Crc.gov.my

International Journal of Impotence Research (2003) 15, 329–336& 2003 Nature Publishing Group All rights reserved Cross-cultural adaptation and validation of the English version ofthe International Index of Erectile Function (IIEF) for use inMalaysia TO Lim1*, A Das1, S Rampal1, M Zaki2, RM Sahabudin3, MJ Rohan4 and S Isaacs5 1Clinical Research Centre, Kuala Lumpur Hospital, Kuala Lumpur, Malaysia; 2Department of Nephrology, Kuala LumpurHospital, Kuala Lumpur, Malaysia; 3Department of Urology, Kuala Lumpur Hospital, Kuala Lumpur, Malaysia; 4Department of Urology, Selayang Hospital, Selangor, Malaysia; and 5Department of Primary Care, Kuala LumpurHospital, Kuala Lumpur, Malaysia We adapted the English International Index of Erectile Function (IIEF) into Malay. This wasdifficult as many sex-related terms do not exist in colloquial Malay. In the pretest, there was nodifficulty with comprehension and all subjects judged the Malay IIEF equivalent to the English IIEF.
After slight modification, a final instrument was evaluated in two studies. Study A included 136subjects. It showed that the instrument had good reliability and discriminant validity. The factorstructure of the English IIEF was not reproducible. Study B included 26 ED subjects who underwentoral sildenafil therapy. The Malay IIEF was sensitive to treatment response. The area under theROC curve of the Malay IIEF-5 was 0.86; the optimal cutoff score has a sensitivity of 85% andspecificity of 75%. The results suggest that the Malay IIEF requires more work, but the Malay IIEF-5has acceptable measurement properties to recommend its use in clinical practice and research.
International Journal of Impotence Research (2003) 15, 329–336. doi:10.1038/sj.ijir.3901009 Keywords: questionnaire; sexual function; erectile function; diagnostic test; cross-culturaladaptation Laboratory-based physiological measures of erec- tile function (EF) such as volumetric plethysmogra-phy (Rigiscan), strain gauge plethysmography and Erectile dysfunction (ED) is defined as the inability erectiometer are not readily accessible in this to achieve or maintain an erection sufficient for country. While they provide objective measure- satisfactory sexual performance.1 A recent preva- ments, they have important design and methodolo- lence survey (unpublished data) in Malaysia esti- gical weaknesses, and may not be the best method to mated 16% of men aged 40 years or older had diagnose ED.5 It is preferable to assess sexual ‘moderate to complete ED’. And the vast majority of function in naturalistic setting with patient self- these men had not sought treatment. Effective report techniques. Recently, the International Index medical treatment for ED has also become available of Erectile Function (IIEF), a self-report composite in recent years, for example, oral drugs like multidimensional measurement scale for male sex- sildenafil (VIAGRA, Pfizer Inc),2,3 tadalafil, varde- ual function has become available. The instrument nafil, oral phentolamine (Vasomax), sublingual was supported by rigorous psychometric, cultural apomorphine (Uprima).4 Thus, ED is probably and linguistic validation.6 This instrument was common, under reported and under diagnosed, developed primarily for research use, and indeed and yet treatable. Doctors in Malaysia, like their the original clinical studies that demonstrated the counterparts elsewhere, need to improve their efficacy of sildenafil2,7 had utilized IIEF score as the efficacy measure. An abbreviated version of the IIEFcomprising only five of the 15 items in the originalinstrument, the so-called IIEF-5, had subsequentlybeen developed as a diagnostic tool for ED tocomplement clinical judgment. The IIEF-5 too was *Correspondence: TO Lim, Clinical Research Centre, supported by rigorous diagnostic validation8 and Level 3, Dermatology Block, Hospital Kuala Lumpur, could reliably be used for the diagnosis of ED and for 50586, Jalan Pahang, Kuala Lumpur, Malaysia.
determining the responsiveness to treatment. As a E-mail: [email protected] 2 September 2002; revised 19 December 2002; result, the IIEF-5 has become very popular in clinical practice and rightly so. Both the IIEF and Cross-cultural adaptation and validation of IIEF in Malaysia its abridged version IIEF-5 are simple and yet have One of the two translators was provided with adequate measurement properties for research and sufficient information to make him aware of the clinical applications, respectively.
