Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Eating disorder diagnostics in the digital era: validation of the Norwegian version of the Eating Disorder Assessment for DSM-5 (EDA-5)

Eating disorder diagnostics in the digital era: validation of the Norwegian version of the Eating... Objective: The Eating Disorder Assessment for DSM-5 (EDA-5) is an electronic, semi-structured interview developed to assess feeding and eating disorders following DSM-5 criteria. The original English version has strong psychometric properties, and previous research has shown high rates of agreement between diagnoses generated by the Eating Disorder Examination (EDE) interview and the EDA-5. The current study aimed to validate the Norwegian version of the EDA-5, and is the first international validation of this diagnostic tool. Method: A total of 91 (87 females and 4 males) adult in- and out-patients were recruited from two of the largest eating disorder clinics in Norway. Diagnoses assigned using the EDA-5 were compared to diagnoses from the EDE interview (v. 17.0D). Results: Results showed that diagnoses assigned using EDE and EDA-5 were identical for 75 (82.4%) of the 91 cases. Among individual diagnostic categories, kappas ranged from moderate (.49) to perfect (1.00) agreement. The majority of discrepant cases occurred between full- and sub-threshold AN and BN. The EDA-5 was significantly quicker to administer compared to the EDE (22 vs. 54 min). Conclusions: The Norwegian EDA-5 can quickly and efficiently generate DSM-5 diagnoses without compromising diagnostic accuracy. It is a promising alternative to existing diagnostic tools, and may help streamline the identification of feeding and eating disorders in clinical settings and in research. Keywords: Eating disorders, Feeding disorders, Assessment, DSM-5, EDA-5, EDE 17.0D, Diagnosis, Interview Plain English Summary results from its first international (Norwegian) validation. We This study compared the agreement between two diagnostic found that the EDA-5 was significantly quicker to administer interviews for feeding and eating disorders; the traditional compared to the EDE interview(22 vs. 54min), andthatthe Eating Disorder Examination (EDE) and the newly developed two instruments generated identical diagnoses in a little over Eating Disorder Assessment for DSM-5 (EDA-5). The EDA-5 82% of the cases. It is therefore a promising alternative to is a web-based diagnostic instrument, and this study presents traditional diagnostic interviews, which are often lengthy and require extensive training to administer. * Correspondence: camilla.lindvall@dahlgren.no Regional Department for Eating Disorders, Division of Mental Health and Addiction, Oslo University Hospital, Ullevål HF, Postboks 4950 Nydalen, 0424 Background Oslo, Norway 2 The principle of early diagnosis and early intervention to Department of Psychology, Bjørknes University College, Oslo, Norway Full list of author information is available at the end of the article optimize disease outcomes is widely accepted in mental © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 2 of 7 health. For individuals with eating disorders, the first (see www.eda5.org), and was developed to minimize par- years of illness appears to offer a critical window for re- ticipant and interviewer burden [8]. It is currently the covery, beyond which outcomes are poorer [1]. It is only existing semi-structured interview guide which as- therefore important to quickly and accurately assess the sesses all DSM-5 feeding and eating disorders. The logic nature of an eating disorder, so that appropriate treat- underlying the EDA-5 relies on an algorithm that selects ment recommendations can be made and possible com- subsequent questions based on answers already ob- plications can be assessed. tained, i.e., it implements diagnostic “skip rules” that In DSM-5, published in 2013, several significant avoid asking questions no longer relevant to making a changes were made to the section describing the eating diagnosis. In 2015, the validity of the EDA-5 was evalu- disorders. The criterion requiring amenorrhea for the ated by comparing diagnoses generated by the EDA-5 to diagnosis of anorexia nervosa (AN) was eliminated, the the results of clinical interviews and the Eating Disorder minimum average frequencies of binge eating and for in- Diagnostic Scale (EDDS) [8]. High rates of agreement appropriate compensatory behaviors were reduced from were observed between the EDA-5 and interviews per- twice to once weekly in the diagnostic criteria for bu- formed by experienced clinicians [7] with kappas ranging limia nervosa (BN), and binge-eating disorder (BED) was from 0.56 (for Other Specified Feeding or Eating Dis- officially recognized. In addition, three disorders that order (OSFED)/Unspecified Feeding or Eating Disorder were described in the section on Disorders Usually First (UFED)) to 0.97 (for BN), supporting the validity of the Diagnosed in Infancy, Childhood, or Adolescence in EDA-5. DSM-IV were combined with the eating disorders in a The aim of the current study was to test the validity of section renamed Feeding and Eating Disorders in DSM- the Norwegian version of the EDA-5, following the ap- 5. These disorders were pica, rumination disorder, and proach of Sysko et al. [8]. feeding disorder of infancy or early childhood; the last was expanded and renamed avoidant/restrictive food in- Methods take disorder (ARFID). Assessment Since the publication of DSM-5, two of the most The eating disorder assessment for DSM-5 (EDA-5) widely used semi-structured interview guides, the Eating The EDA-5 [8] is an electronic, semi-structured inter- Disorder Examination (EDE) v. 17.0D [2], and the Struc- view assessing feeding and eating disorders in adults fol- tured Clinical Interview for DSM Axis 1 Disorders lowing DSM-5 criteria, and the frequencies of salient DSM-5 version (SCID-5) [3], were revised to reflect the behavioral disturbances characteristic of these conditions changes contained in DSM-5. Albeit being widely used, (e.g., the number of objective and subjective binge eating the EDE and the SCID have several limitations. The episodes and compensatory behaviors). The following EDE v. 17.0D lacks items relevant to ARFID, pica and full-threshold diagnoses are assessed by the EDA-5: AN rumination disorder [4], and although the SCID has a (restricting or binge-eating/purging type), BN, BED, module for ARFID, like the EDE, it does not assess pica ARFID, Pica and rumination disorder. In addition, using or rumination disorder. In addition, although the EDE is the EDA-5 the interviewer can assign any of the OSFED available at no cost, individuals who wish to use the EDE diagnoses described in DSM-5 (i.e., OSFED-AN, are required to complete specialized training, with lim- OSFED-BN, OSFED-BED, OSFED-Night Eating Syn- ited accessibility. The administration of the EDE is also drome (NES) and OSFED Purging Disorder (PD)) or time consuming, normally taking between 45 and 90 UFED. As implemented, the EDA-5 relies on an algo- min. The SCID module for the eating disorders is brief, rithm that selects subsequent questions based on an- but is not freely available. Further, the Composite Inter- swers already obtained. Therefore, the number of national Diagnostic Interview (CIDI) [5], a general psy- questions administered varies across individuals, and, chiatric assessment previously used in prevalence studies consequently, so does the length of administration. The [6], and was anticipated to be updated in concert with English version of the EDA-5 was translated into Norwe- the International Classification of Diseases (ICD-11) in gian by a group of researchers and clinicians at Regional 2018 [7]. However, the new version is still not publicly Department of Eating Disorders (RASP) in 2016. The available, and when it launches, it will no longer include translation was a close collaborative process between a section on eating disorders (L. Chaze, personal com- RASP and Eating Disorder Research Unit (EDRU) at munication, January 2017). New York State Psychiatric Institute (NYSPI), Columbia The Eating Disorder Assessment for DSM-5 (EDA-5) University