Amit Upadhyah et al.

Prevalence and Risk factors for Eating Disorders

RESEARCH ARTICLE

PREVALENCE AND RISK FACTORS FOR EATING DISORDERS IN

INDIAN ADOLESCENT FEMALES
Background: Eating disorders (ED) are one of the most common psychiatric problems faced Amit Upadhyah1, Rajesh
by adolescents, and are characterized by a persistent course, comorbid psychopathology, Misra1, Deepak Parchwani2,
medical complications, and elevated mortality. Pankaj Maheria3
Aims & Objective: To assess the prevalence and correlates of eating disorders in a sample of 1 Department of Physiology,
adolescent Indian females. Subharti Medical College, Meerut,
Materials and Methods: In a cross-sectional survey, 120 adolescents females (age: 13-17 Uttar Pradesh, India
years) filled out questionnaires on eating attitudes and behaviors at one independent school. 2 Department of Biochemistry,
ED was measured with the 26-item Eating Attitudes Test (EAT). Participants who scored ≥ 20 Gujarat Adani Institute of Medical
on the EAT were considered to have disordered eating and effect of psychological, behavioral, Sciences, Bhuj, Gujarat, India
and socio-environmental variables in individuals with and without eating disorders, were 3 Department of Anatomy, GMERS
assessed. Medical College, Dharpur-Patan,
Results: Disturbed eating attitudes and behaviors were present in 26.67 % of adolescents girls Gujarat, India
in the sample studied. This group was significantly older, had earlier menarche and lower BMI.
Mean scores and percentage scores on all the scales to assess psychological risk factors were Correspondence to: Amit
found to be significantly higher in the ED group i.e. there were significant associations (p< Upadhyah
0.0001) between elevated EAT scores and dieting behavior, higher drive for thinness and body (dramitupadhyah@ymail.com)
dissatisfaction, external pressures, mood susceptibility of feeding patterns, perfectionism,
occurrence of negative life events and presence and adequacy of emotional support system. Received Date: 02.07.2013
Conclusion: Eating disorders and subthreshold eating conditions are prevalent in a sample of Accepted Date: 04.10.2013
adolescent girls and were strongly associated with various psychological, behavioral, and
socio-environmental domains. Future prospective and experimental studies are warranted to DOI:
advance our understanding of the risk factors to enable better preventive programme 10.5455/njppp.2014.4.041220131
planning.
Key Words: Adolescent Females; Eating Attitude Test; Eating Disorder; Psychological Risk
Factors

INTRODUCTION experimental studies on the risk and maintenance factors

for eating pathology in the western population, there is a
Eating disorders (ED) are one of the most common paucity of such studies in Indian population, more so in
psychiatric problems faced by females, characterized by adolescent girls. There is high prevalence of overweight
chronicity and relapse along with disordered eating and obesity in Indian adolescent females[7],which, coupled
behavior where the patient’s attitude towards weight and with rapid socio-cultural changes puts them at risk for
shape, as well as their perception of body shape, are eating disorders. Thus, we aimed to identify prevalence
disturbed.[1] They are ranked among the ten leading and shared risk factors for disordered eating behaviors in
causes of disability among young women[2], have the Indian adolescent females that could serve as targets for
highest levels of treatment seeking, inpatient integrated preventive interventions. To achieve this aim,
hospitalization, suicide attempts, and mortality of the cross-sectional associations between a range of
most common psychiatric syndromes.[3] Furthermore, psychological, behavioral, and socio-environmental
eating pathology increases the risk for onset of obesity, factors and ED among a sample of adolescent females
depression, and substance abuse.[4] The Diagnostic and were assessed.
Statistical Manual of Mental Disorders (DSM-IV)[5] or the
International Statistical Classification of Diseases and MATERIALS AND METHODS
Related Health Problems (ICD-10)[6] classified ED in
anorexia nervosa, bulimia nervosa and eating disorders The study protocol of this cross-sectional observational
not otherwise specified (EDNOS). Diagnosis of AN and BN study was approved by the Institutional Human Research
is based on strict diagnostic criteria and patients with Ethical Committee and a detailed study overview was
disordered eating not fulfilling all of these criteria are described to parents and participants as an investigation
diagnosed as having EDNOS.[5,6] of adolescent mental and physical health. An active
parental-consent procedure was used to recruit
Although there has been a burgeoning of longitudinal and participants. The study group consisted of 120 adolescent

