10.1.1. Interactions among determinants

 

There are different types of relationships between health determinants. In general, observed associations between different health determinants and outcomes may be causal or non-causal, direct or indirect, and synergistic or antagonistic. For details, reference is made to textbooks of epidemiology and biostatistics (Rosner, 2000; Szklo and Nieto, 2005).

 

Also important to note is that one health determinant can confound the association between another health determinant and the health outcome of interest or act as an effect modifier (Szklo, 2000). Particularly, when performing or interpreting observational studies it is critically important to understand these two concepts and their implications.

 

A confounder is by definition associated with determinant and disease and is not in the causal pathway between the two. Confounding is a threat to validity that should be controlled for by study design and analysis: e.g. matching, adjustment, restriction, or randomization (Szklo, 2000; Szklo and Nieto, 2005).

 

Effect modification is present if the relationship between determinant and disease differs for different levels of another determinant (Rosner, 2000). Frequently the terms effect modification and interaction are used interchangeably, although effect modification is by definition a causal effect whereas interaction may be just a statistical phenomenon. If there is effect modification, stratified instead of adjusted results have to be presented (McClelland and Judd, 1993; Rosner, 2000; Szklo, 2000; Szklo and Nieto, 2005).

 

Examples of interactions among determinants

 

Association within identical group of determinants

Health determinants dealing with lifestyle issues can be divided into three main categories: predisposing, reinforcing, and enabling. The first subgroup mainly consists of individual factors which are influenced by cultural and social reinforcing factors (e.g. peer acceptance and social disapproval). Enabling factors, however, mainly refer to the environment facilitating or creating barriers to health behaviour (Green and Potvin, 2004).

 

Association within subgroup (tobacco use)

Health determinants within the field of tobacco use are manifold: predisposing factors such as knowledge of adverse health effects of tobacco use, attitudes towards tobacco use, biological / genetic factors influencing susceptibility to addiction, enabling factors including costs of tobacco, taxes on tobacco products (e.g. smuggling and low quality cigarettes), availability of tobacco products as well as reinforcing factors such as policies on smoking (e.g. age limits, public smoking, advertising), peer and family attitudes and influences, social support for non-smoking, public opinion towards smoking, and social norms (e.g. smoking during pregnancy, smoking after delivery, social disapproval). Tobacco use as behavioural risk factor is the product of the complex interactions between the described component behaviours and influences. Figure 10.1.3 is also applicable to the “tobacco use” health determinant.

Exposure to second hand smoke and active smoking are closely linked. Parents’ smoking is a powerful influence on the smoking behaviour of their children: Children whose parents smoke and who are therefore exposed to second hand smoke are more likely to smoke in the future (Gidding et al, 1994).

 

Association between subgroups (food choices / nutrition and physical activity)

Both diet and physical activity influence energy balance. Energy input and energy expenditure are direct modulators of body weight. It is critically important to understand the mechanisms by which the interaction between dietary factors and physical activity influences body weight and the development of overweight (Jebb and Moore, 1999; Melzer et al, 2005). High caloric intake combined with low energy output is a behavioural determinant of obesity and ill-health, including cardiovascular disease and other chronic diseases (Burke et al, 1995; Green and Potvin, 2004). Social and environmental factors which increase energy intake and / or reduce physical activity contribute to increases in body weight (Wagner et al, 2005). The amount of energy intake and expenditure as well as patterns of both determinants is relevant. Physical activity is an important interacting lifestyle behaviour and is linked to a whole variety of other major determinants of health (lifestyle, environmental, psychosocial determinants etc) (Green and Potvin, 2004; Wagner and Kirch, 2006).

 

Addition of regular exercise to hypocaloric diet in weight loss significantly decreases body weight and body fat as well as the risk of coronary heart disease (Fagard, 2005; Wood, 1994). There is an ongoing discussion if sequential (non-integrated) or simultaneous interventions targeting physical activity and healthy diet are more effective in reducing behavioural risk factors for cardiovascular and other chronic diseases. In this regard, effectiveness has been reported to differ between subpopulations (e.g. adolescents vs adults, underserved populations vs people of high socio-economic position) (Marcus et al, 2006).