Medical Center, with the process following was developed to provide a guide to conducting a semi- World Health Organization guidelines for translation structured interview to assess whether an individual and adaptation of instruments [9]. Similar to the original meets criteria for an eating disorder according to DSM- interview, the Norwegian version is a computer based, 5 criteria. The EDA-5 is web-based and freely available electronic application (“app”), with diagnostic interview Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 3 of 7 procedures and internal logic identical to those of the a part (average time between the interviews = 1.1 ± 1.8 English version. The interview is available at no cost, days; range 0–10 days), and conducted by different inter- and is accessible via the website www.eda5.org. viewers to avoid contamination. All interviews were con- ducted in-person. Interviewers recorded the length of The eating disorder examination (EDE) v. 17.0D the interviews and completed a checklist where fulfilled The EDE [2] is a semi-structured interview generating diagnostic criteria and assigned DSM-5 diagnoses were operational ED diagnoses, as well providing measures of registered. Inclusion criteria were liberal, with patients the range and severity of ED features. The current ver- being considered eligible if they were medically stable, sion of the interview (v. 17.0D) generates all DSM-5 aged 16 years or above, and provided written consent to full-, subthreshold and unspecified eating disorder diag- participate in the study. No remuneration was offered. noses including AN, BN, BED, OSFED and UFED. The The study was approved by the Norwegian Regional three feeding disorders Pica, ARFID and rumination dis- Committee for Medical and Health Research Ethics (ref. order cannot be assessed using the EDE v. 17.0D. Four 2017/8130) and the Norwegian Data Protection Author- clinically derived subscales assess restraint, eating con- ity at Oslo University Hospital. Diagnostic data was col- cern, shape concern and weight concern. A global sever- lected at the two clinics between November 2017 and ity score is calculated by averaging the four subscales. June 2019. Psychometric studies show sample- and subscale dependent variations in internal consistency ranging Statistical analyses from poor to excellent and adequate concurrent and dis- All statistical analyses were carried out using IBM SPSS criminant validity [10]. The administration of the EDE Statistics Version 25. The sample size (n = 91) was mod- requires significant training, both in the technique of elled on the original validation study [8]. The EDE was interviewing, as well being familiar with the concepts used as a reference instrument in all analyses comparing and rules governing the ratings. The length of adminis- diagnoses. Since the EDE is not designed to assign pica, tration typically ranges from 45 to 75 min [2]. The Nor- rumination disorder or ARFID, reliability rates were not wegian version of the EDE v. 17.0D was used in the calculated for these three diagnoses. Kappas, sensitivity, current study. All EDE items were administered. DSM-5 specificity, negative and positive predictive value and ac- diagnoses were assigned using the diagnostic algorithms curacy was calculated for all assigned diagnoses. The five described in the EDE interview guidelines. latter measures are expressed in percentages. Kappas (ranging from − 1 to + 1) were used to measure diagnos- Procedure tic agreement. A slightly less liberal kappa standard was The first author of this paper (CLD) was trained and su- applied than that reported in the original EDA-5 valid- pervised in the use of the EDA-5 by BTW, and con- ation study (i.e. [11, 12]), with kappa coefficients (κ) be- ducted all EDA-5 interviews at RASP. The EDE ing interpreted as follows κ <0= “No Agreement”, interviews at RASP were conducted by two of the co- .0–.20 = “None”, .21–.39 = “Minimal”, .40–.59 = “Weak”, authors of this paper; a psychiatric nurse (CS) and a .60–.79 = “Moderate”, .80–.90 = “Strong” and > .90 = “Al- highly experienced clinician and senior researcher hold- most perfect” [13]. According to McHugh [14], any ing a PhD in psychology (ØR). CS was trained in the kappa above 0.60 indicates acceptable agreement among EDE assessment by ØR, who also supervised CS raters, whereas little confidence should be placed in re- throughout the data collection period. CLD trained the sults showing kappas below 0.6. In this study, sensitivity head of research and clinical psychologist (KV) and col- (true positives), is defined as the proportion of individ- leagues in the use of the EDA-5 at Modum Bad. All par- uals with a specific EDE diagnosis who were accurately ticipating staff at Modum Bad had prior experience and (i.e., identically) diagnosed using the EDA-5. Specificity formal training in the use of the EDE interview. (true negatives) is the proportion of individuals who did Participants were individuals receiving treatment at not receive a specific EDE diagnosis, who also did not re- one of the two Norwegian tertiary care centers: RASP at ceive that particular diagnosis using the EDA-5. The Oslo University Hospital, Oslo, or the Eating Disorder positive predictive value (PPV) is the probability that an Clinic at Modum Bad Psychiatric Center in Vikersund. individual diagnosed using the EDA-5 received that diag- At RASP, participants were recruited from the out- nosis using the EDE. The negative predictive value patient clinic and two inpatient adult clinics. A team of (NPV) is the probability that an individual, who did not four psychologists, one psychiatrist, two medical doctors receive a certain diagnosis using the EDA-5, did not re- and two psychiatric nurses completed the EDE and the ceive that diagnosis via the EDE. The closer PPV and EDA-5 interviews at Modum Bad. All participants were NPV values are to 1.0 (i.e. 100%), the higher the prob- inpatients. The order of the two interviews was counter- ability that the instrument being validated (in this case, balanced, and intended to occur no more than five days the EDA-5) is doing as good as “gold standard” (in this Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 4 of 7 case, the EDE interview) [14]. Accuracy is the proportion (Mean = 54.0 min, SD = 22.1), t (79) = 12.9, p < .0005 of true results, either true positives or true negatives. It (two-tailed). There were no significant differences be- is calculated as the sum true positives and true negatives tween sites in the time required to conduct the EDE and divided by the sample size (n). Since no OSFED-BED EDA-5 interviews. diagnoses were assigned, sensitivity, negative predictive value or accuracy could not be calculated for these Interview discrepancies diagnoses. Diagnostic distribution using the EDE and the EDA-5 is presented in Table 2. Diagnoses assigned using EDE and Results EDA-5 interviews were identical for 75 (82.4%) of the 91 A total of 101 adult participants were recruited to take cases. Among individual diagnostic categories, kappas part in the study, 57 from RASP and 44 from Modum ranged from moderate (.49) to perfect (1.00) agreement. Bad. Seven of the participants from RASP withdrew be- Kappas, sensitivity, specificity, positive and negative pre- fore having entered the study, and at Modum Bad, three dictive values as well as agreement accuracy are pre- participants were excluded due to interviews taking sented in Table 3. place too far apart (2–6 months). The final sample con- When comparing diagnoses assessed using the EDA-5 sisted of a total of 91 (87 females and 4 males) partici- and the EDE interview, 16 out of 91 patients (17.6%) re- pants. Demographic characteristics and assigned EDE ceived discrepant diagnoses (see Table 4). Twelve of diagnoses are presented in Table 1. Neither the EDE nor these were patients at RASP, and the remaining four the EDA-5 identified cases of OSFED-BED, OSFED-PD, were patients at Modum Bad. OSFED-NES or UFED diagnoses. Similarly, none of the participants received a diagnosis of ARFID, pica or ru- Discussion mination disorder. The label “OSFED Other” was used This study examined the validity