This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

The use of tibolone (Livial, 2.5 mg/day) in the prevention of osteoporotic fractures in postmenopausal women. The two alternative modalities considered were:

commencing treatment at the age of 53 years and continuing it for 5 years;

commencing treatment at the age of 65 years and continuing it for 5 years.

Primary prevention.

Cost-effectiveness and cost-utility analyses.

The study population comprised hypothetical cohorts of women, initially free of osteoporosis-related bone fractures. Four hypothetical cohorts of women were considered, according to the severity of loss in bone mass and age when commencing the treatment:

osteopenic, 53-year-old women;

osteoporotic, 53-year-old women;

osteopenic, 65-year-old women; and

osteoporotic, 65-year-old women.

Osteoporosis was defined as a bone mineral density (BMD) of more than 2.5 standard deviations below the average peak BMD seen in adults. Osteopenia was defined as values of BMD between 1 and 2.5 standard deviations below the average peak level.

The setting was secondary care. The study was performed in Sweden.

The effectiveness evidence was obtained from studies published between 1986 and 1999. The cost data were obtained from studies published between 1996 and 1998. All the costs were converted to 1998 prices, except for the price of tibolone, which was given for 1999.

The effectiveness data were derived from a review of published studies.

A Markov model was used to estimate the long-term costs and health benefits of tibolone therapy. A cycle length of 1 year was used and a time horizon of 25 years was considered. Three types of fractures (hip, vertebral and forearm) were considered at analysis. These fractures were assumed to depend on age and BMD.

The following outcomes were assessed in the review:

the percentages of patients with a hip fracture that healed successfully, that regained only near-normal function, and that remained permanently disabled;

the increase in the mortality risk for the first year following hip fractures for women aged 65 to 74, 75 to 84, and 85 years or older; and

the percentage variation in BMD with time for osteoporotic and osteopenic women commencing treatment with tibolone at age 53 and at age 65.

The utility values for the various health states were also reported. There were other outcomes (e.g. transition probabilities and risk of fractures) that appear to have been used as key parameters in the Markov model, although the authors did not report the results.

Randomised controlled trials (RCTs), double-blind placebo-controlled RCTs, long-term follow-up studies, and Swedish statistical data studies were included in the review.

Not reported.

Not reported.

Not reported.

At least 17 studies were included in the review.

The model parameter estimates were selectively taken from the literature.

The authors stated that the utility values used in the study were similar to those used in other studies. No further differences between the primary studies appear to have been investigated.

Among patients experiencing hip fractures, 50% of them healed successfully, 40% of them regained only near-normal function, and 10% of them remained permanently disabled.

For the cohort of women commencing treatment with tibolone at age 53, the BMD increased by 3.5% over the first 2 years, and remained at that level for another 3 years. For the cohort of women commencing treatment with tibolone at age 65, the BMD increased by 5% over the first two years, and remained at that level for another 3 years.

The increase in mortality rates due to hip fractures was 10% for women aged 65 to 74 years, 20% for women aged 75 to 84 years, and 50% for women aged 85 years or older.

These results, as well as other outcomes that were not reported in the article, were included as key parameters in the Markov model.

The authors made assumptions to derive estimates of effectiveness.

The authors made the following assumptions:

vertebral and forearm fractures required one year of recovery, returning after that time to the pre-fracture health state;

women younger than 65 years did not face an excess mortality risk due to hip fractures;

compliance was assumed to be 100%;

the annual rate of decline of tibolone effectiveness after treatment discontinuation was assumed to be 1% for both osteoporotic and osteopenic women.

The summary measures of health benefit used in the economic analysis were the number of quality-adjusted life-years (QALYs) gained, the number of fractures avoided, and the number of hip fractures avoided with tibolone when compared to the alternative of "doing nothing". The quality of life utility weights for the various health states were taken from Jonsson et al. (see Other Publications of Related Interest). The authors also reported the number of vertebral fractures avoided and the number of forearm fractures avoided with tibolone when compared to "doing nothing". These health benefits were obtained from the Markov model, considering a 25-year period and a cohort of 1,000 women. The QALYs were discounted at a rate of 3%.

The direct costs considered in the analysis were those of the Swedish National Health System. Some of the cost categories included were consultations with primary care physicians, extra home aid, nursing home assistance, hospital emergency room visits, inpatient admission costs, hospitalisations, X-ray and physiotherapy.

The direct cost data were obtained from published studies, therefore, the costs were estimated from actual data. The Markov model and linear interpolation (of some costs) were used to extrapolate the costs to a 25-year timeframe. The average annual costs for the different health states and the degree of recovery (e.g. healthy; with hip, vertebral or forearm fractures; full recovery, near-normal recovery or remaining severely handicapped) were reported, instead of the costs related to the different cost categories.

Discounting was performed using a rate of 3%. The resource quantities and the costs were not reported separately. The Swedish consumer price index was used to convert the costs to 1998 prices. Therefore, the price year was 1998, except for tibolone, whose price was given for 1999.

No statistical analyses of the costs were reported.

No indirect costs were considered in the economic analysis.

Swedish kronor (SEK).

One-way and two-way sensitivity analyses were performed to evaluate the robustness of the results. The parameters varied were tibolone efficacy for osteoporotic women commencing treatment at age 53 (3, 3.5 and 4%) and age 65 (4.5, 5 and 5.5%), the rate of BMD decline after treatment discontinuation (range: 0.5 - 1.5%), patient compliance (compliance rate equal to 75%), unit treatment costs, the discount rate applied (5%), and the duration of treatment (10 years). Therefore, the area of uncertainty investigated was variability in data.

