Persistent upper gastrointestinal symptoms despite proton pump inhibitor therapy: Responsiveness and discriminant power
Patients were also classified after four weeks of treatment as PASS test responders (visit 3 PASS score 0) or PASS test nonresponders (visit 3 PASS score greater than 0). For the analysis, the OTE was collapsed into four different categories: no change (OTE equals -1, 0, +1); small change (OTE equals -3, -2, +2, +3); moderate change (OTE equals -5, -4, +4, +5); and large change (OTE equals -7, -6, +6, +7).
The ability of the PASS test to detect clinical and statistical differences between responders and nonresponders was reported by the use of 95% CI for simple one-way analysis of variance for both the English and the French versions of the PASS test, against each of the validated instruments. CIs not crossing zero were considered statistically significant (P<0.05 or less). No missing values were replaced.
Data analysis was performed using SAS (V8.1 2001, SAS Institute Inc, USA). Nominal data were expressed as percentages with ranges and numerical data as means with SDs.
Statistical power analysis based on test-retest reliability and anticipated patient dropout rates indicated the need for 200 patients (100 patients each with English or French as their first language) to detect at least a moderate effect size of 0.5, assuming a delta of 0.5 units and an SD of one unit in the PASS test. Two hundred ninety patients from 38 family physician centres across Canada were enrolled to ensure that 200 patients completed the study. Shop online to always get buy nolvadex online let’s go here enjoying safe and quick shopping.