Pain questionnaire performance in advanced prostate cancer: comparative results from two international clinical trials

Abstract

To compare pain assessment questionnaires commonly used in advanced prostate cancer trials and to determine the psychometric characteristics and longitudinal relationships by contrasting questionnaire data from two international phase 2 trials. Scores from the Present Pain Intensity (PPI) question of the McGill Pain Questionnaire, the pain intensity scale of the Brief Pain Inventory (BPI), and the Functional Assessment of Cancer Therapy-Prostate (FACT-P) were analyzed using Pearson correlation, intraclass correlation coefficient, and Cronbach's α, respectively. Concordance was evaluated with Cohen's kappa coefficient and McNemar test at baseline (n=224) and two subsequent observations. PPI and FACT-P scores were associated with the BPI score at baseline for Trials 1 and 2: PPI r=0.66 and 0.80, respectively (P<0.001); FACT-P (pain scale) r=-0.76 and -0.82, respectively (P<0.001). However, concordance analysis revealed that the BPI identified pain (score>0) at higher rates than the PPI: at baseline, BPI: 89% (64/72) and 77% (95/124), PPI: 68% (49/72) and 64% (79/124) [Trials 1 and 2, respectively; McNemar test (P<0.001) for both studies]. The FACT-P pain scale identified pain similarly to the BPI pain intensity scale; longitudinal analysis produced comparable findings. All pain scales met standard psychometric acceptability criteria, but the BPI and FACT-P performed better than the PPI. Data suggest the BPI pain intensity and FACT-P pain scales are better than the PPI question at capturing the pain experience among patients with advanced prostate cancer. Additional comparative research is needed in larger population samples.