National Journal of Physiology, Pharmacy & Pharmacology | 2014 | Vol 4 | Issue 2 | 153 – 157
 Amit Upadhyah et al. Prevalence and Risk factors for Eating Disorders

girls from Crosthwaite Girl’s College, Allahabad, UP (after Psychological Evaluation Questionnaire
taking authorization from the Principal, thereafter no
further involvement of the school was required). This is a semi-structured pre-tested questionnaire for the
Participants ranged from 13 to 17 years of age. Inclusion assessment of the specific psychopathology of eating
criteria for enrollment required that participants would disorders and presence or absence of various
have entered puberty (based on self-reported attainment psychological and socio-cultural factors which are known
of menarche). However, participants with preexisting risk factors for eating disorders. It had seven items, each
medical or psychiatric illness severe illness or any form of having variable number of questions- Dieting behavior,
pharmacological intervention, impairment of speech, Body image and Self-esteem, Societal pressures, Mood
hearing, vision, or cognition, age > 18 years, or any variations, Support system, Perfectionism, Negative Life
medical condition that prevented participants from events. The first five items - Dieting behavior, Body image
adhering to the protocol and students showing disinterest and self-esteem, Societal pressures, Mood variations and
were excluded from the study. support system were graded on a continuous three point
scale from zero to two with highest score provided to the
A self-report questionnaire was administered to all most stressful situation, most abnormal behavior or the
enrolled participants. The questionnaire had three parts, response most towards symptomatic direction, a score of
the first part for socio-demographic variables, second one for the intermediate response, and a score of zero
comprised of Eating Attitude Test (EAT) to assess the signifying normal response or absence of the stressor or
abnormalities in eating behavior and the third part had behavior under question. The responses for Perfectionism
questions to assess psychological variables and risk and Negative life events items were either Yes or No. “Yes”
factors. Body image distortion was evaluated using Body signified presence of stressor and was given a score of
Silhouette charts[8] to compare self-reported body shape one, “No” signified absence of stressor and was scored
with investigator reported body shape. A difference of two zero.
or more places on the body silhouette chart in student and
investigator reported body silhouette was taken as Those who completed the questionnaires were then
presence of body image distortion. examined for various anthropometric parameters: Weight
(Kg) and height (meters) were measured (using Omron
Socio-Demographic Questionnaires digital body weight scale HN-286 and SECA 206 wall
mounted metal tapes respectively). Body Mass Index
This questionnaire collected socio-demographic variables (BMI) was calculated by Weight (Kg)/ height squared
regarding their general information i.e. name, age, (m2). Waist circumference was assessed in the standing
address, economic status, education, parental education, position, midway between the highest point of the iliac
birth order, age at menarche, history of any systemic crest and the lowest point of the costal margin in the
illness. midaxillary line. Hip circumference was measured at the
level of the femoral greater trochanter. All
Eating Attitude Test (EAT) anthropometric measures reflect the average of 3
measurements (measured by same person on same
The EAT developed by Garner et al[9] was employed in this
instrument to avoid inter-instrument and inter personal
study for the assessment of attitudinal and behavioral
variation). Age was defined as the age at the time of
dimensions relevant to eating disorders i.e. to distinguish
interview (based on student’s school ID). All assessments
patients with eating disorders from weight-preoccupied,
took place on school campus.
but otherwise healthy, female adolescents. It identifies the
presence of symptoms that are consistent with either a
Statistical Analysis
possible eating disorder or disordered eating and warrant
a complete evaluation. EAT consists of 26 items rated on a Data were expressed as mean ± SD (continuous variables),
six point scale, with a score of 3 assigned to the responses or as percentages of total (categorical variables). Prior to
farthest in the “symptomatic” direction, a score of 2 for the hypothesis testing, data were examined for normality.
immediately adjacent response, a score of 1 for the next Non–normally distributed variables were logarithmically
adjacent response and a 0 score assigned to the three transformed before analysis. Statistical significance was
responses farthest in the “asymptomatic” direction.[10] checked using the Student’s t test. 95 % confidence limits
Higher scores indicate higher disordered eating attitudes were set so that if p value was less than 0.05, the
and behaviors. difference was considered significant.

