City of Hope-Quality of Life Questionnaire-Arabic Version: Assessment of Reliability and Validity

BACKGROUND: Arabic is spoken as a native language by more than 400 million people worldwide. However, there is no specific Arabic language instrument to measure stoma-related quality of life. PURPOSE: This study was designed to assess the validity and reliability of the City of Hope-quality of life-Ostomy Questionnaire (COH-QOL-OQ) Arabic version. METHODS: A cross‑sectional design was used. Intra-class correlation coefficients were calculated to measure reliability, and Pearson’s correlations of an item with its own scale and other scales were scored to evaluate convergent and discriminant validity. Content validity was reviewed by a panel of 5 experts. RESULTS: There were 421 participants with colostomy, ileostomy, or urostomy (239 [56.8%] male and 182 [43.2%] female). All COH-QOL-OQ subscales for the Arabic version demonstrated a high level of internal consistency (Cronbach's α = 0.71–0.87). The initial administration of the confirmatory factor analysis model showed inadequate goodness-of-fit indices (χ² /df = 3.902, NFI = .845, CFI = .880, RMSEA = 0.083). However, after removing item 2 in the social dimension, the final administration of the confirmatory factor analysis model showed significant goodness-of-fit indices (χ² /df = 2.663, NFI = .900, CFI = .935, RMSEA = 0.063). CONCLUSION: The findings suggest that the COH-QOL-OQ Arabic version is a valid and reliable tool to measure quality of life among patients with an ostomy in Saudi Arabia.

Ostomy comes from the Greek word stoma, which means mouth. In medicine, stoma/ostomy is defined as a surgical opening created between the abdominal wall and sections of the colon, ilium, rectum, or bladder to drain body waste due to incontinence of stool or urine, following surgery, or due to chronic inflammation.^1,2 Cancer is the most common indication for an ostomy procedure, although the procedure is also used for other conditions such as inflammatory bowel disease, fecal and urinary incontinence, and diverticular disease.³

The third most common cause of cancer death worldwide is colorectal cancer, and its prevalence is increasing in developing nations, accounting for 11% of all cancer diagnoses.⁴ In 2018, colorectal cancer was considered the second most deadly cancer globally, with approximately 881 000 deaths.⁴ It was the deadliest cancer among males in Saudi Arabia, Oman, and the United Arab Emirates, and it was the deadliest cancer among females in Algeria, Belarus, Japan, Spain, and Portugal.⁴ To the best of the authors’ knowledge, there are no published national statistics for stoma surgery in Saudi Arabia even though colorectal cancer is the most common diagnosis among men (10.6%) and the third most common among women (8.9%).⁵ Colorectal cancer is the second most common cancer in Saudi Arabia, with an incidence of 10.6% in males and 8.9% in females.⁶ The incidence of colorectal cancer is highest in Riyadh, the capital city of Saudi Arabia, where it reached 14.5/100 00 in 2010.⁷ In addition, the prevalence of Crohn disease (regional enteritis) has increased sharply in Saudi Arabia,⁸ and ostomy surgery is used in the management of complications associated with this disease.⁹

Quality of life (QoL) assessment is one of the essential outcome indicators of major surgical procedures and treatment. Health-related quality of life (HRQoL) is a multidimensional concept widely employed to measure the influence of health status on QoL. QoL tools include physical symptoms and well-being, psychological well-being, social well-being, and spiritual well-being.¹⁰ The general QoL instruments (subjective measurement methods of personal satisfaction with general health and well-being) are not sufficiently sensitive to detect the effect of ostomy on QoL among people with a stoma. A multidimensional QoL tool that focuses on the impact of ostomy and is based on areas of concern for this patient population would be beneficial.

