Anda di halaman 1dari 4

Journal of Psychosomatic Research 74 (2013) 175–178

Contents lists available at SciVerse ScienceDirect

Journal of Psychosomatic Research

Eating disorder treatment among women forty and older: Increases in prevalence
over time and comparisons to young adult patients ☆
Diann M. Ackard a, b,⁎, Sara Richter b, Maria J. Frisch c, Deborah Mangham b, Catherine L. Cronemeyer b
Private Practice, 5101 Olson Memorial Highway, Golden Valley, MN 55422, USA
Park Nicollet Melrose Institute, 3525 Monterey Drive, St. Louis Park, MN 55416, USA
Park Nicollet Research Institute, 3800 Excelsior Boulevard, St. Louis Park, MN 55416, USA

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: There is limited information on the prevalence of middle-aged women seeking specialized treatment
Received 17 August 2012 for an eating disorder and whether middle-aged patients are significantly different from young-adult patients.
Received in revised form 17 October 2012 This two-part study sought to identify changes in the past two decades in the prevalence of middle-aged (MA;
Accepted 20 October 2012 40+ years) and young-adult (YA; 18–39 years) women seeking treatment for an eating disorder (ED) and to
identify differences and similarities between both groups.
Methods: For Study 1, all unique female inpatient admissions from 1989 to 2006 were reviewed (n=1,040). For
Anorexia nervosa
Bulimia nervosa
Study 2, women admitted to any treatment level from January–May 2007 were compared, based on age at
Eating disorder not otherwise specified intake admission, on psychological questionnaires and factors relevant to an eating disorder.
Middle aged adult Results: In Study 1, the overall percent of MA women who presented for inpatient ED treatment increased sig-
Older adult nificantly from an average of 4.7% (1989–2001) to an average of 11.6% (2002–2006). In Study 2, at intake, MA
Treatment women were more likely than YA to be married, be older at ED onset and report a longer duration of illness.
Self-esteem, depression, anxiety, ED psychopathology, and BMI were not significantly different between groups.
Conclusions: Findings indicate an increase in the prevalence of inpatient admissions among middle-aged women,
but few differences between middle-aged and younger-aged women at treatment admission. However, the lon-
ger duration of illness among MA warrants in-depth investigation of factors related to resistance to seeking treat-
ment and to existing treatments failing patients, and consideration of tailoring treatment to course of illness.
© 2012 Elsevier Inc. All rights reserved.

Introduction age at eating disorder onset, longer duration of ED illness, greater so-
cial isolation, lower body shape concerns, and less internalization of
There are limited studies examining the prevalence of middle-aged the sociocultural ideal at intake assessment than those presenting
(MA) women seeking treatment for an eating disorder (ED). Scholtz for treatment in young adulthood. There were no differences in anx-
and colleagues (2010) examined all admitted patients to an ED treat- iety or depression scores at admission or on ED outcomes at discharge
ment program within the past decade and found a 1% prevalence of between age groups [3]. In a separate study, issues related to family
women presenting for treatment at 50 years of age or older [1]. Yet, an- (e.g., nonsexual trauma, controlling spouse or partner) and health
other ED treatment program reported a 400% increase of patients were important factors underlying the predisposition to, and mainte-
presenting for treatment at age 40 or older during the same decade [2]. nance of, ED among the those patients 40 years of age and older, but
There is also a dearth of research comparing MA to young adult not those younger than 40 years [4].
(YA) women seeking treatment. In one study comparing patients Thus, this two-part study aims to fill two gaps in the literature re-
35 years old or older at intake to a residential eating disorders treat- garding MA women seeking ED treatment. First, we sought to identify
ment program, those presenting for treatment in midlife had an older the prevalence of inpatient admissions by year and by age at admission
(18–39 years vs. 40 years or older) across an 18-year period for adult
females seeking treatment for an ED (Study 1). We hypothesized an in-
crease in admissions among women 40 and older over time. Second, we
aimed to identify descriptive and psychological differences at intake
☆ Work was conducted at: Park Nicollet Melrose Institute, 3525 Monterey Dr, St. Louis
across outpatient, partial hospital and inpatient admissions between
Park, MN 55416, USA.
⁎ Corresponding author at: 5101 Olson Memorial Highway Suite 4001, Golden Valley,
those aged 18–39 and those 40 years of age and older (Study 2). We hy-
MN 55422, USA. Tel.: +1 763 595 7294x111; fax: +1 763 595 7293. pothesized that MA women seeking treatment would be more psycho-
E-mail address: (D.M. Ackard). logically compromised than their YA peers.