of the Norwegian ver- to group participants whose eating disorder symptoms sion of the web-based diagnostic tool, the EDA-5, in strayed too far from the individual OSFED categories, assigning DSM-5 feeding and eating disorders. In line and deviated significantly from the examples given in with the original study [8], the Norwegian EDA-5 the DSM-5. Site differences were observed for age [F (2, quickly and efficiently generated DSM-5 diagnoses with- 89) = .46, p = .01] (patients recruited from Modum Bad out compromising diagnostic accuracy. As such, it is a were significantly older), but not for gender or BMI. The promising alternative to existing diagnostic tools, and EDA-5 was significantly quicker to administer (Mean = may facilitate the identification of eating disorders in 21.6 min, SD = 8.5) compared to the EDE interview clinical settings as well as in research. Table 1 Demographic characteristics and EDE diagnoses in the full sample (n = 91) and across sites Full sample RASP Modum Bad (N = 91) (n = 50) (n = 41) ** Age (years), mean (SD), range 30.9 (9.8) 28.5 (8.8) 33.8 (10.2) 17–56 17–52 19–56 Body mass index mean (SD), range 21.6 (7.9) 22.4 (9.2) 20.6 (5.8) 12.5–51.4 12.5–51.4 13.3–39.2 Female, n (%) 87 (95.6%) 48 (96.0%) 39 (95.1%) Diagnosis (EDE), n (%) AN 34 (37.4%) 15 (30.0%) 19 (46.4%) AN-R 16 (17.6%) 7 (14.0%) 9 (22.0%) AN-BP 19 (19.8%) 8 (16%) 10 (24.4%) BN 25 (27.5%) 11 (22.0%) 14 (34.1%) BED 6 (6.6%) 5 (10.0%) 1 (2.4%) OSFED 25 (27.5%) 18 (36.0%) 7 (17.1%) OSFED-AN 14 (15.4%) 10 (20.0%) 4 (9.8%) OSFED-BN 4 (4.4%) 3 (6.0%) 1 (2.4%) OSFED-Other 7 (7.7%) 5 (10.0%) 2 (4.9%) No ED 1 (1.1%) 1 (2.0%) 0 (0%) EDE The Eating Disorder Examination v. 17.0D, AN Anorexia Nervosa, AN-R Anorexia Nervosa Restrictive type, AN-BP Anorexia Nervosa Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders, ED Eating Disorder, RASP Regional Department for Eating ** Disorders. = Patients recruited from Modum Bad were significantly older than patients recruited from RASP (p ≤ .01) Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 5 of 7 Table 2 Diagnostic distribution in the full sample (n = 91) using OSFED group. The only two diagnostic categories where the EDE and the EDA-5 there were no discrepancies at all were BED and Diagnosis (%, n) EDE EDA-5 OSFED-BN, partially (due to low N), and possibly underscoring the clarity of the criteria for these categor- AN total 36.2% (33) 51.7% (47) ies. Considering the relatively new inclusion of BED as a AN-R 16.5% (15) 23.1% (21) separate diagnostic entity, these are promising results. AN-BP 19.8% (18) 28.6% (26) The majority of discrepant cases occurred between full- BN 27.5% (25) 20.9% (19) and sub-threshold AN and BN, likely reflecting different BED 6.6% (6) 6.6% (6) time frames used by the interviews in assessing DSM-5 OSFED total 27.5% (25) 19.8% (18) diagnostic criteria A (“Restriction of energy intake rela- tive to requirements leading to a significantly low body OSFED-AN 15.4% (14) 5.5% (5) weight”). DSM-5 does not explicitly specify a time frame OSFED-BN 4.4% (4) 4.4% (4) over which weight should be assessed; the EDA-5 aims OSFED-BED 0% (0) 0% (0) to determine whether an individual has been at a signifi- OSFED-PD 0% (0) 0% (0) cantly low body weight over the last three months, OSFED-NES 0% (0) 0% (0) whereas the EDE focuses only on the current weight. In OSFED-Other 7.7% (7) 9.9% (9) 12 of the 16 discrepant cases, the individual had been at a significantly low weight during the previous three ARFID NA 0% (0) months but was not significantly underweight at the Pica NA 0% (0) time of the interview. If the EDA-5 had focused on Rumination disorder NA 0% (0) current weight, or if the EDE (or the clinician doing the No ED 1.1% (1) 1 (1.1%) EDE) had prompted for lowest weight the previous three EDE The Eating Disorder Examination v. 17.0D, EDA-5 The Eating Disorder months, diagnostic agreement would have risen to Assessment for DSM-5, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP 95.6%. These numbers are in contrast to those in the Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders, OSFED-PD Purging measures’ original validation [8], suggesting that how Disorder, OSFED-NES Night Eating Syndrome, ARFID Avoidant Restrictive Food weight is judged in practice impacts diagnostic distribu- Intake Disorder, ED Eating Disorder, AN total includes AN-R and AN-BP, OSFED total includes OSFED-AN, OSFED-BN, OSFED-BED, OSFED-PD and OSFED-NES tions and rates. Whereas BN and BED diagnoses require a three-month minimum duration of binge eating (for Similar to the sample in the original study [8] the ma- BN and BED) and compensatory behaviors (for BN jority of participants in the present study were adult only), the DSM-5 does not specify the amount of time Caucasian females presenting with BMIs ranging from an individual should have been at a normal weight to be 13 to 51 kg/m . Also, similar to Sysko et al. [8] the high- considered recovered from AN [10]. Nor does it specify est level of agreement between the EDA-5 and the EDE the amount of time an individual should have been (1.00) was found for BN and the lowest kappa (0.62), underweight to be considered fulfilling the AN weight sensitivity (0.50) and PPV (0.89) were found for the criterium. The EDA-5, on the other hand, has Table 3 Agreement of the EDA-5 with the EDE interview (n = 91) Diagnosis κ Sensitivity Specificity Positive predictive value Negative predictive value Accuracy EDE ** AN total .72 1.00 .77 1.00 .72 .85 ** AN-R .77 .94 .92 .99 .73 .93 ** AN-BP .79 1.00 .90 1.00 .72 .91 ** BN .82 .76 1.00 .92 1.00 1.00 ** BED 1.00 1.00 1.00 1.00 1.00 1.00 ** OSFED total .62 .50 1.00 .89 1.00 1.00 ** OSFED-AN .53 .40 1.00 .89 1.00 1.00 ** OSFED-BN 1.00 1.00 1.00 1.00 1.00 1.00 ** OSFED Other .86 1.00 .98 1.00 .78 .98 ** No diagnosis .49 .33 1.00 .98 1.00 1.0 EDA-5 Eating Disorder Assessment for DSM-5, EDE Eating Disorder Examination, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders. AN total includes AN-R and AN-BP, OSFED total includes OSFED- AN, and OSFED-BN. No OSFED-BED, PD or NES cases were diagnosed using the EDE and the EDA-5. EDE is used as the reference assessment in all analyses. **p < .0005 Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 6 of 7 Table 4 Diagnoses assigned using the EDA-5 and the EDE (n = 91) EDA-5 AN-R AN-BP BN BED OSFED-AN OSFED-BN UFED No FED EDE AN-R 15 10 0 0 0 0 0 AN-BP 0 18 00 0 0 0 0 BN 0 5 19 00 0 1 0 BED 0 0 0 6 00 0 0 OSFED-AN 6 2 0 0 5 01 0 OSFED-BN 0 0 0 0 0 4 00 OSFED Other 0 0 0 0 0 0 7 0 No FED 0 0 0 0 0 0 0 1 EDA-5 Eating Disorder Assessment for DSM-5, EDE Eating Disorder Examination, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding or Eating Disorder, FED Feeding and Eating Disorder. Bold figures indicate diagnostic agreement between the EDA-5 and the EDE implemented the same 3-month time frame as required the four cases described above, can be explained by dif- for full threshold BN and BED. The rationale for not ferences in clinical judgment (e.g. case #3), different in- simply using the day-of-evaluation weight is that weight terpretations of diagnostic criteria (e.g. case #2 and #4) can fluctuate greatly over short periods of time, espe- or by different patient reports (e.g. case #1). These are cially in individuals enrolled in nutrition restoration reg- all important factors to consider in the diagnostic imens, including inpatient treatments for AN. Judging process, whether it is for research or clinical purposes. the AN weight criteria over the same duration used to judge the frequencies of behavioral disturbances in BN Strengths and limitations and BED, will presumably decrease diagnostic crossover Strengths of this study include standardized procedural and reduce the number of OSFED cases, the latter being steps for translation and validation, and data collection one of the top priorities of the revisions implemented in at two different sites. Also, high rates of diagnostic DSM-5 [6]. agreement were reached despite variability in inter- In one (case #1) of the four remaining discrepant viewers’ professional degree and specialty, supporting cases, the participant fulfilled BN criteria B through E, the utility of