purposes of IIEF, and the concepts and domains Unfortunately, the validity of IIEF and IIEF-5 are underlying it. This was to ensure reliable restitution only established hitherto in several European lan- of the intended measurement. The other translator guages. There is a need for research to extend its was deliberately blinded to the intent of and validity to the languages commonly used in other concepts underlying IIEF. The translation was then countries. This requires us to translate and cross- contrasted with the one produced under nonblinded culturally adapt the original English instrument such as to minimize the influence of language and The two translations were then back translated other cultural attributes on the subject’s response, into English independently by two translators. The thereby assuring validity of inference. The objec- back translators’ mother tongue was also Malay.
Ideally, it should have been English, but the serviceof such a person was difficult to obtain. However, (1) To translate and adapt the original US English the translator had provided documentary evidence version of IIEF and IIEF-5 into the Malay of fluency in both spoken and written English and of language, such that the translated version is previous translation records. Both back translators crossculturally equivalent to the original ver- were blinded to the intent and concept of IIEF.
A Review Panel was constituted to determine the (2) To determine the psychometric properties of the face validity and crosscultural equivalence of the translated Malay version of the IIEF and the translated instrument. The committee was chaired diagnostic properties of the IIEF-5.
by the principal investigator and comprised of threemedical members with experience in ED and twoother lay members. All members had to demonstrate The Panel was tasked to produce a final version based on both the translated and back-translated The study was conducted in two phases. The first versions of IIEF, as well as the original version. All phase aimed to obtain a crossculturally equivalent versions were considered equally important. The Malay version of the original IIEF. The second phase then determined the measurement properties of theMalay version of both IIEF and IIEF-5 obtained from (a) Modify or reject items and instructions of the translated IIEF; or request for more translationand back translation.
(b) Assure that the translation was fully compre- (c) Verify the crosscultural equivalence between the original English IIEF and the translated version.
The IIEF consists of 15 items that evaluates five The emphasis was on conceptual equivalence distinct domains of the male sexual function. These rather than linguistic or semantic equivalence.
domains and the corresponding IIEF items that tapinto them are EF (questions 1, 2, 3, 4, 5, and 15), Through this iterative process, and by allowing orgasmic function (OF, questions 9 and 10), sexual the Panel to compare the various translated and desire (SD, questions 11 and 12), intercourse back-translated versions of IIEF, it was hoped the satisfaction (IS, questions 6, 7, and 8), and overall translation that was most conceptually equivalent to satisfaction (OS, questions 13 and 14). The IIEF-5 is the original English IIEF could be determined.
a subset of the IIEF and comprises items 2, 4, 5, 7, The Malay version finally determined by the Panel was then subjected to further pretest. Theobjectives were to check for equivalence between the original English IIEF and the final translatedversion as well as to assess comprehension andreadability. This required confirmation that the The original English version of the IIEF was items could be understood without arousing reluc- independently translated into the Malay language tance or hesitation. In all, 14 bilingual volunteers by two qualified translators from the National were recruited for the pretest of the Malay version of Translation Institute. It is stressed that the selection the IIEF. All subjects had provided verbal informed of translators was not based on educational criterion consent, were aged greater than 40 years, had a alone as such persons often may not be culturally condition that predisposed them to ED such as representative. Each translator translated the Eng- hypertension, diabetes mellitus, end-stage kidney lish IIEF into his mother tongue, which was Malay.
failure, and finally had passed both the English and International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in MalaysiaTO Lim et al Malay language fluency tests administered before In the pretest, subjects were requested to complete both the English and Malay versions of the IIEF and The sample size for this study was determined based then were interviewed by a trained interviewer. The on the parameter, the test–retest repeatability. This interviewer used random probe technique to repeat- is measured by the intraclass correlation (r). We edly ask subjects ‘What do you mean?’, and then expect IIEF to have an r of 0.8 in this study,5 and an r probed and encouraged the subject to elucidate his of 0.7 or higher would be acceptable to us. Thus, we or her understanding of the item in an open-ended defined H0: r0 ¼ 0.7 and H1: r1 ¼ 0.8. Using a two- manner. This was to ensure the translated item was sided test as suggested by Walter et al,9 with b ¼ 0.2 understood as having a meaning equivalent to the (80% power) and a ¼ 0.05, a sample size of 117 one in the English IIEF. At the end of the interview, evaluable subjects would be required. Assuming each subject was asked to rate the equivalence of 45% of subjects might refuse to repeat self-admin- each item on a five-point Likert scale (Completely istration of IIEF, a total of 170 subjects would have to equivalent, Almost equivalent, Neutral, Not quite The psychometric properties of the Malay version The outcome of the Phase 1 study described above was a presumed equivalent instrument in the Malaylanguage that had been adapted to the local culture.
