Bulletin, Indian Psychiatric Society, W. Bengal 5(1) 5-9, 1996 Adaptation of Beck Depression Inventory Into Bengali and its Clinical validity. Saugata Basu’, D. Chattopadhyay,** A. Deb,** S. Ash,** J. Samajder,** Z. Deb,* Sutapa Basu* The current study is an attempt to develop a standardized Bengali Beck Depression Inventory and to study its applicability in our clinical setting. The standard procedure was ‘followed whereby a Bengali translation was developed by a group of professionals. This was ‘administered to groups of 10 patients diagnosed as having Depressive disorder (F 32/F 33) ‘and 10 normal subjects. The responses and comments were then used to develop a modified version which was in turn administered to 60 clinically depressed patients (F 32/F 33). Item analysis and split-half reliability were calculted and found to be satisfactory. Severity levels on the basis of scores were calculated. Similarly a retranslated English version was administered to a group of bi-lingual patients with depression (F 32/F 33) to find out statistical validity. To establish clinical validity, scores of 30 normal subjects were conspared with scores of patient group. Both statistical validity and clinical validity were found to be adequate. The Beck Depression Inventory or and its testretest _elinbibty is BLDIL. (Beck, Word, Mendelson, Mock and Erbaugh, 1961) was first developed as an interview schedule where the interviewer read out loudly each item to the patient while the patient read their own copy of the scale, and gave their choice. It is now widely used asa self-rating scale + Itconsists of 21 items, each containing. 4 or 5 statements ranked in order of severity, The patient chooses the statement closest to their present mental state. The split-half reliability is around 0.9 + Clinical Psychologist Psychiatrist. MON Foundation, Keikhali, VILP. Road engoute to Caleutta Airport, Pin: 743 518 approximately 0.75. It has consistently’ been found to correlate well with clinicians ratings of severity of depression, as well as with other scales of depression. BDI is widely used'.to." measure seventy of depression, to find owt efficacy of particular mode of therapiy“including, pharmacotherapy, also in’ cognitive Behaviour Therapy, it is 2 useful tool which can be used directly or indirectly in, the process of therapy. It has the advantage of being useful across a great range of severity levels and in clinical, sub-clinical and student population (Williams, 1992) B.D1. cannot be used to diagnose depression in the absence of 8 prior clinical diagnosis - itis only a measure ofseverity once the clinical diagnosis has been made. This is an important point, as people may have inflated B.D.L. score fora huinber of reasons (eg. bereavement reaction, cheonic low self-esteem). In sense, the generalizability of many research finding on the basis of B.D. may be called into question (Depue and Monroe, 1978 ; Williams, 1992), However, this is a reflection on the use of the scale itself, which remains probably the best all round scale of its type (Williams, 1992). B.D1. is being used widely in India both in clinical settings and research contexts. However, no study describing the applicabitity of 8.1L. in Indian context is available. Vast majority of our clinical population doesn’t have access to English language. This effectively limits the usage of B.D.I.in Bengali speaking population as, there is no Bengali version available. Moreover, in absence of studies, the applicability of B.D1in Indian situation jn general and Bengalee population in Particular is uncertain. Keeping these two problems in view, the present study purports to :— a) Prepare a Bengali adaptation of BDL, b) Find out the clinical applicability of BDL in a group of Bengali apéaking population with a clinical ‘lagnosis of depression ~Materialg'and Methods -— Phase-l: All the items were translated from English to Bengali by 2 mental health professionals independently. These Bengali versions of the scale was shown to 3 other mental health professionals and another person conversant with both the vsaguages viz. Bengali and English. Their comments were incorporated and the first of the scale or D.I. 1 was prepared Ja agreement » "5 persons. 13 5 males and 5 females) diagnosed as having 2 depressive disorder (F 32 or F 33) according to ICD-10 and a group of 10 rormal sjoets ( males an females) tn the first group, patients with psychotic sympions (fae? and F323) were eacluded, The normal subjects were screened| using general Health Questionnaire - 28 or GHQ-28 (Goldberg and Hiller, 1979). Their comments on the language of the scale, ie. D.l.1 were taken into consideration and accordingly structure of the sentences and language were changed. This way second version of the seale or D.L. was prepared. Phase-lII': D.l. was tried out on a purposive sample of 60 patients (30 males and 30 females) diagnosed as having a depressive disorder (F32 and F33) secording to ICD-10. Patients with psychotic symptoms (F32. 