evidence profiles for each intervention for the prespecified clinically important outcomes, which provided summary data on the benefits and safety of each modality. These summaries also included, for each modality examined, the percentage of patients who clinically improved in both control and treatment groups, estimated effect size, number needed to treat, number needed to harm, and a complete quality assessment of the evidence. In addition, the TEP received the final set of clinical scenarios; instructions on the use of the GRADE process (10-13); the ACR White Paper on the use of nonsteroidal antiinflammatory drugs (NSAIDs) (18); AHRQ reports on the use of nonopioid analgesics for OA (19) and treatment of primary and secondary knee OA (20); the EULAR recommendations for the management of OA of the hand, hip, and knee (4-6); the AAOS recommendations for the management of knee OA (8); the OARSI recommendations for the management of OA of the hip and knee (7); the American Heart Association Scientific Statement on the use of NSAIDs (21); and the American College of Cardiology Foundation consensus recommendations on reducing the risk of gastrointestinal (GI) adverse events in patients using antiplatelet and NSAID therapy (22). The TEP also considered drug-specific indications, contraindications, and warnings from product information labels from the US FDA. Next, using a 5-point Likert scale, panelists were asked to use the evidence reports to make a recommendation for each pharmacologic modality as applied to each clinical scenario; nonpharmacologic modalities were evaluated for the base cases of hand, hip, and knee OA, as well as hand OA with involvement of the trapeziometacarpal joint. The scale provided to panelists included the following choices: strong recommendation to use, weak (or conditional) recommendation to use, no recommendation, weak (or conditional) recommendation not to use, and strong recommendation not to use. The strength of a recommendation reflects the quality of the evidence supporting the use of the modality as well as the extent to which one can be confident that desirable effects (i.e., benefits) of an intervention outweigh undesirable effects (i.e., harms). Strong recommendations mean that most informed patients would choose the recommended management and that clinicians can structure their interactions with patients accordingly. Conditional recommendations mean that the majority of informed patients would choose the recommended management but many would not, so clinicians must ensure that patients' care is in keeping with their values and preferences. Based on this initial TEP member feedback, an initial set of recommendations was drafted. Initial voting was done privately with votes submitted electronically using Excel spreadsheets (Microsoft). The TEP met in person in December 2008 to complete the pharmacologic recommendations for hand, hip, and knee OA and the nonpharmacologic recommendations for hand OA. If consensus was not achieved with private voting, further discussion of the modality was conducted at the meeting with open voting and group discussion until consensus was achieved. Subsequently, the TEP met by conference call in September 2009 to complete the nonpharmacologic recommendations for hip and knee OA. Prior to the conference call, the TEP members were provided with summary data on the