Dax T. Varkey, Brendan M. Patterson, R. Alexander Creighton, Jeffrey T. Spang, Ganesh V. Kamath

Rotator cuff tears have a lifetime incidence between 25% and 40% in the United States, but optimum treatment strategies and protocol have not yet been widely accepted. This study evaluated the proportions of patients treated with nonoperative and operative modalities and how this proportion has changed during an 8-year period (2005-2012) among patients with Medicare. Using the PearlDiver patient record database, we identified Medicare patients having been diagnosed with a rotator cuff tear. These patients were then stratified on the basis of treatment with physical therapy, subacromial/glenohumeral injection, or rotator cuff repair. We analyzed the data in regard to standard demographic information, comorbidities, and the Charlson Comorbidity Index. During the study period, 878,049 patients were identified and 397,116 patients had rotator cuff repair. The proportion of patients treated initially with physical therapy dropped from 30.0% in 2005 to 13.2% in 2012, and the subacromial/glenohumeral injection proportion decreased from 6.00% to 4.19% (P < .001). The proportion of patients who had rotator cuff repair increased from 33.8% to 40.4% from 2005 to 2012 (P < .001). Charlson Comorbidity Indexes were significantly lower in operative patients compared with each nonoperative treatment examined. This analysis demonstrates a significant decrease in the initial trial of nonoperative treatment and an increase in the rate of surgery. Patients undergoing rotator cuff repair had fewer comorbidities than those undergoing nonoperative treatments. It also demonstrates that patients who had a trial of injection had a higher incidence of eventual rotator cuff repair compared with the patients with an initial trial of physical therapy.