Neuropathic pain is frequent in the general population, with 7 to 10% of adults presenting with chronic neuropathic pain. To date, the gold standard to evaluate treatments is based on randomized controlled trials. Nonetheless, such design is run on a limited sample and for a limited period. Moreover, many treatments will never be compared directly in sufficiently large and representative populations. A way to overcome several of these limitations is to use real-world data. Indeed, the International Association for the Study of Pain (IASP) includes a special interest group focusing on pain registries and promoting the use of such approaches. In this short narrative review, several of the main chronic pain registries are presented. The strengths and weaknesses of this approach are presented. Indication bias is frequent in observational studies because the choice of treatment is generally influenced by the patients’ characteristics. However, a propensity score can be computed to adjust for these differences. The use of propensity score is briefly explained. Some data specific to neuropathic pain are discussed.