National Journal of Physiology, Pharmacy & Pharmacology | 2014 | Vol 4 | Issue 2 | 153 – 157
 Amit Upadhyah et al. Prevalence and Risk factors for Eating Disorders

RESULTS that group I females were significantly older (p = 0.0210),

had a lower age at menarche (p <0.0001), lower BMI
Using EAT questionnaire as a screening tool for defining (p=0.0321) and lower body weight (p < 0.0001) than the
the population at high risk for an ED, participants (n= other group. However, there was no significant difference
120) were divided into two non-overlapping groups: in waist to hip ratio, birth order, parents working status
Group I or Symptomatic group (n=32; 26.6%) with scores and socio economic status between the study groups. No
≥ 20 {mean score: 24.66 ± 2.42 (range: 20-30)} and Group significant distortion of body image was observed in
II or asymptomatic group, composed of 88 females participants belonging to either of the study groups,
(73.3%) who had scored less than 20 {mean score: 10.22 signifying that these girls are aware of their actual body
± 3.94 (range: 2-18)} on the Eating Attitude Test. size and shape, but it is the cult of thinness that is driving
them to unhealthy eating behaviors.
Table-1: Characteristic of participants with EAT score ≥20 and < 20
Group I (n = 32) Group II (n = 88)
Characteristics Table 2 presents the scales of various psychological and
EAT Score ≥20 EAT Score < 20
Age (years) 15.05 ± 0.86 14.70 ± 0.56* socio-cultural factors which are known risk factors for
Age at Menarche (years) 11.58 ± 0.91 13.43 ± 0.79*
eating disorders. We found a significant difference (p <
Height (m) 1.53 ± 0.05 1.57 ± 0.03*
Weight (Kg) 44.18 ± 4.71 48.26 ± 3.03* 0.0001) between the two groups on all the scales - their
Body Mass Index (Kg\m2) 18.64 + 2.01 19.37 ± 1.38* reported eating of meals, i.e. dieting behavior, had a
Waist Circumference (cm) 59.92 ± 4.03 60.42 ± 3.84
Hip Circumference (cm) 79.46 ± 5.42 80.26 ± 5.21
higher drive for thinness and body dissatisfaction, higher
Waist-Hip Ratio 0.76 ± 0.04 0.76 ± 0.06 perceived external pressures, mood susceptibility of
* p < 0.05
feeding patterns, perfectionism, reported negative life
Table-2: Mean ± SD Scores and % Scores on Psychological Scales events and presence and adequacy of emotional support
Group I Group II system.
(EAT Score ≥20) (EAT Score < 20)
Max.
Scale (n = 32) (n = 88) p Value
Score DISCUSSION
% %
Mean ± SD Mean ± SD
Score Score
Dieting 8 5.28 ± 1.42 66.0 1.47 ± 1.30 18.3 <0.0001
The primary finding of this study, the prevalence of eating
Body Image,
18 11.16 ± 2.92 62.0 2.67 ± 1.98 14.8 <0.0001 disorder symptoms and disordered eating attitudes and
Self Esteem
External behaviors in a sample of adolescent girls came out to be
22 15.66 ± 2.12 71.2 3.33 ± 2.62 15.1 <0.0001
Pressures
Effect of Mood 6 4.22 ± 0.97 70.3 1.88± 1.34 31.3 <0.0001 26.67%. This is comparable with reports from other
Perfectionism 5 3.97 ± 0.86 79.4 1.95 ± 0.95 39.1 <0.0001 investigators[11-16] who had used Eating Attitude Test and
Negative Life reported prevalence of disordered eating attitudes and
12 5.69 ± 1.60 47.4 2.90 ± 1.86 24.1 <0.0001
Events
Support System 7 4.16 ± 0.95 59.4 1.83 ± 1.32 26.1 <0.0001 behaviors to be between 16.5 and 27% in different study
groups.

