Transdiagnostic processes refer to psychological mechanisms understood to precipitate and perpetuate psychopathology across diagnostic categories (Iverach, Menzies, & Menzies, Reference Iverach, Menzies and Menzies2014). Identifying transdiagnostic processes has been a focus of empirical research in recent years following findings that indicate limitations in the effectiveness of traditional disorder-specific interventions (see Craske & Barlow, Reference Craske, Barlow and Barlow2014; Ramsawh, Raffe, Edelen, Rende, & Keller, Reference Ramsawh, Raffa, Edelen, Rende and Keller2009), particularly in addressing comorbid presentations (McManus, Shafran, & Cooper, Reference McManus, Shafran and Cooper2010). It has been argued that transdiagnostic processes provide a parsimonious explanation for the generally high comorbidity rates between mental disorders (Egan, Wade, & Shafran, Reference Egan, Wade and Shafran2011) and that the persistence of symptomatology after disorder-specific treatment can be attributed to transdiagnostic processes that are not addressed in disorder-specific treatments (Brown, Antony, & Barlow, Reference Brown, Antony and Barlow1995; Egan et al., Reference Egan, Wade and Shafran2011; Iverach et al., Reference Iverach, Menzies and Menzies2014). In fact, it has been suggested that identification of transdiagnostic constructs will lead to the development and application of more effective psychological interventions that can address comorbid mental disorders concurrently and prevent the development of further diagnoses post-treatment, compared to disorder-specific interventions (Iverach et al., Reference Iverach, Menzies and Menzies2014). Identifying transdiagnostic processes present in eating disorders in order to improve treatment outcomes is considered particularly important, as the eating disorders are highly comorbid with depressive disorders, anxiety disorders, and personality disorders (Kaye et al., Reference Kaye, Bulik, Thornton, Barbarich and Masters2004; O'Brien & Vincent, Reference O'Brien and Vincent2003; Rosenvinge, Martinussen, & Østensen, Reference Rosenvinge, Martinussen and Østensen2000), and treatment follow-up studies indicate that the majority of patients achieve partial remission at best (Helverskov et al., Reference Helverskov, Clausen, Mors, Frydenberg, Thomsen and Rokkedal2010; Herzog et al., Reference Herzog, Dorer, Keel, Selwyn, Ekeblad, Flores and Keller1999).
Two transdiagnostic processes that have been previously identified are perfectionism (Egan et al., Reference Egan, Wade and Shafran2011) and perceived control (Gallagher, Naragon-Gainey, & Brown, Reference Gallagher, Naragon-Gainey and Brown2014). Cognitive behaviour therapy (CBT) interventions have been developed to target the transdiagnostic processes of perceived control and perfectionism, and these have been shown to lead to recovery from anxiety disorders in outpatient samples (Gallagher et al., Reference Gallagher, Naragon-Gainey and Brown2014) and to significant concurrent decreases in anxiety, depression, and obsessionality in non-clinical samples scoring highly on measures of perfectionism (Pleva & Wade, Reference Pleva and Wade2007). Moreover, a randomised controlled trial conducted by Steele and Wade (Reference Steele and Wade2008) that compared standard, disorder-specific CBT for bulimia nervosa to a transdiagnostic CBT-based intervention for perfectionism in the treatment of bulimia nervosa found that the two interventions yielded similar improvements in symptoms of bulimia nervosa. However, CBT for perfectionism led to greater reductions in additional symptoms of anxiety and depression compared to the disorder-specific intervention (Steele & Wade, Reference Steele and Wade2008). Findings such as these suggest that identification of transdiagnostic processes to inform the development of transdiagnostic treatment approaches may increase the efficacy, efficiency, cost-effectiveness, and generalisability of treatments for mental disorders compared to traditional diagnosis-specific interventions (Dozois, Seeds, & Collins, Reference Dozois, Seeds and Collins2009; Egan et al., Reference Egan, Wade and Shafran2011).
Death Anxiety as a Transdiagnostic Process
Death anxiety has recently been proposed as a potential transdiagnostic process (Iverach et al., Reference Iverach, Menzies and Menzies2014). Death anxiety, sometimes referred to as a ‘dread’ or ‘fear’ of death, is believed to be a universal characteristic that is important for self-preservation but which, when ineffectively managed, may become paralysing (Becker, Reference Becker2014). Terror management theory provides a theoretical account of death anxiety, arguing that much of human behaviour is designed to buffer potential anxiety cued by a sense of mortality, including behaviours that do not appear to be associated with mortality (Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007). For example, it is argued that death anxiety prompts individuals to create meaning and invest in cultural beliefs in order to foster a sense of purpose, self-value, and symbolic immortality to cope with the anxiety and powerlessness evoked by dread of mortality (Greenberg et al., Reference Greenberg, Solomon, Pyszczynski, Rosenblatt, Burling, Lyon and Pinel1992; Hayes, Schimel, Arndt, & Faucher, Reference Hayes, Schimel, Arndt and Faucher2010; Iverach et al., Reference Iverach, Menzies and Menzies2014; Pyszczynski, Greenberg, & Solomon, Reference Pyszczynski, Greenberg and Solomon1999; Routledge, Reference Routledge2012; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007). Indeed, fear of death has been argued to motivate the creation of symbolic language, art and music as a means of transcending the human body (Shaver & Mikulincer, Reference Shaver and Mikulincer2012).
In support of this hypothesis, studies in which participants are primed with mortality cues (questions about the individual's eventual death) or primed with neutral or aversive cues (questions about watching television or dental pain) have shown that people cued with mortality primes uniquely respond by latching onto cultural beliefs and personal worldviews (Solomon, Greenberg, & Pyszczynski, Reference Solomon, Greenberg, Pyszczynski, Greenberg, Koole and Pyszczynski2004). For example, McGregor et al. (Reference McGregor, Lieberman, Greenberg, Solomon, Arndt, Simon and Pyszczynski1998) asked students to write about their political beliefs and then gave them a bogus paragraph allegedly written by a fellow participant either supporting or attacking their political worldview. The participants were then instructed to allocate their fellow participant, who was described as disliking spicy food, an amount of hot sauce. Mortality-primed participants allocated twice as much hot sauce to perceived writers of worldview-inconsistent paragraphs compared to worldview-consistent writers, while non-mortality-primed participants allocated roughly equal amounts. Thus, mortality salience was found to increase aggression towards those who threatened the participants’ worldview.
