SPINE Volume 30, Number 5, pp E123–E127©2005, Lippincott Williams & Wilkins, Inc.

Validation of the Korean Version of the OswestryDisability Index

Dong-Yun Kim, MD,* Sang-Ho Lee, MD, PhD,* Ho-Yeon Lee, MD, PhD,*Hyun-Ju Lee, RN,* Sang-Beom Chang, MD,† Sang-Ki Chung, MD,‡ andHyun-Jib Kim, MD, PhD‡

Study Design. Validation of a translated, culturallyadapted questionnaire.

Objectives. To translate and culturally adapt a Koreanversion of the Oswestry Disability Index (ODI) and to val-idate its use in Korean patients.

Summary of Background Data. The ODI is one of themost widely used and validated instruments for measur-ing disability in spinal disorders. However, no validatedKorean version of the index was available at the time ourstudy was initiated.

Methods. The study was carried out in three phases:the first was translation into Korean and cultural adapta-tion of the questionnaire; the second was a pilot study toassess the comprehensibility of the prefinal version andmodification; the third was a reliability and validity study ofthe final version. The Korean version was tested on 206patients with lumbar spinal disorders who had undergoneoperations at the authors’ institute. Test-retest reliability,internal consistency, concurrent validity, and construct va-lidity were investigated. Follow-up questionnaires were ob-tained from 39 patients at the 3-month postoperative fol-low-up meeting. Differences in the ODI, visual analog scale(VAS), and World Health Organization (WHO) quality of lifeassessment (WHOQOL-BREF) between preoperative andfollow-up questionnaires were evaluated. The correlation ofthe postoperative ODI with the pain rating on a visual ana-log scale and WHOQOL-BREF was also analyzed.

Results. Test-retest reliability was assessed with 88patients in a time interval of 48 hours. The intraclasscorrelation coefficient of test-retest reliability was 0.9167.Reliability estimated by the internal consistency reacheda Cronbach’s alpha of 0.84. The correlation of the preop-erative ODI with the pain rating on a visual analog scale(100 mm) was r � 0.425 (P � 0.0001). The correlationbetween three of the WHOQOL-BREF domains (physicalhealth, psychological health, and environment) and theODI was statistically significant. The correlation coeffi-cient between the ODI and physical health domain of theWHOQOL-BREF was r � �0.48 (P � 0.05). The correla-tions with psychological health and environment do-

mains were low with r � �0.192 and �0.160, respectively,even though statistically significant (P � 0.05). The correla-tion of the postoperative ODI with the pain rating on a visualanalog scale (100 mm) was r � 0.626 (P � 0.0001). Thecorrelation between all four domains of the WHOQOL-BREFand the postoperative ODI was statistically significant.

Conclusions. The results of this study indicate that theKorean version of the ODI is a reliable and valid instrumentfor the measurement of disability in Korean patients withlower back problems. The authors recommend this Koreanversion of the ODI for use in future clinical studies in Korea.

Key words: Oswestry Disability Index, reliability, valid-ity. Spine 2005;30:E123–E127

Condition-specific health status measures are commonlyused as outcome measures in clinical trials and to assesspatient progress in routine clinical practice. The Oswe-stry Disability Index (ODI) has emerged as one of themost commonly recommended condition-specific out-come measures for spinal disorders.1–3 The ODI, developedby Fairbank et al, consists of 10 items that assess the level ofpain and interference with several physical activities.2 Theindex was designed as a measure for assessment of the base-line disability and its change over time. Various drafts of thequestionnaire have been conceived. Version 1.0 of the ques-tionnaire was published in 1980, while the subsequentlyreleased Version 2.0 improved on the original concept andwas validated in a study by a Medical Research Councilgroup in the United Kingdom.4 Version 2.0 is recom-mended for general use by the authors.4

With the increase in the number of multinational andmulticultural research projects, the need to adapt healthstatus measures for use in other than the source languagehas grown rapidly. ODI has been translated into a numberof languages, such as Danish, Dutch, French, German,Spanish, and Swedish.3,4 However, before our efforts, noKorean version of the ODI had been validated. It has beenestablished that a scale or questionnaire cannot be trans-posed directly from one social environment to anotherwithout being revalidated for new, culture-specific condi-tions.5 Thus, the objectives of this study were twofold: totranslate into the Korean language a culturally adapted ver-sion of the ODI, and to then validate it in Korean patients.

