rm{ },}$$

where Pi = 1 if an individual exercises regularly, Λ(.) is the logistic cumulative distribution function, β is a vector of coefficients to be estimated and xi is a vector of individual i’s characteristics.

Participation in cigarette smoking has three outcomes; never smoked (j=0), smoked and quit (j=1), and currently smoking (j=2). We estimated the equation using multinomial logistic regression. Therefore, the model of cigarette smoking participation is

$$\Pr (S_i = j) = {{\ell ^{\beta '_j x_i } } \over {\sum\limits_{k = 0}^2 {\ell ^{\beta '_k x_i } } }},$$

where Si = 0, 1 or 2 according to whether the individual has never smoked cigarettes, has smoked cigarettes and quit, or currently smokes cigarettes, respectively. We estimated coefficients (and odd-ratios) for quitters and current smokers, with ‘never smoked’ as the omitted category.

There is evidence in the literature (Kenkel, Reference Kenkel1991; Leigh and Dhir, Reference Leigh and Dhir1997) of a strong connection between education achievement and health behaviour. To take into account the health behaviours/education interaction we divided the sample into two sub-samples based on education: individuals with primary education or less and individuals with secondary and college education. Similarly, Nicholl et al. (Reference Nicholl, Coleman and Brazier1994) suggest that the health benefits of exercise in terms of disease prevention clearly exceed the costs that would be incurred for adults 45 years and older, but not for younger adults. We therefore divided the sample by age (40 years and older, and 39 years and younger) so we could account for the presence of any health-behaviour/age interactions.

Finally, there is clearly the potential for measurement error in self reporting of behaviours such as exercise and smoking. Since smoking and exercise behaviour are our dependent variables in the regression analysis, measurement error does not appear to be a fatal problem. According to Greene (Reference Greene2000), as long as the regressors are measured properly, measurement error in the dependent variables can be absorbed in the disturbance term of the regression and ignored.

Results

Descriptive statistics are presented in Table 1. Over 21% of the respondents have smoked cigarettes, 8% have smoked and quit, while nearly 14% smoke currently. Men are twice as likely to participate in exercise as women (41% vs. 21%). In terms of family history of chronic disease the most frequently reported disease is high blood pressure (44%) followed by diabetes (34%). Over 60% of respondents have a family member who has been diagnosed with one of the six chronic diseases, with an average of one chronic disease per family. Thirty percent of the respondents reported that they have had a family member die of one of these diseases, while 10% of respondents said they did not know of a family member who had died from any of these diseases.

Smoking status and age interaction (see Tables 2 and 3)

Family history of heart disease is positively associated with quitting smoking among persons who are 40 years and older and persons who are younger than 40 years of age. Individuals less than 40 years of age who have a family member diagnosed with heart disease are likely to have quit smoking. The result has strong statistical significance and relatively large odds-ratio (OR=2.73). Individuals 40 years and older who have had family members die from heart disease are less likely to be current smokers (OR=0.34). On the other hand, individuals 40 years or older who are not aware of their family history are significantly more likely to have been a smoker (OR=5.6) or are currently a smoker (OR=4.45). Ignorance of family history is therefore strongly related to smoking behaviour among persons over 40 years of age.

Table 2.

Logistic regression results persons 40 years and older: smoking status (quitters = 1, smokers = 2, omitted category: never a smoker = 0)

Table 3.

Logistic regression results persons less than 40 years of age: smoking status (quitters = 1, smokers = 2, omitted category: never a smoker = 0)

Exercise and age interaction (see Tables 4 and 5)

The death of a family member from stroke is positively related to participation in physical exercise among individuals who are 40 years and older and among individuals who are less than 40 years of age. The odds-ratio among individuals who are over 40 years of age is slightly larger (OR=1.88) than the odds-ratio among individuals less than 40 years of age (OR=1.68). However, family history of diabetes has a negative association with exercise participation among individuals who are less than 40 years of age (OR=0.55).

Table 4.

Logistic regression results: participation in physical exercise by persons 40 years and older

Table 5.

Logistic regression results: participation in physical exercise by persons less than 40 years of age

Smoking status and education interaction (see Tables 6 and 7)

Some important differences emerge in smoking behaviour by education level. A history of high blood pressure within one’s family appears to be related to smoking status but in somewhat different ways. Individuals with less than secondary education who have a family history of high blood pressure are likely to quit smoking (OR=1.97). Individuals who have secondary education or higher with a family history of diabetes are more likely to have never smoked than to have smoked and quit (OR=0.59), but the result is not statistically significant. Persons whose educational attainment is secondary education or higher are more likely to have smoked and quit if they have a family member diagnosed with heart disease, with fairly large odds (OR=2.5).

Table 6.

