1. 2013;16;569-580Assessment of Neuropathic Pain in Patients with Cancer: The Interobserver Reliability. An Observational Study in Daily Practice
   Prospective Evaluation
   Hans Timmerman, MSc, Irene Heemstra, MD, Annelies Schalkwijk, MSc, Constans Verhagen, MD, PhD, Kris Vissers, MD, PhD, and Yvonne Engels, MD, PhD.

BACKGROUND: Neuropathic pain (NeP) is a burdensome problem in all stages of cancer. Although clinical judgment is accepted as a surrogate for an objective gold standard in diagnosing NeP, no publications were found about its reliability.

OBJECTIVES: Therefore, levels of agreement on the clinical examination of NeP were estimated by calculating kappa-value (?) and percentage of pair wise agreement (PA) to determine the interobserver reliability of diagnosing NeP.

SETTING: The outpatient clinic of medical oncology of the Radboud University Nijmegen Medical Centre.

METHODS: Patients with cancer with potential NeP complaints were recruited from the outpatient clinic of medical oncology. Physicians were recruited from the department of pain and palliative medicine. Physicians and patients were recruited for participation in an observational study in daily practice. Each patient (N = 34) was examined by 2 specialists via independent clinical assessment. All consultations were video recorded.  After each assessment, physicians were asked to indicate the most adequate characterization of the pain: pure NeP, pure nociceptive pain (NoP), mixed pain (MiP), or no pain (NP).

RESULTS: Kappa (K) for the diagnosis of the most adequate pain characterization was 0.50, PA 64.7%. For diagnosing pure NeP k was 0.78 (PA 91.2%), for the NeP component (NeP + MiP)  and NoP component (NoP + MiP), it was respectively 0.52 (PA  76.5%) and 0.61 (PA 82.4%). For the diagnosis on the basis of the grading system between physicians, K was 0.34 (PA 52.9%). The intrarater reliability for the diagnosis of an NeP component on the basis of clinical assessment and the NeP component on the basis of the grading system, for pain specialists K was 0.69 (PA 85.3%) and for palliative care specialists K was 0.61 (PA 79.4%).

LIMITATIONS: The values of K and the PA for the existence of an NeP component are not satisfying and the clinical agreement between physicians around findings from physical examination should encourage a better standardization of the clinical assessment and classification of pain in patients with cancer in respect with the identification of NeP. 

CONCLUSIONS: A substantial level of agreement was found for the diagnosis of pure NeP and a moderate level of agreement for the diagnosis of the NeP component was found, both with a PA = 70%. There was only a fair agreement between the physicians regarding the grading system. However, there was a substantial level of (interrater) agreement for the diagnosis of an NeP component and the outcome of the grading system. The findings in this study also suggest that a better standardization of the clinical assessment and classification of pain in patients with cancer with respect to the identification of neuropathic pain is necessary.