Negative impacts of ostomy have been documented, especially in patients who underwent non–ostomy-related surgery.^11,12 Ostomy often results in ongoing clinical requirements as well as extensive changes in QoL and daily routines that impact physical, psychosocial, social, and spiritual aspects of daily life.¹³ Furthermore, QoL impairment is associated with a number of challenges including psychological issues, dietary restrictions, physical problems, and sexual functioning concerns.^14,15 Therefore, assessing QoL is essential for patients with an ostomy and can assist in disease control as well as reduction or prevention of complications, thus improving QoL.^15,16

The importance of QoL evaluation in patients with an ostomy led to the development and application of several instruments. Different instruments have been constructed to measure HRQoL in this patient population and include the patient’s perspective on stoma care. These instruments include the QOL index, City of Hope‑quality of life‑Ostomy Questionnaire (COH‑QOL‑OQ), Ostomy Adjustment Scale, Ostomy Adjustment Inventory‑23, Coping Strategies Inventory, Survey of Pre‑operative Factors of Ostomy Adjustment and Stoma‑QOL.¹⁰

Nevertheless, most studies have used instruments that are not validated for patients with an ostomy.¹⁶ The COH-QOL-OQ is a disease-specific tool that was first developed in 1983 by investigators at the City of Hope National Medical Center in Duarte, California, and was revised and validated in 2004.¹⁷ COH-QOL-OQ is a comprehensive, multidimensional, and self-reported tool designed to evaluate QoL for patients with an intestinal ostomy (this term includes colostomy and ileostomy. The instrument was developed to assess patients with an ostomy that may be due to carcinoma, including gastrointestinal and bladder carcinoma, and gastrointestinal disease or injury due to trauma.¹⁰ This questionnaire has been translated into several languages including Chinese, Persian, Turkish, Portuguese, Malayalam and Croatian.^{3,10,16,18-20}

In the Kingdom of Saudi Arabia, there is no specific validated and suitable questionnaire for measuring QoL among patients with an ostomy. A standardized QoL tool for this patient population would be beneficial in understanding the physical, psychological, social, and spiritual consequences that may arise after ostomy creation and assist in helping to improve care.^10,15 The purpose of this study was to translate into Arabic the COH-QOL-OQ and evaluate the validity and reliability of this Arabic version when administered to patients with an ostomy in Saudi Arabia.

Validation of the COH-QOL-OQ was conducted as part of a larger research study on factors affecting QoL among people with an ostomy in Saudi Arabia. Permission to conduct the study was obtained from the human research ethics committees at the Royal Melbourne Institute of Technology and the 5 Saudi hospitals included in the study. The authors of the COH-QOL-OQ¹⁰ advised that permission for using or adapting the questionnaire was not required.

Translation. The English COH-QOL-OQ was translated into Arabic by following the World Health Organization forward-back guidelines for translations.²¹ The translation was performed by a urology surgeon, a colorectal surgeon, 2 wound and ostomy care nurses, a nurse, and a psychologist with a PhD qualification, all of whom were native Arabic speakers and proficient in English. Translation was done independently by each. The Arabic version was revised based on recommendations by the participants, with repetition of this process until consensus was met for adequacy of the content. One independent translator translated back the questionnaire into English from the Arabic version. There were minimal differences as a result of the translation process, and these were resolved by agreement in the panel group. For the final step, the English back-translated version was compared to the original COH-QOL-OQ English version to confirm that the full meaning and context of the questionnaire was not lost.

Instrument. The original COH-QOL-OQ is composed of 2 components. The first component consisted of 47 multiple-choice and open-ended questions. These questions related to the patient’s demographic data, work, health insurance, sex, psychological support, clothing, diet, and ostomy care. The second component consisted of 43 items measured on a 10-point Likert scale that were categorized into 4 subscales (items 1–11: physical well-being, items 12–24: psychological well-being, items 25–36: social well-being, and items 37–43: spiritual well-being. It contained some items that needed reverse coding of the responses.^10,19 Scale modification was applied to adapt language, culture, and religion for Saudi patients. Because inherent differences exist between English and Arabic linguistic structure and culture, item wording modification was undertaken. The 47 items of the first component were reduced to 22 in the Arabic version. Four items that covered religious practices was added to the spiritual subscale at the second component. Thus, in the Arabic version, the second component contains 47 items.

Following ethics approval, data were collected through a cross-sectional design from 5 hospitals in Riyadh. Each participant received an information sheet, which explained the purpose of the research, and a consent form, which guaranteed patient anonymity and confidentiality. The participants were informed that they were able to withdraw from the study at any time before submitting the anonymously completed questionnaire. Convenience sampling was used for the validation of the COH-QOL-OQ Arabic version. This included 421 patients with an ostomy who visited an outpatient surgical clinic for routine follow-up care at 1 of 5 hospitals in Riyadh between May 2019 and November 2019. The questionnaire was in hard copy format only, which is a common and convenient method of data collection in Saudi Arabia. Inclusion criteria were being older than 18, having undergone surgical formation of an ostomy, and being a Muslim who speaks Arabic as this tool was translated and modified to appropriate religious practice.