Logistic regression results persons with less than secondary education: smoking status (Quitters = 1, Smokers = 2, Omitted category: never a smoker = 0)

Table 7.

Logistic regression results persons with secondary education or higher: smoking status (Quitters = 1, Smokers = 2, Omitted category: never a smoker = 0)

Exercise and education interaction (see Tables 8 and 9)

Family history of heart disease has differential effects on participation in exercise by education level. There is a positive association between participation in exercise and family history of heart disease among individuals with less than secondary education (OR=2.07). Family history of stroke affects participation in exercise by individuals with less than secondary education but the effect is negative (OR=0.47). In contrast to individuals with less than secondary education, the relationship between family history of heart disease and participation in exercise is negative for individuals with secondary education of higher (OR=0.57). There is also a negative relationship between family history of high blood pressure and exercise participation (OR=0.77). Family history of high cholesterol has a positive association with exercise among individuals with secondary education or higher (OR=2.14).

Table 8.

Logistic regression results: participation in physical exercise by persons with less than secondary education

Table 9.

Logistic regression results: participation in physical exercise by persons with secondary education or higher

Results summary

It is fair to say family history of heart disease and stroke both exert significant influence on health-related behaviours (negative or positive). In three of the four subpopulations persons with a family history of heart disease are likely to have given up smoking. Family history of stroke is also associated with exercise participation in three of the four subpopulations (positive in two subpopulations and negative in one). Family history of diabetes, high blood pressure and cholesterol exert some influence on health-related behaviours but with far less consistency. Finally, for each subpopulation, the family health factors associated with exercise are different from the factors associated with smoking.

Discussion

This paper presented an investigation into whether individuals are more likely to participate in healthy behaviours or avoid unhealthy behaviours if their family has a reported history of diagnosis with, or death from chronic illnesses whose risks are modifiable through participation in (avoidance of) certain behaviours. The major findings are as follows. First, the six chronic illnesses identified are prevalent in the study population. Sixty percent of respondents have a family history of diagnosis with at least one disease, while 30% have family members who died from at least one disease. Second, in general, family history of heart disease and stroke increases the probability of participation in exercise, but also reduces the probability of participation in cigarette smoking.

Overall our results appear to be most closely aligned with the Becker–Grossman theoretical perspective. Individuals appear to adopt certain healthy behaviours if their family health endowment is poor as the Becker–Grossman hypothesis suggests. Furthermore, at least among individuals who are over 40 years of age, those who are not aware of their family health endowment appear, on average, to be more likely to participate in cigarette smoking. However, there are some instances in which family history of chronic disease is associated with lower participation in healthy behaviours. As was suggested earlier, there is the possibility that fatalism influences the way persons respond to possible threats to their health. In addition it is likely that for some persons there is lack of clarity on the contribution of genetic makeup, behaviour and environment to health and the interaction among them (Parrott et al., Reference Parrott, Silk, Weiner, Condit, Harris and Bernhardt2004).

There is evidence that incidences of certain diseases tend to aggregate within a family, even though there are no established genetic markers for those diseases. The use of such information by anyone other than the individual has generated considerable debate, the debate generated at least in part by concerns that such information could be used to exclude some portions of the population from access to health insurance or used to justify higher premiums, even when no disease is actually present. Our results suggest knowledge of one’s family health history can have a significant impact on health behaviours. If individuals modify behaviour sufficiently to compensate for the influence of their family genetic makeup, they may be at no more risk of developing a disease than individuals who engage in unhealthy behaviours and hope to be protected by their genes.

Given the prevalence of diseases in the population it seems that there are significant health benefits to be gained from a healthy lifestyle. Not to mention the possibility of limiting health care costs. One inference that can be drawn from our results is the need for carefully crafted health communication messages that help individuals understand the interplay between family health history, environment and behaviour in the production of their own health. In particular, communication needs to emphasize that health risks can be modified and provide effective ways to modify such risks.

While we believe our results represent an important extension of the general understanding of the interaction between family health endowment and health-related behaviours, they leave a number of questions unanswered. First, does the specific definition of ‘family’ matter for the results? For example, does the occurrence of illness among nuclear family members create a stronger behavioural response than the occurrence of illness among extended family members? Second, does the occurrence of chronic diseases within a family change/create perceptions of familial susceptibility? If so, what factors determine the severity of one’s risk assessment? These questions of course create opportunities for future research.

Footnotes

1 The link between the APOE gene and increased risk of late-onset Alzheimer’ disease is generally accepted, but serious controversy still surrounds whether there are benefits to be derived from clinical APOE testing (Hedgecoe, Reference Hedgecoe2004).

2 To give a concrete example, for a soldier on the battlefield, the risk to health of smoking a cigarette (to be realized at some future date) is infinitesimally small compared to current risks.

