Abstract
The University of Washington Quality of Life (UW-QOL) questionnaire is often used to assess health-related quality of life (HRQOL) of head and neck cancer patients. The aim of this study was to translate the UW-QOL version 4 into the Korean language and to carry out an initial validation study. A recognized methodology for translation of questionnaires was used. The validation study used the final Korean version between March and September 2009. Adult patients were recruited, with a confirmed diagnosis of head and neck cancer, therapy completed and disease-free for at least 1 year. The UW-QOL was successfully translated into Korean. 56 patients completed Korean versions of UW-QOL, the Beck Depression Inventory and the World Health Organization Quality of Life-BREF and various expected correlations were confirmed first between the two UW-QOL subscales (Spearman 0.54 p < 0.001) and then of these subscales with the other concurrent measures. Lower (worse) UW-QOL scores were seen for later stage patients in all 12 domains. The Korean version of UW-QOL is ready for use in the assessment of HRQOL for Korean patients. Validation work needs to be continued to further establish psychometric properties of the questionnaire for use in this population.
In addition to survival following cancer treatment, health-related quality of life (HRQOL) is increasingly recognized as an important outcome parameter . Its importance is highlighted in head and neck cancer, due to its impact on function, emotion and social life . Although several HRQOL questionnaires for head and neck cancer patients exist, the University of Washington Quality of Life (UW-QOL) questionnaire is a popular choice for clinical teams . It is brief, self-administered, multi-factorial, and specific for head and neck cancer patients. The fourth version has 12 domains (pain, appearance, activity, recreation, swallowing, chewing, speech, shoulder, taste, saliva, mood, and anxiety), consisting of two composite scores, one for ‘physical function’ and another for ‘social–emotional function’. Thus physical and social aspects of HRQOL can be evaluated separately. It also contains two global questions, which reflect health-related and overall QOL.
The questionnaire has normative data for comparison in a UK population and can be used as a screening tool in routine clinical practice . It was constructed for English speaking populations and has been translated into several other languages . There is tremendous merit in being able to compare HRQOL outcomes across different cultures; however a careful and psychometrically robust translation into other languages is required . The objective of this study was to translate the UW-QOL questionnaire into the Korean language, to recruit patients into an initial validation study that includes comparison concurrently with two other well established questionnaires, the Korean translation of the Beck Depression Inventory (BDI) and the World Health Organization Quality of Life (WHO QOL)-BREF, and to compare the results with UK patient and normative data.
Materials and methods
Translation process
According to internationally accepted guidelines , the fourth version of the UW-QOL questionnaire was translated from English and was culturally adapted into the Korean language.
First, a forward translation was performed involving three bicultural experts who were native speakers of Korean and who lived in Korea. They were fluent in English and had experience in translating medical questionnaires. Two of these experts independently translated the English UW-QOL questionnaire into Korean, thus creating two separate forward translations. The third expert compared the two translations and any discrepancies were resolved by discussion between the three experts.
The agreed Korean version was then independently back-translated into English by another two bicultural experts, both native speakers of English and fluent in Korean. They had not seen the original English version. They reviewed each other’s translation and resolved any discrepancies. The agreed back-translation was compared to the original English version and any discrepancies resolved by revising the Korean version.
The revised Korean version was back-translated into English again, compared to the original English version with further revision as necessary to the Korean version and this process was repeated until no further discrepancies were found. The final Korean version was achieved.
During the translation process, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) principles were strictly followed . It described the key persons involved in the translation process: ‘Forward translators should be native speakers of the target language and fluent in the source language. It is preferable that forward translators reside in the target country and have experience in the translation of questionnaire. Back translators should be native speakers of the language of the source measure, and fluent in the target language.’ Therefore, forward translation was done by native speakers of Korean. They had lived in the USA for more than 6 years and studied at US universities, so they were fluent in English. They were medical doctors of the Seoul National University Hospital and had experience in the translation of medical questionnaires. Back translation was done by native speakers of English, who had lived in Korea for more than 2 years and were fluent in Korean.