Participants exposed to mortality cues have also been found to endorse more severe penalties to prostitutes (Rosenblatt, Greenberg, Solomon, Pyszczynski, & Lyon, Reference Rosenblatt, Greenberg, Solomon, Pyszczynski and Lyon1989); to take longer and self-report more discomfort when instructed to act disrespectfully towards culturally relevant items (e.g., using a crucifix as a hammer or sifting dye through an American flag; Greenberg, Porteus, Simon, Pyszczynski, & Solomon, Reference Greenberg, Porteus, Simon, Pyszczynski and Solomon1995); to donate more money to charity, especially those that benefit a group with which they identify (Jonas, Greenberg, & Frey, Reference Jonas, Greenberg and Frey2003); to sit closer to same-race confederates and further away from different-race confederates (Ochsmann & Mathy, Reference Ochsmann and Mathy1994); and to increase their driving speed on a driving simulator when driving ability is seen as valuable (Taubman Ben-Ari, Florian, & Mikulincer, Reference Taubman Ben-Ari, Florian and Mikulincer1999) Taken together, this evidence has been interpreted to mean that when primed with mortality cues, people invest in their cultural beliefs to protect themselves from death anxiety and experience greater distress when defying sociocultural norms (Greenberg et al., Reference Greenberg, Porteus, Simon, Pyszczynski and Solomon1995), endorse harsher punishments for perpetrators of sociocultural violations (Rosenblatt et al., Reference Rosenblatt, Greenberg, Solomon, Pyszczynski and Lyon1989), act more violently towards dissimilar others (McGregor et al., Reference McGregor, Lieberman, Greenberg, Solomon, Arndt, Simon and Pyszczynski1998), more favourably towards similar others (Jonas et al., Reference Jonas, Greenberg and Frey2003; McGregor et al., Reference McGregor, Lieberman, Greenberg, Solomon, Arndt, Simon and Pyszczynski1998; Ochsmann & Mathy, Reference Ochsmann and Mathy1994), and attempt to meet personally relevant cultural standards, even when doing so can paradoxically endanger their life (Taubman Ben-Ari et al., Reference Taubman Ben-Ari, Florian and Mikulincer1999).
Death Anxiety and Psychopathology
Following from the view that much of human behaviour is designed to protect individuals from a fear of death, death anxiety has been proposed as the ‘worm at the core’ of psychopathology, or the fundamental transdiagnostic process driving mental disorders (Arndt, Routledge, Cox, & Goldenberg, Reference Arndt, Routledge, Cox and Goldenberg2005). Indeed, it has been argued that symptoms of mental illness result from and are exacerbated by insufficiently buffered death anxiety or maladaptive coping strategies to deal with death anxiety (Iverach et al, Reference Iverach, Menzies and Menzies2014; Maxfield, John, & Pyszczynski, Reference Maxfield, John and Pyszczynski2014; Strachan, Pyszczynski, Greenberg, & Solomon, Reference Strachan, Pyszczynski, Greenberg, Solomon and Snyder2001; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007). For example, dread of death has been defined as a core feature of somatic symptom and related disorders, with patients frequently endorsing fears of bodily failure, pain, separation, and loss of control and power (Noyes, Stuart, Longley, Langbehn, & Happel, Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Starcevic, Reference Starcevic2005). Death anxiety correlates highly with hypochondriasis (Noyes et al., Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Starcevic, Reference Starcevic2005), and it has been found that people with hypochondriasis disclose greater anxiety around death than healthy controls (Kellner, Abbott, Winslow, & Pathak, Reference Kellner, Abbott, Winslow and Pathak1987).
Similarly, death anxiety appears to have an important role in the anxiety disorders and in obsessive-compulsive disorder. Dread of death may be related to separation anxiety disorder, with separation anxiety disorder frequently involving persistent worry about losing a loved one, including through death (American Psychiatric Association, 2013). Death anxiety may also be related to agoraphobia, with a greater number of hypochondriacal concerns and a high rate of death-related catastrophe fears present in those with agoraphobia (Foa, Steketee, & Young, Reference Foa, Steketee and Young1984). Many specific phobias and obsessive-compulsive subtypes focus on injury and death, with common phobias including heights, spiders, snakes, and blood (Iverach et al., Reference Iverach, Menzies and Menzies2014); compulsive hand-washers frequently citing chronic and fatal illnesses such as HIV/AIDS as the driving force behind their anxiety and compulsive behaviours (St Clare, Menzies, & Jones, Reference St Clare, Menzies and Jones2008); and compulsive checkers often claiming they check stoves and power points to avoid fire and death of the self or loved ones (Vaccaro, Jones, Menzies, & St Clare, Reference Vaccaro, Jones, Menzies and St Clare2010). Moreover, in Strachan et al.’s (Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007) series of experiments, priming mortality was found to increase phobic reactions to spider-related stimuli in those with spider phobia but not in those unafraid of spiders; and to increase compulsive hand-washing, measured by the amount of time spent washing hands, and the amount of soap and paper towel used, in compulsive hand-washers. In addition, priming mortality resulted in a decreased amount of time socialising compared to priming other aversive content (e.g., dental pain), particularly in those with greater social anxiety (Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007). As experimentally increasing the accessibility of death-related thoughts temporally corresponds with increases in symptoms, dread of death seems to have a direct effect on mental disorders. Similarly, conceptualisations of panic disorder frequently attribute the precipitation and maintenance of panic attacks to catastrophic misappraisals of bodily sensations as ‘fatal’, with patients commonly reporting they think they are having a heart attack or dying on detection of a physiological change, and evidence suggesting they are hypervigilant to terms that signal physical peril, like ‘pain’ and ‘disease’ (Craske & Barlow, Reference Craske, Barlow and Barlow2014; Hope, Rapee, Heimberg, & Dombeck, Reference Hope, Rapee, Heimberg and Dombeck1990). Moreover, reductions in these catastrophic misappraisals temporally predicts a decline in symptom severity, frequency of panic, avoidance behaviour, and distress in patients with panic disorder (Teachman, Marker, & Clerkin, Reference Teachman, Marker and Clerkin2010). This indicates the potency of fear of death in influencing panic.