(1) Test–retest repeatability was measured by Pear- son’s correlation coefficient and intraclass cor-relation. This was estimated using the ANOVA (2) Internal consistency of the domains and total score were measured by the Cronbach’s a statistic.
Cross-cultural equivalence does not imply equiva- (3) The factorial validity was assessed by a princi- lence of psychometric properties. These properties pal components analysis with varimax rotation may change in the process of translating and and extracting only the first five factors to attempt to reproduce the factor structure of IIEF.
Two different studies were designed to validate (4) Discriminant validity or the ability of IIEF to discriminate between subjects with and withoutclinical diagnosis of ED as determined indepen- (1) Study A enrolled 171 subjects comprising 111 dently by a physician. This was evaluated using normal healthy volunteers from the community (Community sample) and 60 patients attending (5) Treatment responsiveness (sensitivity) was eval- primary care clinics (Clinical sample). All uated by comparing the pretreatment and post- subjects had provided verbal informed consent, treatment domain scores of patients who under- had passed a Malay literacy test (to ensure they go sildenafil therapy and who self-rated as could self-administer the Malay IIEF) and were aged greater than 40 years. After initial screen- (6) Treatment specificity was evaluated by compar- ing, subjects self-administered the Malay ver- ing the pretreatment and post-treatment domain sion of the IIEF. A physician then evaluated the scores of patients who undergo sildenafil ther- subjects to determine the presence of ED based apy and who self-rated as nonresponder at the on medical history, physical examination and objective testing where available, while blinded (7) Finally, the diagnostic performance of IIEF-5 to subjects’ responses to IIEF. At exit from the clinic, subjects were requested to complete theMalay version of the IIEF a second time by self-administration.
(2) Study B enrolled 32 patients who were clini- cally evaluated to have ED and consented to undergo a trial of sildenafil therapy 50 mg 1 hprior to sexual activity (ED sample). After 4 It was fairly easy to obtain a literal word for word weeks of therapy, subjects returned to the clinic translation using formal Malay vocabulary and where they completed the Malay version of IIEF grammatical rules. However, we wanted to strive by self-administration again. They were also for conceptual rather than literal equivalence be- asked a global efficacy question: ‘Did the tween the English IIEF and its Malay version. As it treatment with Sildenafil improve your erec- turned out, colloquial translation and cross-cultural tion?’. Subjects were designated as a ‘responder’ adaptation of the English IIEF into Malay was or ‘non-responder’ based on their response to difficult. Short of resorting to ‘vulgar’ words, many sex-related terms do not exist in colloquial Malay.
International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in Malaysia The words in formal Malay were largely of foreign origin. For example, the words penis and vagina inMalay were of Arabic origin (‘zakar’ and ‘faraj’ respectively). Similarly, the words sex and climax Objection to the word ‘penetration’ as there is no were rendered ‘seks’ and ‘klimaks’, both of English equivalent word in Malay in the sexual context.
origin. Certain key concepts in IIEF like erection, We accept that it be interpreted as synonymous penetration and sex life were also difficult to express colloquially in Malay. Literal translations Objection to the phrase ‘kehidupan seks’ (sex life frequently resulted in items that were awkward and in English) as it is not a commonly used phrase in unnatural. On the other hand, sex-related words and Malay. It was replaced with the phrase ‘kemam- concepts were difficult to express naturally in puan seks’ to connote the capacity to engage in colloquial Malay without sounding vulgar. The Malay version of IIEF finally produced by the Objection to the word ‘ketegangan zakar’ as it was Review Panel was at best a compromise. Many not colloquial. It was decided that ‘kemaluan atau English words had required elaboration using multi-ple words, or otherwise conceptual equivalents ‘batang’ keras’ would have to be added to were used in Malay, which are not always compar- elaborate the meaning of ‘erection’ and to make able to the original English version.
it more colloquial, although admittedly these maysound crude to the better educated.