the EDA-5 across professions and experi- but reported objective binge eating without loss of con- ence with feeding and eating disorders. In addition, min- trol and was assigned an OSFED Other diagnosis in the imal resources were required to train interviewers to use EDA-5. In the EDE, the same patient fulfilled AN cri- of EDA-5. Neither the current study nor that of Sysko teria B and C (but not A since she was normal weight), et al. assessed individuals with ARFID, pica, or rumin- with subjective binge eating and received an OSFED-AN ation disorder, so the validity of the EDA-5 in character- diagnosis. One participant (case #2) received a BN diag- izing those disorders is unknown. In addition, more nosis using the EDE, and an AN-BP diagnosis using the research is needed to investigate the applicability of the EDA-5; the interviewing clinician using the EDE inter- EDA-5 in younger populations, and to determine the ex- view assessed her weight (and BMI which was 18.0) as tent to which these results generalize to males and eth- not being “significantly low” and therefore not fulfilling nically diverse samples. Replication in larger samples of the AN weight criterion. The interviewing EDA-5 clin- OSFED (e.g. in non-clinical samples) would be beneficial ician, on the other hand, assessed the patient’s weight/ to assess the diagnostic reliability of the assessment of BMI as being significantly low, fulfilling criteria for an OSFED subthreshold conditions. This may require more AN diagnosis. One participant (case #3) received an AN- detailed criteria to enhance diagnostic concordance R diagnosis using the EDE, and an AN-BP diagnosis within and between measures. Methodological limita- using the EDA-5; the use of laxatives was interpreted as tions such as the lack of inter-rater and test-retest reli- a compensatory behavior in the EDA-5 but not in the ability should also be noted. EDE. The final (case #4) discrepant case occurred as one participant fulfilled all but one BN criterion (D; self- Conclusion evaluation is unduly influenced by body shape and The EDA-5 is currently the only available semi- weight) and received an OSFED Other diagnosis when structured interview capable of assessing all the feeding assessed using the EDA-5; when using the EDE, the clin- and eating disorders described in DSM-5. The current ician judged that criterion D was met, and assigned the study replicates and extends the report of Sysko et al. [8] patient a BN diagnosis. Discordant diagnoses, such as in documenting that, even after translation to Norwegian Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 7 of 7 and in the hands of clinicians who were not involved in In: Walsh BT, Attia E, Glasofer DR, Sysko R, editors. Handbook of Asssessment and treatment of eating disorders. Arlingston: American Psychiatric its development, the EDA-5 efficiently provides valid Association Publishing; 2016. diagnostic assessments of eating disorders following 5. Kessler RC, Ustun TB. The world mental health (WMH) survey initiative DSM-5 criteria. Although they have not been formally version of the World Health Organization (WHO) composite international diagnostic interview (CIDI). Int J Methods Psychiatr Res. 2004;13(2):93–121. tested, versions of the EDA-5 in Spanish and Turkish 6. Dahlgren CL, Wisting L. Transitioning from DSM-IV to DSM-5: a systematic are also available (see www.eda5.org). review of eating disorder prevalence assessment. Int J Eat Disord. 2016; 49(11):975–97. Abbreviations 7. Walsh TB, Attia E, Glasofer DR, Sysko R, editors. Handbook of teh assessment AN: Anorexia Nervosa; AN-R: Anorexia Nervosa Restrictive subtype; AN- and treatment of eating disorders. Arlington: American Psychiatric BP: Anorexia Nervosa Binge-Purge subtype; ARFID: Avoidant/Restrictive Food Association Publishing; 2016. Intake Disorder; BED: Binge Eating Disorder; BN: Bulimia Nervosa; 8. Sysko R, Glasofer DR, Hildebrandt T, Klimek P, Mitchell JE, Berg KC, et al. The CIDI: Composite International Diagnostic Interview; DSM: Diagnostic and eating disorder assessment for DSM-5 (EDA-5): development and validation Statistical Manual of Mental Disorders; ED: Eating Disorder; EDA-5: Eating of a structured interview for feeding and eating disorders. Int J Eat Disord. Disorder Assessment for DSM-5; EDDS: Eating Disorder Diagnostic Scale; 2015;48(5):452–63. EDE: The Eating Disorder Examination; EDRU: Eating Disorder Research Unit; 9. World Health Organization (WHO). Process of translation and adaptation of ICD: International Classification of Diseases; NES: Night Eating Syndrome; instruments. World Health Organization (WHO); 2016. http://www.who.int/ NPV: Negative Predictive Value; NYSPI: New York State Psychiatric Institute; substance_abuse/research_tools/translation/en/. OSFED: Other Specified Feeding and Eating Disorders; PD: Purging Disorder; 10. Burke NL, Tanofsky-Kraff M, Crosby R, Mehari RD, Marwitz SE, Broadney MM, PPV: Positive Predictive Value; RASP: Regional Department for Eating et al. Measurement invariance of the eating disorder examination in black Disorders; SCID: Structured Clinical Interview for DSM Axis 1 Disorders; and white children and adolescents. Int J Eat Dis. 2017;50(7):758–68. UFED: Unspecified Feeding or Eating Disorders 11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159–74. Authors’ contributions 12. Fleiss JL, Levin B, Paik CM. Statistical methods for rates and proportions, CLD, ØR, KV and CS conducted the majority of the diagnostic interviews. Third Edition. New York: John Wiley & Sons, Inc; 2003. CLD, ØR, BTW and KV analyzed and interpreted the data. All authors read 13. McHugh ML. Interrater reliability: the kappa statistic. Biochem Med. 2012; and approved the final manuscript. 22(3):276–82. 14. Parikh R, Mathai A, Parikh S, Chandra Sekhar G, Thomas R. Understanding Funding and using sensitivity, specificity and predictive values. Indian J Ophthalmol. This work was funded by the Regional Department for Eating Disorders, Oslo 2008;56(1):45–50. University Hospital, Ullevål HF, and by Modum Bad, Vikersund, Norway. Availability of data and materials Publisher’sNote The datasets used and/or analysed during the current study are available Springer Nature remains neutral with regard to jurisdictional claims in from the corresponding author on reasonable request. published maps and institutional affiliations. Ethics approval and consent to participate This study was approved by the Norwegian Regional Committee for Medical and Health Research Ethics (ref. 2017/8130) and the Norwegian Data Protection Authority at Oslo University Hospital. All participants provided written consent to participate in the study. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Author details Regional Department for Eating Disorders, Division of Mental Health and Addiction, Oslo University Hospital, Ullevål HF, Postboks 4950 Nydalen, 0424 Oslo, Norway. Department of Psychology, Bjørknes University College, Oslo, Norway. Department of Psychiatry, College of Physicians & Surgeons, Columbia University, New York State Psychiatric Institute, New York, USA. 4 5 Research Institute of Modum Bad, Vikersund, Norway. Institute of Clinical Medicine, University of Oslo, Oslo, Norway. Received: 14 April 2020 Accepted: 9 June 2020 References 1. Fukutomi A, Austin A, McClelland J, Brown A, Glennon D, Mountford V, et al. First episode rapid early intervention for eating disorders: a two-year follow- up. Early Interv Psychiatry. 2019;14(1):137–41. 2. Fairburn CG, Cooper Z, O'Connor M. The Eating Disorder Examination (EDE) 17.0D. 2018. https://www.credo-oxford.com/7.2.html. 3. First MB, Williams JBW, Karg RS, Spitzer RL. Structured clinical interview for DSM-5® disorders—clinician version (SCID-5-CV). Arlington: American Psychiatric Publishing Inc.; 2016. 4. Thomas JJ, Roberto CA, Berg KC. Assessment measures, then and now: a look Back at seminal measures and a look forward to the brave New World. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Eating Disorders Springer Journals

Eating disorder diagnostics in the digital era: validation of the Norwegian version of the Eating Disorder Assessment for DSM-5 (EDA-5)

Loading next page...
