Validation of the Arabic version of the EORTC Colorectal (CR29) module in Moroccan colorectal cancer patients

Purpose: Quality of life outcomes are a focal endpoint of cancer treatment strategies which can be evaluated through patient-related outcome measures (PROMs). We aim to validate the Moroccan Arabic version of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire for Colorectal cancer patients (EORTC QLQ CR-29). Methods: Following translation to Moroccan Arabic, the QLQ CR-29 was administered to patients treated for colorectal cancer (CRC) in the national institute of oncology. Psychometric properties were tested by measuring Cronbach’s alpha coecient for reliability and Intraclass correlation coecient(ICC) to examine test-retest reproducibility. The multitrait-scaling analysis was performed to demonstrate the validity of the instrument and the ability to discriminate between different patient groups was tested using known-groups comparison. Results: In total, 221 patients were included in our study and 34 patients completed the questionnaire twice. The urinary Frequency scale and Stool Frequency scale had good internal consistency with alpha Cronbach coecients of 0,79 and 0,83 respectively, whereas the coecient was moderately lower for the Blood and Mucus in Stool scale (0,61) and the Body Image scale (0,67). The ICCs ranged from 0,88 to 1 indicating good to excellent reproducibility. In multitrait scaling analyses, the criterion for item convergent and divergent validity was satisfactory. The known-group comparison showed statistically signicant differences between patients according to age, gender, stoma status, tumor location, and radiotherapy. Conclusion: The Moroccan Arabic version of the EORTC QLQ-CR29 is a valid and reliable tool and can be used for research and clinical purposes in Moroccan CRC patients. validity of the EORTC QLQ-CR29 using multitrait scaling analysis. [30] Convergent validity was examined by correlating each item with its own scale with an item-scale correlation of ≥ 0.40 equivalent to high correlation. Divergent validity on the other hand was tested by demonstrating that the item correlated higher with its own scale than with the others. of the EORTC QLQ CR-29 show that it’s a reliable and valid instrument to measure the quality of life of colorectal cancer patients and could be used to complement the EORTC QLQ-C30 in assessing HRQOL. Conducting more transcultural validations and standardizing patient-reported outcome questionnaires, especially in the eld of oncology, will allow us to broadly assess cancer therapy outcomes and weigh the benets against the quality of life impact.

The EORTC QLQ-CR29: The EORTC QLQ-CR29 is a colon and rectum site-speci c quality of life module with 29 items consisting of 4 multi-item scales (body image, urinary frequency, blood and mucus in stool, and stool frequency) and 17 functional/symptomatic single-items (sexual interest, urinary incontinence, dysuria, abdominal pain, buttock pain, bloating, dry mouth, hair loss, taste, atulence, fecal incontinence, sore skin, embarrassment, stoma care problem, impotence or dyspareunia). Among these items, only body image, anxiety, weight, and sexual interest are functional scales.
The eighteenth item (Q18) is an indicator of colostomy/ileostomy construction and following, items are separately arranged for patients with a stoma (Q19-Q25) and without (Q19-Q25) according to symptoms of stool frequency, atulence, fecal incontinence, sore skin and embarrassment while item 25 is speci c for stoma care. Sexual interest, impotence and dyspareunia items are categorized according to gender with the corresponding questions being Q26-Q27 and Q28-Q29 for male and female respondents respectively. All questionnaire items ask about the past week except the ones on sexuality, which request the patients to evaluate the past four weeks. As regards the scoring, the multi-item scales and single items are scored using a 1 to 4 point Likert scale ("not at all", "a little", "quite a bit", "very much") with the highest score representing the best functional status or the worst symptom [14].
Translation process : The translation process followed the European Organisation for Research and Treatment of Cancer (EORTC) guidelines [27]. Two different translators independently translated the questionnaire from English to Moroccan Arabic and a consensual version was attained following the discussion of both translations. Back translation into English was then performed by two different translators who had no prior knowledge of the original English version. An expert committee including surgeons, oncologists, epidemiologists and the four translators reviewed all translations and a provisional version was developed.
To examine patients' understanding of the translated version, the provisional Moroccan Arabic EORTC QLQ-CR29 underwent a preliminary test on a group of 34 patients with colon or rectal malignancies after which the nal version of the instrument was generated.

