Background: The painDETECT questionnaire (PD-Q) has been used widely for the identification of neuropathic pain (NeP); however, the reliability of the English version of the PD-Q has never been investigated.
Objective: This study aimed to determine the reliability of the PD-Q pre- (T0) and immediately post- (T1) clinical consultation and at one-week follow-up (T2).
Methods: We recruited 157 patients attending a Neurosurgery Spinal Clinic and Pain Management Department. Minor changes to PD-Q instructions were made to facilitate patient understanding; however, no changes to individual items or scoring were made. Intraclass correlation coefficients (ICCs) were used to assess the reliability of PD-Q total scores between T0-T1 and T0-T2; weighted kappa (κ) was used to assess the agreement of PD-Q classifications (unlikely NeP, ambiguous, likely NeP) between all time-points. To ensure stability of clinical pain, patients scoring ≤2 or ≥6 on the Patient Global Impression Scale (PGIC) at T2 were excluded from the T0-T2 analysis.
Results: Accounting for missing data and exclusions (change in PGIC score), data for 136 individuals (mean [SD] age: 56.8 [15.2]; 54% male) was available, of whom n = 129 were included in the T0-T1 and n = 69 in the T0-T2 comparisons. There was almost perfect agreement between the PD-Q total scores at T0-T1 time-points (ICC 0.911; 95% CI: 0.882-0.941) and substantial agreement at T0-T2 (ICC 0.792; 95% CI: 0.703-0.880). PD-Q classifications demonstrated substantial agreement for T0-T1 (weighted κ: 0.771; 95% CI: 0.683-0.858) and for T0-T2 (weighted κ: 0.691; 95% CI: 0.553-0.830). Missing data was accounted in 13% of our cohort and over 42% of our patients drew multiple pain areas on the PD-Q body chart.
Conclusion: The English version of the PD-Q is reliable as a screening tool for NeP. The validity of the questionnaire is still in question and has to be investigated in future studies.
Keywords: Clinical assessment; neuropathic pain; reliability; screening tool.