• 30
  • Postmenopausal Osteoporotic Women at High Risk of Fracture: METHODS

Patient Population

Our base case analysis consisted of a hypothetical cohort of 65-year-old-women with low bone mineral density (BMD) (i.e., BMD of 2.5 or more standard deviations [SD] below the young adult mean) and a prevalent vertebral fracture. We assume that 21% of 65-to 69- year-olds fall into this study population.

Model Overview

A fracture incidence-based model of the natural history of osteoporosis was developed to estimate the cost-effectiveness and budget impact outcomes for patients at high risk of fracture treated with risedronate and canadian alendronate therapies. A fracture incidence-based model was selected over a BMD-based model because this type of model uses fracture incidence rates to model fracture occurrence, as opposed to an intermediate clinical endpoint such as changes in BMD over time. The model was developed under the guidance of a committee of academic advisors who had independent control over all design and methodological decisions. In accordance with health economic guidelines, the model was transparent and underwent extensive validation. A modeling approach was chosen for this analysis because of the lack of head-to-head trials comparing the therapies under evaluation in this patient subgroup.

Details of the model design, structure, and assumptions have been described previously. In summary, it is a Markov state-transition model in which patients can move between several short- and long-term health outcome states over time (Figure 1). Long-term health states include Healthy, Healthy Post-Vertebral Fracture, Healthy Post-Hip Fracture, Healthy Post-2nd Hip Fracture, and Death. Short-term health states, which patients enter and leave within any given year, are Vertebral Fracture, Hip Fracture, and 2nd Hip Fracture. These short-term states are used to capture the acute-care costs and decrements in quality of life that are associated with fracture.

Figure 1. Allowable state transitions caused

Figure 1. Allowable state transitions caused by fractures (Fx) based on starting state.

The model permits movement between health states annually according to state-dependent transition probabilities (i.e., age-specific fracture incidence and mortality rates) that were derived from best available observational data. Treatment effects are modeled as a relative risk (RR) reduction in fracture rates, whereas a patient population at increased risk of fracture caused by the presence of a variety of risk factors can be modeled by direct modification of fracture rates via RRs. Outputs include frequency of fractures, health care costs, and quality-of-life impact for each treatment group (i.e., risedronate, al-endronate, no treatment). canadian cialis

The following sections describe the input data used in the model. For all model inputs, the best available data sources were consulted and consisted primarily of published values from clinical trials, economic studies, observational studies, and epidemiological databases. Where uncertainty existed or assumptions were required for specific data inputs, sensitivity analyses were conducted on those input values. The base case values and data sources are provided in Table 1.

Table 1 Parameter Values Used in the Base Case Analysis

Parameter Value
Epidemiological Data
Fracture incidence rates
(per 10,000 patients)


65-69 years 31


70-74 years 49


75-79 years 103


80-84 years 167


85 years and older 255


Relative risk in
target population




All ages 6.41


Health Outcomes Data Year of




Medical costs/fracture
Hip fracture $36,864|


Vertebral fracture* $l,880| $7l
Health utilities (weights)
General population aged 65 0.833|
Year of




Hip fracture 0.18


Vertebral fracture
(no previous hip fracture) 0.l6


Vertebral fracture
(previous hip fracture) 0.55


Therapy Data
Efficacy of untreated fracture (RR)**




Risedronate 0.40


Alendronate 0.49


Therapy cost/day*** Risedronate




Hip Fracture

Age-specific incidence rates for hip fracture were obtained from a study of the general female population in Rochester, Minnesota, which includes both patients with and without PMO. Our study focused on treatment of PMO patients at high risk of fracture; therefore, fracture rates were adjusted to reflect the increased risk of new fractures in patients with low BMD and a prevalent vertebral fracture.