0022-3999/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
176 D.M. Ackard et al. / Journal of Psychosomatic Research 74 (2013) 175–178

Methods the interview version of the Eating Disorders Examination and has
four subscales: Restraint, Eating Concern, Weight Concern, and Shape
Design Concern. Mean scale scores (and standard deviations) among young
Australian adult women are: 1.30 (1.40) for Restraint, 0.76 (1.06) for
For both studies, eligible individuals were adult females 18 years Eating Concern, 1.79 (1.51) for Weight Concern, and 2.23 (1.65) for
of age and older who provided active consent at intake assessment Shape Concern [7]; mean scores (and standard deviations) among
for research-related medical records reviews. Admission criteria were undergraduate women in the United States are similar: 1.29 (1.41) for
based upon the American Psychiatric Association's (APA) Guidelines Restraint, 0.87 (1.13) for Eating Concern, 1.89 (1.60) for Weight
for the Treatment of Eating Disorders and included diagnoses of Concern, and 2.29 (1.68) for Shape Concern [8]. Higher scores indicate
Anorexia Nervosa (AN), Bulimia Nervosa (BN) and Eating Disorder greater pathology across the four subscales. The EDE-Q has been
Not Otherwise Specified (EDNOS) as described in the Diagnostic found to have good concurrent validity and criterion validity among
and Statistical Manual of Mental Disorders, 4th Edition [5]. This study community samples [9] and good internal consistency [9,10].
was reviewed and approved by the Park Nicollet Health Services Institu- The Body Image Assessment (BIA) instrument assesses perception
tional Review Board. Participants were not compensated for their of current and ideal body shape with established reliability and valid-
participation. ity [11]. Participants select one female body silhouette perceived to
represent their current body size (CBS), and one silhouette for ideal
Study 1 body size (IBS); figures range from 1 (thinnest) to 9 (heaviest).
From 1989 to 2006, all unique inpatient admissions at the Park The Beck Depression Inventory-2 (BDI-II; [12]) is a 21-item self-
Nicollet Melrose Institute, St. Louis Park, MN (USA) were captured by report instrument designed to measure depression severity using a
data query of an administrative database containing medical record in- 4-point Likert scale (responses from 0 to 3). Total scores range from
formation for all women aged 18 or older. Data collected were year of 0 to 63, and a higher score indicates a higher level of depression. Clinical
admission, age at admission and ED diagnosis. The number of inpatient severity categories are: 0–13 minimal, 14–19 mild, 20–28 moderate,
admissions ranged from a low of 24 in 1993 to a high of 170 in 2006, for and 29–63 severe [12]. Psychometric properties that assess the reliabil-
a total of 1,040 unique admissions across the 18-year study period. Each ity and validity of the BDI-II for measuring depression have been found
patient was only included in the year of her first inpatient admission to to be strong among adolescent [13] and adult populations [14].
this facility; all subsequent admissions were excluded from the data. Participants also completed the Rosenberg Self-Esteem Scale (RSES;
Most participants were between the ages of 18 and 39, inclusive (n = [15]), a 10-item questionnaire that assesses overall self-esteem and
950; 91%). Diagnostic presentation for the entire sample, as determined self-worth; scores range from 10 to 40, and higher scores indicate great-
by clinical psychiatric interview, was 41.5% AN, 29.1% BN and 29.3% er self-esteem. Reliability and validity estimates have been published
EDNOS. Please note that patients receiving inpatient treatment for elsewhere [16] and the measure demonstrates adequate psychometric
Binge Eating Disorder (BED) are included under EDNOS, although pa- properties.
tients with BED are rarely hospitalized at this facility, and the BED diag- The State-Trait Anxiety Inventory (STAI; [17]) is a 40-item self-
