Search

Thursday, February 21, 2019

Gender Bias in the Diagnosis of Personality Disorders

Are Some Axis II Disorders More Prevalent in One Gender than Another?

 By: THEODORE MILLON and
        Seth Grossman
        Carrie Millon
        Sarah Meagher
        Rowena Ramnath

With so many varied and insidious sources of potential gender bias in the diagnosis of personality disorders that overlap and interact with each other, it is nearly impossible to untangle real differences from artifacts. Bias can enter the equation at any point from the DSM diagnostic criteria themselves, to the clinicians who diagnose patients, to the populations sampled in our empirical research, all generated and maintained within the context of an overarching biased and often misogynistic culture. Where do we begin an attempt at sorting out all of these potential sources?

One body of evidence to consider when pondering this question of bias in diagnosing personality disorders is the prevalence rates of different categories of disorders in both males and females. The DSM-IV (APA, 1994) reports that certain personality disorders— namely, antisocial personality disorder—are more frequently diagnosed in males while borderline, histrionic, and dependent personality disorders are more frequently diagnosed in females. Hartung and Widiger (1998) compiled findings from a variety of sources to determine that more males are diagnosed as paranoid, schizoid, schizotypal, antisocial, narcissistic, and compulsive; and more females are diagnosed as borderline, histrionic, and dependent. Other studies have reported roughly equal numbers of males and females diagnosed as schizotypal personality disorder. However, it seems as though males possess more eccentric/odd symptoms and females possess more ideas of reference, magical thinking, and social anxiety (Roth & Baribeau, 1997).

One often-cited criticism in these prevalence figures is that we have very biased samples of patients. We take samples of convenience at hospitals, the VA, prisons, and the like and generalize the numbers to clinical populations and nonclinical populations alike. We have done a fairly poor job of seeking out representative samples to gather reliable prevalence statistics.

Surprisingly, overall, both men and women are equally as likely to receive a diagnosis of a personality disorder (Kass, Spitzer, & Williams, 1983). An equal prevalence of personality disorder diagnoses across gender as a whole does not preclude the existence of gender bias. Hartung and Widiger (1994) suggest that there is a very real possibility that either men or women may in fact have more personality pathology. Should it be a goal of psychologists to have equal numbers of women and men diagnosed as personality disordered? More specifically, should it be a goal that equal numbers of men and women be diagnosed in each category? Widiger (1998) argues that “the purpose of the DSM-IV is to provide an accurate classification of psychopathology, not to develop a diagnostic system that will, democratically, diagnose as many men with a personality disorder as women” (p. 98). While this may be true, it does not excuse the profession from investigating and correcting the potential sources of bias that are contributing the differences we are observing under our current system.

One consideration that adds more complexity to the issue includes the empirical evidence that has been amassed to suggest that there is a systematic failure on the part of the diagnosticians in adhering to the clinical criteria set forth in the DSM when making diagnoses (refer to Widiger, Corbitt, & Funtowitz, 1994, for a more comprehensive review of this literature). For example, a study by Ford and Widiger (1989) found that clinicians diagnosed females with histrionic personality disorder when they failed to meet DSM criteria for the histrionic personality disorder and instead met criteria for antisocial personality disorder. Further, when clinicians were asked to individually assess each of the DSM antisocial and histrionic criteria, the sex of the patient had no effect. Ford and Widiger assert that this is evidence to suggest that the problems lie not in the criteria themselves, but in clinicians’ failure to adhere to the guidelines.

Other evidence supports this contention of a problem at the level of diagnostic labels, not necessarily in the diagnostic criteria. Sprock (2000) used a method derived from the act-frequency approach to have undergraduate students generate behavioral examples of DSM histrionic personality without regard to sex or sex roles. Then, she had a sample of psychologists and psychiatrists rate the representativeness of the symptoms for either histrionic criteria or histrionic personality disorder. She discovered that “feminine” behaviors were rated as more representative of histrionic personality disorder and somewhat more representative of the histrionic criteria than “masculine” behaviors, supporting the same notion that Ford and Widiger argued: The female sex-role is more related to the label than to the criteria. These same arguments can be made for dependent personality disorder, which includes many traditionally feminine qualities such as putting other people’s needs above your own or relying on a husband to provide an income and home.