Methods

The study was carried out in three phases: the first was trans-lation into Korean and cultural adaptation of the question-naire; the second was a pilot study to assess the comprehensi-bility of the prefinal version and modification; the third was areliability and validity study of the final version.

From the *Department of Neurosurgery and †Department of Ortho-pedic Surgery, Wooridul Spine Hospital, Seoul, Korea, and the ‡Depart-ment of Neurosurgery, Seoul National University Bundang Hospital,Acknowledgment date: December 8, 2003. First revision date: June 21,2004. Acceptance date: July 20, 2004.The manuscript submitted does not contain information about medicaldevice(s)/drug(s).No funds were received in support of this work. No benefits in anyform have been or will be received from a commercial party relateddirectly or indirectly to the subject of this manuscript.Address correspondence and reprint requests to Hyun-Jib Kim, MD,PhD, Department of Neurosurgery Seoul National University, Bun-dang Hospital 300 Gumi-Dong Bundang-Gu, Seongnam, Gyeonggi,Korea; E-mail: jibkim@snu.ac.kr

E123

Translation and Cultural Adaptation. The authors translatedand validated the ODI version 2.0. Two translations were ob-tained; one performed by a professional translator and one bya physician. The translators were bilingual native speakers ofKorean. The professional translator was not informed of thepurpose of the interpretation, nor of the cross-cultural conceptsinvolved in the index. The two translators were instructed toaim for conceptual rather than literal translation. Synthesis ofthe translations was conducted by an expert committee, whichconsisted of two translators, a neurosurgeon, an orthopedicsurgeon, and a methodologist. Consensus regarding the Ko-rean wording was reached by discussion.

The committee agreed on numerous initial modifications ofthe translations. For example, with regard to Section 4 (walk-ing) of the ODI, walking distances described in terms of milesor yards would be unfamiliar to the majority of Koreans. Dis-tances quoted in miles or yards (e.g., “1 mile,” “1/2 mile,”“100 yards”) were converted to kilometers or meters, whichare the familiar method of distance measurement to Koreans(respectively: 1 km, 500 m, 100 m). Further, the term “sociallife” in Section 9 (social life), when loosely translated, has a fewquite different meanings in Korean, so this word required sup-plementary wording such as “get-together, social gathering.”Moreover, in Section 10 (traveling), the term “traveling” usu-ally strictly means “a tour for pleasure or business trip” inKorean. Therefore, the supplementary term “transportation”was added for clarity.

Pilot Study and Modification of the Prefinal Version. Prefinaltranslation was tested by 50 subjects with low back pain toestablish that this version could be understood and that thequestions measured what they were intended to measure. The50 patients included 25 men and 25 women with a mean age of42 years (range: 22–67 years). The interviewer was asked todocument any problems that occurred during administration ofthe questionnaire. Each patient was asked to provide com-ments about the questionnaire and identify any words thatwere difficult to understand at the end of the interview. On thebasis of their comments, the final version was developed by thecommittee (see Appendix). Most of the 50 patients correctlyunderstood the questionnaire. However, only 21 patients(42%) answered Section 8 (pertaining to details of their sex life)of the ODI. In addition, some patients complained they wereembarrassed by this section, especially when they were giventhe questionnaire in front of family members. The committeedecided to omit Section 8 (sex life) from the final version of theKorean version of ODI to make questionnaire administrationsmoother and less compounded by the lower response rate of thissection. Six patients expressed concern about the exact meaning ofthe question in Section 9 (pertaining to social life). Supplementaryterms “hobby and leisure activity” were thus added.

Reliability and Validity Study. The study was carried out withthe final version of the questionnaire, on 206 patients withlumbar spinal disorders who had undergone operations at theauthors’ institute. The patients were given the Korean versionof the ODI on the day they consulted their physician. Eighty-eight patients, who were available with 48 hours time interval,were asked to complete a second questionnaire to assess test-retestreliability before surgery. Each time, they were also given the Ko-rean versions of the visual analog scale (100 mm) for pain, and ashort form of the World Health Organization (WHO) quality oflife assessment (WHOQOL-BREF). At the time of the 3-month

postoperative follow-up, the patients were given the same ques-tionnaire that they had been given before surgery.

WHOQOL-BREF is a quality of life assessment tool devel-oped by WHOQOL group and available in 19 different lan-guages, including Korean. It consists of 26 questions and yieldsfour domain scores that denote an individual’s perception ofquality of life in each particular domain (Table 1).6,7 The fourdomains include physical health, psychological health, socialrelationships, and environment. Higher scores denote a higherquality of life.