nosis was not recognized in the early years of this study. report assessment that includes separate measures of state and trait
anxiety. The total score ranges from 20 to 80 for each scale (trait versus
Study 2 state anxiety), and higher scores indicate greater anxiety. Clinical
To further understand the subset of MA women presenting for treat- severity categories, based on score, are: 20–30 low, 31–40 low average,
ment, data were abstracted from medical charts of all women aged 18 41–48 average, 49–60 high average, 61–70 high, and 71–80 very high
or older (range 18 to 64 years) who completed an intake assessment [17]. Psychometric properties of the STAI are sound, with test–retest
for any level of treatment between January 1, 2007 and May 31, 2007 among male and female high school and college students ranging
(n = 164) at the Park Nicollet Melrose Institute, St. Louis Park, MN from .65 to .86 for trait anxiety and .16 to .62 for state anxiety. This
(USA). Most were Caucasian (n = 141, 86.0%) and reported their marital low level of stability for the state-anxiety scale is expected since re-
status as single (n = 115; 70.1%). The vast majority of women were be- sponses to the items on this scale are thought to reflect the influence
tween the ages of 18 and 39, inclusive (n = 134, 81.7%). Over one-half of whatever transient situational factors exist at the time of testing
received outpatient (n = 86; 53.1%) and the remainder received inpa- [17]. The validity correlations are .80 with the Taylor Manifest Anxiety
tient (n = 73; 45.1%) or partial hospitalization (n = 3, 1.8%) treatment. Scale, .75 with the IPAT Anxiety Scale, and .52 with the Multiple Affect
Diagnostic presentation for the entire sample was 20.7% AN, 23.8% BN Adjective Check List [17].
and 55.5% EDNOS. The Park Nicollet Melrose Institute provides a sepa-
rate program for the treatment of Binge Eating Disorder, therefore Data analysis
these patients were excluded from study sample 2.
The results were stratified into two groups by age at intake assess-
Materials ment (18–39, inclusive, versus 40 years or older). Demographic data
including race, marital status, initial assessment treatment recom-
The following information was collected at intake and later abstract- mendation and ED diagnosis were analyzed using the Chi-Square
ed from the medical chart: race, date of birth, marital status, date of ad- Test while a 2-sample t-test with Satterthwaite's approximation for
mission to the treatment facility, age at admission, treatment milieu unequal variances was used to test for differences in age at onset, dura-
(inpatient, partial hospital, intensive outpatient, outpatient), ED diagno- tion of illness and BMI between groups. Differences between question-
sis (AN, BN, EDNOS), self-reported age at onset of ED, and anthropo- naire scores at intake assessment were evaluated using multivariate
metrically measured height and weight as used to calculate BMI. analyses of variance (MANOVA) adjusting for age of eating disorder
In addition, the following psychological measures, described below, onset, duration of illness and BMI. The significance level was set at
were collected at intake: Eating Disorders Examination—Questionnaire, p b .05. Statistical analyses were conducted with SAS version 9.3.
Body Image Assessment, Beck Depression Inventory II, Rosenberg Self-
Esteem Survey, and State-Trait Anxiety Inventory. Results
The Eating Disorders Examination—Questionnaire (EDE-Q; [6]) is a
Study 1—Prevalence of eating disorder admissions by admission age from 1989 to 2006
41-item self-report questionnaire that asks specific questions per-
taining to the presence and frequency of ED behaviors, thoughts, and Inspection of prevalence rates, stratified by year (1989–2006) and age (18–
feelings about body over the past 28 days. The EDE-Q is derived from 39 years versus 40 years or older), shows that the rates of MA women presenting for
D.M. Ackard et al. / Journal of Psychosomatic Research 74 (2013) 175–178 177

Fig. 1. Percent of women 40 years of age or older admitted for inpatient eating disorders treatment by year of admission.

