The Intellectual Functioning of Eating Disorder Patients JOHN D. RANSEEN, PH.D., AND LAURIE L. HUMPHRIES, M.D. Abstract. This study investigated the intellectual functioning of a large group of eating disorder patients to examine two previously reported findings: (I) this population exhibits above-average general intellectual skills; and (2) a specific pattern of strength in verbal abilities. Standard intellectual testing of 100 consecutive inpatient females with eating disorder diagnoses was performed. Results indicate intellectual performance conforming to a normal distribution with no specific pattern of strengths or weaknesses. J. Am. Acad. Child Adolesc. Psychiatry, 1992,31,5:844-846. Key Words: eating disorder, intelligence, psychological testing. Research has found that the typical eating disorder patient is a young white female of middle- or upper-middle-class background (Crisp, 1980). Clinical lore has held that this typical patient is also intellectually bright and verbal (Bruch, 1979). Nevertheless, there is relatively little research to support this contention. Although studies have tended to find that eating disorder inpatients display intellectual functioning that is average to above-average, the sample sizes have generally been small or nonrandom. Wilbur and Colligan (1981) found an average full-scale IQ of 111 in a sample of 34 anorexic patients admitted to an inpatient residential treatment unit. Gordon et al. (1984) found a sample of 10 inpatient anorexic patients to have an average full-scale IQ of 113. Other inpatient studies using only select Wechsler subscales or other intelligence tests have appeared to support above-average intelligence in these patients (Small et aI., 1983; Smart et aI., 1976; Witt et aI., 1985). In addition to high general IQ, one study found that anorexic patients tend to have relative strength in verbal intellectual skills in comparison with deficient visuospatial reasoning abilities (Maxwell et aI., 1984). Our clinical experience over the past few years led us to question the assumption that eating disorder patients have above-average intellectual functions or any unusual pattern of intellectual ability. Consequently, the present study assessed the general intellectual functioning of a large group of eating disorder inpatients to examine this issue and to assess potential differences between eating disorder diagnostic groups. Method During a 3-year period, 100 consecutive female patients admitted to an eating disorder inpatient unit completed psychological testing, including a standard intelligence test.
Accepted October 15, 1991. Dr. Ranseen is Assistant Professor, Departments of Psychiatry, Neurology, and Psychology, and Dr. Humphries is Associate Professor, Department of Psychiatry, University of Kentucky, Lexington. Supported in part by the NIMH Child and Adolescent Mental Health Academic Award (K07MH00767). Reprint requests to Dr. Ranseen, Department of Psychiatry, Annex tt. Rm 206, U. ofKentucky College ofMedicine, Lexington, KY 4053600. 0890-8567/92/3105-0844$03.00/0©1992 by the American Academy of Child and Adolescent Psychiatry.
This evaluation was performed as part of a routine workup. The Wechsler Intelligence Scale for Children-Revised (WISC-R) (Wechsler, 1974) was completed by patients under age 16 (N = 22) and the Wechsler Adult Intelligence Scale-Revised (WAIS-R)(Wechsler, 1981) by those 16 and older (N = 78). Referral for psychological testing included all first time admissions to the eating disorder unit who had primary diagnoses of anorexia nervosa (N = 38), bulimia nervosa (N = 54) or eating disorder, not otherwise specified (N = 8), based on DSM-III-R (American Psychiatric Association, 1987) criteria. The eating disorder unit is part of a primary care facility within an urban center and large rural catchment area. Results This population has a mean age of 21.5 years and 11.4 years of completed school. The low mean educational level is somewhat misleading as it incorporates a number of younger students, most of whom are at grade level. Further, if one examines the older patients (18 years of age and older, N = 57) we find that 51% are attending or have attended some college, 35% completed high school, and 14% have less than a high school education. U.S. population statistics for women indicate academic achievement of 35% attending some college, 42% completing high school, and 23% achieving less than a high school education (Digest ofEducational Statistics, 1989). Consequently, this population has more education than the norm for women in this country. Our population is somewhat unique in that a fairly large number of these women come from very rural Appalachian areas. Based on county of residence, 43% of our sample resides in a nonmetropolitan area (based on University of Kentucky Agriculture Economy statistics). Further, a rating of metropolitan/nonmetropolitan (1 = metropolitan; 9 = isolated from metropolitan area: Bureau of the Census, 1990) finds that our sample has a mean rating of 3.8 (SD = 2.4, range, 2 to 9). There was no correlation between degree of metropolitan/nonmetropolitan residence and level of education (r = -0.17) or full-scale IQ (r = -0.13). Demographic information and intelligence test scores by diagnostic group are provided in Table 1. As can be seen, the bulimia nervosa group is slightly older and better educated. Nevertheless, all three groups display general intellectual functioning (full-scale IQ) and summary measures of verbal and performance IQ that conform to population norms for the Wechsler scales (mean = 100, SD = 15). One-way J. Am. Acad. Child Adolesc. Psychiatry, 31:5, September 1992
TABLE 1. Mean Age, Education, and IQ Scores by Diagnostic Group Anorexia Nervosa (N = 38) Age Education (years complete) Verbal IQ Performance IQ Full-scale IQ
Bulimia Nervosa (N = 54)
Total = 100) Range
10.77 100.76 98.16 99.47
(2.62) (16.21) (16.85) (16.55)
12.49 101.22 101.72 102.09
(2.41) (13.54) (14.65) (14.27)
9.75 97.75 95.75 96.63
(1.67) (14.11) (14.14) (13.92)
4-20 74-139 72-136 75-133
