|
Book Title: Treating and Preventing Adolescent Mental Health Disorders
> pp. [265]-[269]
UNDEFINED: AUTHORS
Treating and Preventing Adolescent Mental Health Disorders
Print ISBN 9780195173642, 2005
pp. [265]-[269]
Table 13.3
Two-stage Studies of Prevalence of Anorexia Nervosa and Bulimia Nervosa in Community Samples of Adolescents| | | Subjects | Methods | Prevalence Rates (%)d
| | | AN | BN | Study | Country (city or state) | Samplea
| Age (years) | Sample Size (response rate in %) | Screeningb
Interview (response rate in %) | Diagnostic Criteriac
| Point | Lifetime | Point | Lifetime | Whitaker et al., 1990 | United States (county in New Jersey) | 1 | 13–18 | F 2544 M 2564 | (91) | EAT-40 Interview | (75) | DSM-III | F M | | 0.3 | F M | | 2.5 0.2 | Rathner & Messner, 1993 | Italy (Brixen) | 2 | 11–20 | F 517 M 0 | (81) | EAT, ANIS Interview | (88) | DSM-III-R | F | 0.58 | | F | | | Wlodarczyk-Bisaga & Dolan, 1996 | Poland (Warsaw) | 2 | 14–16 | F 747 M 0 | (93) | EAT-26 Interview | (92) | DSM-III-R | F | | | F | | | Graber et al., 2003 | United States New York City | 3 | Mean:16 | F 155 M 0 | | EAT SCID | | DSM-III-R | F | | 3.9 | F | | 3.2 | Szabo & Tury, 1991 | Hungary
(Debrecen) | 2 | 14–18 | F 416 M 119 | (49) | EAT-40, BCDS, ANIS Interview | | DSM-III-R | F— M— | | | F M | | | Steinhausen et al., 1997 | Switzerland (Zurich Canton) | 2 | 14–17 | F 276 M 307 | | EDE-S DISC-P | | DSM-III-R | F M | 0.7 | | F M | 0.05 | | Verhulst et al., 1997 | Netherlands | 4 | 13–18 | F/M 853 | (82) | CBCL DISC-C/-P | (91) | DSM-III-R | F/M | 0.3 | | F/M | 0.3 | | Santonastaso et al., 1996 | Italy (Padova) | 2 | 16 | F 359 M 0 | (91) | EAT-40, BMI Interview | (93) | DSM-IV | F | | | F | 0.5 | 1.0 | Rosenvinge et al., 1999 | Norway (Buskerud) | 2 | 15 | F 464 M 214 | (78) | EDI DSED | | DSM-IV | F M | 0.4 | | F M | 1.1 | |
|
a
1 = entire enrollment of grades 9 through 12 in a single New Jersey county, including public, private and parochial schools; 2 = public school students; 3 = private school students; 4 = national probability sample.
b
ANIS, Anorexia Nervosa Inventory Scale; BCDS, Bulimic Cognitive Distortions Scale; BMI, Body Mass Index; CBCL, Children's Behavior Check List; DISC, Diagnostic Interview Schedule for Children (DISC-P, parent version, DISC-C, child version); DSED, Diagnostic Survey for Eating Disorders; EAT, Eating Attitudes Test; EDE-S, Eating Disorder Examination, screening version; EDI, Eating Disorder Inventory; SCID, Structured Clinical Interview for DSM-IIIR.
c
DSM, Diagnostic and Statistical Manual for Mental Disorders (DSM III, 3rd edition; DSM III-R, 3rd, revised edition; DSM IV, 4th edition).
d
AN, anorexia nervosa; BN, bulimia nervosa; F, female; M, male; Point, point prevalence, based on 6 months (Verhulst et al., 1997), 12 months (Steinhausen et al., 1997), or “current” (i.e., at the time of the interview; Rathner & Messner, 1993; Rosenvinge et al., 1999; Santonastaso et al., 1996; Szabo & Tury, 1991; Wlodarczyk-Bisaga & Dolan 1996).