inpatient treatment for an ED vary considerably (from 0 to 13%), but are increasing Findings from the current study are similar to the population-based
(see Fig. 1). Specifically, the overall percent of MA women who presented for inpatient
results by Hay and colleagues on the increase in ED symptoms among
ED treatment increased from an average of 4.7% in 1989 to 2001 to an average of 11.6%
from 2002 to 2006 (Chi-square = 15.38; p b .0001). Diagnostic presentation was signif- older women over time [18] and high prevalence rate of eating disor-
icantly different among MA versus YA women (see Table 1; overall Chi-square = 9.16, ders and body dissatisfaction as well as concerns about eating, weight
p = .0103); analyses of individual associations show that MA patients were less likely and shape in a nonclinical sample of older women [19–21]. Yet the cur-
than YA patients to present with BN for inpatient treatment; all other associations be- rent study results are novel in that they show an increase in diagnosed
tween diagnoses by age group were nonsignificant.
and hospitalized MA women with ED, underscoring the gravity of disor-
dered eating among this age group. While it is possible that the increase
Study 2—Psychological comparisons between young adult and middle-aged women in women seeking treatment for a diagnosed eating disorder is associat-
In the second study of women receiving ED treatment across outpatient, partial hos-
ed with changes in the demographic composition of the community
pital and inpatient programs, there were no differences between groups on race, treat- surrounding the treatment center, data on catchment area factors
ment milieu or ED diagnosis at admission, but MA women were more likely than YA were not collected as part of the current study; thus, the specific reasons
women to be married (60.0% vs. 16.4%, respectively; pb .0001), to report an older age of for the witnessed increase remain unknown.
ED onset (21.4 years vs. 17.0 years; p=.0362) and to have a longer duration of ED illness
Although results from the current investigation are similar to
(26.5 years vs. 8.1 years; pb .0001) (see Table 2). Body mass index was not statistically
significant between groups (YA=21.1 vs. MA=23.3; p-value=0.1514). There were no other studies in that they show more similarities than differences be-
significant differences between groups on self-esteem, body image perception (current tween MA and YA women at admission for ED treatment [2,3,22], we
or desired), depression, anxiety, or ED cognitions and behaviors or ED-related psychopa- believe that our understanding of the needs of MA women seeking
thology (Wilk's lambda p-value=0.3415) after adjusting for age of ED onset, duration of treatment is in its nascence. For example, given that several studies
illness and BMI.
have found that duration of illness is significantly longer among MA
versus YA patients, what are we to understand about some women's
Discussion resistance to seeking treatment earlier and the factors that maintain
their eating and body image disturbances over decades? Alternatively,
This study sought to explore the prevalence of inpatient admissions among those who did seek earlier treatment, what do we need to learn
for ED treatment by year and by age at admission across an 18-year pe- about their past treatments that have failed to help these women recov-
riod for adult females, and then in a subsequent study to identify de- er? Or, for example, in studies where the age at ED onset was significant-
scriptive and psychological differences between MA and YA women. ly older among the MA versus YA patients (average age at ED onset 23.6
Findings from the first study indicate an increase in the prevalence of in- vs. 16.0 years, respectively) [3], what precipitating factors are relevant to
patient admissions among middle-aged women (40 years old or older) women whose ED symptoms begin later than what is typically reported
seeking treatment for an ED. This mirrors findings from another treat- in the literature?
ment center that reported remarkable increasing prevalence of this
population subset [2]. The second study evaluated differences between
age groups on descriptive and psychological measures at intake and
Table 2
found that MA women seeking ED treatment were more likely to be Study 2: Comparisons at intake evaluation between eating disorder patients by age at
married, have an older age at ED onset and a longer duration of ED ill- intake: means (standard deviations)
ness compared to their younger adult peers. Groups were strikingly
18–39 years old 40 years old or older
similar in other areas of presentation.
Age of ED onset (years) 17.0 (4.0) 21.4 (9.9)
Duration of illness (years) 8.1 (6.6) 26.5 (9.3)
BMI 21.1 (4.8) 23.3 (7.9)
Table 1 Rosenberg self-esteem scale 24.3 (6.7) 24.5 (5.7)
Study 1: Percent of patients in each diagnostic category by age range (18–39 years old Body image assessment—current 4.5 (1.9) 4.5 (2.1)
versus 40 years old and older) Body image assessment—desired 2.5 (0.9) 2.6 (1.0)
Beck Depression Inventory 27.2 (14.7) 27.6 (11.7)
Age category
EDE-Q Dietary Restraint 3.6 (1.8) 3.6 (1.6)
ED diagnosis 18–39 years, inclusive 40 years and older EDE-Q Eating Concerns 3.4 (1.5) 3.9 (1.4)
AN 393 (41.4%) 39 (43.3%) EDE-Q Weight Concerns 3.9 (1.7) 4.0 (1.4)
BN 288 (30.3%) 15 (16.7%) EDE-Q Shape Concerns 4.4 (1.6) 4.8 (0.9)
EDNOS 269 (28.3%) 36 (40.0%) STAI State Anxiety 55.2 (13.7) 58.3 (12.6)
Total 950 90 STAI Trait Anxiety 54.9 (13.3) 56.3 (11.4)
178 D.M. Ackard et al. / Journal of Psychosomatic Research 74 (2013) 175–178

