No outside funding supported this research. Chan is supported by grants from the National Heart Lung and Blood Institute (1R01HL123980 and K23HL102224). Tran, Stockl, Lew, and Solow are employed by Optum. Kao and Caglar were employed by Optum when this study was conducted. Chan serves as an advisor and consultant to OptumRx but received no compensation for work on this manuscript. Stockl is also employed by the Journal of Managed Care & Specialty Pharmacy. Spertus reports personal fees from United Healthcare and grants from Lilly, outside of the submitted work. None of the authors have any other financial conflicts of interest to report. Tran and Chan supervised this study, had full access to all of the data in the study, and take responsibility for the integrity of the data and the accuracy of the data analysis. Study design and concept were contributed by Tran and Chan. Tran and Kao collected the data, with analysis and interpretation performed by all the authors. Statistical analysis was performed by Caglar and Kao, and Tran and Chan drafted the manuscript. All authors were involved in the critical revision of the manuscript.

No abstract

Etanercept+MTX was a cost-effective treatment strategy in the base-case scenario; however, the model was sensitive to parameter uncertainties and ACR response criteria. Although Bayesian methods were used to determine transition probabilities, future studies will need to focus on head-to-head comparisons of multiple TNF-α inhibitors to provide valid comparisons.

The authors have reported a number of corrections to their analysis following the Letter to the Editor by Diamantopoulos et al. [1] These corrections have not affected the conclusions of the article but have resulted in a number of changes to the cost and utility results presented throughout the text, figures and tables. In addition, a wider range of utility values were tested in the one-way sensitivity analyses as the corrections resulted in a wider variance than previously reported. Please refer to Bounthavong et al.'s reply [2] to Diamantopoulos et al. for a more comprehensive explanation of these errors and to the online corrected version for correct cost and utility results.

the models hip fracture was a specific outcome, 94% contained vertebral fractures, and 77% contained wrist/forearm fractures. Eleven models incorporate at least one extraskeletal effect on cost and survival (including breast cancer, coronary heart disease, venous thromboembolism, stroke, and colorectal cancer). Thirty-two (32) of the 48 publications (67%) assume 100% compliance or do not directly mention/ model compliance. The majority of the models take the approach that there was discontinuation and non-compliance in the clinical trials, and that the treatment efficacy rates sourced from the clinical trials are underestimated due to the use of an intention-to-treat paradigm. CONCLUSIONS: The current state of osteoporosis modeling favors a non-cohort Markov approach, with individualized, i.e., microsimulation methodology being increasingly utilized as extraskeletal effects are incorporated. Treatment compliance and extraskeletal effects are extremely important in modeling real-world scenarios, yet they are not incorporated into the majority of the published models. Modeled treatment effectiveness should be properly imputed to account for the intention-to-treat impact of RCT-reported values as well as the reduced benefits of treatment noncompliance.