Health behaviour: cancer screening, blood and organ donation, and opioid (mis)use

The current special issue of Psychology & Health follows on from a similar special issue published in 2017 (Conner & Norman, 2017) in focusing on an important topic for contemporary health psychology, namely health behaviours. The study of health behaviours in terms of prevalence, determinants and the effectiveness of interventions has been and remains an important area of health psychology. These special issues of Psychology & Health bring together state of the art reviews on key health behaviours authored by experts in the field. Each review aims to provide an overview of what is currently known about a specific health behaviour in relation to definitions, impacts on health, prevalence in different groups, key determinants, and effectiveness of interventions, as well as providing future directions for research. As we previously argued (Conner & Norman, 2017), such state of the art reviews can both be sources of information and stimulate new research and so be useful to a broad range of readers, from those with little knowledge of the area to those actively involved with research on these health behaviours. These reviews also highlight what is unique about specific health behaviours, knowledge that can be lost in more general reviews.

The previous special issue of Psychology & Health on health behaviours comprised of reviews of eating a healthy diet (de Ridder, Kroese, Evers, Adriaanse, & Gillebaart, 2017), engaging in physical activity (Rhodes, Janssen, Bredin, Warburton, & Bauman, 2017), binge drinking (Kuntsche, Kuntsche, Thrul, & Gmel, 2017) and smoking (West, 2017). The present special issue adds to this list of health behaviours by providing reviews on cancer screening (Sarma, Silver, Kobrin, Marcus, & Ferrer, 2019), blood and organ donation (Fergusson, Murray, & O’Carroll, 2019) and opioid use and misuse (Bolshakova, Sussman, & Bluthenthal, 2019). The behaviours covered in the special issues are a regular focus of attention in Psychology & Health. For example, recent issues of Psychology & Health have seen papers focusing on physical activity (Janssen & Waters, 2019; Parker, Chipperfield, Perry, Hamm, & Hoppmann, 2019; Robinson, Bisson, Hughes, Ebert, & Lachman, 2019; Smith, Williams, O’Donnell, & McKechnie, 2019), binge drinking (Caudwell, Keech, Hamilton, Mullen, & Hagger, 2019; Davies, 2019; Giese, Stok, & Renner, 2019; Norman, Webb, & Millings, 2019; Tatnell, Loxton, Modecki, & Hamilton, 2019), healthy eating (Moreno-Agostino et al., 2019; Voi & Sainsbury, 2019), screening (Tong & Sopory, 2019), smoking (Scholten, Granic, Chen, Veling, & Luijten, 2019) and drug use (Kougiali, Fasulo, Needs, & Van Laar, 2017).

In the previous special issue editorial (Conner & Norman, 2017) we commented on the complex interrelationship between the definition of health behaviours and their prevalence and impacts on health; the degree of commonality among key determinants; the effectiveness of interventions; and potential directions for future research. In the current editorial we revisit some of these same issues in the context of discussing similarities and differences between health behaviours and the relative value of focusing on individual health behaviours versus multiple health behaviours.

In focusing on expert reviews of individual health behaviours these two special issues of Psychology & Health have sought to emphasise the value of specialist knowledge about the seven individual health behaviours reviewed. This work emphasises what is unique about each of these behaviours, particularly in relation to determinants and effective means to change them. Considering those determinants that may be unique to each behaviour: risk perceptions and cancer fatalism were highlighted as determinants of cancer screening; role identity, moral norms, prosocial emotions and vasovagal reactions were highlighted in relation to blood and organ donation; being male, negative childhood experiences, and homelessness were highlighted in relation to opioid abuse; habit was highlighted as a determinant of a healthy diet; weak academic orientation and drinking alcohol were highlighted as determinants of smoking; and ethnicity and good health status were highlighted as determinants of physical activity. Considering those interventions that may be unique to each behaviour a number emerge: patient education and provider recommendation for cancer screening; resilience training for opioid abuse; altruism focused feedback for blood and organ donation; motivational interviewing for binge drinking; self-regulatory training for eating a healthy diet; self-monitoring for physical activity; and nicotine replacement therapy for smoking.