end p.265
PRINTED FROM Treating and Preventing Adolescent Mental Health Disorders (www.oup.com/amhi-treatingpreventing)
© Copyright Oxford University Press, 2006. All Rights Reserved
those participants whose self-report responses suggest the presence of an eating disorder. The quality of the two-stage method depends in part on the participation rates at each assessment. For example, some data suggest that screening efforts may miss eating disorder cases and thus underestimate the prevalence of eating disorders (Fairburn & Beglin, 1990). The sensitivity of the screening instrument is also important: in an uncommon disorder, the omission of even a few cases because of an insensitive screen produces underestimates of the true prevalence of the disorder. In six studies (summarized in Table 13.4), there was no initial screening; rather, the entire sample participated in a diagnostic interview. In general, eating disorder status was assessed in the context of a comprehensive evaluation of psychiatric disorders, rather than in an interview focused specifically on eating disorders. This approach can be advantageous—because participants are not recruited specifically for a study of eating disorders, participation rates are likely unaffected by the participants' attitudes about eating disorders (e.g., wanting to avoid detection of one's eating disorder). There are also limitations to using a one-step assessment design. These studies generally have lower participation rates than those of two-stage studies, possibly because of a greater subject burden from the requirement that all participants complete the diagnostic interview. Also, in an effort to reduce assessment time and subject burden, interviewers usually employ a branched approach to diagnostic assessment: “gated” questions about the key symptom(s) are required for a diagnosis; if these are answered negatively, any further assessment of symptoms is terminated (Feehan, McGee, Raja, & Williams, 1994). Hence, unless the participant acknowledges the initial question (e.g., voluntary efforts to achieve low weight in the case of anorexia nervosa, and recurrent binge eating in the case of bulimia nervosa), no further information is gathered about eating disorder symptoms. It is unclear whether the clinical presentation of eating disorders in children is less prototypic than that in adults. If it is, then a higher number of cases among children would be missed using a one-step approach. The prevalence rate is the actual number of
cases in a population at a certain point in time. Some studies only report either point prevalence rates or “lifetime” rates, and a few studies report both point-prevalence and lifetime rates. Point prevalence is reported over various time frames, ranging from the present (or the time of interview) to within the last 12 months. Lifetime prevalence can be a problematic term, given that participants may have not yet reached the age of maximum risk for developing an eating disorder. In the studies reviewed in this chapter, lifetime prevalence connotes the number of individuals who ever met criteria for anorexia nervosa or bulimia nervosa. Retrospective reports of age of onset from adult samples suggest that full-syndrome status is not reached until mid-to late adolescence (16–18 years) or even young adulthood (18–21 years). In adult community samples, the mean age of onset of anorexia nervosa ranges from 16 to 19 years (Fairburn, Cooper, Doll, & Welch, 1999; Garfinkel et al., 1996; Walters & Kendler, 1995) and that of bulimia nervosa ranges from 18 to 20 years (Garfinkel et al., 1995; Kendler et al., 1991). Few eating disorders begin before age 10 or after age 25, and the rate of new cases climbs steadily in between those ages (Bushnell, Wells, Hornblow, Oakley-Browne, & Joyce, 1990; Lewinsohn, Striegel-Moore, & Seeley, 2000). Hence, “lifetime” prevalence, especially in young samples, should be adjusted to reflect these age-related patterns.
