The Role of Images on Illness Behaviour: Interdisciplinary Theory,Evidence,and Ideas

Cross-disciplinary understandings of the role of images within illness behaviours

Social and behavioural sciences have contributed significant theory to our understandings of illness behaviours. In the context of the current paper, we adopt a broad definition of illness behaviour which is drawn from the behavioural medicine domain: illness behaviours are considered to be the observable actions or reactions to actual or perceived physical, emotional or mental illness; these can include self-care or the use of healthcare services (Rau & Williams, 2013).

An examination of the contributions across the social and behavioural sciences literatures reveals four similarities. First, although differing terms are used, each highlights the importance of the patient perspective through an emphasis on illness beliefs (e.g., sociology), explanatory models (e.g., anthropology) or illness representations (e.g., psychology). Second, despite varying theoretical and philosophical assumptions and underpinnings, empirical research in each discipline generally uncovers beliefs that categorise across similar dimensions. For example, within medical anthropology, Kleinman’s Kleinman (1987) ‘explanatory model’ encompasses beliefs about diagnosis, cause, and appropriate treatments (Helman, 1998; Kleinman, 1987); while, within health psychology the concept of “illness representations” points to the parallel issues of illness identity, cause, consequences, timeline and cure/control (Leventhal et al., 1980; Leventhal & Nerenz, 1985). A third similarity is the consistent evidence across these disciplines, particularly across psychology and medical sociology, indicating that health beliefs may influence illness experience (Bishop & Converse, 1986; Lobban et al., 2003; Scambler & Hopkins, 1986), illness behaviours (Horne & Weinman, 2002; Ross et al., 2004) and ultimately health outcome (Barsky et al., 2002; Petrie et al., 2007). A final commonality is the cross-discipline emphasis on understanding illness beliefs and experience through an almost exclusive focus on the written or spoken word.

While health messages are almost always communicated through verbal or text-based means, a growing body of evidence is emerging which suggests that these traditional delivery systems may not be the most effective (e.g., Garcia-Retamero & Galestic, 2010; Williams & Cameron, 2009). For example, patient perceptions of risk were found to improve and be less susceptible to framing effects on health information when a visual aid was used (Garcia-Retamero & Galestic, 2010). However, a recent systematic review exploring the efficacy of video-based visual health interventions found variations in the effectiveness of the interventions, with video interventions tailored to the individual or specific condition being more effective than non-tailored video interventions (Tuong et al., 2014). These findings suggest that while images have been used in health promotion campaigns in the same manner as in commercial advertisements, this is often carried out with little or no theoretical consideration in their use and effectiveness.

The current paper argues that if illness behaviour is to be more fully understood, academics across the social and behavioural sciences must acknowledge and encompass the wider forms in which illness is experienced and communicated in contemporary society. We suggest that illness may be experienced and communicated through physical and mental representations (i.e., mental images), and that these may have the capacity to embody and influence both beliefs and emotional sequalae, and thus an array of important illness behaviours and outcomes.

This understanding is imperative to academics, health professionals and others engaged in intervention development in healthcare. In recent years there has been an increase in the use of visual communication at both the societal and population level (Williams & Cameron, 2009) and within healthcare interventions (Murray et al., 2001, 2016; Williams et al., 2012) to convey complex information, influence beliefs and emotions (Williams & Cameron, 2009) and, ultimately, elicit a change in health behaviour. However, at present, the evidence base applied to the development of visual interventions is scant (Williams & Cameron, 2009). This poses a risk to the quality of the interventions themselves and, more importantly, to the messages conveyed to and understandings of patients (Williams & Cameron, 2009).

The current paper therefore aims to review the social and behavioural sciences literature to explore what is known about the role of visual stimuli and images on illness behaviours. The main literatures included within this review are decision sciences, health and behavioural psychology, and medical sociology. Wider literatures and papers are also included but with a less detailed focus than the former.

The current paper will explore theoretical and empirical evidence across the social and behavioural sciences examining the importance of images in illness behaviours. It will discuss some of the dominant health behaviour models relevant to the area, and then examine possible ‘precognitive’ or ‘unconscious’ influences of images on judgement and decision making and illness behaviours. The role of image changes will be discussed in relation to behaviour change and non-change, and how these are determined depending upon the role of symptom interpretation and emotional activation. The impact that this has on health outcomes and health care will then be discussed. Prior to exploring the theoretical and empirical evidence, however, it is important to identify what we mean by “image” within the current paper. This will now be defined.