Jones JM et al[12] had observed that prevalence of

disordered eating and behavior increases gradually
during the adolescence years. Our study found a
statistically significant (p value = 0.0210) difference in the
mean age of two groups (Table 1). Further, mean age at
menarche in Group I subjects was significantly lower (p <
0.0001) than Group II subjects, consistent with previous
findings[17,18], thus substantiating the hypothesis that
earlier age at menarche is a risk factor for development of
Eating Disorder. Persons with their weights and body
Figure-1: Percentage scores on Psychological Scales (A: Dieting
mass indexes in the highest quartile and elevated
Scale; B: Body Image; C: External Pressure; D: Effect of Mood; E: adiposity are considered to be at risk of developing an
Perfectionism; F: Negative Life Events; G: Support System) eating disorder[19,20], as the culturally defined ideal for
Socio-demographic and psychological variables and risk attractiveness currently favors thinness. However, in the
factors were compared between these two groups. Table present study, weight and BMI of Group I subjects were
1 shows the socio-demographic characteristics of the found to be significantly lower than those of Group II
group I and group II participants. By analyzing, we find subjects. This may be due to the fact that this is a cross-

National Journal of Physiology, Pharmacy & Pharmacology | 2014 | Vol 4 | Issue 2 | 153 – 157
 Amit Upadhyah et al. Prevalence and Risk factors for Eating Disorders

sectional study and ongoing disordered eating and dieting infer causality. Another limitation is the use of anonymous
behaviors might have led to decreased body weight and self-report questionnaires to collect data, relying upon the
BMI. honesty of the students. It is possible that biased answers
were collected, in which case the prevalence rate may
The secondary aim of this study was to evaluate the have been under- or overestimated.
psychosocial risk factors for ED in a sample of early
adolescent girls. Mean scores and percentage scores on all Still, taken together, these results provide important
the scales to assess psychological risk factors were found information for clinical practice. The results of this study
to be significantly higher for Group I subjects (Table 2). suggest that the prevalence of disordered eating and
Few studies[12,21] had emphasized the role of dieting unhealthy weight-control behaviors in adolescent
attitude and behaviors in development of future eating population is high, with a multitude of psycho-social
disorders. In our study, scores on Dieting scale were factors contributing to their vulnerability. There is a need
associated with the scores on EAT. Similar pattern was for longitudinal study in Indian setup to isolate the risk
obtained on Body image and self-esteem scale (Table 2), factors and evaluate the possible relationship between
in accordance with another study.[22] Self-esteem based on demanding environments and psychosomatic
weight and shape may be particularly important[23], as it vulnerability. In the interim, our findings convey that
is identified as a core feature of bulimia nervosa in DSM- eating disorders comprise a significant health concern
IV.[5]The scales assessing perceived pressure to be thin among adolescent females and health programs should be
from family, peers, and media also correlated with EAT included for educational services on school campuses in
scores (Table 2). This finding is consistent with previous an effort to alleviate potential risk factors and unhealthy
studies covering different developmental periods[24-26], behaviors and attitudes.
thereby suggesting this effect is robust.
CONCLUSION
The difference in scores on scale assessing effect of mood
on eating behavior, perfectionism score (Table 2) was Eating disorders and subthreshold eating conditions are
found to be statistically significant showing that subjects prevalent in a sample of adolescent girls and were
who are perfectionist and in whom mood controls the strongly associated with various psychological,
feeding pattern are more at risk of developing an ED. behavioral, and socio-environmental domains. Future
Findings on the scale estimating presence of negative life prospective and experimental studies are warranted to
events and other stressors are in conformity with the advance our understanding of the risk factors to enable
importance of negative life events as emphasized by other better preventive programme planning.
studies.[27,28] Another important finding was that deficits
in perceived social support were related with the ACKNOWLEDGEMENT
development of disordered eating. Our results (Table 2)
are consistent with the assertion that acceptance in one’s Authors gratefully acknowledge the assistance of Mrs.
immediate social network might help girls feel more Mamta Mittal, M.A., D.G.P. in the psychological aspects of
positively about themselves and their bodies and render this study.
them more resilient to socio-cultural pressures to be thin.
Results suggest that it might be fruitful to direct greater REFERENCES
attention to the role of social support deficits in
1. Quick VM, Byrd-Bredbenner C, Neumark-Sztainer D. Chronic illness
promoting body image disturbances, especially since and disordered eating: a discussion of the literature. Adv Nutr.
females tend to reach out more to their peers to cope with 2013;4(3):277-86.
2. Mathers CD, Vos ET, Stevenson CE, Begg SJ. The Australian Burden
stress.[29] of Disease Study: Measuring the loss of health from diseases,
injuries and risk factors. Med J Aust. 2000;172,592–6.
3. Nock MK, Green JG, Hwang I, McLaughlin KA, Sampson NA,
Nonetheless, this study has a few limitations. Firstly, Zaslavsky AM, Kessler RC. Prevalence, correlates, and treatment of
sampling may not be representative of all of the lifetime suicidal behavior among adolescents: results from the
adolescent Indian population. Participants were taken National Comorbidity Survey Replication Adolescent Supplement.
JAMA Psychiatry. 2013;70(3):300-10.
from one school, having more or less same socio- 4. Stice E, Hayward C, Cameron R, Killen JD, Taylor CB. Body image
economic status and cultural environment, which may and eating related factors predict onset of depression in female
adolescents: A longitudinal study. J Abnorm Psychol. 2000;109:
reduce the generalizability of our findings. Secondly, the 438-44.
study was cross-sectional and thus, we are not able to 5. American Psychiatric Association. Diagnostic and Statistical