Outside of somatic concerns, the anxiety disorders and obsessive-compulsive disorder, death anxiety appears to be insufficiently buffered in those with trauma and stressor-related disorders. A study by Kesebir, Luszczynska, Pyszczynski, and Benight (Reference Kesebir, Luszczynska, Pyszczynski and Benight2011) investigating survivors of domestic violence in Poland found that while individuals with mild or minimal trauma symptoms tend to exhibit behaviours that reflect distal defences to manage their death anxiety, such as increased worldview defence through harsher judgments of perceived moral transgressions, individuals with more severe post-traumatic stress disorder fail to exhibit the typical increased worldview defence in response to mortality priming. Furthermore, in a study of people exposed to the Ivory Coast civil war, those with a greater number of trauma symptoms reported elevated death-related thoughts in response to mortality cues compared to those with fewer symptoms (Chatard et al., Reference Chatard, Pyszczynski, Arndt, Selimbegović, Konan and Van der Linden2012). Findings such as these suggest that those with post-traumatic stress disorder and other trauma symptoms have difficulty managing death anxiety.
Similarly, those with depressive disorders appear to struggle with dread of death. A study of 135 psychiatric outpatients diagnosed with depressive disorders found death anxiety to be positively correlated with depression severity (Ongider & Eyuboglu, Reference Ongider and Eyuboglu2013). Furthermore, two experiments manipulating mortality salience found that university students who were depressed responded to mortality cues with more defence of their culturally derived values than university students who were not depressed (Simon et al., Reference Simon, Greenberg, Harmon-Jones, Solomon, Pyszczynski, Arndt and Abend1996). This implies that greater mortality concerns may be present in people who are depressed (Iverach et al., Reference Iverach, Menzies and Menzies2014).
Importantly, preliminary treatment studies examining the effects of cognitive behavioural therapy for death anxiety have found that in people with hypochondriasis, reducing death anxiety has been associated with a decrease in hypochondriasis, increased cultural worldview investment and self-esteem, and an improvement in general psychological wellbeing in the form of greater self-reported life satisfaction, the development of life goals, and attempts to live a healthier lifestyle (Furer & Walker, Reference Furer and Walker2008; Hiebert, Furer, McPhail, & Walker, Reference Hiebert, Furer, McPhail and Walker2005). Similarly, Menzies, Menzies, and Iverach (Reference Menzies, Menzies and Iverach2015) highlight two case reports of patients with current obsessive-compulsive disorder and a long history of other psychiatric conditions. Through targeting an underlying fear of death in treatment, these patients experienced a normalisation in mood and significant improvement in their obsessions and compulsions. If treating death anxiety can improve psychopathology and indicators of mental wellbeing, this supports the notion that dread of death is causally related to mental illness.
The findings reviewed above implicate death anxiety as an important transdiagnostic process in somatic disorders (Kellner et al., Reference Kellner, Abbott, Winslow and Pathak1987; Noyes et al., Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Starcevic, Reference Starcevic2005), anxiety disorders (Craske & Barlow, Reference Craske, Barlow and Barlow2014; Foa et al., Reference Foa, Steketee and Young1984; Hope et al., Reference Hope, Rapee, Heimberg and Dombeck1990; Iverach et al., Reference Iverach, Menzies and Menzies2014; St Clare et al., Reference St Clare, Menzies and Jones2008; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007; Teachman et al., Reference Teachman, Marker and Clerkin2010; Vaccaro et al., Reference Vaccaro, Jones, Menzies and St Clare2010), trauma and stressor-related disorders (Chatard et al., Reference Chatard, Pyszczynski, Arndt, Selimbegović, Konan and Van der Linden2012; Kesebir et al., Reference Kesebir, Luszczynska, Pyszczynski and Benight2011), and depressive disorders (Iverach et al., Reference Iverach, Menzies and Menzies2014; Ongider & Eyuboglu, Reference Ongider and Eyuboglu2013; Simon et al., Reference Simon, Greenberg, Harmon-Jones, Solomon, Pyszczynski, Arndt and Abend1996). However, little research into the relationship between death anxiety and eating disorder symptomology has been conducted.
Death Anxiety in the Eating Disorders
Collectively, the eating disorders, specifically anorexia nervosa, bulimia nervosa, and binge-eating disorder, are characterised by persisting disturbances in eating behaviours that culminate in impairments in physical and psychosocial functioning (American Psychiatric Association, 2013). Notably, many people exhibit dysfunctional eating attitudes and behaviours, or eating disorder symptoms, without meeting full clinical criteria for an eating disorder diagnosis (Striegel-Moore et al., Reference Striegel‐Moore, Rosselli, Perrin, DeBar, Wilson, May and Kraemer2009). Although in the past there have been references to death anxiety and preoccupation with death as potential aetiological and maintaining factors in anorexia nervosa (Binswanger, Reference Binswanger, May, Angel, Ellenberger, Mendel and Lyons1945; Langdon-Brown, Crookshank, Young, Gordon, & Bevan-Brown, Reference Langdon-Brown, Crookshank, Young, Gordon and Bevan-Brown1931), in general, there has been a paucity of research into the relationship between death anxiety and the eating disorders, or even general eating disorder symptomology. However, there has been some recent speculation into this link. In their review article, Alantar and Maner (Reference Alantar and Maner2008) proposed that a fear of gaining weight, as present in anorexia nervosa and bulimia nervosa or subclinical eating disorder symptoms, may protect people and serve as a distraction from a fear of death. They additionally posit that greater struggles with dread of death in youth may be a risk factor for eating psychopathology. Meanwhile, in Farber, Jackson, Tabin, and Bachar's (Reference Farber, Jackson, Tabin and Bachar2007) article, which presents a case of a patient with chronic anorexia nervosa and a case of a patient with bulimia nervosa, the authors argue that people with eating disorders frequently have a preoccupation with death and annihilation, and that in the instance of treatment-refractory patients, the role of death anxiety in their individual presentations may need to be addressed.