What do we mean by “image”?

For the purposes of this paper, “image” is defined in a broad sense to include any visually based or mental representation or embodiment. This definition reflects an essentially phenomenological stance that focuses on a principal “sense”. However, given ongoing research and debates within neuropsychology, and the evident role of a visual dimension within classic illness metaphors (Helman, 1985; Sontag, 1978, 2001), we take the “visual” to not only encompass physical sight, but also mental representations of images. We will primarily use ‘image’ in the remainder of the current paper to maintain consistency in the terminology used. However, where appropriate and better fitting with the papers or theories being discussed, we may use the term used in the original paper/theory (e.g., visual stimuli, mental representation).

Precognitive influences on illness behaviours

Precognitive influences on illness behaviours may be defined as those influences which we are not necessarily aware of as having an impact on our judgements, decisions and illness-related behaviours (see, for example, Kahneman, 2011, for a thorough overview of unconscious and conscious decision making). These would not be influences that are consciously recognised by a person and may better be described as instinctual, automatic, intuitive, unconscious or heuristic influences. While it may be argued that all human experiences and influences on our behaviours, and hence our illness behaviours, must be cognitive to some degree, these precognitive influences may be best considered as those influences that affect our actions and beliefs that we are not consciously aware of. Note that the use of the term ‘precognitive’ in the current paper is not the same as those studies which posit that precognition is a predictor of future outcomes; we instead use the term to mean unconscious influences on our decision making, as described in this section.

When considering the complexity of influences on illness behaviours, the influence is not merely on behaviour, but also on cognitions such as judgements and decisions. To best understand the role of precognitive influences on illness behaviours, we will first summarise recent theory on human judgement and decision making and will follow this with consideration for precognitive influences on image interpretation and illness behaviours.

Image changes as a prompt to symptom perception

The role of images in symptom interpretation: Embodied and influenced cognition

Although it is generally assumed that symptom interpretation cognitions are in verbal form, the theory in fact suggests that these are based on underlying abstract concepts. The nature of this form of abstraction is unknown, and potentially unknowable. However, it is possible that images can embody beliefs. For example, a patients’ drawing or even their mental represenattion of a broken arm may embody their beliefs in relation to its possible symptoms, cause, curability and general prognosis. There is recent empirical evidence that links aspects of patient’s drawing of their heart after a myocardial infarction to both subsequent behaviour and physical recovery (Broadbent et al., 2004, 2007). Images can elicit a real, emotional response both when viewed and when imagined or remembered. These emotive effects can become apparent with both concrete imagery (i.e., something “true to life” such as a spider or a house) and abstract imagery (e.g., a metaphorical or other representation of a concept or construct, such as an emotion). Evidence from cognitive psychology has suggested that image representation is dependent on retrieval from memory (Craik & Lockhart, 1972). In their ‘levels of processing’ framework, Craik and Lockhart (1972) posited that memory as a process, with fluidity across encoding and retrieval, constantly reinterpreting and adding meaning to past experiences through integrating new understandings to past experiences. The strength of a memory is considered to be increased when it has first been encoded strongly, and our image representations are strengthened when associated with either concrete imagery (Paivio, 1986) or high emotional arousal (Kensinger, 2009).

The role of images in directly impacting emotions, affecting illness behaviour

Images may also generate a direct, unconscious emotional response independent of any cognition. For example, disgust, which is characterised by a desire to retreat from the stimulus, may be produced by the single sensory experience of viewing an image (Woody & Teachman, 2000). An example of an application of the use of negative emotion associated with images and health behaviours is in the treatment of phobias. Images feature prominently in phobias and anxiety disorders. Presentation of a still image of the feared stimuli can produce both an affective (anxiety, panic) and a biological (increased heartbeat, perspiration) response, even when there is no actual (real) object present. Spontaneous imagery is a common feature of anxiety disorders, and when mental representation of images is compared to verbal descriptions of the threatening stimuli, the imagery can produce higher levels of anxiety and a greater emotional response (Hirsch et al., 2006).