A potential weakness in considering health behaviours individually is that some information is potentially lost in relation to common determinants and means to change behaviour. Across the seven behaviours a number of common determinants and means to change the behaviour emerged. Considering those determinants that may be common to more than one behaviour: knowledge was highlighted in relation to both cancer screening and eating a healthy diet; emotions (positive and negative) were highlighted in relation to cancer screening, blood and organ donation, and eating a healthy diet; addiction and mental health problems were highlighted in relation to opioid abuse and smoking; automatic influences were highlighted in relation to cancer screening and eating a healthy diet; motivation and self-efficacy were highlighted as key determinants of blood and organ donation, eating a healthy diet, binge drinking and physical activity; social influences (injunctive and descriptive norms, social support) were highlighted as key determinants of cancer screening, blood and organ donation, binge drinking, smoking and physical activity; attitudes or expectancies were highlighted as key determinants of cancer screening, blood and organ donation, binge drinking, smoking and physical activity; and socioeconomic status was highlighted as a key determinant of cancer screening, opioid abuse, smoking and physical activity.

Similarly, a number of interventions that may be common across more than one behaviour emerge: using reminders and targeting affect may be effective for changing both cancer screening and blood and organ donation; targeting social influences may be effective for changing both opioid abuse and blood and organ donation; implementation intentions may be effective for targeting cancer screening, blood and organ donation, healthy eating, and physical activity; incentives (financial or other) may be effective for changing cancer screening and smoking; drug therapy may be effective for targeting opioid abuse and smoking.

The value of studying multiple health behaviours

It is worth noting that health behaviours are rarely performed or changed in isolation from other health behaviours or non-health behaviours that fill individuals’ lives. For example, Nudelman, Kalish, and Shiloh (2019) used network analysis to examine which health behaviours were most central (i.e., to co-occur and influence one another). Health behaviours linked to basic physiological needs such as nutrition and sleep were found to be the most central supporting the idea that many health behaviours are inter-related. Indeed as Nudelman and Shiloh (2015, p. 1) note, ‘Notwithstanding the differences between [health behaviours], considering them as completely distinct from each other is also incompatible with evidence’. We believe there are two issues worth commenting on here: how similarity and differences between health behaviours might allow knowledge transfer; how studying multiple health behaviours simultaneously or attempting to change multiple behaviours simultaneously might provide new insights not possible when focusing on single health behaviours.

In relation to the first point there have been a number of attempts to consider distinctions among health behaviours or even taxonomies of such behaviours. The classic distinction between types of health behaviours focuses on health-promoting versus health-risk or impeding behaviours. The former are associated with decreased mortality and morbidity, while the latter are associated with increased mortality and morbidity. Rothman and Salovey (1997) expanded on this distinction to create three types of health behaviour: preventive (those which aim to prevent onset of ill-health), detective (those which aim to detect potential problems) and curative (those which aim to cure or treat a health problem). Another distinction is based on distinguishing among health behaviours on the basis of frequency of occurrence (frequent versus infrequent) with the suggestion that more frequent health behaviours may be more habitual in nature. (McEachan, Lawton, and Conner (2010) used an empirical approach to classify health behaviours along three key dimensions: ‘easy immediate pay-offs’ vs ‘effortful long-term pay-offs’; ‘private un-problematic vs. public and problematic’; ‘important routines vs. unimportant one-offs’. Finally, a recent attempt to produce an interdisciplinary taxonomy of behaviour (including health behaviours) included 250 behaviours classified into nine domains: engaging in learning and applying knowledge; communicating; moving/exercising; engaging in self-care behaviour; engaging in domestic life activities; engaging in interpersonal interactions and relationships; engaging in behaviour related to major life areas; engaging in behaviour related to community, social, and civic life; engaging in mood/state changing activities and behaviour (Larsen et al., 2019).

The extent to which these different ways of classifying health behaviours can help us generalise knowledge from one health behaviour to another mostly remains relatively unexplored. One exception we are aware of in relation to determinants of health behaviours is the meta-analysis of the Reasoned Action Approach (RAA) by McEachan et al. (2016). Here differences between health-protection and health-risk behaviours in relation to RAA determinants were explicitly explored. Experiential attitudes and descriptive norms emerged as significantly stronger predictors of health-risk compared to health-protection behaviours. Other ways of classifying health behaviours might be usefully employed in other meta-analyses of determinants of health behaviours to explore similarities and differences. This can help inform thinking on the role of different dimensions in relation to generating generalisable knowledge about determinants of health behaviours.

