This study was undertaken as part of a longitudinal, prospective study on survival rates and QoL for oral cancer patients at Khartoum Teaching Dental Hospital (KTDH), Sudan from January 2013 to December 2016. KTDH is the main Maxillofacial surgical centre, to which the majority of oral cancer cases from all over Sudan are referred. Ethical approval was obtained from the Ministry of Health Ethics Committee, Sudan. While several QoL instruments exist, the FACT (V-4) QOL instrument was chosen as it has been widely validated and translated into many languages including Arabic, and its reliability had been previously confirmed by many researchers [12,13,14,15]. Furthermore, the questionnaire is short, easily completed and feasible for busy clinical settings. The modular construction of FACT allows comparison across cancer diagnoses while still probing issues specific to oral cancer. It consists of four subscales, including physical, social, functional, and emotional well-being issues. Response options comprised a 5-point Likert scale (‘not at all’, ‘a little bit’, ‘somewhat’, ‘quite a bit’, and ‘very much’, scored from 0 to 4), which indicates ‘how true’ the patient's response over the past seven days, with higher FACT subscale and summary scores denoting better QoL. The additional 12-item Head & Neck (FACT-H&N) specific subscale was not available in Arabic at that time.
Participants Formal sample size calculation was not undertaken, the sample size was determined by the number of patients presenting during the recruitment period. All new patients with histologically proven oral cancer (ICD-10, C00, C02-08, C14.8, C31.1, C41.1), attending KDTH between January 2013-December 2016, who satisfied the inclusion criteria were consented and enrolled in the prospective study. Basic demographic and clinical details were collected as part of this study, including age, gender, marital status, living arrangements, habits, tribe, tumour stage, type and chronic illness profile. Of these, 130 agreed to participate in the QoL study, with 60 participating in the initial pilot study. A single interviewer identified and approached all newly diagnosed patients explained the purpose of the study, and following consent administered and collected all questionnaires. As literacy levels in Sudan are relatively low, the Questerview technique was used where required .
Inclusion criteria Adult patients (18 years and above) with histologically confirmed oral cancer, who were able to speak/understand Arabic and who had not yet received any oral cancer therapy. Patients deemed unable to understand or respond to the QoL questionnaire or unable to speak/understand Arabic were excluded.
Translation process Translation to Arabic and validation steps were undertaken with the advice and consent of the original developers to adapt it for Sudanese oral cancer patients, following standard FACT translation methodology . Steps involved: 1) Forward translations from English to Arabic by two expert translators from Khartoum University, working independently, 2) Reconciliation of these two forward translations by a third translator, 3) Back translation into English by a fourth translator, 4) Review and finalization by a fifth translator, 5) Proof-reading the Arabic FACT-H&N questionnaire by the bi-lingual, bi-cultural researcher, 6) lastly, the Arabic translated FACT-H&N was pre-tested for face validity and content validity, then cross-culturally adapted for the Sudanese context in terms of its conceptual and operational equivalence. Pre-testing was undertaken by the main researcher using semi-structured, face-to-face in-depth interviews with a purposive sample of ten oral cancer patients of different ages, cancer stages and cancer types. These ten patients completed the translated questionnaire, and then answered the Patient Interview Form (PIF), a cognitive debriefing script provided by the original developers, to assess acceptability and to see whether any items were confusing or upsetting. The item on sex satisfaction was considered unacceptable by most patients and had a higher rate of missing data. All other items were deemed clear, acceptable and understandable.
Sixty newly diagnosed oral cancer patients were enrolled in the pilot study; they completed the baseline assessment, and a test-retest seven days later, to assess the feasibility, reliability, validity and internal consistency, i.e., psychometric properties, of the FACT-H&N scale and the patients’ response to the translated Arabic version. FACT summary scores included the general module (FACT-G), head and neck specific module (FACT-H&N), FACT symptom-index score (FHNSI) which is a subset of the specific FACT-H&N module, and the FACT-H&N-Trial Outcome Index (FACT-H&N TOI) which is a combination of the functional, physical and head and neck subscales, and the head and neck symptom index (FHNSI).
Instrument reliability was evaluated by the Test-Retest Reliability at baseline and after seven days using the Intra-Class Correlation Coefficient (ICC), while the internal consistency reliability was evaluated using Cronbach’s α coefficient for the pilot group. Construct validity was assessed using the known group’s validity by the analysis of variance (ANOVA) test: pilot study patients were asked the global health-related QoL item: “how do you rate your health-related quality of life in the past 7 days”? The FACT summary and subscale mean scores were tabulated against the three categories of the global HRQOL item (very poor/poor, average and good/very good).
Furthermore, the patients who took part in the pilot study were included in a longitudinal prospective study. A total of 130 patients were enrolled, and they were assessed for their HRQOL at four data collection points: at diagnosis (baseline), and one, three and 12 months later after commencing treatment. Once again, the cross-sectional construct validity was assessed using known group validity. All 130 patients were categorized at diagnosis (baseline) into early (I/II) or late/advanced disease (stage III/IV). FACT summary and subscale mean scores for early and late stages were compared using the independent t-test. Convergent validity was assessed using the correlation matrices for FACT summary scale and sub-scales using Pearson’s correlation co-efficient.
Missing responses were managed by pro-rating these scores, as recommended in the FACT manual. The sum of each FACT-G module (27 items) was calculated. Derivatives of FACT scores were calculated. All statistical tests were done using SPSS® version 22.0; p < 0.05 was considered statistically significant.