Test-retest reliability was measured by comparing the re-sults of the first and second administration of ODI. Reliabilitywas assessed with intraclass correlation coefficient. Cronbach’salpha was used to evaluate internal consistency. Concurrentvalidity is the comparison of two measures done at the sametime. Concurrent validity was measured by comparing the re-sponses to the ODI with the results of the visual analog scale(VAS). Construct validity is the extent to which the measure-ment corresponds to theoretical concepts concerning the phe-nomenon under study. To test the construct validity, a strongcorrelation between the ODI and the quality of life measuredby the WHOQOL-BREF was hypothesized.

Follow-up questionnaires were obtained from 39 patients atthe 3-month postoperative follow-up meeting. Differences inthe ODI, VAS, and WHOQOL-BREF between preoperativeand follow-up questionnaires were evaluated. The correlationof the postoperative ODI with the pain rating on a visual ana-log scale and WHOQOL-BREF was also analyzed.

Results

Table 2 contains sex, age, educational, and occupationallevel of the study subjects. The study enrolled 206 pa-

Table 1. WHOQOL-BREF Domains (From World HealthOrganization, with permission)

Domain Facets Incorporated Within Domains

1. Physical health Activities of daily livingDependence on medicinal substances and

medical aidsEnergy and fatigueMobilityPain and discomfortSleep and restWork Capacity

2. Psychological Bodily image and appearanceNegative feelingsPositive feelingsSelf-esteemSpirituality/Religion/Personal beliefsThinking, learning, memory and

concentration3. Social relationships Personal relationships

Social supportSexual activity

4. Environment Financial resourcesFreedom, physical safety and securityHealth and social care: accessibility and

qualityHome environmentOpportunities for acquiring new

information and skillsParticipation in and opportunities for

recreation/leisure activitiesPhysical environment

(pollution/noise/traffic/climate)Transport

E124 Spine • Volume 30 • Number 5 • 2005

tients with lumbar spinal disorders who had undergoneoperations at the authors’ institute. The mean age of thepatients was 46 years (standard deviation � 15), and thestudy group comprised 131 men and 75 women. Theirdiagnoses were lumbar disc herniation (146 patients),lumbar stenosis (28 patients), degenerative spondylolis-thesis (9 patients), isthmic spondylolisthesis (13 pa-tients), and discogenic back pain (10 patients). The per-formed operations were microscopic discectomy in 90patients (43.7%), percutaneous endoscopic discectomyin 56 patients (27.2%), transforaminal lumbar interbodyfusion with pedicle screw fixation in 16 patients (7.8%),anterior lumbar interbody fusion with pedicle screw fix-ation in 11 patients (5.3%), artificial disc replacement in4 patients (1.9%), intradiscal electrothermal therapy in 6patients (2.9%), and decompressive laminectomy in 23patients (11.2%).

Reliability estimated by the internal consistencyreached a Cronbach’s alpha of 0.84. Test-retest reliabil-ity was assessed with 88 patients on two occasions sep-arated by a time interval of 48 hours. There was nostatistically significant difference in sex, education, occu-pation, diagnosis, and operation method between thegroup completing the retest (88 patients) and those who didnot (118 patients). Mean age was 42 years (SD � 14.5) inthe retest group and 48.8 years (SD � 14.4) in the groupwithout retest. Difference in age was statistically significant(P � 0.001, Student’s t test). The intraclass correlation co-efficient of test-retest reliability was 0.9167.

Concurrent validity was measured by comparing theresponses to the ODI with the results of the visual analogscale. The correlation of the preoperative ODI with thepain rating on a visual analog scale (100 mm) was r �0.425 (P � 0.0001).

In determining construct validity, a correlation be-tween the ODI and WHOQOL-BREF was hypothesized.

Table 3 summarizes the correlation between four do-mains of the WHOQOL-BREF and ODI. The correla-tion between three of the WHOQOL-BREF domains(physical health, psychological health, and environment)and the ODI was statistically significant. The correlationcoefficient between the ODI and physical health domainof the WHOQOL-BREF was r � �0.48 (P � 0.05). Thecorrelations with psychological health and environmentdomains were low with r � �0.192 and �0.160, respec-tively, even though statistically significant (P � 0.05).

