Abstract
PURPOSE: Chemotherapy-induced peripheral neuropathy (CIPN) is a disabling complication of many chemotherapies. We investigated the feasibility of using health plan claims and administrative data to identify CIPN occurrence by comparing patients who received neurotoxic and non-neurotoxic chemotherapies. METHODS: The sample included over 53,000,000 patients from two regional and one national insurer in the USA (> 400,000 exposed to chemotherapy). Peripheral neuropathy was identified using a broad definition (definition 1) and a specific definition (i.e., drug-induced polyneuropathy code) (definition 2). RESULTS: CIPN incidence as measured by definition 1 within 6 months of chemotherapy initiation was 18.1% and 6.2% for patients who received neurotoxic and non-neurotoxic chemotherapy, respectively (relative risk neurotoxic vs. non-neurotoxic (RR), 2.93 (95% CI, 2.87-2.98)). For definition 2, these incidences were 3.6% and 0.1% (RR, 25.2 (95% CI, 22.8-27.8)). The incidences of new analgesic prescriptions for neurotoxic and non-neurotoxic groups were as follows: gabapentin, 7.1%/1.7%; pregabalin, 0.69%/0.31%; and duloxetine, 0.78%/0.76%. The incidence of CIPN as defined by definitions 1 and 2 was low compared with that of published research studies, but the relative risk of CIPN among patients who received neurotoxic chemotherapies compared with those who received non-neurotoxic chemotherapies was high using definition 2. CONCLUSIONS: These data suggest that as used currently by clinicians, administrative codes likely underestimate CIPN incidence. Thus, studies using administrative data to estimate CIPN incidence are not currently feasible. However, the drug-induced polyneuropathy code is a specific indicator of CIPN in administrative data and may be useful for investigating predictors or potentially preventive therapies of CIPN.