The GSRS and the QOLRAD are two of the most established, validated, reliable and responsive disease-specific instruments available for assessing gastrointestinal symptoms and their impact on patients' daily functioning [37
]. Both questionnaires have been proven to have very good psychometric characteristics when tested in clinical trials in patients with GERD and dyspepsia [20
]. This paper documents the psychometric validation of the Afrikaans, German, Hungarian, Italian, Polish and Spanish translations of the GSRS and the QOLRAD in patients with dyspepsia. The study was conducted mainly in gastroenterology centres and, therefore, results are particular to patients referred for gastroenterological investigation. Because of the standardized methodology used, results from all countries are directly comparable.
Questionnaires were completed in an electronic data capture device (ECD), which ensures that all questions are answered in full. Although ECD technology is becoming an increasing part of clinical trials, it is not yet widely available, and certainly not that widespread in clinical practice. Thus, in South Africa, paper versions of the questionnaires were used for the present study, and this may have affected the results obtained in this country, especially as the use of ECD technology has been shown to increase patient compliance and improve the quality of the data [36
The demographic and clinical characteristics of the patient populations were comparable in the different countries studied in terms of the number of patients recruited, their age, gender, symptom status and medication use. The Afrikaans patient population had the highest incidence of abdominal pain as the predominant dyspeptic symptom as well as the highest proportion of patients with a history of previous peptic ulcer and/or ulcerative reflux esophagitis. This was probably due to the fact that all South African subjects were recruited from a single gastroenterology clinic in a healthcare system that may refer only the most severely affected patients.
In terms of psychometric characteristics, internal consistency was high in all domains of all language versions of the QOLRAD, thus supporting construct validity. Internal consistency was also high in the Indigestion, Diarrhoea and Constipation domains of the GSRS in all language versions. However, although abdominal pain is the typical dyspepsia symptom, internal consistency was only moderate in the Abdominal pain domain of the GSRS. This generally low internal consistency in the Abdominal pain domain probably reflects the difficulty of measuring this often fluctuating symptom, as well as the fact that the GSRS Abdominal pain domain contains only two items. In previous international studies conducted in patients with GERD, internal consistency also tended to be lower in the Abdominal pain domain of the GSRS than in its other domains [30
Test-retest reliability was good for most or all domains in the Hungarian and German translations of the QOLRAD, but only for one domain in the Spanish version and for no domains in the Afrikaans and Polish versions of the QOLRAD. It is not clear why the test-retest reliability was low in these patient populations, although it may be an indication that the frequency, severity and impact of abdominal symptoms were not as stable on a week-to-week basis in these patients as was estimated by the investigators. No test-retest was performed in Italy, so it was not possible to assess test-retest reliability for the Italian version of the QOLRAD. Previous international studies in patients with GERD have yielded acceptable test-retest reliability for the German (ICC: 0.70–0.84), Spanish (ICC: 0.79–0.85) and Afrikaans (ICC: 0.71–0.82) versions of the QOLRAD, but only for some domains of the Polish version (ICC: 0.51–74) of the QOLRAD [30
]. In the GSRS, test-retest reliability was acceptable in the Indigestion and Constipation domains in most language versions. However, reliability was lower in the Abdominal pain domains in all countries assessed. As with internal consistency, the generally low test-retest reliability in the Abdominal pain domain is most likely due to the GSRS Abdominal pain domain containing only two items, as well as the complexity of measuring this variable symptom. It is also likely that at least some of the patients deemed stable by the physicians experienced small changes in their symptoms that were picked up by the PROs.
In terms of construct validity, all GSRS domains correlated with all QOLRAD domains. Correlation was significant for the relevant GSRS Abdominal pain domain and most or all QOLRAD domains in the majority of the different language versions. The GSRS Abdominal pain domain also correlated significantly with the relevant SF-36 Bodily pain domain in all language versions. All QOLRAD domains correlated significantly with the majority of SF-36 domains in most language versions. Both the GSRS and the QOLRAD were able to differentiate between patients whose health status differed with regard to frequency and severity of symptoms, thereby confirming the known-groups validity of the instruments. Known-groups validity has also been confirmed for several translations of the GSRS and the QOLRAD in patients with GERD [30
]. Confirmatory factor analysis has been shown to support the validity of the Scandinavian language versions of the GSRS and the QOLRAD in patients with GERD [41
]. Future studies could use this approach to further assess the validity of the additional language versions presented here.
The relevant GSRS Abdominal pain domain correlated with the HAD anxiety score in most of the language versions assessed, and most of the QOLRAD domains correlated with the HAD anxiety score in all language versions assessed. Dyspepsia tends to be associated with anxiety and depression [58
]. Previous reports are somewhat contradictory on the role of psychological morbidity in dyspepsia symptoms and healthcare-seeking behaviour [1
], suggesting that additional research is required on this potentially important aspect of the disorder.
When comparing the frequency and severity of patient-reported symptoms with those assessed by the physician, correlations were only low to moderate. These discrepancies in patient-reported and observer-reported symptom status have been well-documented [59
]. Our view is that the physician examination needs to be balanced with the patient-reported symptom status when deciding on patient management and outcomes [39
In summary, all language versions of the QOLRAD showed good internal consistency and reliability, although only the German and Hungarian translations were able to demonstrate acceptable test-retest reliability. In general, all language versions of the GSRS were reliable and showed good internal consistency for the Indigestion, Diarrhoea and Constipation domains, but not for the Abdominal pain domain. Overall, the test-retest reliability of the GSRS was not acceptable.