Translation of Malay language Rome III FD Diagnostic questionnaire
The English version Rome III FD Diagnostic questionnaire was translated into the Malay language by 2 independent translators; one a bilingual physician (YYL) and the other, a qualified linguistic expert. The translators then met together with the research coordinator (JS) for reconciliation of differences as outlined in the guidelines issued by the Rome foundation . The revised Malay translated version of the FD questionnaire was then back translated into English by another qualified local linguistic expert who then met with the research coordinator for reconciliation. Further revisions were made to the Malay version if needed during a subsequent reconciliation process involving two linguistic experts, two clinicians and the research coordinator. All revisions made to both forward and backward translations during the reconciliation process were carefully documented in a table showing the source language, target language version, comments and revisions made. The Malay translation was then tested for internal consistency, randomly, in 50 subjects who had a good command of the Malay language.
The Red Flag and the Psychosocial Alarm Questionnaires used in the current study were previously translated into the Malay language and had been validated for use in the local population. The intra-class correlation co-efficient (ICC) for the Red Flags and Psychosocial Alarm Questionnaires was 0962 and 0.994 respectively . Briefly, the Red Flag Questionnaire includes history for the past three months of fever, weight loss, cancer in family members, blood mixed with stool, anaemia and change in bowel habit after age 50. The Psychosocial Alarm Questionnaire has 7 items designed to identify those patients with psychological “problems” and who may benefit from psychological evaluation. There are two questions for anxiety and depression of which the answers, “most or a lot of the time” identified the presence of these disorders with a mean Hospital Anxiety and Depression Scale (HADS) score of 13.3 and 9.0 respectively (compared with HADS mean score of 6.0 and 5.0 in those who answered “occasionally or not at all”) .
Definitions and diagnosis of FD according to Rome III
According to the Rome III definition, FD is characterized by bothersome postprandial fullness, early satiation, epigastric pain and burning for the past 3 months with at least 1 day per week or more with onset of more than 6 months prior to diagnosis and no evidence of structural disease at endoscopy . It is subdivided into two meal-related syndromes namely post-prandial distress syndrome (PDS) and epigastric pain syndrome (EPS). Briefly, PDS is characterized by bothersome post-prandial fullness after ordinary sized meals and or early satiation that prevent finishing a regular meal with frequency of at least several times per week. With EPS, there is intermittent pain or burning discomfort of at least moderate severity, localized to epigastrium and occurs for at least a day per week. The pain or burning discomfort often disappears completely in the same day, not relieved by defecation or flatulence and not fulfilling criteria for gallbladder or sphincter of Oddi disorders.
The Rome III FD questionnaire consists of 18 items with answers in ordinal scale and individual frequency thresholds. Item 3, 5 and 7 consist of questions on frequency of symptoms of post-prandial fullness, early satiation and epigastric pain or burning in the last 3 months with items 3, 5 and 8 for determination of symptom onset for 6 months or longer in a yes or no response. Items 3 and 5 are questions diagnostic for PDS and items 1, 2, 7, 9, 10 and 12 are questions that characterize EPS. Other items from 11 to 18 contain questions to exclude functional gallbladder and sphincter of Oddi disorders.
Cross-sectional survey of FD
The translated questionnaires were subsequently used in a prospective cross-sectional survey for FD involving only ethnic Malays attending a primary care clinic within the university hospital of Universiti Sains Malaysia (USM). The hospital is located in the heart of Kota Bahru, the state capital of Kelantan, with an estimated population of 570,000 and 90% of its racial diverse population consists of ethnic Malays , all embracing the Islamic faith.
Consecutive Malay subjects (18 years old and above) presented at the primary care clinic were screened by the study investigator (NW) for inclusion into the study. Using a systematic 1:2 random sampling approach, those subjects randomized for screening were then called-in for an interview to determine for their study suitability. Subjects who were able to understand and complete the Malay translated questionnaires and also agreed for subsequent upper endoscopy examination were included. Those subjects with a known organic cause for their dyspeptic symptoms including previous peptic ulcer disease, previous endoscopy showing reflux disease and gastro-duodenal diseases, upper gastrointestinal tract cancers and psychiatric illnesses were excluded.
A total of 192 Malay subjects were initially randomized and screened for inclusion into the study. This sample size was calculated with a good power and to allow for multivariate analysis, the two proportion formula (for categorical variables) and the two means formula (for continuous variables) were used to derive the sample size calculated with the PS software. After considering 20% drop-out, a power of 80%, a precision of 0.05 and an estimated proportion of 11.8% , the calculated sample size was 192.
All enrolled subjects were given a questionnaire booklet to answer, usually within 20–30 minutes. The booklet contained Malay translated Rome III FD questionnaire, Red Flags and Psychosocial Alarm Questionnaires as well as socio-demographic questions including age, gender, educational background, occupation, individual income, number of children, smoking habits, drugs taken and past medical history. We have included the Red flags and Psychosocial Alarm Questionnaires in the current study since the association between red flags and psychological symptoms with FD is unknown in this population.
Only one trained interviewer (AW) was involved to maintain consistency in data collection and the interviewer was a clinician experienced with the Rome III questionnaires and Rome III diagnostic criteria. The body mass index or BMI (in Kg/m2) was determined in all study subjects, with height measured using a standard measuring tape (in cm) and weight with a calibrated weighing machine (in Kg) (Seca Deutschland, Hamburg, Germany). Those subjects who fulfil the Rome III criteria for FD were then referred for an upper endoscopy examination to rule out other possible upper gastrointestinal causes of their dyspeptic symptoms.
All upper endoscopy was performed by one experienced endoscopist (NM) who was blinded to the study; without knowing the diagnosis of FD beforehand in study subjects. All subjects were examined for H. pylori infection using CLO-test and histology with biopsies taken from at least two sites including antrum and body.
If there was further suspicion of any other gastrointestinal diseases and or red flags were positive then investigations were organized during subsequent follow-up at the gastroenterology clinic. Subjects with psychosocial alarm symptoms were then referred to psychiatrist if and when needed.
The study was approved by the Human Ethics Committee of USM.
Data and statistical analysis
All data were recorded in mean, standard deviation (SD), frequency and percentage unless otherwise stated. Univariable logistic regression was used to test the association of independent variables (including age, sex, BMI, education level, occupational status, family income level, marital status, smoking status, red flags and psychosocial alarm features) with the presence or absence of FD, these are reported as crude odds ratio (OR) and 95% confidence interval (CI). Variables with statistical significance below 0.05 from univariable analysis were subsequently tested using the multivariable logistic regression analysis and reported as adjusted odds ratio, 95% CI and P value. All analyses were carried out using SPSS version 18.0 (SPSS Inc, IL, Chicago). A P value < 0.05 was considered statistically significant for all analyses, including non-parametric tests.