 
/lp/springer-journals/eating-disorder-diagnostics-in-the-digital-era-validation-of-the-QPRXcbbMCT

References (7)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
eISSN
2050-2974
DOI
10.1186/s40337-020-00310-7
Publisher site
See Article on Publisher Site

Abstract

Objective: The Eating Disorder Assessment for DSM-5 (EDA-5) is an electronic, semi-structured interview developed to assess feeding and eating disorders following DSM-5 criteria. The original English version has strong psychometric properties, and previous research has shown high rates of agreement between diagnoses generated by the Eating Disorder Examination (EDE) interview and the EDA-5. The current study aimed to validate the Norwegian version of the EDA-5, and is the first international validation of this diagnostic tool. Method: A total of 91 (87 females and 4 males) adult in- and out-patients were recruited from two of the largest eating disorder clinics in Norway. Diagnoses assigned using the EDA-5 were compared to diagnoses from the EDE interview (v. 17.0D). Results: Results showed that diagnoses assigned using EDE and EDA-5 were identical for 75 (82.4%) of the 91 cases. Among individual diagnostic categories, kappas ranged from moderate (.49) to perfect (1.00) agreement. The majority of discrepant cases occurred between full- and sub-threshold AN and BN. The EDA-5 was significantly quicker to administer compared to the EDE (22 vs. 54 min). Conclusions: The Norwegian EDA-5 can quickly and efficiently generate DSM-5 diagnoses without compromising diagnostic accuracy. It is a promising alternative to existing diagnostic tools, and may help streamline the identification of feeding and eating disorders in clinical settings and in research. Keywords: Eating disorders, Feeding disorders, Assessment, DSM-5, EDA-5, EDE 17.0D, Diagnosis, Interview Plain English Summary results from its first international (Norwegian) validation. We This study compared the agreement between two diagnostic found that the EDA-5 was significantly quicker to administer interviews for feeding and eating disorders; the traditional compared to the EDE interview(22 vs. 54min), andthatthe Eating Disorder Examination (EDE) and the newly developed two instruments generated identical diagnoses in a little over Eating Disorder Assessment for DSM-5 (EDA-5). The EDA-5 82% of the cases. It is therefore a promising alternative to is a web-based diagnostic instrument, and this study presents traditional diagnostic interviews, which are often lengthy and require extensive training to administer. * Correspondence: camilla.lindvall@dahlgren.no Regional Department for Eating Disorders, Division of Mental Health and Addiction, Oslo University Hospital, Ullevål HF, Postboks 4950 Nydalen, 0424 Background Oslo, Norway 2 The principle of early diagnosis and early intervention to Department of Psychology, Bjørknes University College, Oslo, Norway Full list of author information is available at the end of the article optimize disease outcomes is widely accepted in mental © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 2 of 7 health. For individuals with eating disorders, the first (see www.eda5.org), and was developed to minimize par- years of illness appears to offer a critical window for re- ticipant and interviewer burden [8]. It is currently the covery, beyond which outcomes are poorer [1]. It is only existing semi-structured interview guide which as- therefore important to quickly and accurately assess the sesses all DSM-5 feeding and eating disorders. The logic nature of an eating disorder, so that appropriate treat- underlying the EDA-5 relies on an algorithm that selects ment recommendations can be made and possible com- subsequent questions based on answers already ob- plications can be assessed. tained, i.e., it implements diagnostic “skip rules” that In DSM-5, published in 2013, several significant avoid asking questions no longer relevant to making a changes were made to the section describing the eating diagnosis. In 2015, the validity of the EDA-5 was evalu- disorders. The criterion requiring amenorrhea for the ated by comparing diagnoses generated by the EDA-5 to diagnosis of anorexia nervosa (AN) was eliminated, the the results of clinical interviews and the Eating Disorder minimum average frequencies of binge eating and for in- Diagnostic Scale (EDDS) [8]. High rates of agreement appropriate compensatory behaviors were reduced from were observed between the EDA-5 and interviews per- twice to once weekly in the diagnostic criteria for bu- formed by experienced clinicians [7] with kappas ranging limia nervosa (BN), and binge-eating disorder (BED) was from 0.56 (for Other Specified Feeding or Eating Dis- officially recognized. In addition, three disorders that order (OSFED)/Unspecified Feeding or Eating Disorder were described in the section on Disorders Usually First (UFED)) to 0.97 (for BN), supporting the validity of the Diagnosed in Infancy, Childhood, or Adolescence in EDA-5. DSM-IV were combined with the eating disorders in a The aim of the current study was to test the validity of section renamed Feeding and Eating Disorders in DSM- the Norwegian version of the EDA-5, following the ap- 5. These disorders were pica, rumination disorder, and proach of Sysko et al. [8]. feeding disorder of infancy or early childhood; the last was expanded and renamed avoidant/restrictive food in- Methods take disorder (ARFID). Assessment Since the publication of DSM-5, two of the most The eating disorder assessment for DSM-5 (EDA-5) widely used semi-structured interview guides, the Eating The EDA-5 [8] is an electronic, semi-structured inter- Disorder Examination (EDE) v. 17.0D [2], and the Struc- view assessing feeding and eating disorders in adults fol- tured Clinical Interview for DSM Axis 1 Disorders lowing DSM-5 criteria, and the frequencies of salient DSM-5 version (SCID-5) [3], were revised to reflect the behavioral disturbances characteristic of these conditions changes contained in DSM-5. Albeit being widely used, (e.g., the number of objective and subjective binge eating the EDE and the SCID have several limitations. The episodes and compensatory behaviors). The following EDE v. 17.0D lacks items relevant to ARFID, pica and full-threshold diagnoses are assessed by the EDA-5: AN rumination disorder [4], and although the SCID has a (restricting or binge-eating/purging type), BN, BED, module for ARFID, like the EDE, it does not assess pica ARFID, Pica and rumination disorder. In addition, using or rumination disorder. In addition, although the EDE is the EDA-5 the interviewer can assign any of the OSFED available at no cost, individuals who wish to use the EDE diagnoses described in DSM-5 (i.e., OSFED-AN, are required to complete specialized training, with lim- OSFED-BN, OSFED-BED, OSFED-Night Eating Syn- ited accessibility. The administration of the EDE is also drome (NES) and OSFED Purging Disorder (PD)) or time consuming, normally taking between 45 and 90 UFED. As implemented, the EDA-5 relies on an algo- min. The SCID module for the eating disorders is brief, rithm that selects subsequent questions based on an- but is not freely available. Further, the Composite Inter- swers already obtained. Therefore, the number of national Diagnostic Interview (CIDI) [5], a general psy- questions administered varies across individuals, and, chiatric assessment previously used in prevalence studies consequently, so does the length of administration. The [6], and was anticipated to be updated in concert with English version of the EDA-5 was translated into Norwe- the International Classification of Diseases (ICD-11) in gian by a group of researchers and clinicians at Regional 2018 [7]. However, the new version is still not publicly Department of Eating Disorders (RASP) in 2016. The available, and when it launches, it will no longer include translation was a close collaborative process between a section on eating disorders (L. Chaze, personal com- RASP and Eating Disorder Research Unit (EDRU) at munication, January 2017). New York State Psychiatric Institute (NYSPI), Columbia The Eating Disorder Assessment for DSM-5 (EDA-5) University Medical