The mechanisms by which physical activity influences food intake are complex (Titchenal, 1988; Tappy et al, 2003). The type, intensity, frequency and permanence of activity as well as age, degree of fitness, and body composition play an important role in food intake regulation (Blundell and King, 1998; Melzer et al, 2005). Only few studies have addressed the effects of energy expended in exercise on diet. (Tappy et al, 2003) Within the normal activity range, energy intake is balanced with the activity level. (Melzer et al, 2005). Most evidence suggests that energy intake is more or less resistant to short term (1-3 days) increases in energy expenditure, while increases in physical activity are not automatically followed by equivalent increases in caloric intake (King, 1999; King et al, 1997; Melzer et al, 2005). A weak short term coupling between energy expenditure and energy intake may be due to the fact that behavioural acts of food intake depend partly on environmental contingencies (King, 1998). In addition, only the exposure to high-fat, low-carbohydrate foods, which have a weak effect on satiation, may cause exercise-induced hyperphagia (King et al, 1997). In lean people, in long-term engagement in exercise (≥ 7 days), however, the correlation between energy intake and expenditure increases: lean subjects demonstrate an increase in energy intake. However, obese, untrained people do not tend to balance the extra energy expenditure through increased energy intake. Intense exercise is more influential compared to low-intensity physical activity (King et al, 1997; Melzer et al, 2005; Titchenal, 1988).

 

Selection of macronutrients (carbohydrate, fat, protein) may be altered by changes in energy expenditure. The literature is not unanimous in this regard. Long-term increases in physical activity have been shown to increase the proportion of calories from carbohydrate (King et al, 1997). Highly active children consume a higher percentage of energy as carbohydrate and a lower percentage as fat compared to children with a low activity level (Parsons et al, 1999). Ambler et al (1998) reported, however, that increased levels of activity are associated to increases in fat and decreases in carbohydrate consumption in female, non-obese adolescents. Conflicting results may be explained by the fact that the food intake response to exercise is influenced not only by physiological factors but also by psychological and cognitive factors such as volitional control towards exercise, food-related cognitions, susceptibility to food advertising and reasons for exercising (King, 1999; King et al, 1997; Parsons et al, 1999; Schmitt et al, 2007).

 

Studies investigating interactions between nutrition and physical activity show several limitations. Diet and physical activity are both difficult to measure (Parsons et al, 1999; Titchenal, 1988). Methodological limitations of accurately assessing energy expenditure and intake create an uncertainty about the true interaction between physical activity and diet (King et al, 1997). In addition, only few studies investigate the two risk factors simultaneously (Parsons et al, 1999). Most of exercise intervention studies only consider the energy intake immediately following the exercise (King, 1999). A clear understanding of the interaction between food intake and physical activity is necessary to improve individual and societal health management (Melzer et al, 2005). Inconsistent and conflicting study results reflect the complexity of health determinants interactions and a lack of knowledge in this field (Titchenal, 1988).

 

Food intake is not only influenced by exercise, but also affects the level of engagement in physical exercise. Feelings of fullness after nutrient-rich meals potentially reverse the individual’s intention to exercise. In addition, overweight and obese subjects may feel stigmatized when participating in work-out classes in which the great majority of individuals are lean, fit and good-looking. The family environment and social norms may not be supportive either. Moreover, there are types of physical activity which are not recommended or possible for overweight and obese people as they are harmful to joints or lead to substantial heart rate increases.

 

Figure 10.1.4 illustrates the described associations as well as the integration of nutrition and physical activity in the model of health and disease determinants.

 

Figure 10.1.4. Integration of nutrition and physical activity in the model of health and disease determinants

 

 

Association between different groups of determinants (lifestyle determinants: alcohol consumption and psychosocial determinants)

 

In addition to biological factors and individual knowledge and beliefs, social circumstances critically influence risk behaviour (Green and Potvin, 2004) (Figure 10.1.2). Alcohol consumption is linked to psychosocial determinants: drinking is a social act in many cultures (Rehm et al, 1996); mood management and dependency are other reasons for alcohol consumption; social networks influence individual development of identities and lifestyle attitudes. Social connectedness as well as social isolation (e.g. after migration and resettlement) also contributes indirectly to lifestyle changes. Concerning alcohol consumption, core relationships and social networks / peers may reinforce both healthy and unhealthy behaviour (Babor et al, 1987; Hart, 2004; Maida, 1984). On the other hand, alcohol consumption impacts on psychosocial determinants. For example, declining consumption of alcoholic beverages by men has been shown to reduce domestic violence as well as stressful life events in general and to ameliorate the effort-reward balance (Collins et al, 1997; Hart, 2004).