Follow-up questionnaires were obtained from 39 pa-tients at the 3-month postoperative follow-up meeting.Statistically significant improvement was observed in theODI, VAS, and physical health domain of the WHO-QOL-BREF (Table 4). The ODI score was improvedfrom 41.7 (SD � 17.5) to 18.4 (SD � 13.3). The visualanalog scale showed a decrease from 7.2 (SD � 2.1) to2.4 (SD � 1.9). The physical health domain of the WHO-QOL-BREF showed improvement, increasing from 10.3(SD � 2.7) to 13.2 (SD � 2.4). The correlation of thepostoperative ODI with the pain rating on a visual ana-log scale (100 mm) was r � 0.626 (P � 0.0001). Thecorrelation between all four domains of the WHOQOL-BREF and the postoperative ODI was statistically signif-icant (Table 3).

Discussion

The results of this study indicate that the Korean versionof the ODI is a reliable and valid instrument for themeasurement of disability in Korean patients with lowerback problems. The Korean version of the ODI showedsimilar results to those presented in the original ODIEnglish version with test-retest reliability,2,8,9 internal

Table 2. Demographic Data

Pilot Study Group Validation Study Group

N 50 206Age* 42 (14.5) 46 (14.8)Sex

Male 25 (50) 131 (64)Female 25 (50) 75 (36)

EducationNone 2 (4) 5 (2)Elementary school 8 (16) 48 (23)Middle school 9 (18) 31 (15)High school 15 (30) 64 (31)University 16 (32) 58 (28)

ProfessionBlue collar 9 (18) 34 (17)Intermediary level 8 (16) 31 (15)White collar 9 (18) 25 (12)Retired 3 (6) 11 (5)Housewife 9 (18) 32 (16)Others† 10 (20) 25 (12)NA‡ 2 (4) 48 (23)

* Mean (Standard deviation).† “Others” includes: students, those on disability, and unemployed.‡ Not available.Values in parentheses are percentages.

Table 3. Pearson’s Correlation Coefficients of the ODIwith 4 domains of the WHOQOL-BREF

Domains

Preoperative ODI(n � 206)

Postoperative ODI(n � 39)

Correlation P Correlation P

Physical health �0.480 0.0001 �0.510 0.001Psychological �0.192 0.006 �0.569 0.0001Social relationships �0.090 0.199 �0.366 0.028Environment �0.160 0.022 �0.364 0.025

Table 4. Descriptive Data of ODI, VAS, andWHOQOL-BREF

Preoperative Postoperative P

ODI 41.7 (� 17.5) 18.4 (� 13.3) 0.0001*VAS 7.2 (� 2.1) 2.4 (� 1.9) 0.0001*WHOQOL-BREF

Physical health 10.3 (� 2.7) 13.2 (� 2.4) 0.0001*Psychological 12.7 (� 2.0) 12.8 (� 2.7) NS†Social relationships 13.4 (� 2.2) 13.5 (� 2.4) NSEnvironment 11.4 (� 2.1) 11.9 (� 2.6) NS

*Paired t test.†Not significant.

E125Korean Version of the Oswestry Disability Index • Kim et al

consistency,2,9–11 and correlation with a visual analogscale.8 Although there has not been any comparablestudy, significant correlation was found between theWHOQOL-BREF, especially in the physical health do-main, and the Korean version of the ODI. Correlationbetween the Korean version of the ODI and VAS andWHOQOL-BREF was also significant in postoperativefollow-up questionnaires.

The sex life question in Section 8 was omitted in thisKorean version of the ODI. This question is unaccept-able in some cultures, and has been omitted in certainstudies.12–15 Fujiwara et al reported low response rate(60%) to the sex life question in Japan in their validationstudy of Japanese version of the ODI, although they didnot omit the sex question from the Japanese version ofthe ODI.16 We omitted Section 8 (sex life) from the finalversion of the Korean version of ODI to make question-naire administration smoother and less compounded bythe lower response rate of this section. Although the ODIwas developed originally as a self-administered question-naire, the administration of the ODI by telephone inter-view has been reported.17,18 Regardless, in Korean cul-ture it remains quite difficult to ask questions related tothe sex life of the interviewee.

To the authors’ knowledge, this Korean version of theODI is the first condition specific outcome instrument forspinal disorders to have been validated in Korean. De-velopment and validation of multiple-language versionsof existing validated questionnaires plays a key role instandardizing the outcome measurement and increasingthe statistical power of clinical studies. The results of thisstudy demonstrated that the ODI was successfully cross-culturally translated into Korean without losing the psy-chometric properties of the original versions. Therefore,this Korean version of the ODI seems to be a reliable andvalid outcome measure for assessing functional status ofpatients with spinal disorders. The use of this Koreanversion of the ODI can be recommended in future clinicalstudies in Korea.