These differences in the prevalence rate for females are hardly a one-sided bias on the part of the DSM. It is just as easily arguable that there are male stereotyped behaviors to be found in criteria for personality disorders. For example, the narcissistic personality disorder contains criteria that are traditionally ascribed to healthy male functioning such as an inflated sense of self-importance, a preoccupation with fantasies of unlimited success and power, possessing a sense of entitlement, lacking empathy, and assuming an arrogant manner. Traditionally socialized male characteristics can be seen in the criteria for antisocial personality as well, such as deceitfulness, impulsivity, aggressiveness as evidenced by getting into physical fights, and irresponsibility (APA, 1994).

While neither side of the debate would likely argue that these criteria mirror sex-typed behavior, they would argue as to the significance of this fact. Many feminists would argue that labeling women as personality disordered is an act of punishing women for conforming to the very criteria we ask them to conform to (Landrine, 1989; Walker, 1994). They argue that women are actually in a double bind. If they do not act in a manner that is deferential, dependent, and sexually provocative, they risk becoming social outcasts. If they do, they are branded “mad” or “disordered.” Landrine (1989) asserts:

             The purposes of masquerading gender roles as madness may be (a) to locate falsely within persons all of the ludicrous cognitive and behavioral limitations that actually reside in gender roles and stratification so that (b) to direct our attention—not to changing gender roles or to eradicating gender stratification—but to changing individuals through therapy and to eradicating their ostensible personal problems. (p. 332)

Or worse, it is a way to label victims of sexual abuse or domestic violence as “sick” rather than placing the blame on the perpetrators of these crimes. Given that a huge percentage of women diagnosed as personality disordered have histories of sexual and physical abuse, this argument is convincing (Brown, 1992).

Others (including Widiger, Corbitt, & Funtowitz, 1994) argue that just because these disorders are founded on a biased society that encourages some of these characteristics does not negate the fact that in these extreme forms of expression, they are pathological. If we conceptualize personality disorders as lying on a continuum rather than as categorical, it is justifiable that at these extreme ends, the behaviors, whatever their original source of motivation, are disordered.

There is empirical support for this notion that several personality disorders are exaggerations of normal socialized sex-typed behaviors. Landrine (1987, 1989) proposes what she calls a social-role hypothesis that posits:

              Each personality disorder represents the role/role-stereotype of the specific status group (Sex X Social Class X Marital Status Group) that tends to receive that diagnosis most often, such that the personality disorder categories as a whole represent the role/role-stereotypes of both sexes. Thus, women might receive certain diagnoses more often than men—and vice versa— because the category on question is by and large equivalent to their role.

Based on this social-role hypothesis that a personality disorder might be differentially assigned to males or females when the description is consistent with gender-role stereotypes, Rienzi and Scrams (1991) argued that clinically untrained people should be able to accurately make this distinction as well. If assigning these diagnoses to men and women is a social construction, non-clinicians should fall as victim to these biases as psychologists. In their study of university students, they asked students to assign gender to six descriptions of DSM-III-R personality disorders. Significant agreement was found, with paranoid, antisocial, and compulsive personality disorders being viewed as male and dependent and histrionic personality disorders viewed as female. In another study along this same vein, Rienzi, Forquera, and Hitchcock (1994) asked undergraduates to assign either a label of male or female to gender-ambiguous vignettes and found a similar bias to labeling the narcissistic vignette as a male and the dependent vignette as female.

