A multitude of reports have delineated the risks of using non-steroidal anti-inflammatory drugs but have not been totally congruent. Meta-analyses of randomized controlled trials sometimes concur regarding gastrointestinal risk and cardiovascular risk but rarely report a balance of these risks for any one drug. Benefits measured in these studies are usually not reported. Observational data sets, supposedly reflective of 'real world' patients, do not always agree with the randomized controlled trial reports. Clinicians need assessments measuring the balance of harms and benefits so that better decisions based on their patients' unique risk factors can be reached.