While few studies have directly examined the relationship between death anxiety and eating disorder symptomology, there is some empirical evidence to suggest dread of death may underlie the omnipresent overvaluation of weight and shape and the associated cognition and behaviours in disordered eating (Giles, Reference Giles1995; Goldenberg, Arndt, Hart, & Brown, Reference Goldenberg, Arndt, Hart and Brown2005). Goldenberg et al. (Reference Goldenberg, Arndt, Hart and Brown2005) conducted three experiments with non-clinical samples and found that women who were primed for mortality salience perceived themselves as further away from their ideal thinness, restricting their consumption of high-calorie nutritious foods more than women primed for dental pain. Further, Giles (Reference Giles1995) examined 31 women who met criteria for anorexia nervosa and who had not experienced any purging episodes in the past three months, and 31 control participants matched for age, gender, ethnicity, and socioeconomic status. Those with anorexia nervosa had significantly greater death anxiety towards themselves and others compared to controls. This supports a relationship between eating disorder symptomology and death anxiety.
It may seem paradoxical to suggest death anxiety as a transdiagnostic process in the eating disorders, given that anorexia nervosa has the highest mortality rate of all mental disorders (Farber et al., Reference Farber, Jackson, Tabin and Bachar2007), and that people with eating disorders frequently deny concerns about their life-threatening behaviours (Farber, Reference Farber2000). However, it has been argued that those with eating disorders and those with subclinical eating disorder symptoms turn to body weight and shape for identity and meaning, to defend against feelings of powerlessness and an inability to control their environment (Bruch, Reference Bruch1978; Egan et al., Reference Egan, Wade, Shafran and Antony2014; Russell, Halasz, & Beumont, Reference Russell, Halasz and Beumont1990; Slade, Reference Slade1982). As death anxiety can instil feelings of powerlessness and lack of control, and individuals act in ways that provide identity and meaning to buffer an underlying dread of death (Iverach et al., Reference Iverach, Menzies and Menzies2014; Noyes et al., Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007; Yalom, Reference Yalom2008), death anxiety may play an important role in eating psychopathology. Circumstantially, the fact that low self-esteem is a core feature of multiple eating disorders (Fairburn, Cooper, & Shafran, Reference Fairburn, Cooper and Shafran2003) and elevated self-esteem can act as a buffer for death anxiety (Greenberg, Reference Greenberg, Shaver and Mikulincer2012; Greenberg et al., Reference Greenberg, Solomon, Pyszczynski, Rosenblatt, Burling, Lyon and Pinel1992) may suggest that individuals with disordered eating attitudes and behaviours may be particularly vulnerable to death anxiety mismanagement. Further, as thinness is culturally valued in Western societies, people may be existentially motivated to control their body weight and shape (Goldenberg et al., Reference Goldenberg, Arndt, Hart and Brown2005). As the physical body represents mortality, applying cultural standards of weight and shape to it may convert the body from a reminder of death to a protection against death anxiety (Goldenberg, Pyszczynski, Greenberg, & Solomon, Reference Goldenberg, Pyszczynski, Greenberg and Solomon2000). Indeed, two experiments by Goldenberg, McCoy, Pyszczynski, Greenberg, and Solomon (Reference Goldenberg, McCoy, Pyszczynski, Greenberg and Solomon2000) found that when mortality was made salient, those who derived self-worth from their appearance increased their identification with aspects of their physical bodies for which they believed they were successfully meeting cultural standards, and reported a greater desire for bodily activity, specifically the physical aspects of sexual intercourse. This implies that the body can become a form of defence against dread of death when sociocultural standards of beauty are met, and people, including those with disordered eating, may be motivated to value and attempt to control their weight and shape in order to buffer an underlying fear of death. Though a connection between death anxiety and eating psychopathology may initially appear counter-intuitive, a preliminary study thoroughly examining whether death anxiety as a transdiagnostic process extends to eating disorder symptomology is warranted.
Death Anxiety, Perfectionism and Eating Disorders
Notably, prior transdiagnostic processes investigated in the context of disordered eating include low self-esteem, interpersonal difficulties, mood intolerance, and perfectionism (Fairburn et al., Reference Fairburn, Cooper and Shafran2003). Perfectionism, in particular, which is characterised by the continual striving to reach high standards and the over-dependence of self-worth on achievement (Egan et al., Reference Egan, Wade and Shafran2011; Egan, Wade, Shafran, & Antony, Reference Egan, Wade, Shafran and Antony2014; Riley & Shafran, Reference Riley and Shafran2005), has been well established as a transdiagnostic process in eating disorders. Indeed, perfectionism has prospectively been shown to predict the emergence of bulimic symptoms in female students (Steele, Corsini, & Wade, Reference Steele, Corsini and Wade2007), and retrospective reports suggest childhood perfectionism is linked to the development of eating disorders, including bulimia nervosa and anorexia nervosa (Fairburn, Cooper, Doll, & Welch, Reference Fairburn, Cooper, Doll and Welch1999; Fairburn, Welch, Doll, Davies, & O'Connor, Reference Fairburn, Welch, Doll, Davies and O'Connor1997; Southgate, Tchanturia, Collier, & Treasure, Reference Southgate, Tchanturia, Collier and Treasure2008). In fact, perfectionism has previously been argued to be the fundamental transdiagnostic process driving all psychopathology. Indeed, findings suggest that perfectionism is typically elevated in people with eating disorders, obsessive-compulsive disorder, and major depressive disorder compared to healthy controls, that greater pretreatment levels of perfectionism are indicative of a poorer prognosis, and that treating perfectionism yields large reductions in symptoms of anxiety, depression, obsessionality, and disordered eating (Egan et al., Reference Egan, Wade and Shafran2011; Egan et al., Reference Egan, Wade, Shafran and Antony2014). However, the emerging evidence for death anxiety as a fundamental transdiagnostic process in somatic disorders (Kellner et al., Reference Kellner, Abbott, Winslow and Pathak1987; Noyes et al., Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Starcevic, Reference Starcevic2005), anxiety disorders (Craske & Barlow, Reference Craske, Barlow and Barlow2014; Foa et al., Reference Foa, Steketee and Young1984; Hope et al., Reference Hope, Rapee, Heimberg and Dombeck1990; Iverach et al., Reference Iverach, Menzies and Menzies2014; St Clare et al., Reference St Clare, Menzies and Jones2008; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007; Teachman et al., Reference Teachman, Marker and Clerkin2010; Vaccaro et al., Reference Vaccaro, Jones, Menzies and St Clare2010), trauma and stressor-related disorders (Chatard et al., Reference Chatard, Pyszczynski, Arndt, Selimbegović, Konan and Van der Linden2012; Kesebir et al., Reference Kesebir, Luszczynska, Pyszczynski and Benight2011) and depressive disorders (Iverach et al., Reference Iverach, Menzies and Menzies2014; Ongider & Eyuboglu, Reference Ongider and Eyuboglu2013; Simon et al., Reference Simon, Greenberg, Harmon-Jones, Solomon, Pyszczynski, Arndt and Abend1996), and treatment follow-up studies showing that the majority of patients with eating disorders only achieve partial remission when perfectionism is addressed in treatment, with many patients developing other diagnoses (Helverskov et al., Reference Helverskov, Clausen, Mors, Frydenberg, Thomsen and Rokkedal2010; Steinhausen, Reference Steinhausen2009; Turner, Marshall, Stopa, & Waller, Reference Turner, Marshall, Stopa and Waller2015), challenges the argument that perfectionism is the fundamental transdiagnostic process for all mental disorders, including eating disorders.