Center, with the process following was developed to provide a guide to conducting a semi- World Health Organization guidelines for translation structured interview to assess whether an individual and adaptation of instruments [9]. Similar to the original meets criteria for an eating disorder according to DSM- interview, the Norwegian version is a computer based, 5 criteria. The EDA-5 is web-based and freely available electronic application (“app”), with diagnostic interview Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 3 of 7 procedures and internal logic identical to those of the a part (average time between the interviews = 1.1 ± 1.8 English version. The interview is available at no cost, days; range 0–10 days), and conducted by different inter- and is accessible via the website www.eda5.org. viewers to avoid contamination. All interviews were con- ducted in-person. Interviewers recorded the length of The eating disorder examination (EDE) v. 17.0D the interviews and completed a checklist where fulfilled The EDE [2] is a semi-structured interview generating diagnostic criteria and assigned DSM-5 diagnoses were operational ED diagnoses, as well providing measures of registered. Inclusion criteria were liberal, with patients the range and severity of ED features. The current ver- being considered eligible if they were medically stable, sion of the interview (v. 17.0D) generates all DSM-5 aged 16 years or above, and provided written consent to full-, subthreshold and unspecified eating disorder diag- participate in the study. No remuneration was offered. noses including AN, BN, BED, OSFED and UFED. The The study was approved by the Norwegian Regional three feeding disorders Pica, ARFID and rumination dis- Committee for Medical and Health Research Ethics (ref. order cannot be assessed using the EDE v. 17.0D. Four 2017/8130) and the Norwegian Data Protection Author- clinically derived subscales assess restraint, eating con- ity at Oslo University Hospital. Diagnostic data was col- cern, shape concern and weight concern. A global sever- lected at the two clinics between November 2017 and ity score is calculated by averaging the four subscales. June 2019. Psychometric studies show sample- and subscale dependent variations in internal consistency ranging Statistical analyses from poor to excellent and adequate concurrent and dis- All statistical analyses were carried out using IBM SPSS criminant validity [10]. The administration of the EDE Statistics Version 25. The sample size (n = 91) was mod- requires significant training, both in the technique of elled on the original validation study [8]. The EDE was interviewing, as well being familiar with the concepts used as a reference instrument in all analyses comparing and rules governing the ratings. The length of adminis- diagnoses. Since the EDE is not designed to assign pica, tration typically ranges from 45 to 75 min [2]. The Nor- rumination disorder or ARFID, reliability rates were not wegian version of the EDE v. 17.0D was used in the calculated for these three diagnoses. Kappas, sensitivity, current study. All EDE items were administered. DSM-5 specificity, negative and positive predictive value and ac- diagnoses were assigned using the diagnostic algorithms curacy was calculated for all assigned diagnoses. The five described in the EDE interview guidelines. latter measures are expressed in percentages. Kappas (ranging from − 1 to + 1) were used to measure diagnos- Procedure tic agreement. A slightly less liberal kappa standard was The first author of this paper (CLD) was trained and su- applied than that reported in the original EDA-5 valid- pervised in the use of the EDA-5 by BTW, and con- ation study (i.e. [11, 12]), with kappa coefficients (κ) be- ducted all EDA-5 interviews at RASP. The EDE ing interpreted as follows κ <0= “No Agreement”, interviews at RASP were conducted by two of the co- .0–.20 = “None”, .21–.39 = “Minimal”, .40–.59 = “Weak”, authors of this paper; a psychiatric nurse (CS) and a .60–.79 = “Moderate”, .80–.90 = “Strong” and > .90 = “Al- highly experienced clinician and senior researcher hold- most perfect” [13]. According to McHugh [14], any ing a PhD in psychology (ØR). CS was trained in the kappa above 0.60 indicates acceptable agreement among EDE assessment by ØR, who also supervised CS raters, whereas little confidence should be placed in re- throughout the data collection period. CLD trained the sults showing kappas below 0.6. In this study, sensitivity head of research and clinical psychologist (KV) and col- (true positives), is defined as the proportion of individ- leagues in the use of the EDA-5 at Modum Bad. All par- uals with a specific EDE diagnosis who were accurately ticipating staff at Modum Bad had prior experience and (i.e., identically) diagnosed using the EDA-5. Specificity formal training in the use of the EDE interview. (true negatives) is the proportion of individuals who did Participants were individuals receiving treatment at not receive a specific EDE diagnosis, who also did not re- one of the two Norwegian tertiary care centers: RASP at ceive that particular diagnosis using the EDA-5. The Oslo University Hospital, Oslo, or the Eating Disorder positive predictive value (PPV) is the probability that an Clinic at Modum Bad Psychiatric Center in Vikersund. individual diagnosed using the EDA-5 received that diag- At RASP, participants were recruited from the out- nosis using the EDE. The negative predictive value patient clinic and two inpatient adult clinics. A team of (NPV) is the probability that an individual, who did not four psychologists, one psychiatrist, two medical doctors receive a certain diagnosis using the EDA-5, did not re- and two psychiatric nurses completed the EDE and the ceive that diagnosis via the EDE. The closer PPV and EDA-5 interviews at Modum Bad. All participants were NPV values are to 1.0 (i.e. 100%), the higher the prob- inpatients. The order of the two interviews was counter- ability that the instrument being validated (in this case, balanced, and intended to occur no more than five days the EDA-5) is doing as good as “gold standard” (in this Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 4 of 7 case, the EDE interview) [14]. Accuracy is the proportion (Mean = 54.0 min, SD = 22.1), t (79) = 12.9, p < .0005 of true results, either true positives or true negatives. It (two-tailed). There were no significant differences be- is calculated as the sum true positives and true negatives tween sites in the time required to conduct the EDE and divided by the sample size (n). Since no OSFED-BED EDA-5 interviews. diagnoses were assigned, sensitivity, negative predictive value or accuracy could not be calculated for these Interview discrepancies diagnoses. Diagnostic distribution using the EDE and the EDA-5 is presented in Table 2. Diagnoses assigned using EDE and Results EDA-5 interviews were identical for 75 (82.4%) of the 91 A total of 101 adult participants were recruited to take cases. Among individual diagnostic categories, kappas part in the study, 57 from RASP and 44 from Modum ranged from moderate (.49) to perfect (1.00) agreement. Bad. Seven of the participants from RASP withdrew be- Kappas, sensitivity, specificity, positive and negative pre- fore having entered the study, and at Modum Bad, three dictive values as well as agreement accuracy are pre- participants were excluded due to interviews taking sented in Table 3. place too far apart (2–6 months). The final sample con- When comparing diagnoses assessed using the EDA-5 sisted of a total of 91 (87 females and 4 males) partici- and the EDE interview, 16 out of 91 patients (17.6%) re- pants. Demographic characteristics and assigned EDE ceived discrepant diagnoses (see Table 4). Twelve of diagnoses are presented in Table 1. Neither the EDE nor these were patients at RASP, and the remaining four the EDA-5 identified cases of OSFED-BED, OSFED-PD, were patients at Modum Bad. OSFED-NES or UFED diagnoses. Similarly, none of the participants received a diagnosis of ARFID, pica or ru- Discussion mination disorder. The label “OSFED Other” was used This study examined the validity of the