The Authors Respond

DISCLOSURES
Funding for the original study referred to in this letter was received through the PhRMA Foundation Value Assessment Challenge Award. The authors have no conflicts of interest to declare.


■■ The Authors Respond
We appreciate the interest in our recent study and the opportunity to respond to the critiques by Hiligsmann, Reginster, and Silverman (henceforth HiRS). 1 Before we respond to comments raised by HiRS, we wanted to acknowledge that our paper was originally written as a full-length research paper for the PhRMA Foundation Award; however, it was submitted to JMCP as a Viewpoints article with 2 other papers that also won a PhRMA Foundation Award. As a result, some details of our discrete-event simulation (DES) model and sensitivity analysis were shortened but remained concise to accommodate the journal restrictions in terms of number of tables and figures, as well as total number of words allowed.
First, HiRS highlighted a lack of lifetime-horizon consideration in our DES model. It is important to note that our study examined monotherapy options for women with postmenopausal osteoporosis (PMO) based on the current treatment guidelines. 2,3 Within the treatment options, the longest duration of treatment was 5 years, recommended for oral bisphosphonate (alendronate). An assumed linear, gradual offset of fracture-reduction benefits (treatment effects) would occur over the subsequent 5 years after treatment discontinuation. [4][5][6] Therefore, it was reasonable, clinically and economically, to use a 10-year time horizon rather than a lifetime horizon where treatment effects were unrealistically overestimated beyond the possible time horizon. Furthermore, we carefully examined the economic evaluation study of sequential therapy with abaloparatide/alendronate (ABL/ALN) as an illustration of its advocacy of lifetime horizon. 7 In the study, HiRS made an exaggerated assumption where treatment effect of ABL value of patient-value modeling, we would recommend that policymakers interpret study conclusions with caution. In addition, we urge researchers to be more transparent regarding model assumptions and data and to conduct extensive univariate and probability sensitivity analyses. By providing recommendations and minimum criteria for an economic evaluation in osteoporosis and an osteoporosis-specific checklist for reporting, our recent guidelines could help to improve the reporting, quality, and reliability of economic evaluations. 2 remained constant (i.e., 100% fracture-reduction benefits) for an additional 5 years after discontinuation in order to show that sequential therapy was always cost-effective.
Second, we agree that sequential therapy, specifically teriparatide followed by alendronate, might be a possible option. Nevertheless, the current treatment guidelines did not recommend sequential therapy of ABL/ALN. Furthermore, no clinical evidence directly compares sequential therapy with monotherapy treatments recommended in the current treatment guidelines; thus, it is not appropriate to add sequential therapy of ABL/ALN suggested by HiRS as another treatment option.
Third, HiRS claimed that some of our model data were not reported, such as side effects, effect of medication adherence, and treatment effects after discontinuation. On the contrary, costs of most common side effects, including hypercalcemia, nausea, and gastrointestinal events, were accounted for and reported in Table 1. In addition, the real-world adherence rates of all recommended treatments, as well as the assumed fracture reduction benefits after discontinuation, were acknowledged and discussed under the subheading "Treatment Effects and Real-World Adherence Rates" of our paper.
Finally, while the main purpose of our Viewpoints article was to propose and promote an alternative patient-level modeling approach using DES, we also introduced in brief the DES model for the current treatment guidelines in women with PMO. It was unavoidable that some details of the model description and sensitivity analysis were not reported in the article.
In the end, we stand by our analysis. HiRS's comments, as well as exaggerated assumptions in their own referenced model, only strengthen the confidence in our results.