Prevalence Studies of Anorexia Nervosa and Bulimia Nervosa: Major Findings
Full-Syndrome Anorexia Nervosa
Full-syndrome current anorexia nervosa is not uncommon among adolescent girls, with rates ranging from 0% to 0.9%, depending on the study. Epidemiologically, it is undetectable among boys. It is difficult to discern a clear trend in prevalence rates for girls because of the considerable variations in methodology used across studies. Those studies in which no girls were identified with a diagnosis of current anorexia nervosa included relatively young samples (ages 11 to 16 years; McKnight Investigators, 2003; Santonastaso, Zanetti, Sala, & Favaretto, 1996; Wlodarczyk-Bisaga & Dolan, 1996). Rathner and
end p.266
PRINTED FROM Treating and Preventing Adolescent Mental Health Disorders (www.oup.com/amhi-treatingpreventing)
© Copyright Oxford University Press, 2006. All Rights Reserved
Table 13.4
Interview-based Studies of Prevalence of Anorexia Nervosa and Bulimia Nervosa in Community Samples of Adolescents| | | Subjects | Methods | | Study | Country (City or state) | | Age (years) | Sample Size (response rate in %) | | | AN | BN | Point | Lifetime | Point | Lifetime | | United States (Oregon) | 1 | 15–18 | F 797 M 711 | (61) | K-SADS | DSM-III-R | F M | | 0.74 | F M | 0.49 | 1.6 0.14 | | New Zealand (Dunedin) | 2 | 21 | F 469 M 492 | (84) | DIS | DSM-III-R | F M | 0.90 0.0 | | F M | 1.5 0.4 | | | Germany (Munich) | 3 | 14–24 | F 1488 M 1533 | (71) | M-CIDI | DSM-IV | F M | 0.30 | 1.0 0.1 | F M | 0.7 | 1.7 | Johnson, Cohen, Kasen, et al., 2002 a
| United States (New York) | 4 | 9–23 | F 366 M 351 | | DISC-I | DSM-IV | F M | 0.28 | | F M | 4.0 0.28 | | McKnight Investigators, 2003 | United States (Arizona, California) | 1 | 11–14 | F 1103 M 0 | | MEDE | DSM-IV | F | | | F | 0.37 | | | Great Britain | 5 | 5–15 | F/M 10,438 | DAWBA | DSM-IV | F/M 0.1–0.4 (current eating disorder) | | | |
|
a
Rates reported reflect the number of respondents who either at early adolescence (9 to 19 years) or mid-adolescence (11 to 23 years) assessment met criteria.
b
1 = school students; 2 = entire birth cohort 4/72–3/73; 3 = city residents; 4 = representative sample of two counties; 5 = national sample.
c
DAWBA, Development and Well-Being Assessment; DIS, Diagnostic Interview Schedule; DISC-I, Diagnostic Interview Schedule for Children; K-SADS, Kiddie Schedule for Affective Disorders and Schizophrenia; M-CICI, Munich-Composite International Diagnostic Interview; MEDE, McKnight Eating Disorder Examination.
d
DSM, Diagnostic and Statistical Manual for Mental Disorders (DSM III-R, 3rd, revised edition; DSM IV, 4th edition).
e
AN, anorexia nervosa; BN, bulimia nervosa; F, female; M, male; point, point prevalence, based on 3 months (McKnight Investigators, 2003), 12 months (Johnson, Cohen, Kasen, et al., 2002a; Lewinsohn et al., 1993; Newman et al., 1996; Wittchen et al., 1998), or “current” (i.e., at the time of the interview; Emerson, 2003).