An alternative view is that perfectionism may be related to death anxiety. Similar to terror management theory's proposition that individuals strive to create self-purpose and self-value to buffer death anxiety, perfectionistic individuals usually turn to an area of life that offers them a sense of accomplishment and control, such as weight and shape in the eating disordered patient (Egan et al., Reference Egan, Wade, Shafran and Antony2014). In this sense, there is conceptual overlap between death anxiety and perfectionism in fastidiously striving to create meaning. Interestingly, in a series of randomised, controlled trials, Kesebir (Reference Kesebir2014) found that humility, defined as a willingness to accept one's limits and low levels of self-focus — which may be understood as the opposite of the high standards and focus on personal achievement inherent in perfectionism — was inversely associated with death anxiety. Specifically, priming humility decreased death anxiety, and only individuals low in humility experienced increased death anxiety in response to mortality cues (Kesebir, Reference Kesebir2014).
In line with Strachan et al. (Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007), it is possible that perfectionism is an unreliable or maladaptive method of coping with death anxiety that causes psychopathology by inefficiently buffering dread of death. Indeed, perfectionism, when high standards and goals are met, can produce positive emotional consequences, such as bolstered self-esteem and feelings of self-efficacy, while failing to meet goals culminates in self-criticism, shame, guilt, and low mood (e.g., Bieling, Israeli, & Antony, Reference Bieling, Israeli and Antony2004; DiBartolo, Li, & Frost, Reference DiBartolo, Li and Frost2008; Rhéaume et al., Reference Rhéaume, Freeston, Ladouceur, Bouchard, Gallant, Talbot and Vallières2000). Based on the theoretical connection between death anxiety and perfectionism, it is possible that perfectionism is not a fundamental transdiagnostic process but merely a product of, or a mechanism to deal with, an underlying dread of death. If death anxiety is the fundamental transdiagnostic process across psychiatric disorders, death anxiety should be a better predictor of mental health outcomes such as disordered eating than perfectionism, a well-established transdiagnostic process in eating psychopathology.
Aim
The aim of the current study was to examine the relationship between death anxiety, perfectionism and eating disorder symptoms in a general population sample. Specifically, it was hypothesised that:
-
• death anxiety will be positively associated with eating disorder symptomology;
-
• death anxiety will be positively associated with perfectionism;
-
• death anxiety will be a better predictor of eating disorder symptomology than perfectionism.
As high self-esteem acts as a buffer against death anxiety and low self-esteem is related to more disordered eating, self-esteem was included in the study as a control variable and it was hypothesised that:
-
• self-esteem will be negatively associated with death anxiety;
-
• self-esteem will be negatively associated with eating disorder symptomology.
Method
Participants
Participants were recruited from an advertisement promoted on the social media website Facebook, as well as on the Centre for Eating and Dieting Disorders website. Participants were Australian residents of at least 18 years of age and were not required to have a diagnosed eating disorder. A total of 164 participants, 132 (80.5%) females and 32 (19.5%) males aged from 18 to 71 years (M = 33.55; SD = 15.45) completed the study. A total of 65 participants (39.6%) self-reported a current mental disorder diagnosis, and 25 participants (15.2%) self-reported a current eating disorder diagnosis.
Measures
Background questions
Participants were asked to report their age, gender, whether they had a current diagnosed mental disorder, and whether they had a current diagnosed eating disorder.
Eating Attitudes Test — 26 (EAT-26; Garner, Olmstead, Bohr, & Garfinkel, Reference Garner, Olmsted, Bohr and Garfinkel1982)
The EAT-26 is a shortened 26-item version of the 40-item EAT, which was designed to screen for eating disorder symptoms. For each item, participants rate their agreement on a 6-point scale ranging from always to never. Excluding the final item, which is reverse-scored, each always item receives a score of 3; each usually item yields a score of 2; each often item is scored as a 1; and each item rated sometimes, rarely or never is scored as a 0, so that total scores range from 0 to 78. Higher scores on the EAT-26 indicate greater eating disturbance, with a score of 20 or above indicating need for referral for further assessment. The EAT-26 is highly correlated with the EAT (r = .98) and has good test–retest reliability (Carter & Moss, Reference Carter and Moss1984), adequate internal consistency (Cronbach's alpha = .86; Gleaves, Pearson, Ambwani, & Morey, Reference Gleaves, Pearson, Ambwani and Morey2014), and good convergent validity (Doninger, Enders, & Burnett, Reference Doninger, Enders and Burnett2005). The EAT-26 additionally has good sensitivity and adequate specificity for eating disorders (Garner et al., Reference Garner, Olmsted, Bohr and Garfinkel1982).
