Concordance between telephone survey classification and face-to-face structured clinical interview in the diagnosis of generalized anxiety disorder in Hong Kong
Introduction
Although telephone interviewing has steadily replaced face-to-face interviewing for survey data collection, such a shifting of research mode has occurred to a lesser extent in the field of psychiatric epidemiology (Holbrook, Green, & Krosnick, 2003). Widely cited community psychiatric surveys are still based on face-to-face structured interviews and have been mostly carried out in Western societies that can afford this mode of research. Notwithstanding certain inherent limitations, telephone surveys of mental disorders confer several advantages such as affordability, quick turnaround time, wide geographic coverage, anonymity and, arguably, facilitation of disclosure about stigma-sensitive information (Evans, Kessler, Lewis, Peters, & Sharp, 2004; Fenig, Levav, & Kohn, 1993; Ohayon & Hong, 2006; Rohde, Lewinsohn, & Seeley, 1997). Therefore, it is worthwhile to explore their utility in scenarios where face-to-face interviews are not feasible, such as middle and low income countries with scarce resources for mental health research (Sharan, Levav, Olifson, de Francisco, & Saxena, 2007).
Although some researchers believed that survey satisficing and social desirability bias were more likely to occur during telephone than face-to-face interviewing (Holbrook et al., 2003), various psychiatric studies found good agreement between both modes of interviewing when the same instruments for assessment were used (Allen, Cull, & Sharpe, 2003; Aziz & Kenford, 2004; Brar et al., 2002; Evans et al., 2004; Fenig et al., 1993; Kobak et al., 1997; Pridemore, Damphousse, & Moore, 2005; Wells, Burnman, Leake, & Robins, 1988). Nonetheless, these studies, mostly carried out in the U.S., were often based on small clinical samples of patients who were familiar with the researchers. They also tended to use rating scales instead of algorithm-based diagnostic instruments that could identify specific mental disorders. Consequently, although these studies demonstrated reliability of the methods used, they were less relevant for methodological comparison in the context of large-scale telephone surveys, especially in non-U.S. populations.
Population-based telephone surveys of mental disorders were fewer in number than clinical studies, and their results were usually not reappraised by subsequent face-to-face interviews. Nonetheless, such surveys as conducted in Western communities generally revealed prevalence estimates comparable to those of surveys using face-to-face structured interviews (Eisen et al., 2004; Ohayon, 2007; Ohayon, Priest, Guilleminault, & Caulet, 1999). The situation for Asian communities, however, appeared different. For example, although most face-to-face surveys using structured diagnostic schedules in Chinese communities showed very low rates of mental disorders (Chen et al., 1993; Shen et al., 2006), several telephone surveys of anxiety and depressive disorders in Hong Kong have revealed prevalence estimates comparable to those found in Western societies (Lee et al., 2005, Lee et al., 2007a, Lee et al., 2007b). Likewise, with a high participation rate of 91.4%, a telephone survey of the general population in South Korea revealed that major depressive disorder was significantly more common than what was found in previous Korean surveys based on the face-to-face mode of interviewing (Ohayon & Hong, 2006). Authors of these Asian surveys speculated that the anonymity of telephone interviewing could reduce stigma, thereby facilitating symptom expression and more accurate prevalence estimates to be found. This possibility is of cross-cultural interest given that psychiatric stigma is severe in Asian societies (Yang et al., 2007). In contrast to several studies showing that Western people often would prefer face-to-face over telephone interviews (Evans et al., 2004; Holbrook et al., 2003), stigma may render Asian people less tolerant of face-to-face interview with “strangers” in their home than telephone survey. Nonetheless, the suggestion remains speculative until the validity of diagnoses made from the telephone instruments is substantiated by clinical reappraisal interviews. In fact, the alternative possibility that telephone surveys over-estimate the prevalence of common mental disorders in Asian communities cannot be ruled out.