analyses of variance reveal no significant differences between diagnostic groups on these summary measures. As the WISC-R and WAIS-R involve different normative groups, Table 2 provides scaled and summary IQ scores for these tests. The WISC-R group (younger than age 16) does exhibit summary IQ scores that are slightly above average (fullscale IQ = 106.4). This may be misleading, however, as the most recent revision and renorming of the WISC (WISC III; Wechsler, 1991) indicates that the WISC-R summary scores are on average 5 points above the renormed scale. Cumulative frequencies of verbal-performance IQ split across the entire population (Table 3) reveal no tendency for these patients to have a higher verbal IQ or to be more likely to exhibit a verbal over performance IQ discrepancy. These patients display virtually the same likelihood of having a large verbal-performance IQ split, as described in the WAIS-R standardization sample (Matarazzo and Herman, 1984).
Discussion These data suggest that eating disorder patients display a range of general intellectual functioning that roughly conforms to the population at large. This tends to refute a common assumption, supported by a few earlier studies, that eating disorder patients are intellectually brighter than the normal population. This study is similar to those previously reviewed that also tested intellectual functioning in populaTABLE 2. Mean WAIS-R and WISC-R Subscale and Summary IQ Scores WAIS-R (N Information Digit Span Vocabulary Arithmetic Comprehension Similarities Picture Completion Picture Arrangement Block Design Object Assembly Digit symbol (Coding) Verbal IQ Performance IQ Full-scale IQ
Eating Disorder, NOS (N = 8)
8.50 9.84 9.29 8.88 10.19 9.61 9.09 9.61 9.08 9.18 10.40 99.44 98.29 99.16
(2.43) (2.59) (2.92) (2.47) (2.64) (2.95) (2.67) (2.58) (2.71) (2.87) (2.52) (14.13) (14.27) (14.79)
WISC-R (N = 22) SD
9.59 11.31 10.65 11.55 11.35 11.l4 10.36 9.77 10.45 11.06 12.62 105.86 105.86 106.41
(2.56) (2.82) (2.72) (3.25) (3.17)
(3.50) (3.51) (2.84) (3.23) (4.14) (2.75) (15.35) (17.18) (15.33)
J. Am. Acad. Child Adolesc. Psychiatry, 31:5, September 1992
tions of inpatient eating-disordered females. It was, however, a much larger sample than these previous studies and involved consecutive admissions rather than a select group of patients. Consequently, it would seem that this sample more accurately describes the eating disorder population. The primary methodological problem with the current study is the lack of a comparison group of noneating-disordered inpatient females. Such a group might well exhibit even lower general intellectual functioning than the eating disorder patients. This, however, would not negate the conclusion that eating disorder patients do not necessarily constitute a group of patients with above average intellectual skill. The previously mentioned studies that formally assessed intellectual functioning in these patients were all completed a number of years ago. One explanation for the discrepancy between our results and these previous studies is the finding that the demographics of eating disorder have been changing over the past 20 years (Crisp, 1980; Pope et al., 1987). Although in the past these patients were generally found among white, upper-middle-class, urban families (Bruch, 1979), it has become evident that they now come from much more varied backgrounds. It is not clear whether this demographic change reflects an increase and actual broadening of the types of individuals susceptible to this disorder or may simply reflect an increased recognition of eating disorder in populations not previously examined. In any case, these demographic changes may account for the normal distribution of intelligence described in this study. It was quite evident that our population includes a wide mix of patients from urban and rural areas of all socioeconomic classes. Interestingly, although we often view our rural patients as academically deprived (and perhaps less intelligent), there was nothing to suggest any relationship between rural residence and either low intelligence or fewer years of education. In fact, patients coming from the most rural and isolated counties in Appalachia exhibited an average level of intelligence. These data also suggest that there are no fundamental differences in intelligence scores across categories of eating disorders. What is remarkable is the similarity in scores between these diagnostic groups. There was also no support for the contention that there may be some fundamental asymmetry of cognitive functions (verbal comprehension versus visuospatial skills) in eating disorder patients, at least as measured by intelligence tests. In fact, Matarazzo and Herman (1984), found that 23.3% of the normal population
RANSEEN AND HUMPHRIES TABLE
3. Cumulative Frequency of Eating Disorder Patients Verbal-Performance IQ Split (N = /00) Verbal> Performance (%)
Performance> Verbal (%)
will exhibit a verbal-performance IQ split greater than 13 points, which is quite similar to our findings that 26% of these eating disorder patients exhibited this magnitude of difference. It is possible that more detailed neuropsychological testing might reveal some specific cognitive strengths and deficits in these patients. It is interesting to note that, although there was no asymmetry of general cognitive functions, as measured by an intelligence test, our population did exhibit the same relative strength (Digit Span, Digit Symbol) and weakness (Information) on intelligence test subscales, as has previously been described in studies that conducted formal evaluation (Gordon et al., 1984; Witt et al., 1985). This suggests a relative strength in rote tasks involving immediate recall and perceptual-motor speed, but surprising weakness in general knowledge. In summary, this relatively large sample of eating disorder patients contradicts previous research with smaller, and often nonrandom, samples that indicated that eating disorder patients tend to be brighter and have better developed verbal skills. Although many of these young women are bright and verbal, many are not. This is of some clinical relevance since, as noted in a study by Dura and Bornstein (1989), eating disorder patients' perfectionistic strivings are often apparent in achievement goals that appear quite inconsistent with actual intellectual ability.