Statistical analysis. Statistical analyses were conducted using both the IBM SPSS software version 25 and AMOS software version 22. The frequency distributions (counts and percentages) were tabulated for all questions including demographic data characteristics. The level of significance for all statistical tests was P < .05. Cronbach’s alpha was used to assess the internal consistency of the Arabic version of the COH-QOL-OQ.

The validity of the COH-QOL-OQ translated instrument was examined based on several indicators. Face validity indicates that all relevant aspects of the measured phenomena are covered in the instrument based on the opinion of medical experts.^9,15 Face validity was examined by 6 medical experts who agreed by consensus that items covering all aspects of QoL and were suitable, relevant, and understandable for Arabic speakers who are of the Muslim faith.

Convergent and discriminant validity were measured by analyzing item–subscales correlation, following the principle that each subscale score must be correlated to a higher degree with its own subscale score and to a lower degree with scores of other subscales. Because the first author and translation group validated the COH-QOL-OQ Arabic version, only confirmatory factor analysis was used to confirm the 4-factor model of subscales. The model was an HRQoL multidimensional concept aimed at assessing the impact of ostomy on QoL. This model is measured by 5 core dimensions that ask about physical, psychological, social, and spiritual status as well as Islamic religious practice.

Demographic characteristics. A total of 500 surveys were distributed, and 421 questionnaires were completed (response rate of 84%). Most participants were male (56.8%) with ages ranging from 18 to 75 years. The majority of participants had a colostomy (50.6%), and most of these had undergone ostomy surgery due to colon cancer (35.2%). The majority of participants were married before ostomy surgery (70.5%). Divorced status increased after ostomy surgery (2.4% [n = 10] before surgery compared with 5.7% [n = 24] after surgery). Most participants were between 30 and 39 years of age (25%) (Table 1).

Reliability. All subscales of the Arabic version of the COH-QOL-OQ indicated excellent internal consistency (0.95–0.96), and the total scale of the COH-QOL-OQ had a high Cronbach’s alpha (.98) (Table 2).

Content validity. To verify the scale validity, the item–total correlation was calculated for individual items; the total score for each item was compared with its own subscale and with the other subscales (Table 3). The convergent validity coefficients, the correlation between measurements of the same trait obtained using different methods of measurement, and conversely discriminant validity coefficients consist of 2 types. The first type includes correlations between measures of different traits obtained using the same measurement method, while the second type includes correlations between measures of different traits obtained using different measurement methods.²²

Previous research has shown that the accepted correlation of items with their own subscales should be > 0.40.³ All 4 subscales of the COH-QOL-OQ Arabic version show evidence that the items met the criterion of convergent validity. On the physical subscale of the COH-QOL-OQ Arabic version, item 9 had the lowest correlation with the total score on the scale. On the psychological subscale, item 19 had the lowest convergent validity. For the social and spiritual subscales, items 26 and 42 had the lowest validity with their own subscales.

In relation to discriminant validity, an item’s score must have a lower correlation with the total scores of other dimension subscales than with its own total score to prove noticeably smaller in magnitude than convergent validity coefficients. For the Arabic sample, all subscales appear to meet this criterion. In addition, coefficients of correlation scores between the subscales are estimated to be lower than the coefficients of the internal consistency for each subscale (Pearson's coefficients ranging from 0.54 to 0.90; all P < .01). The highest item correlation retrieved was between social and spiritual subscales (0.78–0.87). Table 4, Part 1, Part 2, and Part 3, shows each individual item correlation with total COH-QOL-OQ and its own subscale.