Norwegian ver- to group participants whose eating disorder symptoms sion of the web-based diagnostic tool, the EDA-5, in strayed too far from the individual OSFED categories, assigning DSM-5 feeding and eating disorders. In line and deviated significantly from the examples given in with the original study [8], the Norwegian EDA-5 the DSM-5. Site differences were observed for age [F (2, quickly and efficiently generated DSM-5 diagnoses with- 89) = .46, p = .01] (patients recruited from Modum Bad out compromising diagnostic accuracy. As such, it is a were significantly older), but not for gender or BMI. The promising alternative to existing diagnostic tools, and EDA-5 was significantly quicker to administer (Mean = may facilitate the identification of eating disorders in 21.6 min, SD = 8.5) compared to the EDE interview clinical settings as well as in research. Table 1 Demographic characteristics and EDE diagnoses in the full sample (n = 91) and across sites Full sample RASP Modum Bad (N = 91) (n = 50) (n = 41) ** Age (years), mean (SD), range 30.9 (9.8) 28.5 (8.8) 33.8 (10.2) 17–56 17–52 19–56 Body mass index mean (SD), range 21.6 (7.9) 22.4 (9.2) 20.6 (5.8) 12.5–51.4 12.5–51.4 13.3–39.2 Female, n (%) 87 (95.6%) 48 (96.0%) 39 (95.1%) Diagnosis (EDE), n (%) AN 34 (37.4%) 15 (30.0%) 19 (46.4%) AN-R 16 (17.6%) 7 (14.0%) 9 (22.0%) AN-BP 19 (19.8%) 8 (16%) 10 (24.4%) BN 25 (27.5%) 11 (22.0%) 14 (34.1%) BED 6 (6.6%) 5 (10.0%) 1 (2.4%) OSFED 25 (27.5%) 18 (36.0%) 7 (17.1%) OSFED-AN 14 (15.4%) 10 (20.0%) 4 (9.8%) OSFED-BN 4 (4.4%) 3 (6.0%) 1 (2.4%) OSFED-Other 7 (7.7%) 5 (10.0%) 2 (4.9%) No ED 1 (1.1%) 1 (2.0%) 0 (0%) EDE The Eating Disorder Examination v. 17.0D, AN Anorexia Nervosa, AN-R Anorexia Nervosa Restrictive type, AN-BP Anorexia Nervosa Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders, ED Eating Disorder, RASP Regional Department for Eating ** Disorders. = Patients recruited from Modum Bad were significantly older than patients recruited from RASP (p ≤ .01) Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 5 of 7 Table 2 Diagnostic distribution in the full sample (n = 91) using OSFED group. The only two diagnostic categories where the EDE and the EDA-5 there were no discrepancies at all were BED and Diagnosis (%, n) EDE EDA-5 OSFED-BN, partially (due to low N), and possibly underscoring the clarity of the criteria for these categor- AN total 36.2% (33) 51.7% (47) ies. Considering the relatively new inclusion of BED as a AN-R 16.5% (15) 23.1% (21) separate diagnostic entity, these are promising results. AN-BP 19.8% (18) 28.6% (26) The majority of discrepant cases occurred between full- BN 27.5% (25) 20.9% (19) and sub-threshold AN and BN, likely reflecting different BED 6.6% (6) 6.6% (6) time frames used by the interviews in assessing DSM-5 OSFED total 27.5% (25) 19.8% (18) diagnostic criteria A (“Restriction of energy intake rela- tive to requirements leading to a significantly low body OSFED-AN 15.4% (14) 5.5% (5) weight”). DSM-5 does not explicitly specify a time frame OSFED-BN 4.4% (4) 4.4% (4) over which weight should be assessed; the EDA-5 aims OSFED-BED 0% (0) 0% (0) to determine whether an individual has been at a signifi- OSFED-PD 0% (0) 0% (0) cantly low body weight over the last three months, OSFED-NES 0% (0) 0% (0) whereas the EDE focuses only on the current weight. In OSFED-Other 7.7% (7) 9.9% (9) 12 of the 16 discrepant cases, the individual had been at a significantly low weight during the previous three ARFID NA 0% (0) months but was not significantly underweight at the Pica NA 0% (0) time of the interview. If the EDA-5 had focused on Rumination disorder NA 0% (0) current weight, or if the EDE (or the clinician doing the No ED 1.1% (1) 1 (1.1%) EDE) had prompted for lowest weight the previous three EDE The Eating Disorder Examination v. 17.0D, EDA-5 The Eating Disorder months, diagnostic agreement would have risen to Assessment for DSM-5, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP 95.6%. These numbers are in contrast to those in the Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders, OSFED-PD Purging measures’ original validation [8], suggesting that how Disorder, OSFED-NES Night Eating Syndrome, ARFID Avoidant Restrictive Food weight is judged in practice impacts diagnostic distribu- Intake Disorder, ED Eating Disorder, AN total includes AN-R and AN-BP, OSFED total includes OSFED-AN, OSFED-BN, OSFED-BED, OSFED-PD and OSFED-NES tions and rates. Whereas BN and BED diagnoses require a three-month minimum duration of binge eating (for Similar to the sample in the original study [8] the ma- BN and BED) and compensatory behaviors (for BN jority of participants in the present study were adult only), the DSM-5 does not specify the amount of time Caucasian females presenting with BMIs ranging from an individual should have been at a normal weight to be 13 to 51 kg/m . Also, similar to Sysko et al. [8] the high- considered recovered from AN [10]. Nor does it specify est level of agreement between the EDA-5 and the EDE the amount of time an individual should have been (1.00) was found for BN and the lowest kappa (0.62), underweight to be considered fulfilling the AN weight sensitivity (0.50) and PPV (0.89) were found for the criterium. The EDA-5, on the other hand, has Table 3 Agreement of the EDA-5 with the EDE interview (n = 91) Diagnosis κ Sensitivity Specificity Positive predictive value Negative predictive value Accuracy EDE ** AN total .72 1.00 .77 1.00 .72 .85 ** AN-R .77 .94 .92 .99 .73 .93 ** AN-BP .79 1.00 .90 1.00 .72 .91 ** BN .82 .76 1.00 .92 1.00 1.00 ** BED 1.00 1.00 1.00 1.00 1.00 1.00 ** OSFED total .62 .50 1.00 .89 1.00 1.00 ** OSFED-AN .53 .40 1.00 .89 1.00 1.00 ** OSFED-BN 1.00 1.00 1.00 1.00 1.00 1.00 ** OSFED Other .86 1.00 .98 1.00 .78 .98 ** No diagnosis .49 .33 1.00 .98 1.00 1.0 EDA-5 Eating Disorder Assessment for DSM-5, EDE Eating Disorder Examination, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding and Eating Disorders. AN total includes AN-R and AN-BP, OSFED total includes OSFED- AN, and OSFED-BN. No OSFED-BED, PD or NES cases were diagnosed using the EDE and the EDA-5. EDE is used as the reference assessment in all analyses. **p < .0005 Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 6 of 7 Table 4 Diagnoses assigned using the EDA-5 and the EDE (n = 91) EDA-5 AN-R AN-BP BN BED OSFED-AN OSFED-BN UFED No FED EDE AN-R 15 10 0 0 0 0 0 AN-BP 0 18 00 0 0 0 0 BN 0 5 19 00 0 1 0 BED 0 0 0 6 00 0 0 OSFED-AN 6 2 0 0 5 01 0 OSFED-BN 0 0 0 0 0 4 00 OSFED Other 0 0 0 0 0 0 7 0 No FED 0 0 0 0 0 0 0 1 EDA-5 Eating Disorder Assessment for DSM-5, EDE Eating Disorder Examination, AN Anorexia Nervosa, AN-R Restrictive type, AN-BP Binge-eating/Purging type, BN Bulimia Nervosa, BED Binge Eating Disorder, OSFED Other Specified Feeding or Eating Disorder, FED Feeding and Eating Disorder. Bold figures indicate diagnostic agreement between the EDA-5 and the EDE implemented the same 3-month time frame as required the four cases described above, can be explained by dif- for full threshold BN and BED. The rationale for not ferences in clinical judgment (e.g. case #3), different in- simply using the day-of-evaluation weight is that weight terpretations of diagnostic criteria (e.g. case #2 and #4) can fluctuate greatly over short periods of time, espe- or by different patient reports (e.g. case #1). These are cially in individuals enrolled in nutrition restoration reg- all important factors to consider in the diagnostic imens, including inpatient treatments for AN. Judging process, whether it is for research or clinical purposes. the AN weight criteria over the same duration used to judge the frequencies of behavioral disturbances in BN Strengths and limitations and BED, will presumably decrease diagnostic crossover Strengths of this study include standardized procedural and reduce the number of OSFED cases, the latter being steps for translation and validation, and data collection one of the top priorities of the revisions implemented in at two different sites. Also, high rates of diagnostic DSM-5 [6]. agreement were reached despite variability in inter- In one (case #1) of the four remaining discrepant viewers’ professional