Kally and Cumella reported on 100 women in residential ED treat- oversight of the study details and helped to revise the article. D.M.
ment and grouped their patients by age at ED onset (b40 years versus served as Principal Investigator of Study 1 and helped to draft the
40 years or older). Their research explored precipitating factors of article. C.C. coordinated data for both studies and helped to draft the
EDs and more general psychopathology; however, only issues related article. All authors read and approved the first and revised drafts of
to current family (e.g., nonsexual trauma, spousal/partner control) the article.
and health issues were found to be important precipitants for the
older-onset compared to the typical-onset patients [4]. With respect
Funding source
to tailoring treatment for this population, Forman and Davis, in their
sample of women seeking residential treatment, found that older pa-
These studies were supported in part by the Park Nicollet Institute.
tients (35 years old or older at admission) rated a grief and loss group
However, the Park Nicollet Institute was not involved in the study de-
and family therapy as significantly more helpful than younger patients
sign, the collection, analysis and interpretation of data, the writing of
(under 35 years of age) [3], although this same study found no differ-
the report, or in the decision to submit the article for publication.
ences between age groups on outcome measures. Are interpersonal
and relationship issues core factors that maintain eating disturbances
in mature women? Continued research is significantly warranted to References
understand how best to serve MA women seeking treatment. Promising
[1] Scholtz S, Hill LS, Lacey H. Eating disorders in older women: does late onset an-
avenues of future works include encouraging treatment at an earlier age orexia nervosa exist? Int J Eat Disord 2010;43:393-7.
so as to shorten duration of illness and discouraging the development of [2] Cumella EJ, Kally Z. Profile of 50 women with midlife-onset eating disorders. Eat
food/body “relationships” over connections with trusted others. Disord 2008;16:193-203.
[3] Forman M, Davis W. Characteristics of middle-aged women in inpatient treat-
ment for eating disorders. Eat Disord 2005;13:231-43.
Strengths and limitations [4] Kally Z, Cumella EJ. 100 Midlife women with eating disorders: a phenomenologi-
cal analysis of etiology. J Gen Psychol 2008;135:359-77.
There are several strengths to these studies. A chart review allows [5] American Psychiatric Association. Diagnostic and statistical manual for mental
disorders. 4th ed. Washington, DC: American Psychiatric Association; 2004.
researchers to obtain a broad range of descriptive and other infor- [6] Fairburn CG, Cooper Z. The Eating Disorder Examination. In: Fairburn CG, Wilson
mation routinely administered at intake evaluation in an efficient GT, editors. Binge eating: nature, assessment and treatment. 12th edition. New York:
manner, and offers the least disruption to the patient. An additional Guilford Press; 1993.
[7] Mond JM, Hay PJ, Rodgers B, Owen C. Eating Disorder Examination Questionnaire
strength is the availability of psychometrically sound measures com- (EDE-Q): norms for young adult women. Behav Res Ther 2006;44:53-62.
pleted at intake evaluation across a range of constructs including, but [8] Luce KH, Crowther JH, Pole M. Eating Disorder Examination Questionnaire
not limited to, ED. However, chart review studies are limited to the (EDE-Q): norms for undergraduate women. Int J Eat Disord 2008;41:273-6.
[9] Mond JM, Hay PJ, Rodgers B, Owen C, Beaumont PJ. Temporal stability of the Eating
