Arabic translation, adaptation and modification of the dialysis symptom index for chronic kidney disease stages four and five

Permission was granted by Weisbord to modify and then translate the DSI into Arabic. First, two symptoms (depression and nocturia) and three free fields were added. Both depression [14] and nocturia [15,16] are highly prevalent in CKD, but are not routinely included in symptom measures. The additional free fields were included to capture any possible symptoms that were missing from the DSI. Then two symptom dimensions (severity and frequency) were also added to the DSI [2,12,13]. All symptom dimensions (frequency, severity and distress) were rated on a 0-10 point scale because this type of response scale is easier for patients to use (i.e. similar to the commonly used numerical pain rating scale), provides a wider range of possible scores, and increases the statistical analyses that are available. In the original DSI, patients were asked to consider symptoms they had experienced in the previous week. However, we extended the time frame to four weeks to reflect the chronic nature of CKD symptoms. For example, some symptoms may occur often only during dialysis (e.g. cramps and dizziness) and may not be persistent over 4 weeks. These modifications created a comprehensive instrument, re-named (with permission from Weisbord) the CKD Symptom Burden Index (CKD-SBI), to assess symptoms in different stages of CKD. The time needed to complete the modified instrument ranged between 15 to 30 minutes.

The CKD-SBI includes four symptom dimensions (prevalence, distress, severity and frequency) across 32 symptoms (see Additional file 1). The prevalence scale assesses the presence or absence (Yes/No) of each symptom with a higher overall score indicating a higher symptom occurrence. The possible range of prevalence scores is 0-32. Distress, severity and frequency are each measured using a 0 to 10 scale for each of the symptoms. Participants rate the “distress” subscale from none to very much, “severity” from none to very severe, and “frequency” from never to constant. A maximum total score for each of these scales is 320 (range 0-320). The total CKD-SBI score is calculated by summing the scores of each CKD-SBI subscale (prevalence, distress, severity and frequency) and then multiplying the result by 0.1008 (constant number). The total score for the CKD-SBI scale therefore ranges between 0-100, where 100 indicates the highest symptom burden possible.

We used the Brislin back-translation method [14,17]; the most frequently used and accepted method for effective translation equivalence of instruments [18]. The method begins with a bilingual expert translating the instrument from its source language (English) into the target language (Arabic). Consultations with an expert panel are an important stage to ensure clarity, detect linguistic mistakes and ensure cross-cultural equivalence [19]. Then a blind back-translation (without accessing the original source language version) from the target language to the source language should be undertaken by another bilingual translator [18]. The back-translated version is then compared with the original version. If there are errors in the meaning in the blind back-translated version after comparison with the original version, the process of translation is repeated for the sections that include errors in the meaning [20]. Figure 1 summarises the steps in the Brislin’s method.

Translation processes were undertaken in Saudi Arabia using bilingual (Arabic/ English), certified translators. Formal written Arabic language was used as this form of Arabic is used in all Arabic-speaking countries; thus increasing the usefulness of the instrument to other CKD populations outside of Saudi Arabia. The Arabic expert panel involved three nephrology nurses, two nephrologists, a social worker and two primary school teachers. The goal of the panel was to: 1) ensure clarity, readability and linguistic appropriateness; 2) review of grammatical style and comprehensibility; and 3) maintain cultural equivalence. The primary school teachers ensured suitable literacy level of the CKD-SBI for people with limited literacy skills (which will be tested in the pilot study). The expert panel recommended several minor linguistic and grammatical modifications, correction of spelling, and clearer instruction statements. A blind back-translation from the Arabic version into English was then performed by another bilingual certified translator independently of the first translator and without access to the original English CKD-SBI. The procedure that was used to compare the back-translated version with the original version involved a new independent professional translator, two bilingual members of the expert panel and the first author. There were no errors identified in meaning after comparing both versions.

Content validity index (CVI) testing also occurred during this phase using the expert panel and a professional translator from a Saudi hospital. Individually they were asked to rate their level of agreement with the relevance and equivalence of each item between the Arabic and English versions on a four-point scale (1 = ‘not relevant’ to 4 = ‘very relevant’) to evaluate content validity [21]. A score of 3 or 4 indicates that the translated item is valid and equivalent to the English version [21]. The total CVI was 0.98 with items ranging from 0.88-1.00, which indicated that the CKD-SBI had good content validity.

Siaki [22] recommends pretesting of a translated instrument as a method to determine equivalence between the target language and source language. We conducted a pilot study of both versions of the CKD-SBI using bilingual university students and measured test-retest reliability, intra-rater reliability and internal consistency. As symptoms are also experienced by normal healthy populations but in a lesser frequency than those with CKD, using university students was appropriate when assessing language equivalence. Several studies have used this method when assessing language equivalence [23-25]. The pilot study involved anonymous completion of the English and Arabic versions of the CKD-SBI, on two separate occasions, one week apart. Demographic information (i.e. age, gender) was not relevant to determining equivalence and was not collected. An Arabic announcement was applied to the university notice board, and a snowball sampling technique was used to recruit participants during April-May, 2013. Ethics approval was obtained from the Queensland University of Technology (QUT) prior to commencement.

The main study used a cross-sectional design to evaluate the psychometric properties of the CKD-SBI in Arabic-speaking CKD patients who were either receiving dialysis or not. A convenience sample of CKD patients from three hospitals in Saudi Arabia were recruited during July 2013 to February 2014. Initially treating clinicians identified potential eligible participants - adults with CKD (eGFR < 30 mls/min/1.73 m²) or currently receiving either haemodialysis or peritoneal dialysis, willing to participate, and able to communicate in Arabic. Exclusion criteria were cognitive impairment that would preclude voluntary, informed consent, and those with critical conditions. Ethical approval was obtained from King Abdulaziz University Hospital, Jeddah Research Centre, and QUT Human Research Ethics Committees. All participants provided written informed consent. Data was collected through in-person distribution in the nephrology clinics (non-dialysis CKD patients) and the kidney dialysis centres (dialysis patients). Participants completed the newly translated Arabic version of the CKD-SBI (32 items) along with the Kidney Disease Quality of Life short form (KDQOL-36™). The KDQOL-36™ has 36 items and is available in Arabic [26]. The total score of the KDQOL-36 ranges from 0-100, with higher scores indicating a better QOL. Demographic and renal characteristics were also collected. Davies et al.’s [27] co-morbidity index was used to estimate the number of co-morbidities for the sample. This index consists of seven domains of co-morbid conditions, including malignancy (active, non-cutaneous); ischaemic heart disease, peripheral vascular disease, left ventricular dysfunction, diabetes mellitus (type 1 or 2); systemic collagen vascular disease; or other significant pathology. The total score is determined by summing the score of each co-morbid domain. The scores are then graded into three risk groups: grade 0 is a zero score, grade 1 (1-2 domains), or grade 2 (3-7 domains).

Data for the pilot study and the main study were entered into IBM SPSS Statistics version 21 in two different data sets and reported in the results section separately. For the pilot study, the percentage of agreement between the original English and Arabic versions of the CKD-SBI was calculated to examine equivalence. Kappa statistics were used to assess the inter-rater agreement between the English and Arabic versions for categorical data. Intra Class Correlation coefficient and Wilcoxon signed-ranks test for matched pairs (numerical scales) were used to determine the consistency of CKD-SBI scores. The ICC coefficient was used to estimate the inter-rater reliability of the CKD-SBI. Pearson’s correlation coefficients were also calculated to assess the test-retest reliability. Cronbach’s alpha reliability coefficients were calculated to determine the internal consistency.