Developmental antecedents of anxiety sensitivity

Abstract

Anxiety sensitivity (AS) has been implicated in the etiology of anxiety and depressive disorders. Nonetheless, little is known about the developmental antecedents of AS or about its role as a mediator of emotional distress. To investigate these issues, 249 university students completed a battery of self-report measures. Regression analyses and partial correlations were used to evaluate hypotheses. Analyses revealed that exposure to parental threatening, hostile, and rejecting behaviors (1) predicted overall AS and (2) appeared differentially related to AS factors. Additionally, AS acted as a mediator between parenting and both current emotional distress and history of emotional disorder symptoms. These data support the notion that parental behaviors hypothesized as etiologic in the development of emotional disorders may also be etiologic in the development of AS and that AS may serve as a mediator between childhood experience and emotional distress.

Introduction

The expectancy model of fear and anxiety posits that anxiety sensitivity (AS), that is, concern over anxiety sensations such as shakiness, nervousness, and lack of concentration, helps to explain fearful behavior in any situation during which such sensations might be encountered (Reiss, 1991; Reiss & McNally, 1985). Congruent with these implications, much empirical research has focused on relationships between AS and anxious distress (Taylor, 1999). Indeed, increased levels of AS have been found among persons with a wide range of anxiety disorder diagnoses, including panic disorder with and without agoraphobia, posttraumatic stress disorder, generalized anxiety disorder, obsessive–compulsive disorder, and social phobia, as well as among nonclinical persons with panic symptomatology (Brown & Cash, 1990; Lau, Calamari, & Waraczynski, 1996; Rabian, Peterson, Richters, & Jensen, 1993; Stewart, Knize, & Pihl, 1992; Taylor, Koch, & McNally, 1992). Moreover, several studies have found that AS prospectively predicts panic attacks and anxious symptomatology (Maller & Reiss, 1992; Schmidt et al., 1997, Schmidt et al., 1999). Thus, AS may play a role in the development of pathological anxiety.

Of late, research has also begun to focus on AS in forms of distress bearing some relationship with anxiety; most of this research has focused on AS in depression. Such studies have indicated that AS is a concomitant of major depression and depressive symptomatology (Otto, Pollack, Fava, Uccello, & Rosenbaum, 1995; Taylor, Koch, Woody, & McLean, 1996). Moreover, AS has also been shown to prospectively predict depressive symptomatology (Schmidt et al., 1997). Thus, AS may be important in the development of depression as well as anxiety.

Despite evidence that AS plays a role in both anxiety and depressive disorders, and the possibility that this role may be etiologic in nature, little is known regarding the developmental antecedents of this sensitivity. Indeed, only one study to our knowledge has investigated such antecedents. In that study, Watt, Stewart, and Cox (1998) (1) examined the ability of childhood operant and vicarious learning experiences to predict overall AS among university students and (2) compared the childhood operant and vicarious learning experiences of university students with high, moderate, and low levels of overall AS. They found that parental reinforcement of sick role behavior in response to both childhood anxiety symptoms and childhood cold symptoms, observation of parental uncontrolled behavior (e.g., due to being drunk or angry), and observation of parental sick role behavior in response to parents’ own anxiety symptoms predicted overall AS. Additionally, students with high AS retrospectively reported: (1) more parental reinforcement of sick role behavior in response to their childhood anxiety symptoms than moderate and low AS students, (2) more parental reinforcement of sick role behavior related to their childhood cold symptoms than moderate and low AS students, (3) more frequent observation of parental sick role behavior in response to parents’ own anxiety symptoms compared to low AS students, and (4) more frequent observation of parental uncontrolled behavior compared to moderate AS students. Thus, both operant and vicarious learning processes were implicated in the development of AS. Given such findings, further investigation of childhood experiences in the development of AS seems warranted.

In addition to traditional learning theory perspectives, one potentially useful framework for understanding development of AS is Bowlby, 1969/1982, Bowlby, 1973, Bowlby, 1980 attachment theory. Bowlby (1969/1982) suggested that an attachment relationship consisting of proximity-maintaining behaviors by the human infant and the behaviors of his or her primary caregivers typically forms during the first year of the infant’s life, with the goals of security and protection of the infant. He further suggested that from infancy forward, interactions within the context of these ongoing childhood attachment relationships contribute to the development of people’s internal working models about the world and how it operates. Moreover, Bowlby hypothesized that the quality of interactions within the context of these childhood attachment relationships could influence persons’ psychological health, presumably through these interactions’ effects on working models.

