|
Indian Journal of Cancer, Vol. 44, No. 4, October-December, 2007, pp. 147-154 Original Article Validation of the University of Washington quality of life questionnaires for head and neck cancer patients in India D'cruz AK, Yueh B, Das AK, Mcdowell JA, Chaukar DA, Ernest AW Head and Neck Service, Department of Surgery, Tata Memorial Hospital, Dr. E. Borges Road, Parel, Mumbai - 400 012
Code Number: cn07027 Abstract Quality of life (QOL) is a multidimensional construct capturing the subjective wellbeing of patients in physical, emotional, functional and social domains. Available work on post treatment QOL have only been made in western literature and less in Indian literature.Aims : To translate the UWQOL into both Hindi and Marathi and psychometrically validate the translation in HandN cancer patients in Indian population. Settings and Design : A prospective study was done at the Tata Memorial Hospital for patients who were treated for H and N cancers. Materials and Methods : 147 patients were enrolled from January to April 2005. The study was carried out in two phases. Patients were given translated versions of the UWQOL and EORTC QOL questionnaires preoperatively, 15 days postoperatively and then three months postoperatively. Results : Both the Hindi and Marathi translations had strong internal consistency (Cronbach's alpha=0.7971 and 0.7839). UWQOL composite scores correlated well with the global questions on overall QOL in both the Hindi (r=0.69) and Marathi (r=0.66) translations and also with Tstage. QOL scores were worse three months postoperatively than preoperatively and for patients undergoing surgery that violated the mucosa. A strong correlations was observed (r>0.50) between all similar domains on the UWQOL and EORTC HandN35 except the saliva item on the Marathi translation, where r<0.50, but Pvalues were significant. Conclusions : The Marathi and Hindi versions of the UWQOL appear to be valid and reliable instruments for assessing the QOL in Indian population and will be a vital tool for achieving greater insight into the short and the long term QOL. Keywords: Head and neck cancer, quality of life, questionnaires, validity Introduction Increasing attention has been given in recent years to the importance of quality of life (QOL) in assessing outcomes of patients treated for head and neck (H and N) cancers. QOL is a multidimensional construct capturing the subjective well-being of patients in physical, emotional, functional and social domains. Although therapeutic options may provide similar survival rates, QOL outcomes between treatment modalities may vary significantly. It is therefore ideal for pre--therapy discussions to include consideration of optimal post--therapy QOL outcomes. Although substantial efforts have been made in the western literature to document post-therapy QOL, less has been documented in the Indian literature, despite the prevalence of H and N cancer in the Indian population. [1],[2],[3],[4],[5] A variety of scales have been developed, with the most widely used and accepted scales being the European Organization for Research and Treatment of Cancer's Quality of Life Questionnaire (EORTC-H and N-35), the Functional Assessment of Cancer Therapy (FACT H and N) scale and the University of Washington Quality of Life questionnaire (UW-QOL). An instrument suitable for routine clinical use should be brief and impose minimal patient burden. The EORTC QOL questionnaire is the most widely used instrument and has been validated in various languages. We were the first to validate the same in Hindi and Marathi [6] the two vernacular languages most widely spoken by our patients. Our findings did suggest that this instrument was suitable in our patient population; however these scales are lengthy, complex and time consuming consisting of 35 (H and N specific) and 30 (Global) questions, respectively. The UW-QOL is brief, consisting of 12 questions and is commonly used by many centers globally. The aim of this project was to translate the UW-QOL into both Hindi and Marathi and psychometrically validate the translation in head and neck cancer patients in India and compare the same with the previously validated EORTC questionnaire. Materials and Methods This study was carried out in two phases. We performed a linguistic and cultural translation of the UW-QOL from English to the Indian languages in Phase 1 and then prospectively validated the translated scales at Tata Memorial Hospital (Parel, India) in Phase 2. Translation process The translation process began with a forward translation of the original US English questionnaire into both Hindi and Marathi [Figure - 1a, 1b],[Figure - 2a, 2b]. Two native Hindi and Marathi speakers with bi-cultural expertise (US and Indian) were used for each forward translation. After translation, a bi-lingual coordinator performed a review of each of the translations and, along