Pain is very common throughout the world and is an increasing problem in the ageing population. Non‐steroidal anti‐inflammatory drugs (NSAIDs) are widely prescribed to treat pain and many are also available without prescription, or over the counter. These drugs are effective painkillers, but they can also have severe adverse effects, particularly on the upper gastrointestinal (GI) tract. Therapeutic decisions should be made using the best available evidence and there is a growing body of evidence showing that the new specific cyclooxygenase‐2 (COX‐2) inhibitors, or coxibs, are effective pain killers that do not cause GI harm. The risks associated with the use of NSAIDs are substantial, with a 1 in 1200 chance of dying from a major GI adverse effect after 2 months of NSAID therapy. These risks increase with age and are avoidable. The costs associated with the prevention and treatment of NSAID‐induced GI adverse effects can more than double the cost of the original therapy and should be included when costing NSAID interventions. Taking these costs into account, the expense of switching from a conventional NSAID to a coxib is relatively modest. Compared with other interventions that society may be willing to consider to prevent one death, such as those for the rail (£15 million) and road (£100 000) networks in the UK, the cost of preventing one death by switching to a coxib is much lower, with a high estimate being £20 000–30 000, which is in line with the accepted benchmarks for the cost‐effectiveness of medical interventions.