Initial validation study using the Korean version
A prospective validation was performed with the final Korean version of UW-QOL questionnaire between March 2009 and September 2009. The study was approved by the institutional review boards at the Seoul National University Hospital. Patients were enrolled into the study from the Seoul National University Hospital Otorhinolaryngology clinic. Inclusion criteria were: at least 18 years of age; confirmed diagnosis of head and neck cancer; therapy complete with at least 1 year disease-free survival.
After giving informed consent patients were asked to complete various questionnaires at the end of their regular scheduled clinic visit. Questionnaires included the Korean version of the UW-QOL, the Korean validated form of BDI , and the Korean version of WHO QOL-BREF .
Scoring for UW-QOL domains is scaled to so that a score of 0 represents the worst possible response and 100 the best possible response. Physical and social–emotional function scores were each obtained from the mean of six domains. The domains relevant to physical function are appearance, swallowing, chewing, speech, taste and saliva. The domains relevant to social–emotional function are pain, activity, recreation, shoulder, mood and anxiety. Three separate global questions about QOL were also scaled from 0 to 100.
BDI is one of the most widely used self-report instruments for measuring the severity of depression . The questionnaire contains 21 questions about emotional, cognitive, motivational, physiological and other symptoms. Each item is scored 0–3, reflecting how the patient has felt over the past week. The BDI total score is obtained from summing the item scores with higher scores reflecting more severe depression. In Korea, a score of 16 is suggested as the optimal cut off-point for the screening of depression . The patients were divided into two categories according to their BDI score: patients without depression (BDI score of less than or equal to 16) and those who may have or who definitely have depression (BDI score of more than 16). The UW-QOL function scores were compared between patients with and without depression. Correlations between BDI and UW-QOL function scores were also analyzed.
WHO QOL-BREF is a multidimensional measure which comprises important aspects of life . It contains 24 items within four domains: physical health; psychological; social relationships; and environmental. A WHOQOL-BREF domain score is calculated by multiplying the mean score of all the items within that domain by four, and each domain is scaled from 0 to 100, with higher scores reflecting better QOL. Correlations between WHO QOL-BREF scores and UW-QOL function scores were analyzed.
The medical records of the enrolled patients were reviewed to obtain information on patient and clinical characteristics including tumour site, stage, treatment and medical co-morbidity data.
Statistical analyses
The authors examined internal consistency (Cronbach’s α), and looked for a value of 0.70 or higher . Evidence towards construct validity is obtained by a questionnaire behaving according to hypothesized relationships. The authors made five initial hypotheses: the physical and social–emotional function scores of the UW-QOL questionnaire should correlate reasonably well with the global question scores of the UW-QOL questionnaire; patients without depression according to the BDI score should have better UW-QOL social–emotional function scores; UW-QOL physical function scores should correlate reasonably well with the physical health scores of the WHOQOL-BREF; UW-QOL social–emotional function scores should correlate reasonably well with BDI score and with the psychological and social relationship scores of the WHOQOL-BREF; UW-QOL physical and social–emotional function scores should show little or no correlation with the environment scores of the WHOQOL-BREF.
The Spearman correlation coefficient was used to assess the level of correlation between measures. The Mann–Whitney test was used to compare UW-QOL function scores between patients with and without depression (BDI), between early and late staging, and between treatment groups with regard to UW-QOL domains, subscales and global scores. The statistical analysis was performed using version 12.0 of the SPSS statistical program (SPSS, Chicago, IL, USA) for Windows.