Another line of research lends an additional slant on this argument. Some theorists have hypothesized that histrionic personality disorder and antisocial personality disorder are actually “expressions of the same latent disposition” (Hamburger, Lilienfeld, & Hogben, 1996, p. 52). What we observe superficially are only the gender-stereotyped behaviors of the same underlying pathology. As Widiger and Spitzer (1991) argue, the differences we see in the distribution of personality disorders may be due to etiological factors such as different sex hormones that influence the final expression of pathology. Histrionic and antisocial personalities may be an example of such an effect. They may both represent expressions of the same pathology, and we are misconceptualizing them as two separate entities because their superficial expressions (the symptoms they outwardly express) are different.

Yet another area of the literature to contemplate when considering gender bias in diagnosing personality disorders is the area of criminal behavior. The once hard-and-fast rule that men display more criminal and antisocial behavior is rapidly changing. Historically, female crime rates have been incredibly small and considered an aberration. In the late 1890s, Lombroso proposed what he called the “masculinity hypothesis” to explain women who committed crimes. He believed that excessive body hair, wrinkles, and an abnormally shaped head were all outward signs of the female criminal. Perhaps because of Lombroso’s belief that female criminals are masculine-like and, therefore, somehow unnatural, it has been widely believed that female criminals are somehow sicker than male criminals. Lombroso even quotes an Italian proverb, “Rarely is a woman wicked, but when she is she surpasses the man,” which illustrates this belief (Lombroso & Ferrero, 1916, p. 147). Clearly, the prototype for an antisocial was considered
a male antisocial. A female antisocial is somehow only a poor cousin to the prototype or “real” antisocial.

Conclusions
After weighing all perspectives and sources of evidence, it is clear that there are problems on multiple levels. Our conceptualizations are fuzzy, our samples are biased, our measures are biased, and our clinicians fall prey to their own biases. The question of what to do is even more difficult to address. At the theoretical level, we have great room for improvement. A logical place to start is at the DSM level. However, tales of how DSM committees work are horrifying (see Caplan, 1991, for details on how the self-defeating personality diagnosis was retained in the DSM). What we put faith in as an unbiased and scientific system for diagnosing mental disorders is often a very political and even random set of criteria. To add further fuel to the fire, the vast majority of DSM authors and committee members on personality disorder workgroups have been male. Eighty-nine percent of DSM-III personality disorder workgroup members (8 of 9) were male, 84% for DSM-III-R (32 of 38), and 78% for the DSM-IV (7 of 9; Widiger, 1998). Future revisions of the DSM need to draw on the resources of a far wider clinical base than the 1,000 individuals involved in the production of the DSM-IV. Additionally, if our clinical samples are biased, as a community, we need to make the investment in solid epidemiological research to determine real prevalence rates, even if we are using our flawed diagnostic criteria. It is difficult to obtain a clear picture of where to fix a problem if we do not know how pervasive the problem is.

The fairly convincing evidence that clinicians do not adhere to the existing criteria when making diagnoses is also troubling. How to tackle these failures is a complex issue. Perhaps raising awareness of biases is a first step. If clinicians have taken the time to introspect a little on why they assume histrionics are female and narcissists are male, they may pause to consider alternative diagnoses. The DSM-IV contains only one sentence about this issue buried within the general discussion of personality disorders:

          Although these differences in prevalence probably reflect real gender differences in the presence of such patterns, clinicians must be cautious not to overdiagnose or underdiagnose certain Personality Disorders in females or in males because of social stereotypes abouttypical gender roles and behaviors. (APA, 1994, p. 632)

Perhaps another solution would be to include reminders throughout the DSM, for example, in the histrionic section: “If a female client is presenting with X, Y, and Z symptoms, make sure to consider a diagnosis of antisocial personality disorder as well.” While these suggestions are merely stopgap measures to address the larger issues of how our diagnostic criteria and, on a grander scale, our entire mental health system are biased against women, they are a step toward ameliorating the negative effects such biases can have on our mental health.

References

Personality Disorders in Modern Life, second edition, 2000, 2004 by John Wiley & Sons, Inc.

No comments:

Post a Comment