Original ArticleAccuracy of a Diagnostic Algorithm to Diagnose Breakthrough Cancer Pain as Compared With Clinical Assessment

ContextBreakthrough cancer pain (BTCP) is a heterogeneous condition, and there are no internationally agreed standardized criteria to diagnose it. There are published algorithms to assist with diagnosis, but these differ in content. There are no comparative data to support use.ObjectivesTo compare the diagnostic ability of a simple algorithm against a comprehensive clinical assessment to diagnose BTCP and to assess if verbal rating descriptors can adequately discriminate controlled background pain.MethodsPatients with cancer pain completed a three-step algorithm with a researcher to determine if they had controlled background pain and BTCP. This was followed by a detailed pain consultation with a clinical specialist who was blinded to the algorithm results and determined an independent pain diagnosis. The sensitivity, specificity, and positive and negative predictive values were calculated for the condition of BTCP. Further analysis determined which verbal pain severity descriptors corresponded with the condition of controlled background pain.ResultsThe algorithm had a sensitivity of 0.54 and a specificity of 0.76 in the identification of BTCP. The positive predictive value was 0.7, and the negative predictive value was 0.62. The sensitivity of a background pain severity rating of mild or less to accurately categorize controlled background pain was 0.69 compared with 0.97 for severity of moderate or less; however, this was balanced by a higher specificity rating for mild or less, 0.78 compared with 0.2.ConclusionThe diagnostic breakthrough pain algorithm had a good positive predictive value but limited sensitivity using a cutoff score of “mild” to define controlled background pain. When the cutoff level was changed to moderate, the sensitivity increased, but specificity reduced. A comprehensive clinical assessment remains the preferred method to diagnose BTCP.