 |
 |
|
|
Sex bias in research has consistently led to descriptions of psychiatric disorders in APA's Diagnostic and Statistical Manual of Mental Disorders (DSM) that result in some diagnoses being incorrectly applied more often to women than to men and vice versa, according to Thomas Widiger, Ph.D., research coordinator for the DSM-IV.
Widiger spoke in June in Toronto at the APA annual meeting symposium "Critical Issues in Psychiatric Nosology."
In a critique of various presentations, symposium discussant Robert Spitzer, M.D., who chaired the APA task forces on DSM-III and DSM-III-R and was a consultant for DSM-II, praised Widiger's analysis as "elegant."
Widiger is a clinical psychologist at the University of Kentucky. He and a colleague reviewed 254 studies published in the Journal of Abnormal Child Psychology from 1987 through 1994. They found that 70 studies across all childhood disorders were confined to just one sex, and that 99.6 percent of those studies were of boys. Seventy of the 254 studies were on ADHD; fewer than half of the ADHD studies included any girls, and the remaining 36 were confined to boys.
While this could represent the true sex differential, Widiger is skeptical. "If the appearance or presentation of ADHD does vary with gender, then we do wonder whether we are focusing too heavily on this male variant."
Another researcher analyzed studies of substance abuse or dependence in 17 key English-language journals published in 1992. Of the 601 studies that concerned humans, only 6 percent had study populations that were predominantly female, said Widiger. "The males were the majority, and often substantially, in 73 percent of the studies."
Another review examined every study on schizophrenia published in the American Journal of Psychiatry, Archives of General Psychiatry, Journal of Abnormal Psychology, and Journal of Nervous and Mental Disease from 1985 though 1989. Of the total of 189 studies, only 8 percent had more women than men, 75 percent had more men than women, and 25 percent were confined to men. What is striking about this finding is that schizophrenia is "said to occur equally in males and females," Widiger commented.
"All of these reviewers and many others suggest that this disproportionate representation of gender in research is having a real and significant effect on how we understand these disorders," he said.
Biased samples can have a major impact on the development of diagnostic criteria. The current understanding of somatization disorder is based almost entirely on research with females, said Widiger. All of the 1,116 persons with somatization disorder studied for DSM-III were women. Among the diagnostic criteria for this diagnosis was one group of "female reproductive symptoms, which include four symptoms that no male could possibly have," Widiger noted. The authors of DSM-III recognized this problem and reduced the threshold for the diagnosis from 14 to 12 symptoms when the diagnosis was applied to men. But this reduction was not based on any research on men, Widiger said. In DSM-III-R the authors decided that "the differential threshold was arbitrary and uninformed" and revised the symptom list so that the number of symptoms required for the disorder was the same for men and women. But in trying to offset the bias of having four symptoms that only women could have, they added one that only males could have.
"That is an interesting approach-trying to address one bias by adding a complementary bias," remarked Widiger.
In DSM-IV the criteria set was simplified into four groups, but one group continued to be confined to "largely female reproductive symptoms," Widiger pointed out. When this bias was recognized, a male-biased item-erectile or ejaculatory dysfunction-was added. Again, this change was "not informed by any research on somatization disorder in males."
The DSM-IV field trials on somatization disorder again included only women. The manual states that the disorder is rare in men. While this could be true, said Widiger, "this opinion is not shared by researchers using the ICD-10 criteria set in other cultures," he noted.
A similar bias, but toward males, exists in the DSM criteria for conduct disorder. In DSM-III-R, for example, the conduct disorder criteria included "forcing someone into a sexual activity." Very few girls with conduct disorder commit rape, observed Widiger. "Conduct disorder will be expressed differently in girls, but the criteria appear to be weighted heavily [toward] how the disorder appears in boys."
In DSM-IV, the authors were concerned with this differential and tried to address it, he noted. The section "Specific Culture, Age, and Gender Features" in DSM-IV notes that boys with conduct disorder "frequently exhibit fighting, stealing, vandalism, and school discipline problems," whereas girls "are more likely to exhibit lying, truancy, running away, substance use, and prostitution."
If this is accurate, it is necessary to ask whether "we should be using the same criteria set for boys and girls," said Widiger. "Perhaps prostitution should replace rape when diagnosing conduct disorder in girls."
But some have opposed this suggestion, he added. The danger of setting up two sets of criteria is that it could result in girls' being diagnosed with a disorder at a lower threshold of dysfunctional behavior than boys. Widiger suggested that instead of setting up parallel criteria for the same disorder, it might be better to diagnose girls with a different disorder.
DSM-IV task force chair Allen Frances, M.D., stated that "histrionic personality disorder as defined in DSM-IV may continue to represent an exaggerated version of stereotypical feminine traits and may place too little emphasis on items and examples that [tap the] macho male version," which involves the expression of exaggerated masculine traits. Remarked Widiger, "In his own way, Stanley Kowalski in A Streetcar Named Desire is just as histrionic as Blanche."
The problem with masculinizing histrionic personality disorder is that it "changes and distorts the disorder; it's no longer histrionic personality disorder," Widiger said.
Widiger recommended that, for DSM-V, "a gender differences advisory committee" be established, similar to DSM-IV's cultural advisory committee. Such a gender advisory committee has been created for DSM-IV as it undergoes revision, said Widiger. (These revisions are not considered extensive enough to call for a DSM-IV-R, but will be included in future printings of DSM-IV.)
Widiger further recommended that there be "epidemiological research with probability-based sampling of community populations." Clinical samples may be too biased by referral issues to provide valid estimates of differential sex prevalence rates, he said.
Both men and women should be represented in research, and where a disorder appears to be mainly associated with one sex, an effort should be made to sample the other sex when evaluating that disorder, said Widiger. The National Institute of Mental Health is already following this policy in deciding whether to fund field trials for diagnostic criteria, he noted.
It is hard to know what to do when gender-biased criteria are identified in DSM, Widiger said. "Frankly, I expect that many, many-probably the vast majority of diagnostic criteria-will be differentially valid for males and females, because, one, gender does affect the presentation and manifestation of many disorders, and two, many disorders are inherently gender related. Separate criteria may make sense only for those disorders that are not gender related."
Either way, "if separate criteria are developed, or gender-related disorders are accepted, their thresholds for diagnosis should be equalized, that is, [researchers should] establish a gender-neutral standard of clinically significant function and impairment."