with the translators, reached an iterative consensus. The consensus version of both the Hindi and Marathi translations were sent to new translators who performed a back translation from each of the Indian languages to English. The coordinator finally amalgamated the two English back translations into a final questionnaire that was then compared to the original English version to ensure concordance with not only the translation but also nuances in meaning. Psychometric validation of the translated questionnaires Study Population. We enrolled patients from the Tata Memorial Hospital who had received treatment for cancers of the head and neck. The Tata Memorial Hospital is a tertiary cancer site with the largest volume of H and N cancer in India. All participants were adult patients with thyroid and H and N squamous cell carcinoma who were presenting for pre-operative evaluation. Subjects were required to speak and read Hindi or Marathi. Patients were recruited between January and April 2005. Data Collection. Enrollees (n=147) were given a packet of self--administered questionnaires during a routine pre-operative consultation. The packet included translated versions of the UW-QOL and EORTC H and N35. Patients were asked to complete this packet pre-operatively, 15 days post-operatively and then three months post-operatively. An additional cohort of patients (n=35) were asked to return a second UW-QOL within 7-10 days, which was returned by mail, in order to evaluate the test--retest reliability of the scale. This interval was chosen because enough time had elapsed to prevent patients from remembering their responses to the first administration of the scale, but not enough time to allow clinically meaningful change to have occurred in the interim. Patients did not receive remuneration. The medical records of these patients were abstracted to obtain data on demographics, cancer site, stage and treatment. Scoring of UW-QOL and EORTC H and N-35 Each item (question) on the UW-QOL (version 4) is scored from 0 to 100. A higher score is indicative of better quality of life. A 'composite' score is obtained by calculating the mean of the 12 items. There are also 3 'global' questions which are analyzed separately. Individual items on the EORTC H and N35 module were also scored from 0 to 100. A higher score on functional scale indicates higher QOL while a higher score on symptom scale indicates worst QOL. Data analysis Reliability was established by measuring both internal consistency (Cronbach's alpha) and test--retest reliability (intra-class correlation coefficient). Internal consistency is considered good if alpha approximates 0.70 but does not exceed 0.90, because values over 0.90 imply the presence of redundant items. [7] Test--retest reliability was measured with the intra-class correlation coefficient (ICC), which is more rigorous than Pearson's correlation coefficient r because it considers not just the strength of the correlation but also systematic variations. [8] The 'validity' of a scale depends on whether the scale truly measures what it purports to measure. Three forms of validity-content, criterion and construct-are typically sought. Content validity has been ensured by the translation process, as the authors have made sure that the translated scale touches on issues relevant to an Indian H and N population. Criterion validity asks whether the scale compares favorably to a gold standard, but this is typically less relevant in quality of life research since there is no "gold standard" measure of QOL. Construct validity therefore receives the bulk of attention. One popular method for establishing construct validity is to compare the scale's scores against other variables ('concurrent validity'). For example, one might theorize that scores should be worse in patients with advanced stage disease. In this study, we evaluated the relationship of UW-QOL composite scores with: T-stage, global UW-QOL scores and relevant EORTC H and NQ35 subscales. Finally, we evaluated the transition scores (difference between pre-operative and post-operative) scores, theorizing that patients with operations involving the mucosa (as opposed to neck and thyroid procedures) experience the greatest diminishments in QOL. We also calculated sought to evaluate the magnitude of transition scores, to evaluate the sensitivity of the translated scales to clinically important change. [9] All statistical analyses were performed using with Stata for Windows (Version 8.2, College Station, Texas, USA). Results The UW-QOL was successfully translated into both Hindi and Marathi [Figure - 1a, 1b],[Figure - 2a, 2b]. Of the 147 patients who participated in the validation phase, 86 completed the Hindi version and 61 completed the Marathi version. The mean age of participants was 50.5, with the majority of patients being male (74.5%). Other baseline characteristics