Clinically, we have also observed a number of our patients who appear reasonably bright in conversation whose tested intellectual and academic skills are actually quite modest or below average. Often these young women are not doing particularly well in school, but this is assumed to be the result of emotional problems and difficulties related to their eating disorder. Further, they express vocational goals that seem unrealistic, given their tested abilities. For instance, it has not been unusual for us to obtain an IQ of 90 in a young woman doing marginal college work whose firm goal is to attend medical school. Falling short of these implausible expectations is clearly a major stressor in the life of this individual. Issues of separation from family are often high-
lighted in the treatment of eating disorder patients. This sometimes overlooks the difficult developmental tasks involved with finding an appropriate vocation. Restructuring of such expectations can be a difficult and painful task for the patient, but it may be quite necessary if goals are grossly out of line with abilities. Formal testing of intellectual ability and other cognitive skills can help formulate realistic goals, particularly as the assumption that eating disorder patients are bright and capable is not always supported by the test findings. References American Psychiatric Association (1987), Diagnostic and Statistical Manual ofMental Disorders, 3rd edition-revised DSM-ll/-R. Washington, DC: American Psychiatric Association. Bruch, H. (1979), The Golden Cage. Cambridge: Harvard University Press. Bureau of the Census, (1990), Statistical Abstracts of the United States, Washington, DC: Government Printing Office. Crisp, A. H. (1980). Anorexia Nervosa: Let Me Be. London: Academic Press. Digest of Education Statistics (1989), Washington, DC: U.S. Dept. of Commerce, Bureau of the Census. Dura, J. R. & Bornstein, R. A. (1989), Differences between IQ and school achievement in anorexia nervosa. J. Clin. Psychol., 45:433435. Gordon, D. P., Halmi, K. A., & Ippolito, P. M. (1984), A comparison of the psychological evaluation of adolescents with anorexia nervosa and of adolescents with conduct disorders. J. Adolesc., 7:245266. Matarazzo, T. D. & Herman, D. O. (1984), Base rate data for the WAIS-R: Test-retest reliability and VIQ-PIQ differences. J. Clin. Neuropsychol., 6:351-366. Maxwell, J. K., Tucker, D. M., & Townes, B. D. (1984), Asymmetric cognitive function in anorexia nervosa. Int. J. Neurosci., 24:37-44. Pope, H. G., Champoux, B. S., & Hudson, J. I. (1987), Eating disorder and socioeconomic class: Anorexia nervosa and bulimia in nine communities. J Nerv. Ment. Dis., 175:620-623. Small, A., Madero J., Teagno, L., & Ebert M. (1983), Intellect, perceptual characteristics and weight gain in anorexia nervosa. J. Clin. Psychol., 39:780-782. Smart, D. E., Beumont, P. J. V., & George, G. C. W. (1976), Some personality characteristics of patients with anorexia nervosa. Br. J. Psychiatry, 128:57-60. Wechsler, D. (1974), WISC-R: Manual. San Antonio, Texas: The Psychological Corporation. - - (1981), WAlS-R: Manual. New York: The Psychological Corporation. - - (1991), WISC-ll/ Manual. San Antonio, Texas: The Psychological Corporation. Wilber, C. J. & Colligan, R. C. (1981), Psychologic and behavioral correlates of anorexia nervosa. Developmental and Behavioral Pediatrics, 2:89-92. Witt, E. D., Ryan, C., & Hsu, L. K. G. (1985), Learning deficits in adolescents with anorexia nervosa. J. Nerv. Ment. Dis., 173:182184.
J.Am. Acad. Child Adolesc. Psychiatry, 31:5, September 1992