References

Armstrong, K., Weber, B., FitzGerald, G., Hershey, J.C., Pauly, M.V., Lemaire, J., Subramaian, K., and Asch, D.A. (2003),‘Life insurance and breast cancer risk assessment: adverse selection, genetic testing decisions, and discrimination’, American Journal of Medical Genetics, 120A: 359–364.CrossRefGoogle ScholarPubMed
Barrow, C. (1998). Family in the Caribbean: Themes and Perspectives, Kingston: Ian Randle Publishers.Google Scholar
Becker, G.S. (1965), ‘A theory of the allocation of time’, Economic Journal,75:493–517.CrossRefGoogle Scholar
Birch, S. and Stoddart, G. (1991), ‘Incentives to be healthy: an economic model of health-related behaviour’, in Lopez-Casasnovas, G(ed.), Incentives in Health Services, Barcelona: Springer-Verlag.Google Scholar
BrownH.S., III H.S., III, Pagan, J.A., and Bastida, E. (2005), ‘The impact of diabetes on employment: genetic IVs in a bivariate probit’, Health Economics, 14:537–544.CrossRefGoogle Scholar
Chen, S., Shogren, J.F., Orazem, P.F., and Crocker, T. D. (2002), ‘Prices and health: identifying the effects of nutrition, exercise, and medication choices on blood pressure’, American Journal of Agricultural Economics, 84: 990–1002.CrossRefGoogle Scholar
Chiappetta, T.O. (2005), ‘Managing healthcare costs’, Public Personnel Management,34:313–320.CrossRefGoogle Scholar
Chodick, G., Heymann, A.D., Wood, F., and Kokia, E. (2005), ‘The direct medical cost of diabetics in Israel’, European Journal of Health Economics, 6: 166–171.CrossRefGoogle ScholarPubMed
Clarke, E. (1999) (1957), My Mother who Fathered Me, Kingston: The Press University of the West Indies.Google Scholar
Coulter, C.H. (2006), ‘The employer’s case for health management’, Benefits Quarterly, 22: 23–33.Google ScholarPubMed
Dickie, M. and Gerking, S. (1997), ‘Genetic risk factors and offsetting behaviour: the case of skin cancer’, Journal of Risk Uncertainty, 15: 81–97.CrossRefGoogle Scholar
Dow, W.H., Holmes, J., Philipson, T., and Sala-i-Martin, X. (1999), ‘Longevity complementarities under competing risks’, American Economic Review, 89(5):1358–1371.CrossRefGoogle Scholar
Ganz, M.L.(2001), ‘Family health effects: complements or substitutes’, Health Economics,10:699–714.CrossRefGoogle ScholarPubMed
Gorin, S.S. (2005), ‘Correlates of colorectal cancer screening compliance among urban Hispanics’, Journal of Behavioural Medicine, 28(2):125–137.CrossRefGoogle ScholarPubMed
Greene, W.A. (2000), Econometric Analysis, 4th edition, New Jersey: Prentice Hall.Google Scholar
Grossman, M. (1972), ‘On the concept of health capital and the demand for health’, Journal of Political Economy, 80(2): 223–255.CrossRefGoogle Scholar
Hedgecoe, A. (1996), ‘Genetic catch-22: Testing, risk and private health insurance’, Business and Professional Ethics Journal, 15(2):69–85.CrossRefGoogle ScholarPubMed
Hedgecoe, A. (2004), The Politics of Personalized Medicine: Pharmacogenetics in the Clinic, Cambridge: Cambridge University PressCrossRefGoogle Scholar
Kenkel, D. S. (1991), ‘Health behavior, health knowledge and schooling’, The Journal of Political Economy, 99(2): 287–305.CrossRefGoogle Scholar
Lee, M. and Kobayashi, S. (2001), ‘Proportional treatment effects for count response panel data: effects of binary exercise on health care demand’, Health Economics, 10:411–428.CrossRefGoogle ScholarPubMed
Lee, J.M., Turini, M., Botteman, M.F., Stpehens, J.M., and Pashos, C.L. (2004), ‘Economic burden of head and neck cancer: a literature review’, European Journal of Health Economics, 5: 70–80.CrossRefGoogle ScholarPubMed
Leigh, J.P. and Dhir, R. (1997), ‘Schooling and frailty among seniors’, Economics of Education Review, 16(1): 45–57.CrossRefGoogle Scholar
Lemaire, J., Subramanian, K., Armstrong, K., and Asch, D.A. (2000), ‘Pricing term insurance in the presence of a family history of breast or ovarian cancer’, North American Actuarial Journal, 4(2): 75–87.CrossRefGoogle Scholar
Liang, W., Yuan, E., Mandelblatt, J.S., and Pasick, R. (2004), ‘How do older Chinese women view health and cancer screening? Results from focus groups and implications for intervention’, Ethnicity and Health, 9(3): 283–304.CrossRefGoogle Scholar