National Journal of Physiology, Pharmacy & Pharmacology | 2014 | Vol 4 | Issue 2 | 153 – 157
 Amit Upadhyah et al. Prevalence and Risk factors for Eating Disorders

Manual of Mental Disorders (DSM IV). Washington, DC: American and early sexual activity are associated with bulimic-type eating
Psychiatric Association 1994. pathology in middle adolescence. J Adolesc Health.
6. World Health Organization (WHO). The ICD-10 Classification of 2001;28(4):346-52.
Mental and Behaviour Disorder. Clinical Description and Diagnostic 19. Vardar E, Erzengin M. The prevalence of eating disorders (EDs) and
Guidelines. Geneva, Washington DC: WHO; 1992. comorbid psychiatric disorders in adolescents: a two-stage
7. Kumar N, Goyal J, Parmar I, Shah V. Prevalence of overweight and community-based study. Turk Psikiyatri Derg. 2011;22(4):205-12.
obesity in affluent adolescent girls in Surat city, western India. Int J 20. Musssap AJ. Waist-to-hip ratio and unhealthy body change in
Med Sci Public Health. 2012;1(1):2-4 women. Sex Roles. 2007;56(1):33-43.
8. CM Bulik, TD Wade, AC Heath, NG Martin, Stunkard AJ, Eaves LJ. 21. Fairburn CG, Cooper Z, Doll HA, Davies BA. Identifying Dieters Who
Relating body mass index to figural stimuli: population-based Will Develop an Eating Disorder: A Prospective, Population-Based
normative data for Caucasians. Int J Obes Relat Metab Disord. Study. Am J Psysciatry. 2005;162:2249-55.
2001;25(10):1517-24. 22. Bener A Tewfik I. Prevalence of overweight, obesity and associated
9. Garner DM. Garfinkel PE. The Eating Attitudes Test: An index of the psychological problems in Qatar’s female population; Obes Rev.
symptoms of anorexia nervosa. Psychol Med. 1979;9: 273-9. 2006;7(2):139-45.
10. Garner DM. Eating Attitude Test© (EAT-26): Scoring and 23. Kotecha PV, Patel SV, Mazumdar VS, Baxi RK, Misra S, Shah MB, et
Interpretation. (cited May 26, 2013). Available from URL: al. Body Mass Index (BMI) and Self-Perception of Weight and
http://www.eat-26.com/Docs/EAT-26IntpretScoring-Test-3-20- Height among school going adolescents in urban Vadodara, India.
10.pdf Int J Med Sci Public Health. 2013; 2:745-748.
11. Anstine D Grinenko D. Rapid screening for disordered eating in 24. Field AE, Camargo CA, Taylor CB, Berkey CS, Roberts SB, Colditz GA.
college-aged females in the primary care setting. J Adolesc Health. Peer, parent, and media influences on the development of weight
2000;26(5):338-42. concerns and frequent dieting among preadolescent and
12. Jones J M Bennett S, Olmsted MP, Lawson ML, Rodin G. Disordered adolescent girls and boys. Pediatrics. 2001;107:54-60.