As images alone are capable of producing powerful emotional, cognitive, and associated biological responses, it is unsurprising that they may influence behaviour. Kleinman et al. (1978) reported the case of a 60-year-old woman, admitted to hospital and diagnosed as having ‘water in her lungs’. The woman proceeded to act bizarrely, repeatedly vomiting and urinating in her bed. It later transpired that this woman was the wife and daughter of plumbers, and believed the inner workings of the body to be akin to a system of inter-connected pipes. Her explanatory model/illness representation thus indicated quite logically to her that frequent expulsion of fluids from the body would cure the problem of water in her lungs. In this example the influence of her visual perception of the body as a system of pipes on her illness behaviour is clear. This is a relatively old example, but one which is helpful in providing clarity over the varied and anatomically incorrect representations that people may hold.

While there is some evidence for improved knowledge and anatomical accuracy in patients’ images in illness representations (Broadbent et al., 2019), there remains idiosyncratic differences in patient’s drawings of their illnesses which may help to predict their health outcomes (Broadbent et al., 2019; Petrie & Weinman, 2012). For example, Petrie and Weinman (2012) describe the drawings made at three timepoints by two patients who had underwent heart bypass surgery; while both patients drew increasingly ‘healthy/healing’ pictures of hearts, one also focused also on the heart rhythm improving, which Petrie and Weinman (2012) highlight is an important indicator for recovery. In their systematic review of 101 papers published exploring patient drawings of illnesses, Broadbent et al. (2019) found evidence indicating that where patients had drawn larger images and more organ damage, outcomes and health perceptions were worse. Improvements to patient-practitioner understandings of the patient’s image representations of their illness may therefore be helpful in improving outcomes and targeting psychoeducation where required to help improve illness behaviours related to recovery (Petrie & Weinman, 2012).

Although there is evidence to support the use of images and imagery for behaviour change, there has been relatively little consistent attention in health research; with the majority of the research being confined almost exclusively to the investigation of children’s views concerning illness (Guillemin, 2004). However, the investigation of adults’ mental image representations of illness is an exciting emerging field of research, with advocates of its value scattered across several disciplines (e.g., Broadbent et al., 2004; Cross et al., 2006; Guillemin, 2004).

Due to the infancy of the work in this area, there are several important and fundamental questions which have not yet been addressed:

To what degree does the experience of illness encompass images, for whom, and during which illnesses?

It is also important to consider the theories and literature on images in illness behaviours discussed in relation to health care: how do images affect the experience of illness, the behaviours associated with it, and how might they affect health outcomes? The proceeding section will explore these issues.

Illness experience

An individual’s experience of illness symptoms can be influenced by their image representations associated with the symptoms. To illustrate this, consider the case of indigestion versus tapeworm. A person experiencing a mild but persistent stomach upset, eventually consults a doctor to be told that it is: a) mild indigestion; or b) a tapeworm - an infestation of the digestive tract with the tapeworm parasite. This is a ribbon shaped creature which can come from a cow, pig, sheep, dog, cat and other animals. Tapeworm eggs are passed in the stools of a person who is infected and spread through water or surfaces contaminated with faeces. Despite both conditions presenting identical symptoms, the images associated with the latter elicits entirely different cognitions and emotions than those that would be associated with the more common condition of indigestion. The two conditions are therefore experienced in very different ways, despite the physiological symptoms being similar.

Illness behaviour

Image representations can also affect an individual’s illness behaviours. For example, Williams et al. (2007) examined adherence to non-medicinal treatments in children with cystic fibrosis, and found that their visualisation of mucus building up in their lungs, rather than any verbalised beliefs, was the chief motivator for them to perform chest physiotherapy. Similarly, Kleinman et al. (1978)afore-mentioned case study of the woman who was admitted to hospital with ‘water in her lungs’ and who had attempted to cure this through repeated vomiting and urination clearly illustrates that visualisation can influence illness behaviours, as the patient’s image of the body as a system of interconnected pipes dictated her behaviour as she attempted to cure her illness (i.e., excessive expulsion of liquids).