The final Malay version of IIEF questionnaire can beobtained from the author on request.
In all, 14 bilingual subjects participated in thepretest. Their mean s.d. age was 54 (8.5) years. Allsubjects had no difficulty understanding the trans- A total of 171 subjects were enrolled into Study A lated instrument and found it readable. Their and 32 subjects into Study B. However, only 136 of verbatim responses showed that they were able to the Study A subjects had evaluable data (no missing discern the essential meaning being conveyed.
data and completed repeat IIEF administration), and Many suggested alternative words or phrases to only 26 subjects completed follow-up in Study B.
improve the instrument presentation or readability.
Table 1 shows the characteristics of these subjects.
More than 90% of the subjects (range 93–100%) All the study subjects had to pass a Malay Literacy rated 14 of the 15 items in the Malay version of the Test to be eligible for the study. This test was not IIEF as ‘Almost Equivalent’ or ‘Completely Equiva- meant to measure the degree of literacy, but was to lent’ to the original English version. Of the subjects, ensure the subjects could self-administer the Malay 86% rated item seven as equivalent. The correlation IIEF. The mean age was about 54 in all groups. The between their scores on the original English instru- clinical sample had high prevalence of comorbid ment and the translated Malay version was uni- diseases such as hypertension and diabetes. Simi- formly high (r ranging from 0.7371 to 1). Out of 15 larly, high proportion of subjects (range 47–57%) items, eight had r greater than 0.9, and only three had clinically diagnosed ED. All subjects in Study B items had r between 0.7 and 0.8. The correlation had to have ED to be eligible for inclusion.
coefficient for the total score was 0.98.
We reviewed the many suggestions put forward by the subjects. Taking into consideration the fre-quency of similar suggestions among subjects, andin the light of their rating of equivalence and Table 2 shows the internal consistency and test– correlation of their scores, three suggestions were retest repeatability of the Malay IIEF and IIEF-5. It is Table 1 Study designs and baseline characteristics of subjects enrolled in the studies International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in MalaysiaTO Lim et al clear that both measures are uniformly high for all extracted using principal components analysis with domains, for both IIEF overall and IIEF-5 scores.
The expected structure of five distinct domains is not clearly present. The eigenvalue is concentrated on the first factor, while the remaining four factorsextracted have eigenvalue less than 1. Factor 2 of theMalay version of IIEF correspond with the OS Table 3 shows the ability of the IIEF to discriminate domain of the original IIEF, while factor 3 corre- between subjects clinically evaluated to have ED spond with SD domain, and Factor 4 with OF and those who do not. Undoubtedly, the domain domain. Factor 1 contains a mixture of loadings scores of the IIEF and the IIEF-5 were able to from both EF and IS domains. The intercorrelation discriminate between patients with clinical ED and among the five domains confirm that the EF and IS those without. The differences in scores were most domain scores are highly correlated (r ¼ 0.9). The obvious in the EF domain score, and the IIEF-5 other domain scores are also highly correlated with Table 6 shows the mean change in scores for each Table 4 shows the results of the attempt to reproduce domain, total IIEF score and IIEF-5 score, before and the factor structure of the original IIEF. Factor after treatment among responder to sildenafil treat- loadings and eigenvalues of the first five factors ment. Significant changes in mean scores were Table 2 IIEF domain characteristics: reliability Table 3 IIEF domain characteristics: discriminant validity Principal components analysis with varimax rotation of 15 items of the Malay version of IIEF: factor loadings and eigenvaluesa aItems with high loadings within each factor are boldfaced.
International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in Malaysia aThe highest intercorrelation is boldfaced.
Table 6 IIEF domain characteristics: sensitivity negative), or conversely that maximizes the overall correct classification rates (both true positive and true negative). Table 7 shows the sensitivity, specificity and the correct classification rate of IIEF-5 for a range of scores from 16 to 22. The optimal cutoff point is 17, which has a sensitivity of 85%, specificity of 75% and the highest overallcorrect classification rate of 80% among all otherpossible choice of cutoff points.