Confirmatory factor analysis. The confirmatory factor analysis (CFA) is a statistical method that can be used to identify the fit to a predefined factor structure of an observed data set. It checks precise a priori assumptions between the latent variables identified and the data observed.²³ Factor analysis techniques such as CFA use an intercorrelation matrix among the items to find a high covariation pattern. This technique helps researchers to reduce a set of related items to a factor that can mainly explain the changes among them.^24,25 Whereas many statistical tests (such as t-test, ANOVA, or F test) employ a single indicator to make a conclusion, CFA uses a group of fit indices to make an inference because the chi-square test of CFA is sensitive to sample size. When the sample size of a CFA is large, the chi-square test is not reliable.²⁶ Therefore, Jöreskog and Sörbom²⁷ suggested a series of goodness-of-fit (GOF) indices that indicate to what degree a proposed model will fit with empirical data. The current study used the following indices to judge GOF: chi-square/degrees of freedom (chi-square/df) < 5, GOF ≥ .90, root mean square error of approximation (RMSEA) ≤ .08, comparative fit index (CFI) ≥ .90,^25,26 and factor loading > 0.70.²⁸

Model fitness. A 4-factor model was developed in AMOS, and estimates were calculated in terms of covariance, residuals, standardized regression, and modification indices. The first model was nearly fit (see values in Table 5). The initial administration of the CFA model showed inadequate GOF indices. The chi-square/df was 3.902, CFI = .880, normed fit index (NFI) = .845, and RMSEA = 0.083. Some of these indices, particularly CFI, do not fully meet the acceptable level but are approximately close to the required threshold (Table 5). Consequently, a strategy was needed to achieve an adequate GOF for the theoretical model. Hair et al²⁸ suggested removing items with weight < .70 loading, which represents an improvement over the original model. This modification indicates that item weight < .70 loading should be removed, and then the GOF indices values would be predicted to be acceptable. To make the model fit, item 2 from social well-being was removed due to the lower weight (0.46) on its factor. Furthermore, covariance was created among residuals with modification indices > 4.0. After doing maximum possible changings into the model, it was fitted into the threshold as shown in Table 6. The final 4-factor model of COH-QOL-OQ is shown in Figure 1.

The final administration of the CFA model showed significant GOF indices. The chi-square/df was 2.663, CFI ≥ .935, and RMSEA = .063. In conclusion, the indices provide sufficient support to claim that COH-QOL-OQ is a 4-structural construct including physical, psychological, social, and spiritual components that has 46 total items. This also indicates that it is a valid questionnaire to measure QoL among ostomy patients in Saudi Arabia.

There are no validated and reliable tools to assess QoL for Arabic-speaking patients undergoing ostomy. This is increasingly significant given that existing QoL assessment tools in general are sensitively insufficient to detect a specific effect of ostomy surgery on QoL.¹⁷ There have been other studies^29,30 that have used COH-QOL-OQ as an instrument for measuring HRQoL among ostomy patients and have confirmed it to be a reliable survey assessment instrument. However, there have been no studies assessing QoL among patients with an ostomy in Saudi Arabia or other Arabic-speaking populations. The current study sought to meet this need by the translation and validation of the COH-QOL-OQ instrument.

The reliability and validity of a measurement tool is essential. According to LoBiondo-Wood and Haber,³¹ reliability is an indicator of a measurement tool’s ability to consistently deliver the same yresults, whereas a measurement tool’s validity indicates the extent to which the measurement tool measures the design it was proposed to measure. Cronbach’s alpha coefficient is the most commonly reliability measure used in research studies and therefore was used in this study. Based on the Cronbach’s alpha for each of the COH-QOL-OQ subscales for the Arabic version, this instrument demonstrated validity and reliability. The overall score of .987 indicated that internal consistency was extremely good for the scale. Each dimension’s alpha was also calculated. Field²³ also recommended that the alpha for each subscale be calculated separately. The internal consistency of the subscales for each dimension of the COH-QOL-OQ Arabic version showed that the internal consistency of each subscale was also very good. These findings are similar to other translations of the COH-QOL-OQ tool, including the Malayalam,³ Iranian,¹⁶ Croatian,¹⁰ and Chinese.¹⁹ This indicates a satisfactory level of reliability.