eating attitudes and behaviors in teenaged girls; a school based 25. Garrusi B, Baneshi MR. Eating disorders and their associated risk
study. CMAJ. 2001;165(5):547-52. factors among Iranian population - a community based study. Glob
13. Goyal RK, Shah VN, Saboo BD, Phatak SR, Shah NN, Gohel MC, et al. J Health Sci. 2012;5(1):193-202.
Prevalence of overweight and obesity in Indian adolescent school 26. Gonçalves JD, Moreira EA, Trindade EB, Fiates GM. Eating disorders
going children: its relationship with socioeconomic status and in childhood and adolescence. Rev Paul Pediatr 2013;31(1):96-103.
associated lifestyle factors. J Assoc Physicians India. 2010;58:151- 27. The McKnight Investigators. Risk factors for the onset of eating
8. disorders in adolescent girls: results of the McKnight longitudinal
14. Palma-Coca O, Hernández-Serrato MI, Villalobos-Hernández A, risk factor study. Am J Psychiatry. 2003; 160(2):248-54.
Unikel-Santoncini C, Olaiz-Fernández G, Bojorquez-Chapela I. 28. Taylor CB, Bryson S, Celio Doyle AA, Luce KH, Cunning D, Abascal
Association of socioeconomic status, problem behaviors, and LB, et al. The Adverse Effect of Negative Comments about Weight
disordered eating in Mexican adolescents: results of the Mexican and Shape From Family and Siblings on Women at High Risk for
National Health and Nutrition Survey 2006. J Adolesc Health. Eating Disorders. Pediatrics. 2006;118(2):731-8.
2011;49(4):400-6. 29. Thaker RB, Verma AP. A study of perceived stress and coping styles
15. Swanson SA, Crow SJ, Le Grange D, Swendsen J, Merikangas KR. among mid adolescents. Natl J Physiol Pharm Pharmacol
Prevalence and correlates of eating disorders in adolescents. 2014;4:22-25. Published online first. DOI:
Results from the national comorbidity survey replication 10.5455/njppp.2014.4.200620131
adolescent supplement. Arch Gen Psychiatry. 2011;68(7):714-23.
16. Prisco AP, Araújo TM, Almeida MM, Santos KO. Prevalence of eating Cite this article as: Upadhyah AA, Misra R, Parchwani DN, Maheria
disorders in urban workers in a city of the northeast of Brazil. Cien PB. Prevalence and risk factors for eating disorders in Indian
Saude Colet. 2013;18(4):1109-18. adolescent females. Natl J Physiol Pharm Pharmacol 2014; 4:153-
17. Ong KK Northstone K, Wells JC, Rubin C, Ness AR, Golding J, et al. 157.
Earlier mother's age at menarche predicts rapid infancy growth Source of Support: Nil
and childhood obesity. PLoS Medicine. 2007;4(4):132 Conflict of interest: None declared
18. Kaltiala- Heino R Rimpelä M, Rissanen A, Rantanen P. Early puberty

National Journal of Physiology, Pharmacy & Pharmacology | 2014 | Vol 4 | Issue 2 | 153 – 157