information that is included in the patient's chart; thus, missing Disorders Examination Questionnaire. Int J Eat Disord 2004;36:195-203.
data, such as the number of BED patients within the EDNOS diagnos- [10] Luce KH, Crowther JH. The reliability of the Eating Disorder Examination Self-
tic category for Study 1, cannot be obtained and further questions Report Questionnaire Version (EDE-Q). Int J Eat Disord 1999;25:349-51.
[11] Williamson DA, Davis CJ, Bennett SJ, Goreczny AJ, Gleaves DH. Development of a
cannot be asked without additional patient consent. Finally, the ex-
simple procedure for assessing body image disturbances. Behav Assess 1989;11:
clusion of BED patients in Study 2 limits our understanding of poten- 433-46.
tial differences at intake assessment between YA and MA patients [12] Beck AT, Steer RA, Brown GK. Manual for the Beck Depression Inventory—II. San
Antonio, TX: Psychological Corporation; 1996.
presenting for BED treatment.
[13] Osman A, Kopper B, Barrios F, Gutierrez P, Bagge C. Reliability and validity of the Beck
Depression Inventory—II with adolescent psychiatric inpatients. Psychol Assess
Conclusions 2004;16:120-32.
[14] Robinson JP, Shaver PR, Wrightsman LS. Measures of personality and social psycho-
logical attitudes. San Diego, CA: Academic Press; 1991.
The percentage of inpatient admissions for middle-aged women [15] Rosenberg M. Society and the adolescent self-image. Princeton, NJ: Princeton Univer-
has increased over an 18-year time period. Although there are more sity Press; 1965.
similarities than differences between middle-aged and young adult [16] Fleming J, Courtney B. The dimensionality of self-esteem: hierarchical facet model
for revised measurement scales. J Pers Soc Psychol 1984;46:404-42.
female patients, there is also a significant dearth in our understanding [17] Spielberger CD. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting
of the factors that serve to precipitate and maintain eating distur- Psychologists Press; 2003.
bances among this population. Future research is indicated, particu- [18] Hay PJ, Mond JM, Buttner P, Darby A. Eating disorder behaviors are increasing: find-
ings from two sequential community surveys in South Australia. Public Libr Sci
larly in the areas of early intervention and treatment outcome. 2008;3:e1541.
[19] Mangweth-Matzek B, Rupp CI, Hausmann A, Assmayr K, Mariacher E, Kemmler G,
Competing interest statement et al. NEver too old for eating disorders or body satisfaction: a community study
of elderly women. Int J Eat Disord 2006;39:583-6.
[20] Gagne DA, Von Holle A, Brownley KA, Runfola CD, Hofmeier SM, Branch KE, et al. Eat-
All authors have completed the Unified Competing Interest form ing disorder symptoms and weight and shape concerns in a large web-based conve-
and declare that they have no competing interests to report. nience sample of women ages 50 and above: results of the gender and body image
(GABI) study. Int J Eat Disord 2012;45(7):832-44.
[21] Runfola CD, Von Holle A, Trace SE, Brownley KA, Hofmeier SM, Gagne DA, et al.
Authors contributions Body dissatisfaction in women across the lifespan: results of the UNC-SELF and
Gender and Body Image (GABI) studies. Eur Eat Disord Rev 2012, http://dx.doi.
D.A. served as Principal Investigator of Study 2 and contributed to org/10.1002/erv.2201.
[22] Guerdjikova AI, O'Melia AM, Mori N, McCoy J, McElroy SL. Binge eating disorder in
the design of the study and drafted the article. S.R. participated in the elderly individuals. Int J Eat Disord 2012;45(7):905-8.
design of the study and performed the statistical analysis. M.F. provided