Health outcomes

Health outcomes have also been shown to be influenced by patients’ image representations of their conditions. Broadbent et al. (2004) asked patients who had suffered a myocardial infarction (MI) to draw their hearts in a pre- and post-MI state. This study found that the patients’ drawings of their heart pre- and post- MI were not only accurate indicators of underlying beliefs about their heart attack, but also a better predictive marker of time taken to return to work than any clinical indicator. Those who perceived less damage to their heart recovered more quickly, regardless of the extent of actual damage.

The need to address patients’ images is also important due to the potential for inaccuracy. Inaccurate images nourish inaccurate and potentially damaging misunderstandings of illness. The breadth of inaccuracy in lay understandings of anatomy was demonstrated by Boyle (1970) in his comparison of doctors’ and patients’ location of various organs within a human silhouette. In his study, Boyle (1970) asked patients and doctors to identify within a picture of a human silhouette where the kidneys and heart were located. Surprisingly, the identified locations of these organs were widely varied; clearly indicating that we cannot assume that lay members of the population have a strong understanding and knowledge about their internal bodily functioning or even the positioning of major organs.

External images pertaining to illnesses and the body undoubtedly play a role in subjective visualisations of illness. In clinical settings patients observe images informally in the form of sketches during consultations, and formally through use of x-rays, ultrasounds, and endoscopic cameras. Visual language or metaphors used by health care professionals in interactions with patients may also have a significant impact (Harrow et al., 2008).

The health service in the UK readily and increasingly employs imagery of the body in its growing number of health promotion campaigns, but their use is largely atheoretical (Williams & Cameron, 2009), despite a wealth of theoretical models attempting to define and explain health and illness behaviours existing in the health, psychology and sociology literature (e.g., Theory of Planned Behaviour, Leventhal’s Self-Regulatory Model, Health Belief Model). This is in a context of increasing emphasis on shared decision-making, and a post-modern culture whose population’s everyday lives are increasingly dominated by the visual and who interpret the world and find meaning in more visual ways (Mirzoeff, 1999). Within this visually-focussed society, Government spending on health promotion campaigns involving visual media will almost certainly increase. A strong theoretical premise to justify additional spending and to maximise the effectiveness of health promotion campaigns is therefore crucial.

The backbone of verbal and textual health interventions is usually the spectrum of behavioural theory, but the applicability of these theories to image based interventions is unknown. The lack of image-appropriate theory is evident in the fact that the UK’s Department of Health consulted the public via a purpose-created website, over which graphic images would encourage smoking cessation if introduced onto cigarette packets. The consultation followed evidence from Canada indicating the potential effectiveness of pictorial warnings (Department of Health, 2006), though the intricacies of their impact – how, why, which features – have yet to be determined.

Notwithstanding their atheoretical nature, in the broader health context it has been established that image based interventions can be effective. Petrie et al. (2002) carried out a Randomised Control Trial of an intervention for MI patients, using drawings to explain pathophysiology and symptoms and addressing patients’ misconceptions. Compared to controls (who received standard MI educational materials), the intervention group reported increased preparedness to leave hospital, significantly fewer angina symptoms at three-month follow-up, and they returned to work significantly faster after their MI. To further illustrate, Shahab et al. (2007) carried out a Randomised Control Trial investigating an image based intervention to encourage smoking cessation. Smokers in the intervention group received an ultrasound photograph of their own artery showing atherosclerotic plaque alongside an image of a healthy artery, while controls received verbal information standard to the clinic’s normal practice. The intervention group showed increased perceived susceptibility to atherosclerosis, higher rates of engagement in smoking cessation behaviours, and reported increased intention to stop smoking when followed up at four weeks; although the latter effect was only reported by those with higher levels of self-efficacy. An earlier, similar study by Bovet et al. (2002) employed the same technique with a sample of 153 smokers and found an increased quit-rate at six-months among the group which received ultrasound photographs of their plaques, compared to controls. These findings demonstrate the potential for effective image based interventions in improving health behaviours.

To improve these interventions, inform future development, and target them appropriately, an underpinning theory is required. Of primary concern is the exposure of the complex inter-relationships between the following variables: socioeconomic status and gender, culture, image representations, external images, illness perceptions, cognitions, emotions, and behaviour.