A CFA was undertaken to assess the validity of the COH-QOL-OQ tool to test the factor structure of the predetermined measurement model. For this study, there was good convergent validity and, similarly, all scores of discriminant validity were less than convergent validity, which indicates a lack of overlap between numerous constructs in measuring the same trial. Anaraki et al¹⁶ and Mayadevi et al³ evaluated the convergent and discriminant validity of the COH-QOL-OQ based on the items’ correlations with the total scale and other scales and concluded that in the Iranian and Malayalam COH-QOL-OQ instruments, respectively, there was a satisfactory correlations level, and that each subscale measured a single trait. However, Konjevoda et al¹⁰ demonstrated that some items of COH-QOL-OQ Croatian version illustrated less than ideal correlation with their own subscale and that some items had moderate to high levels of correlation to other subscales.

A CFA also was used to test the validity of the attribute groupings of the COH-QOL-OQ instrument. The final step of validity is administration of the CFA model to show significant GOF indices. The chi-square/df was 2.663, CFI =.935, NFI= .900, and RMSEA = .063. A Chinese validation of the COH-QOL-OQ by Gao et al¹⁹ employed 32 items in the CFA to validate their results from exploratory factor analysis. In addition, in their Croatian validation of the COH-QOL-OQ, Konjevoda et al¹⁰ performed CFA according to the original structure of 43 items classified into 4 subscales, and the CFA model did not indicate good model fit. In the present study, the modified reduced model proved a good fit with 4 factors regarding reliability for all scales. For validation of the COH-QOL-OQ Chinese version, a reduced model presented a good fit for a 4-factor model, with satisfactory levels of QoL scale reliabilities for all dimensions except the spiritual well-being scale.¹⁹ A less-than-perfect GOF was shown in the Croation version, which may have been due to the specific sample.¹⁰ In the current study, the final administration of the CFA model showed significant GOF indices. This indicates that it is a valid questionnaire to measure QoL among patients with an ostomy in the Saudi Arabian population.

Participants in this study were mainly Saudi, which may affect the generalizability of the results and how other Arabic-speaking populations perceive the translated version of the questionnaire. Some populations may find the current version difficult to understand because the Middle Eastern world has different dialects. Multicenter studies are needed in different Middle Eastern countries that have Muslim and Arabic populations to confirm these results and their generalizability. In addition, the backward translation was made by a bilingual translator whose first language is Arabic; at the time, it was not possible to find translators whose first language was English. Moreover, as none of the patients spoke English, testing both the English and Arabic versions of instruments with bilingual participants was not possible.

Improving disease-specific QoL is an emerging research priority.²⁷ Saudi Arabia lacks studies investigating QoL in patients with an ostomy. Research into this area would benefit Saudi patients.²⁸ A study of this kind would enable identification of factors influencing QoL among patients with an ostomy. The COH-QOL-OQ is a specific and sensitive instrument that provides further insight into the issues of most concern to this patient population. The current study translated the COH-QOL-OQ to Arabic to verify it for future research studies evaluating QoL of Muslim patients with ostomies in Saudi Arabia. To the best of the authors’ knowledge, this is the first translation of the COH-QOL-OQ into Arabic.

The findings from this study validate the first Arabic version of the COH-QOL-OQ and will assist in the assessment of HRQoL of Muslim patients with an ostomy in Saudi Arabia. A strength of this study is the diversity of data, including demographic characteristics such as age and sex as well as disease diagnosis and inclusion of several medical care facilities. Further studies utilizing the Arabic version of the COH-QOL-OQ may assist in improving QoL of this patient population and offer more understanding of their needs.

Ms Alenezi is a doctoral candidate and Dr Livesay is an associate professor, Discipline of Nursing, Royal Melbourne Institute of Technology (RMIT), Melbourne, Australia. Dr Kimpton is a senior lecturer, Chiropractic and Exercise Sciences, RMIT, Melbourne, Australia. Dr McGrath is a senior lecturer, Discipline of Nursing, RMIT, Melbourne, Australia. Dr Bedaiwi is consultant, Urology Department, Prince Sultan Military Medical City, Riyadh, Saudi Arabia. Dr Khan is an assistant professor, Department of Occupational Health, Faculty of Public Health and Health Informatics, Umm Al-Qura University, Saudi Arabia.

Address all correspondence to: Aishah Alenezi, 113/ 13 Bush Blvd, Mill Park, VIC, Australia; email: f4aisha@hotmail.com.