end p.267
PRINTED FROM Treating and Preventing Adolescent Mental Health Disorders (www.oup.com/amhi-treatingpreventing)
© Copyright Oxford University Press, 2006. All Rights Reserved
Messner ( 1993) initially observed a 0.39% rate for anorexia nervosa, but found additional cases for a total rate of 0.58%, and reported that none of the current cases identified in their Italian sample were under age 15 years. In a randomly selected population sample of 15-year-old girls in Southern Norway, by contrast, about 4 in 100 current anorexia nervosa cases were found (Rosenvinge, Borgen, & Boerresen, 1999). An earlier study from the Mayo Clinic had a prevalence rate of 0.48% among 15-to 19-year-old girls, a rate higher than that for any other age group (Lucas, Beard, O'Fallon, & Kurland, 1991). In a large study of 14-to 24-year-old Germans, Wittchen, Nelson, and Lanchner ( 1998) oversampled 14-and 15-year-olds, which may explain why only 0.3% of current anorexia nervosa cases were identified among the female participants, despite the wider age range of the study sample. A relatively large number of girls with current anorexia nervosa (0.7%) was observed in a Zurich Canton sample of 14-to 17-year-olds. In this study, confirmatory interviews were conducted with a parent, rather than with the child, as in most other studies (Steinhausen, Winkler, & Meier, 1997). Denial is a hallmark of anorexia nervosa, and future studies are needed to determine whether parental reports result in better detection of anorexia nervosa than that from child self-report only. Finally, the highest point prevalence rate (past 12 months), 0.9%, was reported in a New Zealand study (Newman et al., 1996) and was based on diagnostic interviews when the girls were 21 years old. This well-controlled, longitudinal study of an entire birth cohort of children born between April 1972 and March 1973 had previously reported prevalence rates for major mental disorders, but not eating disorders, based on assessments at ages 11 years (Anderson, Williams, McGee, & Silva, 1987), 13 years (Frost, Moffitt, & McGee, 1989), 15 years (McGee et al., 1990), and 18 years (Feehan, McGee, & Williams, 1993; Feehan et al., 1994). It is unclear whether rates from earlier assessments would have been similar to those found in the other large-scale studies conducted in the United States (Johnson, Cohen, Kasen, & Brock, 2002; Lewinsohn et al., 1993) or the large Munich study (Wittchen et al., 1998) in which a majority of girls were younger than 21 years. For girls, lifetime rates (ever having met criteria for anorexia nervosa) range considerably, with rates as low as 3 in 100 (Whitaker et al., 1990) in a study using DSM III criteria (which are more restrictive for anorexia nervosa than later editions of the DSM) and as high as 3.9 (Graber et al., 2003).
Partial-Syndrome Anorexia Nervosa
Only a few studies have provided data about partial anorexia nervosa, and when such data are reported, typically rates for partial anorexia nervosa have exceeded those for full-syndrome anorexia nervosa (McKnight Investigators, 2003; Rathner & Messner, 1993; Wittchen et al., 1998). Wittchen and colleagues ( 1998) found an additional 1.3% of females and 0.4% of males who fell “just short one DSM IV criterion” (p. 116) for having ever experienced anorexia nervosa, for a combined lifetime rate of partial anorexia nervosa or anorexia nervosa of 2.3% in females and 0.5% in males. In the young sample from the McKnight study, five girls (0.37%) met criteria for current partial anorexia nervosa, defined as meeting all but the amenorrhea criteria (McKnight Investigators, 2003, p. 249). Despite the use of quite liberal criteria for partial anorexia nervosa (“at least one DSM IV criterion not met,” p. 385), Rosenvinge and colleagues ( 1999) found no cases of partial anorexia nervosa in their sample of 15-year-olds. Not surprisingly, rates of partial anorexia nervosa also seem to be correlated with age of the sample, with higher rates found in samples of older girls. For example, Rathner and Messner ( 1993) found no partial anorexia nervosa among the younger girls (ages 11 to 14 years), but found a rate of 1.3% among the older girls (15 to 20 years).