2 and F33.3), past history suggestive of any psychotic illness, history suggestive of any organic involvement, any other simultaneous psychiatric diagnosis were excluded. All the patients hailed from urban setting. All were Hindus, studied at least til Class Vill, could ‘read and write Bengali Majority were married (HU) and from nuclear families (75%). The mean age was 332 (+ 7.94) with a range 19 years to 45 years. For the purpose of item-analysis, item-total correlations were done on all items of D.l, Results are given in Table-l. (See Table} Phase-IV : For the purpose of finding out reliability, split-half reliability. was done on all 21 items of D.L. (Tabie-th, (See Table 11) Phase-V : For finding out se level on the basis of scores of DL, total Tange of scores (12-55) were divided into 3 groups : Group-t (first 20 scores, ie,, 12- 22, N = 24): Croup-ll (next 20 scores, ie mt verity,Table-I 2 Item total correlation of Dal. (df = 58) tems Correlation Coefficient(r) A. Sadness ao B. Pessimism at C. Sense of failure an D. Dissatisfaction 28° "EB Guilt “an F. Expectation of Punishment aus GC. Self-Dislike so" H. Self-accusation an 1. Suicidal ideas os J. Crying aa“ K. Irritability an L. Social Withdrawal an M. Indecesiveness 38 N. Body image change 38 ©. Work retardation 54 P. Insomnia 3st QQ Fatiguabitity 4g R. Anorexia aor S. Weight loss are T. Somatic preoccupation 36 U. Loss of Libido san + Significant at 05 level + Significant at 01 level. Group Il and Group-lll. Results are given in Table-Lil & IV. (Gee Table I & IV) Phase-V1 : Validity - For finding out statistical validity, D.L. or Bengal version of B.D.L. was back translated into English by one person who was conversant in both English and Bengali ; experienced in translation work and was not involved in the present study. Then BDJ. and | translated version of Dal. were basically e | gue bicd ent om 1b? -| feKenls Thee 1b paknle” Bengali speaking, but they were convers: in English. as well. The inclusion ad ‘exclusion criteria for screening the patient were same as on Phase-lIl. To eliminate” ‘effect of performance or progressive error. Intra-subject counter-balancing method (abba) was used. Correlation cvefficient between B.D. ‘and D.l. scores was calculated, Result is given in Table-V. (See Table V)Split half reliability of DL. ; = (N=60) 81 Table-II Means and standard deviations of total scores of DI. for Group-l ind Group-Il Group-t (N= 24) Group-ll (N= 24) Significance 1683 +38 27.79 £33 t= 1064 . p< 001 Table-IV Means and standard deviations of total scores of D1 for Group-ll and Group-Ith Group-It (N= 24) Group-Ill (N = 24) Significance 2779 £33 4091 + 69 t= 776 p< oot Table-V Correlation coefficient between B.D and translated version of Dl. Scores (N=16) 97 Table-VI Means and standard deviations of total scores of D.L of normals and group-1 “Normals (N'=30) Group-I (N=24) Significance oe 45 4 26 16.83 +38 t= 47 p< ou Phase-VI : For estabiishing clinical This normal group was compared validity, DL was tried out on 30 normal with Croup of Phased or matin subjects (15 males and 15 females), depressed groupon total scoreot D1 with screened through GHQ-28 and the help of test. Results are given in comparable with clinical group weresame —Table-VI as in Phase, (GeeTable VI)Se, ‘gesults and Discussion :— Table-I * Suggests that each item of Di. significantly contributes to total scare as correlation coefficient of each item with total core is significant. So all the items of Dil. were retained for the final version, Table-II Suggests that D.lisa highly reliable clinical tool as split-half reliability is very high. This value is comparable with the value (0.9) found by Beck etal (1961), Table-IIl & TablelV indicate that Di. can-be used as measure of severity levels of depression as Croup-1 significantly differs from Group-ll and Group-Il_ significantly differs from Group-Il. this suggests that Group-t is midly depressed group witha range 12-22, Group-Il is moderately depressed group with a range 23-32, and Group-ill & severely depressed group ‘with, scores more than 32. The present findings are somewhat different from the earlier findings by Murphy, Simons, Wetzel and Lustman (1984) who used the following definition of severity levels: not depressed, 0-9 ; mildly depressed, 10-15 ; moderately depressed, 16-24 : severely depressed, 25 and above. However, Williams (1992) commented that these cut. offs are somewhat arbitrary, and differ between researchers. So on the basis of Present findings, these ranges can be used in stratifying severity in Bengali speaking clinical population. Table-V_ reveals a significantly high correlation coefficient between B.D.L. and translated version of DL. scores. This suggests B.D.L and Dil. can measure the effective in measuring,different aspects of depression as B.D. Vable-VI_ suggests that normals scored significantly lower than mildly depressed group, i.e. Group-l. This indicates that Di. can. discriminate between even clinically mildly depressed group and normals. CO CONCLUSION — The study show that D1. can be used asa standardized Dengali version of Deck Depression Inventory. Dil valid instrument and it measus aspect of depression as effectively as BDL; DA. can discriminate between normals and even mildly depreseed Patients and Dl. can be used to measure levels of severity of depression reliatle and res different (This paper was presented by D. ‘Chattapadhyay at the Annual Conference of the Eastern Zonal Branch of Indian Psychiatric Societwheld at Bhubaneswar ‘on 18 October 1995,] REFERENCES ;— - Beck AT; Ward CH: Mendeloon, M.; Mock, J.£. and Erbaugh, J.K. (1961), An inventory for measuring depression, Archives of General Psychiatry, 4,561-571 Depue, R.A. and Monroe, 5.4. (1978). Learned helplessness inthe Perspective af the depresave disorders Conceptual and definitional - issues, Joumal of Abnormal Psychology, 87, 2-20, Freeman, A. (1994) Depression : 4. Cognitive therapy approach - a viewers normal. Newbridge Professional Programs : New York Goldberg, D.P. and Hiller, VE. (1978). & scaled version cf the Cene:si Health Questionnaire Medicine, 9, 139-116, sycholag:eo% ‘urphy, G.E, Simons, a.D.; Weisel, R.D. and Lustman, P.J. (1984), Copnve therapy and pharmacotherary : Singleang together in the treatment ci depression Archives of General Psychiatry, 4l, 3 Williams, J.M.G.. (1992). The psychological treatment of depression Routledge: London,