Budgetary Impact Analysis of Denosumab in a US Health Plan
Published Online: October 03, 2013
Anju Parthan, PhD; Nicholas P. Emptage, MAE; Douglas C. A. Taylor, MBA; Hema N. Viswanathan, PhD; Nicole Yurgin, PhD; Bradley Stolshek, PharmD; Karen M. Clements, ScD; Charles Y. Tao, MA; and Milton C. Weinstein, PhD
In the United States, 26% of women 65 years and older and 50% of women 85 years and older are affected by postmenopausal osteoporosis; moreover, 1 out of 2 women aged more than 50 years will experience a fracture in her lifetime.1 Over 1.5 million fractures per year are attributable to osteoporosis; these fractures result in 500,000 hospitalizations, 800,000 emergency department visits, 2.6 million physician visits, 180,000 nursing home placements, and $12 billion to $18 billion in direct healthcare costs in the United States each year.2 In addition to cost, fractures also result in severe functional disability and may negatively affect a patient’s activities of daily living and psychological status.2,3
Several pharmacologic agents are available to treat osteoporosis. The most common are bisphosphonates including alendronate, risedronate, ibandronate, and zoledronic acid. Other treatments include teriparatide, a form of parathyroid hormone, and raloxifene, a selective estrogen receptor modulator. These treatments have been shown to be effective in reducing the risk of fractures.4 However, many of these treatments are associated with high rates of side effects,5 and patients frequently complain about the inconvenience of administration.6 Rates of discontinuation are high for many of these treatments, and effi cacy may be reduced due to lack of adherence to prescribed regimens.5 One study estimated that as many as 50% of patients discontinue osteoporosis therapy within a year of initiation.6
Denosumab is a fully human monoclonal antibody, administered once every 6 months as a subcutaneous injection, which specifi cally targets receptor activator of nuclear factor kappa-B ligand (RANKL), a key mediator of the resorptive phase of bone remodeling. Denosumab was approved in June 2010 by the US Food and Drug Administration for the treatment of osteoporosis in postmenopausal women at high risk for fracture, defi ned as a history of osteoporotic fracture, multiple risk factors for fracture, or failure of or intolerance to other available osteoporosis therapy.7 The objective of this study was to estimate the 3-year budgetary impact of adding denosumab to the formulary of a health insurance payer in the United States as a treatment option for women with osteoporosis at high risk for fracture.
We assessed the budgetary impact of denosumab by comparing total healthcare costs in a scenario without denosumab (in which denosumab was not introduced) with costs in a scenario with denosumab (in which denosumab was introduced). The analyses were performed from the perspective of a hypothetical US health insurance plan, whose enrollees include both patients with traditional commercial insurance benefi ts (ie, those younger than 65 years) and those with Medicare Advantagebenefits (ie, those 65 years and older). A 3-year time horizon was used to refl ect the budgetary period of interest to payers.
Decision analysis methods were used to construct and estimate a Markov model that used six 6-month cycles to characterize 3 years of osteoporosis treatment in the identified population.
The model structure, shown in Figure 1, was previously used in an analysis by Jönsson and colleagues8 of the cost-effectiveness of denosumab for the treatment of postmenopausal osteoporosis. A similar structure was also used in other studies that evaluated the cost-effectiveness of bisphosphonates and strontium ranelate in postmenopausal osteoporosis.9-11 The health states in our Markov model are also aligned with the health states in the reference model for osteoporosis as described in a review paper by Zethraeus and colleagues12; these authors recommend using a reference model for costeffectiveness analyses in osteoporosis in order to ensure the comparability of the analyses and to serve as a toolf or cross-validation of present and future cost-effectiveness models.
The Markov model consists of 7 health states (Figure 1). The assumptions regarding transitions between the health states were adapted from the study by Jönsson and colleagues.8 All patients began in the well health state. In each cycle, a patient might sustain a hip, vertebral, or other fracture; remain healthy; or die. After 6 months in any fracture state, the patient had a risk of sustaining a new fracture or dying. After 1 year in a given fracture state, the patients could (1) sustain a new fracture; (2) move to the postfracture state (either after hip fracture or after vertebral fracture, depending on the previous health state); (3) move back to the well state (“other” fracture patients only); or (4) die. The patient remained in the postfracture state (shown as a circular arrow) if she did not die or sustain a new fracture. Following the hierarchical structure of previously published models referenced above, the model tracked the history of either prior hip fracture or vertebral fracture, but not both. Once a patient experienced a hip fracture, the patient remained in the post–hip fracture state for life; however, patients in the post–vertebral fracture state who subsequently experienced a hip fracture moved to the post–hip fracture state. Patients in the other-fracture state could, in subsequent cycles, experience a hip fracture, a vertebral fracture, the same health or become well and transition to that specific state.
The model used age-specific background fracture incidence13- 15 rates in 1-year age intervals for hip, vertebral, and other fractures, and for the risk of going to a nursing home following a hip fracture.14,16 The 3 main components of this budgetary impact model were the population component, the product utilization component, and the economic component, described below in detail.
The target population of the analysis (Figure 2 and Table 117-22) included postmenopausal, osteoporotic women 50 years and older who were at high risk for fracture. In order to calculate the size of the population of interest, we used US Census data to estimate the proportions of plan members aged 50 to 64, aged 65 and older, and the proportion of females within these age groups.17 The proportion of women currently treated for osteoporosis was estimated from an analysis of Thomson Reuters MarketScan Commercial and Medicare Supplemental Databases.18 This data source was considered appropriate for the study because it had a good representation of both commercially and Medicare-insured patients (about 56% of patients were
PDF is available on the last page.