Rosenberg Self-Esteem Scale (RSES; Rosenberg, Reference Rosenberg1965)
The RSES consists of 10 statements that measure self-esteem using a 4-point Likert scale, ranging from strongly agree to strongly disagree. Scores for each positively keyed item range from 3 (strongly agree) to 0 (strongly disagree), with total scores on the RSES ranging from 0 to 30. Higher scores on the RSES indicate higher self-esteem, or a greater perception of self-worth, with healthy self-esteem indicated by a score between 15 and 25. The RSES is a simple, widely used measure, with recognised face validity (Sinclair et al., Reference Sinclair, Blais, Gansler, Sandberg, Bistis and LoCicero2010), structural and predictive validity, internal consistency and test–retest reliability (Schmitt & Allik, Reference Schmitt and Allik2005; Torrey, Mueser, McHugo, & Drake, 2000).
Almost Perfect Scale — Revised (APS-R; Slaney, Rice, Mobley, Trippi, & Ashby, Reference Slaney, Rice, Mobley, Trippi and Ashby2001)
The APS-R is a 23-item measure of perfectionism, consisting of three factors (High Standards, Order, and Discrepancy). Items are scored on a 7-point Likert scale that ranges from 1 (strongly disagree) to 7 (strongly agree). The APS-R possesses good construct validity, and adequate overall internal consistency, with Cronbach's alpha for each subscale ranging from .82 to .92 (Slaney et al., Reference Slaney, Rice, Mobley, Trippi and Ashby2001). Higher scores overall indicate greater perfectionism.
Death Anxiety Scale (DAS; Templer, Reference Templer1970)
The DAS is a forced choice, 15-item measure of death anxiety, where each positively keyed item is scored 1 (true) or 0 (false), such that total scores on the DAS range from 0 to 15. Translated into multiple languages, it has been established as a reliable and valid measure of death anxiety, yielding good convergent validity, and adequate test–retest correlations and internal consistency estimates (Abdel-Khalek, Reference Abdel-Khalek2004; Royal & Elahi, Reference Royal and Elahi2011; Tomás-Sábado & Gomez-Benito, Reference Tomás-Sábado and Gómez-Benito2002). Higher scores on the DAS are indicative of greater death anxiety.
Procedure
The study was approved by the Navitas Professional Institute Human Research Ethics Committee. Participants completed all measures through an online survey that they accessed through a link from an advertisement. Participants read information about the study and those who consented to participate confirmed that they were current Australian residents over 18 years of age before being directed to the survey. Participants first entered their age and gender and then completed the EAT-26, the Rosenberg Self-Esteem Scale, the APS-R, and the DAS in that order. Participants were asked whether they had a currently diagnosed mental disorder, whether they had a currently diagnosed eating disorder, and were then debriefed. As an incentive for completing the study, participants were offered to go into a draw to receive one of four AU$100 JB Hi-Fi vouchers.
Data Analysis
All analyses were run using the IBM SPSS Statistics v23.0 Graduate Pack.
Results
The correlations between the DAS, EAT-26, RSES, age, gender, and self-reported mental and eating disorder diagnoses are presented in Table 1. As predicted, DAS scores were positively correlated with EAT-26 scores, self-reported mental and eating disorder diagnoses, and negatively correlated with RSES scores. Additionally, as expected, EAT-26 scores were highly correlated with self-reported eating disorder diagnoses and negatively correlated with RSES scores; APS-R scores were positively correlated with self-reported mental and eating disorder diagnoses and negatively correlated with RSES scores; and RSES scores were negatively correlated with self-reported diagnoses.
TABLE 1 Descriptive Statistics and Zero-Order Correlations for Gender, Age, EAT-26 Scores, RSES Scores, DAS Scores, APS-R Scores, Self-Reported Mental Disorder Diagnostic Status and Self-Reported Eating Disorder Diagnostic Status
Note: SRMD = self-reported mental disorder; SRED = self-reported eating disorder.
a Point biserial correlation between dichotomous and interval scales. Dichotomous variable coding: Gender (Female = 1, Male = 0); SRMD (Yes = 1, No = 0); SRED (Yes = 1, No = 0).
bNumber coded as ‘1’; percentage.
** Correlation is significant at the .01 level (two-tailed); *Correlation is significant at the .05 level (two-tailed)
To examine the contribution of death anxiety to eating disorder symptomology compared to perfectionism and self-esteem, a hierarchical multiple regression was conducted with EAT-26 scores as the dependent variable, and DAS, APS-R and RSES scores, as well as age, as predictors. Only variables that were correlated with EAT-26 were included as predictors (Tabachnick & Fidell, Reference Tabachnick and Fidell2014). The sample size (N = 164) was above the minimum recommended by Tabachnick and Fidell (Reference Tabachnick and Fidell2014) for multiple regression analysis with four predictor variables (minimum N = 108). The data were screened for outliers, multicollinearity, normality, linearity, and homoscedasticity of residuals, following the guidelines of Tabachnick and Fidell (Reference Tabachnick and Fidell2014). The EAT-26 violated the parametric assumption of homoscedasticity. Using a square-root transformation eliminated heteroscedasticity; however, this made the results of the regression difficult to interpret. As repeating the multiple regression analysis with the transformed EAT-26 variable yielded the same pattern of results as the analysis with the untransformed variable, and Tabachnick and Fidell (Reference Tabachnick and Fidell2014) note that heteroscedasticity does not invalidate a regression analysis but weakens uncovered relationships, the analysis with the untransformed variable is presented here for ease of interpretation.
Age was entered at Step 1, the RSES and APS-R were entered at Step 2, and the DAS was entered at Step 3. As shown in Table 2, at step 1, age accounted for 3.6% of the variance in EAT-26 scores and contributed significantly to the regression model, F(1, 162) = 6.01, p < .05. Introducing the RSES and APS-R at step 2 explained an additional 30.8% of the variance in EAT scores, and this change in R 2 was significant, F(2, 160) = 37.55, p < .001. Adding the DAS at step 3 explained an additional 7.1% of the variance in EAT-26, and this change in R 2 was significant, F(1, 159) = 19.15, p < .001. The overall multiple regression model was significant, with DAS, APS-R, and RSES scores and age explaining 41.4% of the variance in EAT-26 scores, F(4,159) = 28.12, p < .001. As shown in Table 2, in the final model only the DAS and RSES significantly predicted EAT-26 scores, uniquely explaining 7.1% and 7.9% of the variance in EAT-26 scores respectively. Controlling for the APS-R, RSES, and age for every extra point score on the DAS, EAT-26 scores increased by 1.01, b = 1.01, SEb = .23, β = .29, p < .001 [95% CI = .55, 1.46]. Similarly, controlling for the DAS, APS-R and age, for every extra point score on the RSES, EAT-26 scores decreased by 0.76, b = −.76, SEb = .16, β = −.37, p < .001 [95% CI = −1.08, −.44]. Thus, DAS was positively associated with EAT-26, and RSES was negatively associated with EAT-26.