of our cohort are summarized in [Table - 1]. We note that a greater percentage of patients completing the Hindi questionnaires were male and had tumors of the oral tongue and pyriform sinus. Both the Hindi and Marathi translations had strong internal consistency (Cronbach's alpha=0.7971 and 0.7839, respectively, using data collected before surgery). Test--retest reliability was also excellent (ICC=0.9069 for Marathi; 0.8964 for Hindi); the raw data are shown in [Figure - 3]. There was also strong evidence of construct validity. UW-QOL composite scores correlated very well with the global question on overall QOL in both the Hindi (r=0.69) and Marathi (r=0.66) translations. There was strong relationship between the composite and global scores, with better composite scores associated with better global scores, as expected [Figure - 4],[Figure - 5]. UW-QOL composite scores also correlated with T--stage, with patients with early stage disease having better composite scores both pre-operatively and post-operatively [Figure - 6],[Figure - 7],[Figure - 8],[Figure - 9]. As expected, QOL scores are worse three months after surgery than pre-operatively, since the recovery time for surgical patients usually requires one year. [10] Because no composite score from the EORTC H and N35 is calculated, only comparisons between relevant domains between the UW-QOL and the H and N35 are relevant [Table - 2]. In general, strong correlations were observed (r>0.50) between all similar domains, including pain, swallowing, speech and saliva items on the UW-QOL. The only exception was with the saliva item on the Marathi translation, where r< 0.50, but P -values were still statistically significant. Both translated versions were also sensitive to clinically important change, such as an operation in the head and neck region. The effect size of the change score was 0.68 in the Marathi population and 0.70 with the Hindi population. For reference, an effect size of 0.5 is considered representative of moderate change and an effect size of 0.8 is considered large. [11] By excluding thyroid and salivary tumors (since the UW-QOL is aimed at patients with squamous carcinoma of the H and N), the effect sizes are even larger (0.81 and 0.73, for Marathi and Hindi patients, respectively). We also found that greater decrements were observed for patients undergoing surgery that violated the mucosa, as expected [Table - 3]. Discussion Disease specific health status scales not only an important tool that help to monitor an individual patient's health status after treatment, but also are useful to elucidate important differences in QOL for within and between diverse populations. In this study, we have successfully translated the UW-QOL into Hindi and Marathi and have separately validated these scales in Indian head and neck patients. The resulting scales maintain the brevity of the original UW-QOL scale. The design of the scale enabled easy administration and resulted in little burden on our patients. We are also encouraged by the scale's performance, since it not only appears to appropriately measure QOL in our patient population, but also seems to be sensitive to clinically meaningful change. Limitations We have tested the translated scales on patients seen exclusively at the Tata Memorial Hospital in Parel, India. Although we do not anticipate that responses from patients in the rest of India would vary systematically from ours, we do note that generalizing the rest of the Indian population requires further study. It is also important to note the small sample size used in the test--retest phase of the validation. Finally, it is also important to recognize that more data are required in order to achieve greater understanding of benchmark values for H and N patients. Conclusions Nonetheless, the Marathi and Hindi versions of the UW-QOL appear to be valid and reliable instruments. Because of its brevity and ease of administration, we believe it can be rapidly implemented into general routine clinical practice to monitor a patient's H and N cancer-specific QOL and will be a vital tool for achieving greater insight into the short- and long -term QOL of our individual and collective patients. Furthermore, emerging translations of this scale into Brazilian, Portuguese, [12] Swahili, Spanish and several Asian languages will facilitate greater understanding of how QOL differs in developing countries around the world.Acknowledgement We thank Dr. Siddharth A. Shah for his kind help in compiling and submission of the manuscript.References
Copyright 2007 - Indian Journal of Cancer The following images related to this document are available:Photo images[cn07027t2.jpg] [cn07027f7.jpg] [cn07027t3.jpg] [cn07027f9.jpg] [cn07027f1a.jpg] [cn07027f8.jpg] [cn07027f6.jpg] [cn07027f5.jpg] [cn07027t1.jpg] [cn07027f4.jpg] [cn07027f2a.jpg] [cn07027f3.jpg] [cn07027f1b.jpg] [cn07027f2b.jpg] |
|