Study population and data collection :
Patients were prospectively recruited from the national oncology institute during the period between November 2019 and January 2020. Patients aged over 18 years with a pathologically con rmed cancer of the colon and/or the rectum who had undergone surgery at least 6 months prior to the enrollment in the study were included. Patients were excluded if they were unable to understand the questionnaire, presented cognitive and/or medical complications that hindered the interview completion and the submission of an uncompleted questionnaire. Participants were either approached during follow up visits or contacted via telephone. Patient's characteristics were reported according to age, gender, stoma status, cancer localisation (colon vs rectum), neoadjuvant radiochemotherapy and adjuvant chemotherapy.
As the sample size determination for psychometric validation studies lacks clear recommendations [28], we determined the required sample by allocating a number of observations 5 to 10 times greater than the variables.
[29] Accordingly, the sample needed size ranged between 150 and 300 participants.

Statistical Analysis:
The scores for the QLQ-CR29 and the QLQ-C30 questionnaires were linearly converted into 0 to 100 point scores according to the standard EORTC guidelines. [26] Descriptive statistics were generated through mean, median, standard deviation (SD), and oor and ceiling effects and age was categorized in 3 groups: < 40 years ; 41-65 years and > 65 years.
There are two different levels of reliability; internal consistency and reproducibility.
Internal consistency reliability was determined using Cronbach's alpha coe cient with a score greater than 0,7 considered acceptable, above 0,8 was good and higher than 0,9 was considered excellent.
As for reproducibility, a random subgroup of patients was selected to retake the QLQ CR-29 questionnaire after 7 to 14 days from the rst interview to examine the test-retest reliability. The results of the two measurements were assessed using the Intraclass correlation coe cient (ICC) and an ICC score of 0,7 or higher was considered acceptable.
We tested the construct validity of the EORTC QLQ-CR29 using multitrait scaling analysis.
[30] Convergent validity was examined by correlating each item with its own scale with an item-scale correlation of ≥0.40 equivalent to high correlation. Divergent validity on the other hand was tested by demonstrating that the item correlated higher with its own scale than with the others. Concurrent validity was examined by comparing the scores of the QLQ-CR29 and the QLQ-C30 using Pearson's correlation.
Clinical validity was assessed using known group comparison through the Mann Whitney U test to examine the QLQ-CR29' ability to differentiate clinically distinct patients. Subgroups were categorized according to : age (<65 years vs ≥65 years ), gender (male vs female), stoma status (permanent vs no stoma), tumor site (colon vs rectal ) and neoadjuvant radiotherapy (no vs yes). All statistical analyses were performed using SPSS 26.0 (SPSS Inc., Chicago,IL, USA). Statistically signi cant results were de ned with a P < 0.05 .

Patients Characteristics:
The sociodemographic and clinical characteristics of the patients enrolled in the study are detailed in Table 1. In total, 221 patients completed the questionnaire among which 123 were males and 98 were females with a mean age of 55,6 ±12,7years. Seventy eight (35,9%) participants had colonic cancer and 138 (64,1%) had rectal cancer, while 50 patients had a stoma (22,6%). Missing items were only associated with sexual problems with a miss rate of 9% for males and 23% for females. Table 2 summarizes the distribution of the EORTC QLQ CR-29 and QLQ-C30 scores. The mean score for the different dimensions of the QLQ CR-29 ranged from 16,44 to 75,56 with the items "Hair loss" and "Weight" scoring the lowest and highest respectively. The percentage of respondents at oor was high (>50%) in 12 areas while the percentage of respondents at ceiling was high (>50%) in 1 item. The range of scores was broad in twenty-one dimensions except the bags change which ranged from 0 to 83.

Reliability :
The internal consistency of the EORTC QLQ-CR29 reached the 0.7 criterion showing good consistency for the urinary frequency scale (0,79) and stool frequency scale (0,83), while for the blood and mucus (0,615) and the body image (0,672) scales the alpha Cronbach coe cient was slightly below the criterion (0.7). The Cronbach's alpha coe cient was higher for patients without stoma compared to those with a stoma, except for the body image scale (0,64 with vs 0,69 without) which indicates higher reliability for patients without a stoma. More details are shown in Table 3.
The reproducibility (test-retest reliability) of the Arabic version of the QLQ-CR29 was tested using the intraclass correlation coe cient (ICC) after administering the instrument twice to 34 patients. For each item, the ICCs ranged from 0,889 to 1 indicating good to excellent reproducibility.

Construct validity :
All items exceeded the 0,40 criterion for item-scale convergent validity. Similarly, items correlated better with their own scales than with others which shows good divergent validity. Details of the multitrait scaling analysis are shown in Table 3.