TABLE 2 Sequential Regression Model for Variables Predicting Disordered Eating
Note: **Coefficients are significant at the 0.01 level (two-tailed); *Coefficients are significant at the 0.05 level (two-tailed).
Post-hoc, a simultaneous multiple regression was run with only age and APS-R as predictors to investigate whether APS-R was a significant predictor of EAT-26 when DAS and RSES were excluded from the regression model. Together, age and APS-R accounted for 22.0% of the variance in EAT-26, F(2, 161) = 22.77, p < .001. APS-R was a significant predictor of EAT-26, uniquely explaining 18.5% of the variance in EAT-26.
Discussion
This study investigated the relationship between death anxiety, disordered eating, and perfectionism. Based on the previous findings supporting death anxiety as a transdiagnostic construct that may be involved in eating disorder symptomology and that perfectionism may be an unreliable method of coping with death anxiety, it was predicted that death anxiety would be positively associated with eating disorder symptomology and perfectionism, and that death anxiety would be a better predictor of eating disorder symptomology than perfectionism. As expected, it was found that greater death anxiety was associated with greater eating disturbance, greater perfectionism, and self-reports of a current mental disorder or eating disorder. Moreover, after controlling for self-esteem, age and perfectionism, death anxiety remained a significant predictor of eating disorder symptomology. Therefore, the current findings are consistent with the view that death anxiety is a transdiagnostic process that underlies symptomology associated with disordered eating (Giles, Reference Giles1995; Goldenberg et al., Reference Goldenberg, Arndt, Hart and Brown2005; Iverach et al., Reference Iverach, Menzies and Menzies2014), as well as symptomology associated with somatic disorders (Kellner et al., Reference Kellner, Abbott, Winslow and Pathak1987; Noyes et al., Reference Noyes, Stuart, Longley, Langbehn and Happel2002; Starcevic, Reference Starcevic2005), anxiety disorders (Craske & Barlow, Reference Craske, Barlow and Barlow2014; Foa et al., Reference Foa, Steketee and Young1984; Hope et al., Reference Hope, Rapee, Heimberg and Dombeck1990; Iverach et al., Reference Iverach, Menzies and Menzies2014; St Clare et al., Reference St Clare, Menzies and Jones2008; Strachan et al., Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007; Teachman et al., Reference Teachman, Marker and Clerkin2010; Vaccaro et al., Reference Vaccaro, Jones, Menzies and St Clare2010), trauma and stressor-related disorders (Chatard et al., Reference Chatard, Pyszczynski, Arndt, Selimbegović, Konan and Van der Linden2012; Kesebir et al., Reference Kesebir, Luszczynska, Pyszczynski and Benight2011), and depressive disorders (Iverach et al., Reference Iverach, Menzies and Menzies2014; Ongider & Eyuboglu, Reference Ongider and Eyuboglu2013; Simon et al., Reference Simon, Greenberg, Harmon-Jones, Solomon, Pyszczynski, Arndt and Abend1996).
Perfectionism has been proposed as a fundamental transdiagnostic process in major depressive disorder, obsessive-compulsive disorder, and the eating disorders (Egan et al., Reference Egan, Wade and Shafran2011; Egan et al., Reference Egan, Wade, Shafran and Antony2014), and interventions specifically targeting perfectionism have been found to be effective in treating bulimia nervosa and co-occuring symptoms of anxiety and depression in a clinical sample (Steele & Wade, Reference Steele and Wade2008), and in treating anxiety, depression and obsessionality in a non-clinical sample (Pleva & Wade Reference Pleva and Wade2007). Similarly, self-esteem has been implicated as a transdiagnostic process in the eating disorders (Fairburn et al., Reference Fairburn, Cooper and Shafran2003). Importantly, while perfectionism was associated with disordered eating and with self-reports of a current mental or eating disorder in the present study, perfectionism was not a significant, independent predictor of eating disorder symptomology when self-esteem, age, and death anxiety were controlled. Thus, the current findings are inconsistent with the wealth of literature concerning perfectionism as a key transdiagnostic process particularly implicated in disordered eating. However, self-esteem was a significant, independent predictor of eating disorder symptomology when age, death anxiety, and perfectionism were controlled consistent with earlier research (e.g., Fairburn, Cooper, & Shafran, Reference Fairburn, Cooper, Shafran and Fairburn2008).
Since perfectionism was found to be significantly correlated with death anxiety and self-esteem in this study but was not a significant predictor of disordered eating with self-esteem and death anxiety in the model, this suggests that perfectionism was not uniquely accounting for variability in disordered eating. To our knowledge, no study to date has examined the unique contribution of perfectionism, self-esteem, and death anxiety in predicting eating disorder symptomology, and it is possible that previously reported findings concerning the relationship between effects of perfectionism on psychopathology (e.g., Egan et al., Reference Egan, Wade and Shafran2011; Egan et al., Reference Egan, Wade, Shafran and Antony2014; Southgate et al., Reference Southgate, Tchanturia, Collier and Treasure2008; Steele et al., Reference Steele, Corsini and Wade2007) are a result of the omission of death anxiety from these studies. If, in line with Strachan et al. (Reference Strachan, Schimel, Arndt, Williams, Solomon, Pyszczynski and Greenberg2007), perfectionism is a maladaptive method of coping with death anxiety through fastidiously striving to create meaning, and death anxiety is the fundamental transdiagnostic process underlying psychopathology (Arndt et al., Reference Arndt, Routledge, Cox and Goldenberg2005; Iverach et al., Reference Iverach, Menzies and Menzies2014; Maxfield et al., Reference Maxfield, John and Pyszczynski2014), then arguably when death anxiety is not directly assessed and included in analyses, perfectionism may be acting as a proxy measure of death anxiety. Similarly, the effectiveness of treatments aimed at perfectionism in reducing symptomology (Egan et al., Reference Egan, Wade and Shafran2011; Egan et al., Reference Egan, Wade, Shafran and Antony2014; Pleva & Wade, Reference Pleva and Wade2007; Steele & Wade, Reference Steele and Wade2008) may be due to these interventions reducing the use of ineffective methods to cope with death anxiety. Consistent with this interpretation, the findings of the post-hoc simultaneous multiple regression reported here indicated that perfectionism was a significant predictor of disordered eating when self-esteem and death anxiety were not included in the model, but that adding death anxiety and self-esteem eliminated the effects of perfectionism on disturbed eating.