Sensitivity, specificity, and the correct classification rate Translating and adapting the English IIEF intoMalay Our experience with translating and adapting the English IIEF into Malay was more difficult than had been the experience described with other Asian languages such as Thai and Chinese.10 In retrospect,this perhaps should not have surprised us. Whilethe experience of sex is undoubtedly universal andmost sex-related concepts exist at least implicitly in observed in all domains. The change was most all cultures, it is also undeniable that cultures differ salient in the EF score, and the IIEF-5 score.
in the way the experience of sex is conceived and Only two patients out of the 26 patients who expressed. Malay is perhaps not an especially rich enrolled for the sildenafil treatment trial were language for expressing sex-related matters. Many of nonresponders. The sample size was insufficient the key words and concepts referred to in the IIEF for determining the specificity of IIEF. Hence, results simply do not exist in the same highly developed forms as it does in English. The equivalent words orconcepts in formal Malay were themselves of Arabic or English origin, the colloquial form often simplydid not exist. When some of these words exist, theyrefer to the subject somewhat obliquely rather than Figure 1 shows the ROC curve for IIEF-5. While the in a direct fashion, for example, sexual intercourse graph suggests that the IIEF-5 does diagnostically means literally ‘to join in one body’ in Malay, or it distinguish between ED and non-ED, the area under has moral overtones beyond the word itself, for the curve is only 0.86. We can use the curve in example, sex organ is referred to as ‘shame’ literally determining the ‘optimal’ cutoff point on IIEF-5 in Malay. Some of the colloquial equivalent in score to be used for diagnosing ED. ‘Optimal’ here Malay may be regarded as vulgar by the better refers to the point that minimizes the overall educated. For example, the key concept of erection misclassification rates (both false positive and false in terms of comprehension and naturalness was best International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in MalaysiaTO Lim et al rendered ‘batang keras’ in Malay (literally a stiff rod sions. A similar lack of subtle distinction is found with the response format. Many subjects haddifficulty with the distinction among ‘extremelydifficult’, ‘very difficult’ and ‘difficult’. This ac- counted for the attenuated difference in scorebetween subjects with ED and those without, andsimilarly the attenuated response in IIEF score to The Malay IIEF turns out to be as reliable as the original IIEF. This needs to be emphasized in view The Malay IIEF-5 did diagnostically distinguish of the difficulties encountered in adapting the between ED and non-ED subjects, although again English IIEF into Malay. Further, sex is a taboo with reduced diagnostic performance. The AUC of subject and may thus result in anxiety among 0.86 may compare unfavorably with the AUC of 0.97 respondents. This is compounded by the fact that reported for the English IIEF,8 however, it is good most developing countries, which include Malaysia, are not ‘questionnaire sophisticated’. This was Finally, previous attempts in translating and especially so with the community sample in this adapting the English IIEF into other languages,10 study among whom, many were not familiar with while successful had stopped short at that. The the use of pencil-and-paper data collection. That the translated instruments were pressed into service Malay IIEF could achieve such high reliability under whether for clinical practice or for clinical trial. It is such circumstances was indeed remarkable.
assumed that the translated IIEF ‘inherits’ the The validity of the Malay IIEF is also to some psychometric properties of the original IIEF. This extent supported. It does discriminate between study shows that this is generally an unsafe subjects with and without ED and it is sensitive to assumption. In view of the difficulty in achieving response to treatment; although somewhat attenu- equivalence with the English IIEF as shown by this ated in scores compared with the English IIEF. For study, it may be tempting to develop language- or example, the mean difference in the EF domain culture-specific measures. However, it is still pre- score between subjects with clinical ED and those ferable to have a common instrument for measure- without was 15.1 for the English IIEF6 and only 8.1 ment to enable comparability of results across for the Malay IIEF. Similarly, the mean change in culture and to enable crosscultural studies, which score in the EF domain among responders (sensitiv- can contribute significantly to our understanding of ity) was 12.8 for the English IIEF,6 while it was only cultural notions underlying our universal experi- 6.0 for the Malay IIEF. Interestingly, the mean change in EF domain scores observed in trials ofsildenafil3,11 conducted in Asia was about 11, whichwas similar to those observed in US12 and Eur- opean13 trials. On that basis, a recent review of IIEF4suggests that the IIEF (or at least its EF domain) was More work is needed to fine tune the Malay IIEF to a robust measure in different countries and cultures.
