1. 2015;18;E323-E331Opioid Therapy in the Treatment of Chronic Pain Conditions in Germany
   Epidemiology
   Andreas Werber, MD, Ursula Marschall, MD, Helmut Lhoest, MD, Winfried Hauser, MD, Babak Moradi, MD, and Marcus Schiltenwolf, MD.

BACKGROUND: Primarily used for treating malignant pain, opioids are recently applied for chronic non-tumor pain. For the lack of evidence based strategies from long-term studies, opioids are discussed controversially, esp. considering cost-benefit. OBJECTIVES: The purpose of this study is to evaluate characteristics in prescribing opioids for tumor and non-tumor pain conditions. STUDY DESIGN: Cost effectiveness study/observational study SETTING/METHODS: Health insurance data of a German statutory health insurance company (N = 6.800.000, data acquisition from 2006 to 2010) was evaluated by assigning opioid prescriptions to certain pain related diagnoses using CART (Classification And Regression Tree) segmentation analysis. Age- and gender-specific characteristics of prescriptions were calculated. RESULTS: The number of prescriptions of opioid prescriptions increased linearly. Prescriptions of mild opioids were decreasing for non-tumor pain, but increasing for tumor pain, while the number of prescriptions of strong opioids was increasing both for tumor and nontumor pain. Differences occurred in terms of duration and kind of the preferred substances, including the considerations of common contraindications (e.g. somatoform disorders). The majority of strong opioids being prescribed for non-tumor pain were fentanyl pain patches for 40 to 45 year old males with average annual costs of 1833 Euros per patient. Out of 21000 patients with somatoform pain disorder, 44.4% were treated with opioids (20.7% with mild, 23.7% with strong opioids). LIMITATIONS: The results did not consider if the opioid medication was actually taken by the patients. Another difference in terms of representativeness might occur since the gender distribution varies between the official statistical data and data collected by the health insurance company. Because of the acquisition of the data, no conclusions about possible correlation of pain syndromes and educational and social classes are possible. Tumor patients who received an opioid prescription for non-tumor pain could not be excluded. CONCLUSIONS: While the overall expenditure of the health insurance companies increased, it remains unknown which patient is receiving a particular opioid medication. Prescribing behavior was often not consistent with common indications and contraindications.