Key Points:

● The Oswestry Disability Index was cross-culturally translated into Korean and validated.● The Korean version of the Oswestry DisabilityIndex had acceptable reproducibility, concurrentvalidity, and construct validity.● The results of this study indicate that the Koreanversion of the ODI is a reliable, valid instrument forthe measurement of disability in Korean patientswith lower back problems.

AcknowledgmentsThe documentation of translation, back translation, andthe results of the reliability and validity study were re-ported to the original developer, Dr. Jeremy Fairbank.The authors thank Dr. Jeremy Fairbank for granting per-mission to translate the Oswestry Disability Index intoKorean and appraisal of this Korean version of the ODI.The authors are also grateful to In-Kyung Cho, Hye-RanLee, In-Sook Cho, and Jihee Hwang for their invaluableassistance on this project.

References

1. Deyo RA, Battie M, Beurskens AJ, et al. Outcome measures for low backpain research: a proposal for standardized use. Spine 1998;23:2003–2013.

2. Fairbank JC, Couper J, Davies JB, et al. The Oswestry low back pain disabil-ity questionnaire. Physiotherapy 1980;66:271–273.

3. Roland M, Fairbank J. The Roland-Morris Disability Questionnaire and theOswestry Disability Questionnaire. Spine 2000;25:3115–3124.

4. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine 2000;25:2940–2953.

5. Beaton DE, Bombardier C, Guillemin F, et al. Guidelines for the process ofcross-cultural adaptation of self-report measures. Spine 2000;25:3186–3191.

6. Min SK, Kim KI, Suh SY, et al. Development of Korean Version of WorldHealth Organization Quality of Life Scale (WHOQOL). J Korean Neuropsy-chiatr Assoc 2000;39:78–87.

7. The WHOQOL Group. Development of the World Health OrganisationWHOQOL-BREF Quality of Life Assessment. Psychol Med 1998;28:551–558.

8. Gronblad M, Hupli M, Wennerstrand P, et al. Intercorrelation and test-retestreliability of the Pain Disability Index (PDI) and the Oswestry DisabilityQuestionnaire (ODQ) and their correlation with pain intensity in low backpain patients. Clin J Pain 1993;9:189–195.

9. Kopec JA, Esdaile JM, Abrahamowicz M, et al. The Quebec Back PainDisability Scale: conceptualization and development. J Clin Epidemiol 1996;49:151–161.

10. Fisher K, Johnson M. Validation of the Oswestry low back pain disabilityquestionnaire, its sensitivity as a measure of change following treatment andits relationship with other aspects of the chronic pain experience. PhysiotherTheory Pract 1997;13:67–80.

11. Strong J, Ashton R, Large RG. Function and the patient with chronic low-back-pain. Clin J Pain 1994;10:191–196.

12. Coleman RE, Houston S, Purohit OP, et al. A randomized phase II study oforal pamidronate for the treatment of bone metastases from breast cancer.Eur J Cancer 1998;34:820–824.

13. Poussa M, Schlenzka D, Seitsalo S, et al. Surgical treatment of severe isthmicspondylolisthesis in adolescents: reduction or fusion in situ. Spine 1993;18:894–901.

14. Purohit OP, Anthony C, Radstone CR, et al. High-dose intravenous pamidr-onate for metastatic bone pain. Br J Cancer 1994;70:554–558.

15. Tyrrell CT, Bruning PF, May-Levin F, et al. Pamidronate infusions as single-agent therapy for bone metastases: a phase II trial in patients with breastcancer. Eur J Cancer 1995;31A:1976–1980.

16. Fujiwara A, Kobayashi N, Saiki K, et al. Association of the Japanese Ortho-paedic Association Score with the Oswestry Disability Index, Roland-MorrisDisability Questionnaire, and Short-Form 36. Spine 2003;28:1601–1607.

17. Hupli M, Sainio P, Hurri H, et al. Comparison of trunk strength measure-ments between two different isokinetic devices used at clinical settings. J Spi-nal Disord 1997;10:391–7.

18. Mayo NE. Letter. Spine 1995;20:1535–1536.

E126 Spine • Volume 30 • Number 5 • 2005

Appendix

E127Korean Version of the Oswestry Disability Index • Kim et al