Congruent with the idea that maladaptive attachment interactions might lead to psychological distress, Bowlby, 1973, Bowlby, 1980 identified two sets of caregiver behaviors that might contribute to the development of anxiety and depression: threatening behaviors and hostile, rejecting behaviors. He suggested that threatening behaviors, that is, behaviors that indicate a caregiver may not be physically or emotionally available when needed, are particularly characteristic of anxious individuals’ childhood attachment relationships. These threats can take many forms, including threats to stop loving a child if (s)he does not behave, threats to abandon a child, and threats to leave the home, become ill, or commit suicide (Bowlby, 1973). Bowlby, 1980, Bowlby, 1988 further suggested that hostile, rejecting behaviors, that is, behaviors that imply a caregiver certainly will not be available when needed, are particularly characteristic of depressed individuals’ childhood attachment relationships. Such hostility and rejection is embodied by a child being told (s)he is unlovable, inadequate, and incompetent, and, unlike caregiver threatening behavior, typically includes constant, active rebuffs of the child’s bids for responsive behavior. In sum, Bowlby suggested childhood attachment experiences that might differentially lead to anxiety and depression.

We believed that Bowlby’s ideas about the kinds of caregiver behavior that may lead to anxiety and depression might also contribute to knowledge about development of AS. AS can be conceptualized as a type of internal working model that appears to develop early in life, preceding the onset of psychological distress. Consequently, experiences with primary caregivers would seem to have particular importance in the development of AS. Thus, to better understand the origins of AS and expand upon the work of Watt et al. (1998), the current study investigated the role of parental threatening, hostile, and rejecting behavior in the development of AS.

We examined two main sets of hypotheses. First, based on findings that high overall AS has been related to both anxiety and depression (Otto et al., 1995, Taylor et al., 1992, Taylor et al., 1996), we predicted that both sets of negative parental behavior (i.e., threatening behavior and hostile, rejecting behavior) would be related to the development of overall AS. Second, based on the work of Taylor et al. (1996), we predicted differential relationships between parental behavior and AS factors.

In their study of relationships between AS, depression, and anxiety, Taylor et al. (1996) found three factors of AS: (1) fear of publicly observable anxiety symptoms such as fears of shaking, (2) fear of anxiety-related bodily sensations, such as fears of nausea, and (3) fear of losing control over one’s thoughts, such as worrying one is mentally ill. Moreover, these investigators found that the first two factors were closely associated with anxiety symptomatology, whereas the latter factor was closely associated with depressive symptomatology. Thus, we believed that the latter factor (fear of losing control over one’s thoughts) might be best predicted by parental hostile, rejecting behaviors, hypothesized by Bowlby (1980) as an antecedent of depression. Furthermore, we believed that the first two factors (fear of publicly observable anxiety symptoms and fear of anxiety-related bodily sensations) might be best predicted by parental threatening behaviors, hypothesized by Bowlby (1973) as an antecedent of anxiety.

Providing that expected relationships were found between parenting and AS scores, we planned to examine a third set of hypotheses. Specifically, we were interested in the extent to which AS and its factors might mediate relationships between parenting and both current and past anxious and depressive symptoms. Indeed, Bowlby, 1973, Bowlby, 1980 theorizing about differential antecedents of anxiety and depression, research findings that AS acts as a vulnerability factor for the development of anxiety and depression (Maller & Reiss, 1992; Schmidt et al., 1997, Schmidt et al., 1999), and Taylor et al.’s (1996) findings of unique relationships between AS factors and anxiety and depression suggest specific mediational relationships between parenting, AS, and both current and past symptoms. For example, overall AS would be expected to mediate relationships between parental threatening behavior and current and past anxiety symptoms, while fear of losing control over one’s thoughts would be expected to mediate relationships between parental hostile, rejecting behavior and current and past depression symptoms. We planned to examine mediational hypotheses when predicted relationships between parenting and AS were found.

To investigate our hypotheses, university students completed self-report measures assessing exposure to negative parenting behaviors during childhood, AS, current anxious and depressive symptoms, and history of anxiety and depressive disorder symptoms. Multiple regression analyses and partial correlations were used to examine the hypotheses.