شاخص های ابعادی از شدت اختلال اضطراب فراگیر برای DSM-V

For DSM-V, simple dimensional measures of disorder severity will accompany diagnostic criteria. The current studies examine convergent validity and test–retest reliability of two potential dimensional indicators of worry severity for generalized anxiety disorder (GAD): percent of the day worried and number of worry domains. In study 1, archival data from diagnostic interviews from a community sample of individuals diagnosed with one or more anxiety disorders (n = 233) were used to assess correlations between percent of the day worried and number of worry domains with other measures of worry severity (clinical severity rating (CSR), age of onset, number of comorbid disorders, Penn state worry questionnaire (PSWQ)) and DSM-IV criteria (excessiveness, uncontrollability and number of physical symptoms). Both measures were significantly correlated with CSR and number of comorbid disorders, and with all three DSM-IV criteria. In study 2, test–retest reliability of percent of the day worried and number of worry domains were compared to test–retest reliability of DSM-IV diagnostic criteria in a non-clinical sample of undergraduate students (n = 97) at a large west coast university. All measures had low test–retest reliability except percent of the day worried, which had moderate test–retest reliability. Findings suggest that these two indicators capture worry severity, and percent of the day worried may be the most reliable existing indicator. These measures may be useful as dimensional measures for DSM-V.

DSM-IV (American Psychiatric Association, 2000) significantly revised diagnostic criteria for generalized anxiety disorder (GAD) as it was written in DSM-III-R (American Psychiatric Association, 1987). Although diagnostic reliability has improved since these revisions, GAD has one of the lowest rates of inter-rater reliability of any disorder listed in the DSM-IV, at least within clinical samples (Wittchen, Lachner, Wunderlich, & Pfister, 1998). Worry is the defining feature of GAD, and is characterized as “excessive” and “difficult to control”. The vagueness of these descriptors may explain the reliability issues with GAD. For example, a determination of what level of worry is excessive likely differs substantially among patients as well as raters. A more concrete and specific indicator of GAD severity may improve inter-rater reliability. Due to the likely inclusion of dimensional measures of disorder severity alongside diagnostic criteria in DSM-V, it will be useful to identify simple, brief measures of GAD severity that can be easily reported by patients (Helzer, 2008). The goal of this report is to evaluate two potential indicators of GAD severity that are more specific than the current criteria and are amenable to self report: percent of the day spent worrying and number of worry domains. Several studies have shown that inter-rater reliability for the diagnosis of GAD is poor. Brown, Di Nardo, Lehman, and Campbell (2001) assessed inter-rater reliability of all anxiety disorders within a treatment-seeking sample. Two separate interviewers assessed symptoms, using the Anxiety Disorders Interview Schedule-IV (Brown, Di Nardo, & Barlow, 1994) approximately two weeks apart. Of all the anxiety disorders, GAD had the lowest inter-rater reliability, with a kappa of .67 compared to .72 for panic disorder, .86 for specific phobia, .77 for social phobia and .85 for obsessive–compulsive disorder. The primary source of unreliability between the two interviewers was identified for each participant using a rating system, and the most frequent source of discrepancy from one interview to the next was a difference in patient reports of their symptoms. The authors suggest that vagueness of the diagnostic criteria (i.e., “excessive worry”) may make it difficult for patients to reliably report on their symptoms.

It is clear from the low test–retest reliability of the GAD diagnosis that improvements are warranted in DSM-V. For DSM-V, reliable and simple dimensional measures of GAD severity will be required, however the current measures are not well defined and therefore, unreliable. Results from the current study suggest that number of worry domains and percent of the day worried may be considered good measures of GAD severity that would be easy for participants to report. They correlate moderately with other indicators of GAD severity and correlate strongly with current DSM-IV diagnostic criteria for GAD. These measures correlate particularly well with the excessiveness and uncontrollability criteria, suggesting that they may be measuring the same constructs. However, identifying the amount of the day spent worrying and the most frequent topics of worry may be a more concrete way for individuals with GAD to define and report the severity of worry than rating the excessiveness or uncontrollability on a Likert scale. In a recent study, Berle et al. (2011) created a 3-item version of the Penn state worry questionnaire by identifying items most related to DSM criteria and using factor loadings. Two of the three final items are as follows: “Many situations make me worry” and “I worry all the time”. Participants rate the extent to which they agree with these statements. These items correspond to the breadth of worry (number of worry domains) and frequency of worry (percent of the day worried) as assessed in the current study. Therefore, further research that compares the three item Penn state worry questionnaire with number of worry domains and percent of the day worried may be beneficial for identifying the best dimensional measure for DSM-V. Overall, current diagnostic criteria for GAD demonstrated poor test–retest reliability, which is consistent with low reliability of the GAD diagnosis in clinical samples. Percent of the day worried however demonstrated the highest test–retest reliability of all indicators, suggesting that percent of the day worried may be useful for identifying the severity of GAD for DSM-V. Further research that examines test–retest reliability in a clinical population will help identify whether percent of the day worried may be a better measure of worry severity than existing measures. Percent of the day worried is easy for participants to report, relates strongly to DSM-IV diagnostic criteria and other indicators of GAD severity, and demonstrates the highest test–retest reliability in a non-clinical sample. It therefore shows promise as a potential dimensional measure of GAD severity in DSM-V.