A similar approach might prove useful in relation to examining the effectiveness of individual behaviour change techniques (BCTs; Michie et al., 2013) across health behaviours classified according to the above dimensions (Epton, Currie, & Armitage, 2017; Harkin et al., 2016). For example, showing that goal planning is an effective way to change a range of health protection behaviours might be informative in developing a viable intervention to target a new health protection behaviour. This would be a different approach to many meta-analyses of BCTs that have focused on looking at which individual or combination of BCTs are effective for changing a specific behaviour (e.g., Michie, Abraham, Whittington, McAteer, & Gupta, 2009; Prestwich et al., 2016).

In relation to the second point, studying multiple health behaviours simultaneously, it is notable that a number of recent studies have sought to predict (e.g., Conner, McEachan, Lawton, & Gardner, 2017) or change (Wilding, Conner, Prestwich, Lawton, & Sheeran, 2019) multiple behaviours simultaneously. Such predictive studies have the advantage of exploring the multiple behaviours that individuals are engaged in at any one time, taking advantage of multilevel analyses, and assessing relationships within individuals rather than the usual focus of examining such relationships across individuals. For example, while the typical study of individual health behaviours assesses whether the person with the strongest intention will be the most likely to engage in the behaviour, a multiple behaviour study focuses on whether a behaviour an individual has the strongest intention to perform is the one he/she is most likely to engage in. Multi-behaviour studies have been used in the last few years to explore a number of moderators of the intention-health behaviour relationship such as goal priority (Conner et al., 2016), degree of reasoned action (Sheeran & Conner, 2019), self-identity (Carfora, Caso, Sparks, & Conner, 2017), socio-economic status (Schüz, Brick, Wilding, & Conner, 2019) and intelligence (Sniehotta, Presseau, Allan, & Araujo-Soares, 2016). More pertinent to the issue of identifying whether health behaviours have common and unique determinants is a study of the RAA by Conner et al. (2017). This showed that within the same sample of individuals, intentions, capacity, autonomy, experiential attitude and descriptive norms (from the RAA) were significantly stronger predictors of engaging in risk compared to protection health behaviours, that past behaviour was a significantly stronger predictor of protection compared to risk health behaviours, while there were no differences for instrumental attitude and injunctive norms. Such information can guide whether knowledge can generalise from one health behaviour to another, and potentially inform studies of the determinants of novel protection and risk behaviours.

Work targeting multiple health behaviours with similar interventions (i.e., based on the same or similar BCTs) could also provide valuable information about the similarities and differences between health behaviours. Multi-behaviour interventions have been suggested as a practical way to improve overall health outcomes, although it is not clear that such interventions necessarily use the same BCTs in relation to each health behaviour targeted and are equally effective for all health behaviours. For example, Wilding et al. (2019) used the question-behaviour effect to change multiple health behaviours in the same sample of individuals but reported that the intervention was more effective in increasing health protection behaviours than in reducing health risk behaviours. A recent review of multi-behaviour interventions in the domains of smoking, diet and physical activity (Wilson et al., 2015), concluded that interventions targeting a moderate number of recommendations (two to four) produced the highest level of change. The review also concluded that interventions targeting change in a single behaviour may be more appropriate in shorter interventions. In contrast, in their review of online health behaviour interventions, Webb, Joseph, Yardley, and Michie (2010) reported that interventions that targeted multiple health behaviours had smaller effects on health behaviour than those targeting single health behaviours. The focus on multiple health behaviours may dilute intervention effects on individual health behaviours. Gaining a better understanding of the value of individual versus multi-behaviour health interventions is a priority for research in this area, particularly where the effectiveness of the same BCT across more than one health behaviour in the same population is assessed.

Conclusions

The study of health behaviour represents an important area of health psychology with the potential to make an important contribution to efforts to improving health. Detailed study of individual health behaviours, such as those included in this special issue, can provide important insights related to unique determinants and effective ways to change that behaviour. Nevertheless such work can be usefully complemented by work on multiple health behaviours that might help inform us about whether we can generalise knowledge from one particular health behaviour to another both in terms of the predictors of, and interventions to change, health behaviour.

Mark Conner
University of Leeds
Paul Norman
University of Sheffield