 

Predisposing psychosocial factors of alcohol consumption include early drinking experiences, expectations about the effects of alcohol consumption, heavy social drinking, psychological stress, low self-esteem, and a family history of alcohol dependence. Reinforcing psychosocial factors mainly refer to family and peer influences. The availability of alcoholic and non-alcoholic drinks as well as the cost of alcohol are examples of enabling factors and barriers of alcohol use (Green and Potvin, 2004).

 

Parental drinking affects the environment and quality of life of growing children possibly leading to social, psychological and somatic problems which persist into adulthood. The association between drinking and aggressive behaviour has been shown in many studies, but there is still no evidence of a causal relationship. Other social, personality and genetic factors are influential (Gmel and Rehm, 2003; Milgram, 1993; Taylor and Chermack, 1993). The evidence linking male alcohol consumption to intimate partner violence is weak – partly due to publication bias and low study quality (Gil-González, 2006). Additional psychosocial consequences of alcohol consumption include interpersonal conflicts, family disruption, failure to fulfil social roles, concern about drinking, victimization as well as reduction of anxiety, stress, and worries (Bondy, 1996; Pohorecky, 1991; Thakker, 1998). Alcohol consumption is also important in the development of psychiatric disorders (social phobia, social anxiety) and suicidal behaviour (Blow et al, 2004; Burke and Stephens, 1999; Kushner et al, 1990; Marshall, 1994; Miller et al, 1991; Miller et al, 1992; Sher, 2006; Milgram, 1993).

 

Family structure and family life influence alcohol consumption. It has been reported that alcohol plays an adaptive role in family life: marital satisfaction is higher and rates of depression are lower in families with higher levels of alcohol consumption. These associations seem to be stronger in steady drinkers compared to binge drinkers (Jacobs and Wolin, 1989). Children of alcoholics are at an increased risk of alcoholism in adulthood. This intergenerational continuity has mostly been studied in problem drinking fathers and their sons and may be mediated by disturbed family relationships / family disharmony. The evidence to conclude that the offspring of problem drinkers is at an increased risk of difficulties related to self-esteem, life satisfaction, relationship formation, antisocial and drug-using behaviour is not sufficient yet (Velleman, 1992a; Velleman, 1992b; Wiers et al, 1994). For the initiation of drinking, parental influence is more important than peer behaviour. Single-parent families, the perception of excessive drinking in the family environment and especially weak family bonds encourage frequent and excessive drinking among adolescents. In late adolescence and early adulthood, however, peer influence becomes a predominant predictor of alcohol consumption (Kuntsche and Jordan, 2006; Kuntsche and Kuendig, 2006; Milgram, 1993).

 

The likelihood of an individual engaging in alcohol use depends on the stress level and the individuals’ resources, ability to cope and drug protective competence. The latter includes social influence, sociability, self-worth, control/responsibility, intimacy, nurturance, goal and spiritual directedness (Brady and Sonne, 1999; Lindenberg et al, 1998; Milgram, 1993).

 

Stress is also considered as a major contributor to continuation of alcohol misuse and to relapse in recovering alcohol misusers. Again, heavy drinking may be the cause or the consequence of stress and stressful life events. Life events can both increase and decrease alcohol consumption depending on the type of event, degree of personal involvement, gender and age (Veenstra et al, 2006). In general, the association between alcohol consumption and stress is complex and modified by age and gender (Brady and Sonne, 1999; Pohorecky, 1991).

 

Figure 10.1.5. illustrates examples of the described associations.

 

Figure 10.1.5. Examples of psychosocial determinants which are affected and affect alcohol consumption

 

 

 

 

 

 

 

 

For what concerns interventions on alcohol consumption, it is critically important not to underestimate the resources as well as the barriers originating from social networks to the development of healthy behaviour / behaviour change. In history, the rising prevalence of companionate marriages was most influential in reducing alcohol consumption in British men. Social relationships within the family or household strongly influence lifestyle and have great health promotion potential (Hart, 2004).

 

Europe is observing an increasing prevalence of high-risk sexual behaviour. Condom use can be challenged by several social, cultural and economic factors. One potential explanation for poor motivation in condom use and decreased perception of risk may be alcohol consumption prior to having sex.

Numerous studies have reported the correlation of excessive drinking and risky sexual behaviour (Thompson et al, 2005). Being away from home is also associated with concurrency of partnership and an increase in risk behaviours.