Concurrent Validity :
Correlations between the scales of the QLQ-CR29 and QLQ-C30 were low (r < 0.40). However, some areas with more related content showed higher correlations (r > 0.40), namely body image and social functioning. The abdominal pain scale also had a good correlation with the QLQ-C30 pain scale and stoma care problems were correlated to the global quality of life scale. In addition, most functional scales of the QLQ-CR29 were positively correlated with functional scales of the QLQ-C30 and negatively correlated with symptom scales of the QLQ-C30, while most symptom scales of the QLQ-CR29 were positively correlated with symptom scales of the QLQ-C30 and negatively correlated with functional scales of the QLQ-C30 as detailed in Table 4.

Clinical Validity:
The EORTC QLQ-CR29 was able to distinguish patients based on differences between known groups. Table 5 and Table 6 Differences in the scores of patients with stoma were noted as they presented signi cantly more anxiety and body image issues. Males with stoma reported higher symptom scores for the "impotence" scale.
The participants with rectal cancer had worse QoL than those with colon cancer and male patients with rectal cancer had signi cantly higher symptom scores for atulence, fecal incontinence, sore skin around the anus, stool frequency, defecation problems, and sexual dysfunction.
In addition, patients who received neoadjuvant radiotherapy had signi cantly higher symptom scores and more problems related to blood and mucus, buttock pain, bloating, stoma care problems, atulence, fecal incontinence, sore skin, stool frequency, embarrassment and defecation problems. Furthermore, the QLQ-CR29 showed differences between age groups with younger patients found to suffer more from defecation problems, stool frequency and embarrassment.

Discussion
Health-related quality of life (HRQL) in CRC is an important component in both day to day practice and clinical research, therefore the proper assessment of patients' HRQOL is crucial [31]. This study showed that the Arabic version of the EORTC QLQ-CR29 questionnaire has good internal consistency, test-retest reliability and validity and is therefore valid and reliable to assess the quality of life of Moroccan colorectal cancer patients.
The internal consistency of the Arabic EORTC QLQ CR-29 demonstrated satisfactory results for the urinary frequency scale and stool frequency scale, with higher reliability scores for patients without a stoma which is similar to the Chinese validation [20]. As regards the blood and mucus and the body image scales, the alpha Cronbach coe cients were acceptable which was the case in other similar studies. [18,32] On the other hand, as suggested by Arraras et al [18], some differences may be due to the fact that the EORTC original validation was conducted on an international sample with high variance, while the Spanish validation concerned a more homogenous sample which may impact the alpha Cronbach coe cient.
The ICCs of our study were all greater than 0.8, thus indicating good to excellent reproducibility for both single item and multi-item scales. The Reliability coe cients were higher in our study than those reported by the Dutch validation [17] and mostly similar to those in the original psychometric validation study.
[15] As such, the Moroccan Arabic translation of the QLQ CR-29 is a stable instrument.
The multitrait analysis con rmed the structure of all scales, which proves that the Moroccan Arabic translation of the QLQ-CR29 has a valid construct.
In the assessment of construct validity, correlations between the scales of the QLQ-C30 and the QLQ CR-29 were mostly low (<0.4) indicating that the two questionnaires measure different concepts. Few areas with related content had higher correlation scores which was expected given the similar concepts of these particular scales. Nonetheless, the results show that the two questionnaires are independent.
In terms of clinical validity, we found less signi cant differences related to stoma status than the original study [15]. Moreover, patients with colon cancer had a better function and fewer symptoms, including sexual interest in males and stool frequency as opposed to rectal cancer. Interestingly, patients with rectal cancer and a stoma experienced more embarrassment with borderline signi cance (p=0.053). When comparing age groups, younger patients reported worse symptoms than older patients. Similar results were reported by the Dutch and Spanish Validation studies [17,18]. In addition, the particularities of the Moroccan population may be contributing to elderly patients' display of better resilience, QoL satisfaction, relatively better acceptance and the aforementioned results. Consequently, the QLQ-CR29 was found to discriminate between age groups.
A higher missing data rate was registered for sexual dimensions compared to others as patients were more reticent about answering sex-related questions which makes their interpretation more di cult. Similar observations were made in the Chinese and Iranian studies, which hindered discussions regarding sexual activity and even ostomy. [20,22] Nonetheless, providing explanations to patients when answering the questionnaire was noted to help. In our context, this issue may be explained by the cultural and religious particularities of the Moroccan population where sexual practices are taboo.
[33] More studies addressing this problem should be conducted to determine the reliability and validity of the CR-29 in evaluating the sexual aspects of QoL for patients according to cultural contexts. This study has some limitations, one of which is the limited sample size of patients. However, the minimum sample size was set at a 150 and other EORTC QLQ-CR29 validation were performed on a smaller population. Self-administration was not possible due to the high level of illiteracy in our context; consequently, patients received the help of one of the investigators who was in charge of reading the questions and different options for the answer.