From a trusted source, the Cochrane Collaboration. The authors reviewed 13 Cochrane Reviews published up to February, 2017, on the treatment of acute and chronic pain with a range of topical analgesics.
Most reviews compared the topical painkillers with topical placebo.
They were able to show with moderate to high quality evidence that topical NSIADs (diclofenac Emulgel, ketoprofen gel, piroxicam gel, diclofenac Flector plasters, and other diclofenac plasters) are effective (50% pain reduction in 20-50% of people), over 1 week, in the treatment of strains and sprains and (topical diclofenac and ketoprofen) to a lesser extent (50% pain reduction in 10-20% of people), over less than 6 to 12 weeks, in the treatment of hand and knee osteoarthritis. A single application of high-concentration topical capsaicin was shown with moderate quality evidence to have limited efficacy (50% pain reduction in 8.3% people), over 8-12 weeks, in the treatment of postherptic neuralgia.
There was low or very low quality evidence showing limited efficacy of other topical preparations leading to the conclusion that there is no good evidence to support any other topical pain killer for any other painful condition.
In the treatment of acute conditions topical NSAIDs caused no more systemic or local adverse events than placebo. In the treatment of chronic conditions topical capsaicin (itching or rash) and diclofenac, but not ketoprofen, were associated with more local adverse events than placebo.
The take-home message here is that specific topical painkilling preparations may be effective in the treatment of specific acute and chronic conditions with the authors high-lighting that the exact formulation used being possibly of critical importance.
The authors also high-lighted that the analgesic efficacy of topical analgesics is not just about rubbing them in, despite what is commonly believed. It may though go towards explaining the strong placebo effects noted in the reviews with the placebo preparations resulting in 50% or more improvements in pain in 20-57% of acute and 23-50% chronic conditions.