Results
Translation
After the first round of translation the Korean version of the UW-QOL was not notably different from the original English version. One discrepancy was revealed when translating the word ‘prescription medicine (e.g. morphine)’. Patients can obtain non-steroidal anti-inflammatory drugs without prescription in the USA, but not in Korea. Thus ‘prescription medicine’ meant general painkiller, not opioid painkiller in Korea. The authors translated ‘prescription medicine (e.g. morphine)’ into ‘opioid painkiller (e.g. morphine)’. After other minor corrections, a consensus version was reached. The translation process was repeated until no discrepancies between the original and back-translated versions were found, and only then was the final Korean version of the UW-QOL questionnaire achieved.
Initial validation study
56 patients (56 of 60 patients approached; 93%) were enrolled into the initial validation study. The patients’ characteristics are listed in Table 1 . 18 (32%) patients had surgery alone, 17 (30%) had surgery and radiotherapy, 7 (13%) had primary concomitant chemoradiation, 6 (11%) had radiotherapy alone, 5 (9%) had a combination of surgery, radiation, and chemotherapy, and 3 (5%) had a combination of radiation and chemotherapy.
| Characteristics | Total, n = 56 |
|---|---|
| Age, year | |
| Mean (range) | 61 (35–81) |
| Sex (%) | |
| Male | 43 (77) |
| Female | 13 (23) |
| Tumour subsite (%) | |
| Nasopharynx | 3 (5) |
| Hypopharynx | 4 (7) |
| Oropharynx | 10 (18) |
| Larynx | 15 (27) |
| Oral cavity | 11 (20) |
| Salivary gland | 3 (5) |
| Thyroid | 5 (9) |
| Paranasal sinus | 1 (2) |
| MUO | 3 (5) |
| Double primary | 1 (2) |
| T stage (%) | |
| T1 | 26 (46) |
| T2 | 15 (27) |
| T3 | 8 (14) |
| T4 | 4 (7) |
| Tx | 3 (5) |
| N stage (%) | |
| N0 | 28 (50) |
| N1 | 9 (16) |
| N2 | 15 (27) |
| N3 | 1 (2) |
| Nx | 3 (5) |
| TNM stage (%) | |
| I | 17 (32) |
| II | 5 (9) |
| III | 15 (28) |
| IV | 16 (30) |
The physical and social–emotional function scores of the Korean version of UW-QOL questionnaire ranged from 55.0 to 100 and from 54.2 to 100, with mean (SD) scores of 86.0 (±12.9) and 85.2 (±10.2), respectively. There was moderate correlation between the physical and social–emotional function scores of the UW-QOL questionnaire. The Spearman correlation coefficient between the physical and social–emotional function scores of UW-QOL questionnaire was 0.54 ( p < 0.001).
The internal consistency of the subscales was measured by Cronbach α values of 0.63 for physical function, 0.57 for social–emotional function, and 0.73 overall. In regards to construct validity, the authors first tested the association between the physical and social–emotional function scores of the UW-QOL and of these with the global single question scores of the UW-QOL, expecting trends that patients with higher scores on the global questions would have higher scores on the physical and social–emotional function scores. This hypothesis was confirmed statistically though the correlations were relatively weak ( Table 2 ). Second, two groups, categorized by BDI score were compared based on the hypothesis that patients without depression had better UW-QOL social–emotional function scores. This hypothesis was confirmed statistically by the Mann–Whitney test ( p = 0.009, Fig. 1 ). Third, the authors expected reasonably strong correlation between the physical function scores of the UW-QOL questionnaire and the physical health domain scores of WHO QOL-BREF, and the Spearman correlation though statistically significant was weak ( Table 3 ). Fourth, the authors expected that the UW-QOL social–emotional function scores would correlate reasonably well with BDI score and the psychological domain and social relationship domain of the WHOQOL-BREF. The Spearman correlations were all statistically significant but while the correlation with BDI was moderately strong the correlations with the WHO domains were weak ( Table 3 ). No strong correlation between UW-QOL physical and social–emotional function scores and the environment domain scores of WHOQOL-BREF was expected and this was confirmed by weak statistically non-significant correlations ( Table 3 ).