achieve equivalence with the English IIEF. However, In our opinion, the discrepancy in the sensitivity despite this we feel confident in recommending the property of the Malay IIEF (or its EF domain) reflects IIEF-5 for applications in clinical practice and the differences between clinical trial setting and the clinical research such as clinical trials, where community and primary care setting in which our emphasis is on the erectile component of sexual study was conducted. Similar issues have been function rather than overall male sexual function.
raised concerning the validation and utility of IIEF-5 The diagnostic performance of the Malay IIEF-5 is acceptable and it is sufficiently sensitive to treat- However, the Malay IIEF is also clearly not quite measuring the same thing as the original EnglishIIEF. In particular, the factor structure has changed.
The large eigenvalue for the first factor suggests a rather general measure of sexual function, withoutclear division into five distinct domains. There iscurrently no standard guideline for the crosscultural We thank Pfizer Inc. for funding this study in part.
adaptation and validation of questionnaire instru-ment. While one may not agree that factor structureought to be completely reproducible before one may infer crosscultural equivalence, some degree ofsimilarity is to be expected. Our results show thatthe Malay IIEF has rendered the male sexual 1 NIH Consensus Development Panel on Impotence. Impotence.
function as largely a one-dimensional concept, 2 Goldstein I et al. Oral Sildenafil in the treatment of erectile rather than the finer distinction into five dimen- dysfunction. New Engl J Med 1998; 338: 1397 – 1404.
International Journal of Impotence Research
Cross-cultural adaptation and validation of IIEF in Malaysia 3 Tan HM et al. Asian Sildenafil efficacy and safety study 10 Leung KF, Giroudet C. Cross-cultural adaptation of the (ASSESS-1). Urology 2000; 56: 635 – 640.
International Index of Erectile Function (IIEF) (abstract). Int J 4 Rosen RC, Cappelleri JC, Gendrano III, N. The International Index of Erectile Function (IIEF): a state-of-the-science review.
11 Chen J et al. Clinical efficacy of sildenafil in patients on Int J Impot Res 2002; 14: 226 – 244.
chronic dialysis. J Urol 2001; 165: 819 – 821.
5 Rosen RC. Sexual function assessment in the male: physiolo- 12 Goldstein I et al. Oral sildenafil in the treatment of erectile gical and self-report measures. Int J Impot Res 1998; 10 (Suppl dysfunction. New Engl J Med 1998; 338: 1397 – 1404.
13 Meuleman E et al. Dose-escalation study to assess the efficacy 6 Rosen RC et al. The International Index of Erectile Function and safety of sildenafil citrate in men with erectile dysfunc- (IIEF): a multidimensional scale for assessment of erectile dysfunction. Urology 1997; 49: 822 – 830.
14 Vroege JA. Letter to the Editor. The sexual health inventory for 7 Moreland RB, Goldstein I, Traish A. Sildenafil, a novel men (IIEF-5). Int J Impot Res 1999; 11: 177.
inhibitor of phosphodiesetrase type 5 in human corpus 15 Vroege JA. Letter to the Editor. The sexual health inventory for cavernosum smooth muscle cells. Life Sci 1998; 62: PL-309 – men (IIEF-5): reply to Cappelleri and Rosen. Int J Impot Res 8 Rosen RC et al. Development and evaluation of an abridged 5- 16 Rosen RC, Cappelleri JC. Correspondence (reply). The sexual item version of the International Index of Erectile Function health inventory for men (IIEF-5): reply to Vroege. Int J Impot (IIEF-5) as a diagnostic tool for erectile dysfunction. Int J Impotence Res 1999; 11: 319 – 326.
17 Cappelleri JC, Rosen RC. Correspondence (reply). The sexual 9 Walter SD, Eliasziw M, Donner A. Sample size and optimal health inventory for men (IIEF-5). Int J Impot Res 1999; 11: designs for reliability studies. Stat Med 1998; 17: 101 – 110.
International Journal of Impotence Research

Source: http://www.crc.gov.my/documents/Journal/IIEF.pdf