One reason for the frequent failure to examine the validity of telephone survey instruments used in general population surveys is the practical difficulty of re-contacting the respondents with no prior relationship to the researchers for subsequent face-to-face interviews. Telephone surveys are typically anomalous and brief. Respondents are not as personally engaged as during face-to-face interviews and usually do not expect, or presumably prefer not, to be re-contacted. In the present study, we attempt to overcome this difficulty by re-interviewing in-person a sub-sample of the same respondents who previously took part in a telephone-based general population survey of generalized anxiety disorder (GAD) in Hong Kong. Our primary aim is to examine the concordance between telephone survey classification and face-to-face structured clinical interview diagnosis of GAD in the context of a large epidemiological survey.
Section snippets
Method
Hong Kong has a population of 6.9 million people and 96.3% of its 2,198,000 households have one or more telephones (Hong Kong Census and Statistics Department, 2005). Using a telephone-based methodology similar to a previous survey of GAD in Hong Kong which had not included a second phase of face-to-face clinical interview (Lee et al., 2007a), data collection for the present telephone survey was completed during July 10–21, 2006. Two thousand and five respondents were successfully interviewed
Measures
The telephone survey instrument consisted of items about sociodemographic characteristics and the DSM-IV-TR symptoms of GAD. It classified respondents as having GAD in the previous 1 year if:
- (i)
they reported
- [a]
being habitually prone to worries/anxiousness, and/or
- [b]
a period of time in the previous 1 year when they always experienced worries/anxiousness, and/or
- [c]
being always worried/anxious about one or more matters (namely, health, study, work, family, relationships, finance, or no reason at all) in the
- [a]
Concordance between dichotomous telephone survey classification and SCID interview diagnosis
Because of over-sampling as described above, 22 of the 100 interviewees (22%, S.E. = 4.16) who underwent clinical interviews had been classified as having GAD in the telephone survey. SCID interviews indicated that 57 of these 100 interviewees had a diagnosis of GAD (57%, S.E. = 4.98). As shown in Table 1, the telephone survey identified 29.8% of those diagnosed as GAD by SCID interviews, whereas the SCID interviews identified 77.3% of those classified as GAD by telephone survey. Fig. 1 showed the
Telephone interview instrument versus face-to-face SCID interview
To our knowledge, this paper is the first attempt to study the concordance of face-to-face clinical interviews and telephone instrument interviews of the same subjects who took part in a large general population survey of GAD. The overall results indicate that the telephone interview method used did not over-estimate the prevalence of GAD when compared to face-to-face clinical interviews. We found a significant degree of discordance occurred between clinical SCID diagnosis and telephone survey
References (38)
- et al.
Diagnosing major depression in medical outpatients: acceptability of telephone interviews
Journal of Psychosomatic Research
(2003) - et al.
Can clinical features of bipolar-I disorder be assessed reliably on the telephone?
Journal of Affective Disorders
(2002) - et al.
Interpretation biases in social anxiety: response generation, response selection, and self-appraisals
Behavior Research and Therapy
(2007) - et al.
Twelve-month prevalence, correlates, and treatment preference of adults with DSM-IV major depressive episode in Hong Kong
Journal of Affective Disorders
(2007) Epidemiology of depression and its treatment in the general population
Journal of Psychiatric Research
(2007)- et al.
The prevalence of depressive disorders in the United Kingdom
Biological Psychiatry
(1999) - et al.
Obtaining sensitive information from a wary population: a comparison of telephone and face-to-face surveys of welfare recipients in the United States
Social Science & Medicine
(2005) - et al.
Broadening the definition of generalized anxiety disorder: effects on prevalence and associations with other disorders in the National Comorbidity Survey Replication
Journal of Anxiety Disorders
(2007) - et al.
Use of mental health services for anxiety, mood, and substance disorders in 17 countries in the WHO World Mental Health Surveys
The Lancet
(2007) - et al.
Culture and stigma: adding moral experience to stigma theory
Social Science & Medicine
(2007)