degree and specialty, supporting cases, the participant fulfilled BN criteria B through E, the utility of the EDA-5 across professions and experi- but reported objective binge eating without loss of con- ence with feeding and eating disorders. In addition, min- trol and was assigned an OSFED Other diagnosis in the imal resources were required to train interviewers to use EDA-5. In the EDE, the same patient fulfilled AN cri- of EDA-5. Neither the current study nor that of Sysko teria B and C (but not A since she was normal weight), et al. assessed individuals with ARFID, pica, or rumin- with subjective binge eating and received an OSFED-AN ation disorder, so the validity of the EDA-5 in character- diagnosis. One participant (case #2) received a BN diag- izing those disorders is unknown. In addition, more nosis using the EDE, and an AN-BP diagnosis using the research is needed to investigate the applicability of the EDA-5; the interviewing clinician using the EDE inter- EDA-5 in younger populations, and to determine the ex- view assessed her weight (and BMI which was 18.0) as tent to which these results generalize to males and eth- not being “significantly low” and therefore not fulfilling nically diverse samples. Replication in larger samples of the AN weight criterion. The interviewing EDA-5 clin- OSFED (e.g. in non-clinical samples) would be beneficial ician, on the other hand, assessed the patient’s weight/ to assess the diagnostic reliability of the assessment of BMI as being significantly low, fulfilling criteria for an OSFED subthreshold conditions. This may require more AN diagnosis. One participant (case #3) received an AN- detailed criteria to enhance diagnostic concordance R diagnosis using the EDE, and an AN-BP diagnosis within and between measures. Methodological limita- using the EDA-5; the use of laxatives was interpreted as tions such as the lack of inter-rater and test-retest reli- a compensatory behavior in the EDA-5 but not in the ability should also be noted. EDE. The final (case #4) discrepant case occurred as one participant fulfilled all but one BN criterion (D; self- Conclusion evaluation is unduly influenced by body shape and The EDA-5 is currently the only available semi- weight) and received an OSFED Other diagnosis when structured interview capable of assessing all the feeding assessed using the EDA-5; when using the EDE, the clin- and eating disorders described in DSM-5. The current ician judged that criterion D was met, and assigned the study replicates and extends the report of Sysko et al. [8] patient a BN diagnosis. Discordant diagnoses, such as in documenting that, even after translation to Norwegian Dahlgren et al. Journal of Eating Disorders (2020) 8:30 Page 7 of 7 and in the hands of clinicians who were not involved in In: Walsh BT, Attia E, Glasofer DR, Sysko R, editors. Handbook of Asssessment and treatment of eating disorders. Arlingston: American Psychiatric its development, the EDA-5 efficiently provides valid Association Publishing; 2016. diagnostic assessments of eating disorders following 5. Kessler RC, Ustun TB. The world mental health (WMH) survey initiative DSM-5 criteria. Although they have not been formally version of the World Health Organization (WHO) composite international diagnostic interview (CIDI). Int J Methods Psychiatr Res. 2004;13(2):93–121. tested, versions of the EDA-5 in Spanish and Turkish 6. Dahlgren CL, Wisting L. Transitioning from DSM-IV to DSM-5: a systematic are also available (see www.eda5.org). review of eating disorder prevalence assessment. Int J Eat Disord. 2016; 49(11):975–97. Abbreviations 7. Walsh TB, Attia E, Glasofer DR, Sysko R, editors. Handbook of teh assessment AN: Anorexia Nervosa; AN-R: Anorexia Nervosa Restrictive subtype; AN- and treatment of eating disorders. Arlington: American Psychiatric BP: Anorexia Nervosa Binge-Purge subtype; ARFID: Avoidant/Restrictive Food Association Publishing; 2016. Intake Disorder; BED: Binge Eating Disorder; BN: Bulimia Nervosa; 8. Sysko R, Glasofer DR, Hildebrandt T, Klimek P, Mitchell JE, Berg KC, et al. The CIDI: Composite International Diagnostic Interview; DSM: Diagnostic and eating disorder assessment for DSM-5 (EDA-5): development and validation Statistical Manual of Mental Disorders; ED: Eating Disorder; EDA-5: Eating of a structured interview for feeding and eating disorders. Int J Eat Disord. Disorder Assessment for DSM-5; EDDS: Eating Disorder Diagnostic Scale; 2015;48(5):452–63. EDE: The Eating Disorder Examination; EDRU: Eating Disorder Research Unit; 9. World Health Organization (WHO). Process of translation and adaptation of ICD: International Classification of Diseases; NES: Night Eating Syndrome; instruments. World Health Organization (WHO); 2016. http://www.who.int/ NPV: Negative Predictive Value; NYSPI: New York State Psychiatric Institute; substance_abuse/research_tools/translation/en/. OSFED: Other Specified Feeding and Eating Disorders; PD: Purging Disorder; 10. Burke NL, Tanofsky-Kraff M, Crosby R, Mehari RD, Marwitz SE, Broadney MM, PPV: Positive Predictive Value; RASP: Regional Department for Eating et al. Measurement invariance of the eating disorder examination in black Disorders; SCID: Structured Clinical Interview for DSM Axis 1 Disorders; and white children and adolescents. Int J Eat Dis. 2017;50(7):758–68. UFED: Unspecified Feeding or Eating Disorders 11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159–74. Authors’ contributions 12. Fleiss JL, Levin B, Paik CM. Statistical methods for rates and proportions, CLD, ØR, KV and CS conducted the majority of the diagnostic interviews. Third Edition. New York: John Wiley & Sons, Inc; 2003. CLD, ØR, BTW and KV analyzed and interpreted the data. All authors read 13. McHugh ML. Interrater reliability: the kappa statistic. Biochem Med. 2012; and approved the final manuscript. 22(3):276–82. 14. Parikh R, Mathai A, Parikh S, Chandra Sekhar G, Thomas R. Understanding Funding and using sensitivity, specificity and predictive values. Indian J Ophthalmol. This work was funded by the Regional Department for Eating Disorders, Oslo 2008;56(1):45–50. University Hospital, Ullevål HF, and by Modum Bad, Vikersund, Norway. Availability of data and materials Publisher’sNote The datasets used and/or analysed during the current study are available Springer Nature remains neutral with regard to jurisdictional claims in from the corresponding author on reasonable request. published maps and institutional affiliations. Ethics approval and consent to participate This study was approved by the Norwegian Regional Committee for Medical and Health Research Ethics (ref. 2017/8130) and the Norwegian Data Protection Authority at Oslo University Hospital. All participants provided written consent to participate in the study. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Author details Regional Department for Eating Disorders, Division of Mental Health and Addiction, Oslo University Hospital, Ullevål HF, Postboks 4950 Nydalen, 0424 Oslo, Norway. Department of Psychology, Bjørknes University College, Oslo, Norway. Department of Psychiatry, College of Physicians & Surgeons, Columbia University, New York State Psychiatric Institute, New York, USA. 4 5 Research Institute of Modum Bad, Vikersund, Norway. Institute of Clinical Medicine, University of Oslo, Oslo, Norway. Received: 14 April 2020 Accepted: 9 June 2020 References 1. Fukutomi A, Austin A, McClelland J, Brown A, Glennon D, Mountford V, et al. First episode rapid early intervention for eating disorders: a two-year follow- up. Early Interv Psychiatry. 2019;14(1):137–41. 2. Fairburn CG, Cooper Z, O'Connor M. The Eating Disorder Examination (EDE) 17.0D. 2018. https://www.credo-oxford.com/7.2.html. 3. First MB, Williams JBW, Karg RS, Spitzer RL. Structured clinical interview for DSM-5® disorders—clinician version (SCID-5-CV). Arlington: American Psychiatric Publishing Inc.; 2016. 4. Thomas JJ, Roberto CA, Berg KC. Assessment measures, then and now: a look Back at seminal measures and a look forward to the brave New World.

Journal

Journal of Eating DisordersSpringer Journals

Published: Jul 9, 2020

There are no references for this article.