The RR associated with a prevalent vertebral fracture was adjusted based on the prevalence of the risk factor in the general female population to estimate the RR in our target population compared to the general female population. For example, if the RR of hip fracture in a 65-year-old population with a previous vertebral fracture, compared to a population without a previous vertebral fracture, is 1.9 and the prevalence ofvertebral fracture in 65-year-olds is 14.3%, then the RR of hip fracture in patients with a previous vertebral fracture, compared to the general population, is 1.7 (= 1.9/[0.143 X 1.9 + 0.857 X 1] ).

The RR reported per SD decrease in BMD was also adjusted to account for the proportion of women in the age-specific general population with BMD at or below that of the target population. For example, with increasing age, the average BMD of the general population decreases, resulting in a greater proportion of the population with a T-score less than or equal to -2.5. As this occurs, the fracture rate increases, resulting in a smaller RR of fracture between the target population and the general population.

Following these adjustments, the RRs were combined multi-plicatively with the population fracture rates to approximate rates of hip and vertebral fracture in a population with these risk factors. The ten-year and lifetime risks of hip fracture in the target population were estimated to be 21.2% and 62.5%, compared with 3.6% and 18.3% for a 65-year-old in the general population. Ideally, the predicted lifetime risk for the target population would be compared to published risks for the same populations. However, there are limited data on lifetime risk in women at high risk of fracture. A report by the World Health Organization provided lifetime risks for a 50-year-old woman with various levels of BMD measured at the hip. In the portion of the population with the lowest BMD (i.e., 0.80 to 0.89, 0.70 to 0.79, 0.60 to 0.69 and <0.60 g/cm2) the lifetime risk of fracture was estimated to be between 40% and 51% (40%, 47%, 50%, and 51%, respectively). Calculation of T-scores from this source is not possible because the mean and SD for the BMD of young adults is not reported for the measurement instrument used and can vary substantially between machines.

Lifetime risks for 50-year-old populations with low BMD could not be predicted by the model because of a lack of data required to calculate the RR of hip fracture of 50-year-olds with low BMD. For these calculations, RR values and risk factor prevalence data were obtained from the Study of Osteoporotic Fracture database (Dennis Black, personal communication). Age-specific BMD data were taken from the National Health and Nutrition Examination Survey (NHANES III) study. Fracture rates were adjusted to account only for risk factors in the starting cohort; no further adjustments were made for fractures occurring during the model.

Efficacy values (RR of hip fracture) of 0.40 and 0.49 were used for patients treated with risedronate, respectively, and were obtained from randomized clinical trials with patient populations similar to the base case cohort. The adjusted age-specific fracture rates were multiplied by these values to estimate fracture rates in the treated cohorts. Excess mortality in the year of hip fracture was considered in the model.

The first-year costs following a hip fracture were assumed to be $36,864, and included expected costs of acute inpatient care (hospital facility, inpatient physician, emergency room, readmissions), long-term care (skilled nursing facility, disability-related care) and outpatient care (rehabilitation care, outpatient physician visits, home health care). The hip fracture cost estimates were derived from previous studies, updated to the year 2000, using the medical care component of the consumer price index. These included acute care hospital costs of $16,293 for women aged 65 to 74 ($18,131 updated to the year 2000).

Expected hip fracture costs in subsequent years were assumed to be $3,832 and included disability-related costs of $1,165 and long-term care costs in skilled nursing facilities of $2,667. The latter cost component was derived by multiplying the annual cost of nursing home care ($38,431) by the probability of a hip fracture patient requiring permanent nursing care (7%). viagra soft

Utility weights were applied to each health state to allow for the calculation of quality-adjusted life years (QALYs). Utilities reflect how quality of life in a health state is valued on a scale from 0 (death) to 1 (perfect health). The age-specific utility weight for 65- to 69-year-old women in the general population was 0.833. This utility value was reduced upon occurrence of a fracture by 0.18 in the year of a hip fracture and 0.09 in all subsequent years. For women with both hip fracture and vertebral fracture, the utility reduction was 0.55 in the year of the hip fracture and 0.09 in all subsequent years.

Online Pharmacy