Lu, L., Macdonald, A., and Wekwete, C. (2006), ‘Premium rates based on genetic studies: how reliable are they?’, Working Paper No. 06/3 (www.ma.hw.ac.uk/ams/girc/publications.php).Google Scholar
Macdonald, A. (2003), ‘Genetics and insurance: what have we learned so far?’, Scandinavian Actuarial Journal, 4: 324–348.CrossRefGoogle Scholar
Macdonald, A., Waters, H.R., and Wekwete, C.T. (2005), ‘A model for coronary heart disease and stroke with applications to critical illness insurance underwriting I: the model’, North American Actuarial Journal, 9(1): 13–40.CrossRefGoogle Scholar
Manning, W.G., Keller, E.B., Newhouse, J.P., Sloss, E.M., and Wasserman, J. (1991), The Costs of Poor Health Habits, Cambridge, MA: Harvard University Press.CrossRefGoogle Scholar
Nabkasorn, C., Miyai, N., Sootmongkol, N., Junprasert, S., Yamamoto, H., Arita, M., and Miyashita, K. (2005), ‘Effects of physical exercise on depression, neuroendocrine stress hormones, and physiological fitness in adolescent females with depressive symptoms’, European Journal of Public Health, 16: 179–184.CrossRefGoogle ScholarPubMed
Nicholl, J.P., Coleman, P., and Brazier, J.E. (1994), ‘Health and healthcare costs and benefits of exercise’, PharmacoEconomics, 5(2): 109–122.CrossRefGoogle ScholarPubMed
Ogloblin, C. and Brock, G. (2003), ‘Smoking in Russia: the “Marlboro Man” rides but without “Virginia Slims” for now’, Comparative Economic Studies, 45:87–103.CrossRefGoogle Scholar
Ory, C., Vanderplas, A., Dezii, C., and Chang, E. (2005), ‘Congestive heart failure: attributable costs within the managed care setting’, Journal of Pharmaceutical Finance, Economics and Policy, 14(2): 87–97.CrossRefGoogle Scholar
Osann, K.E. (1991), ‘Lung cancer in women: the importance of smoking, family history of cancer, and medical history of respiratory disease’, Cancer Research, 51:4893–4897.Google ScholarPubMed
Parrott, R., Silk, K., Weiner, J., Condit, C., Harris, T., and Bernhardt, J. (2004), ‘Deriving lay models of uncertainty about genes’ role in illness causation to guide communication about human genetics’, Journal of Communication, 54(1) 105–122.CrossRefGoogle Scholar
Payne, K.A., Kuybrechts, K. F., Craig Green, T. J., and Klittich, W.S. (2002), ‘Long-term cost-of-illness in stroke: an international review’, PharmacoEconomics, 20:813–825.CrossRefGoogle ScholarPubMed
Schwartz, A.G., Yang, P., and Swanson, G.M. (1996), ‘Familial risk of lung cancer among nonsmokers and their relatives’, American Journal of Epidemiology, 144(6): 554–562.CrossRefGoogle ScholarPubMed
Sharf, B.F., Stelljes, L.A., and Gordon, H.S. (2005), ‘A little bitty spot and I’m a big man: Patients’ perspectives on refusing diagnosis of treatment for lung cancer’, Psycho-oncology, 14: 636–646.CrossRefGoogle Scholar
Tabarrok, A. (1994), ‘Genetic testing: an economic and contractarian analysis’, Journal of Health Economics, 13: 75–91.CrossRefGoogle ScholarPubMed
Thorpe, K.E. (2005), ‘The rise in health care spending and what to do about it?’, Health Affairs, 24: 1436–1445.CrossRefGoogle Scholar
Figure 0

Table 1.Descriptive statistics

Figure 1

Table 2.Logistic regression results persons 40 years and older: smoking status (quitters = 1, smokers = 2, omitted category: never a smoker = 0)

Figure 2

Table 3.Logistic regression results persons less than 40 years of age: smoking status (quitters = 1, smokers = 2, omitted category: never a smoker = 0)

Figure 3

Table 4.Logistic regression results: participation in physical exercise by persons 40 years and older

Figure 4

Table 5.Logistic regression results: participation in physical exercise by persons less than 40 years of age

Figure 5

Table 6.Logistic regression results persons with less than secondary education: smoking status (Quitters = 1, Smokers = 2, Omitted category: never a smoker = 0)

Figure 6

Table 7.Logistic regression results persons with secondary education or higher: smoking status (Quitters = 1, Smokers = 2, Omitted category: never a smoker = 0)

Figure 7

Table 8.Logistic regression results: participation in physical exercise by persons with less than secondary education

Figure 8

Table 9.Logistic regression results: participation in physical exercise by persons with secondary education or higher