Full-Syndrome Bulimia Nervosa
Reported prevalence rates for current bulimia nervosa in girls vary considerably, and with the possible exception of age of the sample, these variations do not seem to reflect methodological differences across the studies. Only two studies found any current male cases and the reported rates were very low (Johnson, Cohen, Kasen, et al., 2002; Wittchen et al., 1998). Rates for females
end p.268
PRINTED FROM Treating and Preventing Adolescent Mental Health Disorders (www.oup.com/amhi-treatingpreventing)
© Copyright Oxford University Press, 2006. All Rights Reserved
(0.5% or less) were reported to be low in several European countries (e.g., Italy: Rathner & Messner, 1993; Switzerland: Steinhausen et al., 1997; Hungary: Szabo & Tury, 1991; the Netherlands: Verhulst et al., 1997; Poland: Wlodarczyk-Bisaga & Dolan, 1996) but not uniformly so (e.g., Norway: Rosenvinge et al., 1999; Germany: Wittchen et al., 1998). Studies reporting relatively higher point prevalence rates have tended to include older participants. Specifically, 12-month prevalence rates were 0.7% in the female Munich sample (ages 14 to 24; Wittchen et al., 1998), 1.5% in the female Dunedin sample (age 21; Newman et al., 1996), and an astonishingly high 4% in the U.S. sample of students from New York state (Johnson, Cohen, Kasen, et al., 2002). In this study, any student who had met criteria during either early adolescence or mid-adolescence (spanning the ages 9 to 23 years) was counted. Johnson, Cohen, Kasen, and Brook have published several reports on eating disorders in their sample, and in one of these they indicated that in early adolescence (mean age 13.8, range 9–19 years), the 12-month prevalence for bulimia nervosa among girls was 1.4% (Johnson, Cohen, Kotler, Kasen, & Brook, 2002). It further bears noting that in general, rates were higher in the studies that employed a one-time assessment strategy rather than the more common two-stage approach. Consistently, the studies found that bulimia nervosa is very rare among adolescent boys (Lewinsohn et al., 1993; Wittchen et al., 1998; Woodside et al., 2001). In most studies, there were no cases of current or lifetime male bulimia nervosa, and when such cases were detected, they were far less common than female cases.
Partial-Syndrome Bulimia Nervosa
As has been observed for anorexia nervosa, partial-syndrome bulimia nervosa is more common than full-syndrome bulimia nervosa (Johnson, Cohen, Kasen, et al., 2002; McKnight Investigators, 2003; Newman et al., 1996; Rathner & Messner, 1993; Rosenvinge et al., 1999; Wittchen et al., 1998). Rathner and Messner ( 1993) found no full-syndrome bulimia nervosa cases and reported that the two girls with partial
bulimia nervosa (0.39%) were among the older girls (>15 years). In the study conducted by the McKnight Investigators ( 2003), 1.2% of girls met criteria for partial bulimia nervosa, for a combined rate of partial bulimia nervosa and bulimia nervosa of 1.6%. A fairly inclusive definition of partial bulimia nervosa was used: all cases in which either binge eating and purging were less frequent than that required by DSM-IV or subjective binges occurred in the presence of at least weekly purging were reported (McKnight Investigators, p. 249). Using another broad definition, the presence of “all but one of the required symptoms,” Wittchen and colleagues ( 1998) identified 1.5% female partial bulimia nervosa cases and 0.6% male partial bulimia nervosa cases. The combined rate of partial bulimia nervosa and bulimia nervosa was 3.2% in females and 0.6% in males. Finally, in what may be the most inclusive definition, “at least one DSM IV criterion not met,” Rosenvinge at al. ( 1999) detected 1% of girls but no boys with partial bulimia nervosa. Although it is clear that a broader definition of bulimia nervosa results in increased rates, the lack of a systematic definition of partial bulimia nervosa makes it impossible to draw any conclusions about the prevalence of more broadly defined bulimia nervosa in adolescents.
Prevalence of Binge Eating Disorder
Only three studies have reported specifically on the prevalence of binge eating disorder (defined mutually exclusively from partial bulimia nervosa or eating disorder not otherwise specified). The McKnight Investigators ( 2003) reported that 0.59% of the 11-to 14-year-old girls in their study met 3-month point prevalence criteria for binge eating disorder. Using 12-month prevalence criteria, Johnson, Cohen, Kasen, et al. ( 2002) identified no male binge eating disorder cases and reported 0.55% of female binge eating disorder cases. The highest rates were found by Rosenvinge and colleagues ( 1999), who reported that 1.5% of the 15-year-old school girls and none of the boys studied met point prevalence (time frame not specified) for binge eating disorder. As described earlier, this particular study found relatively high rates for all eating disor
end p.269
doi:10.1093/9780195173642.003.0014
|
|
|
|