The finding that self-esteem is a unique predictor of disturbed eating challenges the assumption that death anxiety is the key transdiagnostic process accounting for psychopathology (Arndt et al., Reference Arndt, Routledge, Cox and Goldenberg2005; Iverach et al., Reference Iverach, Menzies and Menzies2014), at least for disordered eating. Indeed, the findings of the present study provide support for death anxiety and self-esteem as related but independent predictors of eating disturbance. Self-esteem has been proposed as a transdiagnostic process for disordered eating (Fairburn et al., Reference Fairburn, Cooper, Shafran and Fairburn2008). Low self-esteem is related to increased eating disorder severity (Beren & Chrisler, Reference Beren and Chrisler1990), the presence of early-onset anorexia nervosa (Lask & Bryant-Waugh, Reference Lask and Bryant‐Waugh1992), bulimia nervosa (Cooper & Fairburn, Reference Cooper and Fairburn1993), binge-eating disorder (Mitchell & Mussell, Reference Mitchell and Mussell1995), obesity (de Zwaan et al., Reference de Zwaan, Mitchell, Seim, Specker, Pyle, Raymond and Crosby1994), abnormal eating attitudes (Fisher, Schneider, Pegler, & Napolitano, Reference Fisher, Schneider, Pegler and Napolitano1991) and concerns with body image (Foster, Waddan, & Vogt, Reference Foster, Wadden and Vogt1997). There is also evidence that self-esteem is important in depressive disorders (Orth, Robins, & Meier, Reference Orth, Robins and Meier2009; Sowislo, Orth, & Meier, Reference Sowislo, Orth and Meier2014) and in obsessive-compulsive disorder, social phobia, specific phobia, and overanxious disorder (Maldonado et al., Reference Maldonado, Huang, Chen, Kasen, Cohen and Chen2013). Low self-esteem is also associated with greater anxiety or affective comorbidity in people diagnosed with schizophrenia (Karatzias, Gumley, Power, & O'Grady, Reference Karatzias, Gumley, Power and O'Grady2007). Self-esteem, along with death anxiety, merits further investigation as a unique and fundamental transdiagnostic construct beyond the eating disorders.
Several limitations of this study must be recognised. First, the study was cross-sectional so that no causal inferences can be made. Second, while many participants (15.2%) self-reported eating disorder diagnoses, there was no independent verification of these diagnoses and the sample must be considered a non-clinical sample. A large, strictly clinical sample would strengthen the present findings. Third, as participants completed the questionnaires online, the identity of respondents could not be verified and it is possible that responses were not honest. However, there would be little motivation for an individual to complete the survey dishonestly, and it is also possible that the anonymity of the online methodology reduced the likelihood of response bias towards socially acceptable responses. Further research is also needed to clarify the status of perfectionism, self-esteem, and death anxiety as transdiagnostic processes accounting for psychopathology. If, as argued here, death anxiety and self-esteem largely account for the effects of perfectionism, then this implies that transdiagnostic interventions should focus on these variables rather than perfectionism. As implied by Iverach et al. (Reference Iverach, Menzies and Menzies2014), better targeted treatments have the potential to assist those who do not respond to standard treatments. Since death anxiety has only recently gained attention as a transdiagnostic process, it is still unclear how death anxiety would be best targeted in the treatment of eating disorders. Menzies et al. (Reference Menzies, Menzies and Iverach2015) speculate that several therapies that focus on increasing a personal sense of meaning and replacing maladaptive coping strategies for dealing with death anxiety could be used to address an underlying dread of death in people with mental disorders. Menzies et al. (Reference Menzies, Menzies and Iverach2015) specifically emphasise existential psychotherapy, which has recently been used to treat bipolar disorder (Goldner-Vukov, Moore, & Cupina, Reference Goldner-Vukov, Moore and Cupina2007), panic disorder (Randall, Reference Randall2001), and depression (Stalsett, Gude, Ronnestad, & Monsen, Reference Stålsett, Gude, Rønnestad and Monsen2012); existential-humanistic therapies, such as dignity therapy and meaning-centred therapy, which have been found to be moderately successful in alleviating death anxiety in terminally ill patients (Breitbart et al., Reference Breitbart, Rosenfeld, Pessin, Kaim, Funesti-Esch, Galietta and Brescia2000; Chochinov, Hack, Hassard, & Kristjanso, Reference Chochinov, Hack, Hassard and Kristjanson2004); and cognitive behavioural approaches drawing on elements of existential and existential-humanistic therapies to build a sense of agency and meaning, as well as helpful coping strategies, which have previously been found to improve mental wellbeing (Furer & Walker, Reference Furer and Walker2008; Hiebert et al., Reference Hiebert, Furer, McPhail and Walker2005). Further research into the most effective death anxiety interventions would be beneficial.
In sum, death anxiety has been recently proposed to be the ‘worm at the core’ of psychopathology. This study was the first to examine the role of death anxiety in eating disorder symptomology compared to the recognised transdiagnostic processes, perfectionism, and self-esteem. Replication of the current study, both with a large general population sample and clinical samples, should be undertaken to strengthen the present results, and extending the methodology to compare death anxiety, perfectionism, and self-esteem as predictors of treatment response may be warranted. Longitudinal, prospective studies investigating whether greater death anxiety and lower self-esteem predict the later development of eating or other mental disorders; and randomised controlled trials comparing the effectiveness of treatments directly addressing perfectionism versus those addressing death anxiety and those addressing core low self-esteem could be of particular value.