The numbers of events per 1,000 osteopenic women, commencing treatment at age 53 and considering a 25-year period, were as follows:

131 hip fractures with "no treatment" versus 92 with tibolone (i.e. 39 hip fractures avoided with tibolone);

581 combined fractures with "no treatment" versus 494 with tibolone (i.e. 87 fractures avoided with tibolone);

16.47 discounted QALYs with "no treatment" versus 16.50 with tibolone (incremental number of QALYs with tibolone would be 0.03).

The numbers of events per 1,000 osteoporotic women, commencing treatment at age 53 and considering a 25 year-period, were as follows:

195 hip fractures with "no treatment" versus 152 with tibolone (i.e. 42 hip fractures avoided with tibolone);

749 combined fractures with "no treatment" versus 655 with tibolone (i.e. 95 fractures avoided with tibolone);

16.42 discounted QALYs with "no treatment" versus 16.46 with tibolone (incremental number of QALYs with tibolone would be 0.04).

The authors did not report separately the number of fractures per 1,000 women for the cohort of osteoporotic and osteopenic women initiating treatment at age 65. The numbers of QALYs obtained by osteoporotic women starting treatment at age 65 were 13.59 with "no treatment" versus 13.62 with tibolone. A 3% discount rate was used to discount the QALYs.

For 53-year-old osteopenic women, the total treatment costs were SEK31,335 with "no treatment" versus SEK32,837 with tibolone. Therefore, the incremental cost of tibolone when compared with "no treatment" was SEK1,502.

For 53-year-old osteoporotic women, the total treatment costs were SEK48,399 with "no treatment" versus SEK48,284 with tibolone. The net saving for tibolone when compared with "no treatment" was SEK115.

For 65-year-old osteoporotic women, the total treatment costs were SEK79,148 with "no treatment" versus SEKk81,661 with tibolone.

The total treatment costs for 65-year-old osteopenic women were not reported.

The costs and benefits were combined using incremental cost-effectiveness ratios (ICERs) that measured the net cost of switching from no intervention to tibolone treatment relative to the net improvement in QALYs. Other ratios, such as the incremental costs per fracture avoided and per hip fracture avoided, were also reported where appropriate. For the cohort of osteopenic women starting treatment with tibolone at age 53, the ICER was SEK49,408 per QALY gained when compared with "no treatment". Tibolone was a dominant strategy when compared with "no treatment" for the cohort of osteoporotic women commencing treatment at age 53. For the cohort of women starting treatment with tibolone at age 65, the ICER was SEK94,593 per QALY gained if they were osteopenic, and SEK3,074 per QALY gained if they were osteoporotic, when compared with "no treatment".

The most sensitive key parameters appeared to be the rate of BMD decline following treatment discontinuation, the cost of treating hip fractures, the duration of treatment and the discount rate. The only parameter that made the incremental cost per QALY gained for both osteoporotic and osteopenic women higher than SEK200,000 (the value commonly considered reasonable in Sweden) was tibolone treatment lasting for 10 years.

The results of the study suggested "the administration of tibolone to women with osteopenia and osteoporosis for the prevention of osteoporosis-related bone fractures is cost-effective in Sweden relative to a policy of no treatment, especially when treatment is initiated around the onset of menopause (around 53 years of age) and treatment is administered for a duration of 5 years".

The comparator chosen was a policy of no intervention, which allowed the active value of tibolone treatment to be evaluated. However, the authors mentioned that hormone replacement therapy (HRT) is commonly administered to postmenopausal women in order to protect against fractures, although its associated side effects mean that women may not comply with the treatment. Therefore, it would have been useful to have considered HRT as another alternative comparator. You must decide whether there is a commonly used health technology in your own setting for protecting postmenopausal women against fractures.

The authors did not state that a systematic review of the literature had been undertaken. The effectiveness estimates were selectively taken from literature that was most relevant to the Swedish setting. The authors reported that the mortality risks associated with hip fractures were obtained from studies that included women from the general population, and not only osteopenic or osteoporotic women. This approach may have overestimated the effectiveness of the treatment and may have led the true mortality in osteoporotic and osteopenic women to be underestimated, as acknowledged by the authors.

The authors made assumptions to derive estimates of effectiveness. Some of these assumptions appear to have been justified with reference to the medical literature, but not all because there was a lack of available data. The authors also reported that since some effectiveness parameters (e.g. the decline in the efficacy rates of tibolone after discontinuation of treatment) had not been estimated before, this may have introduced uncertainty into the reliability of the conclusions. A positive aspect of the study was that efforts were made to adjust the effectiveness data to the Swedish population of osteoporotic and osteopenic women.

The estimation of benefits was modelled. The authors reported that the Markov model used to estimate the benefits and costs was suitable for the analysis because osteoporosis presents a sequence of well-defined health states with clear transitions between them. Therefore, this model appears to have been appropriate to derive a measure of health benefit. As the authors stated, one of the measures used to combine the costs and benefits (i.e. the cost per hip fracture avoided) may have underestimated the benefits of tibolone therapy because it excludes the benefits of vertebral and forearm fractures avoided. They justified the use of this summary measure because it allowed the results to be compared with those from other studies. They also stated that some health benefits were not considered in the analysis, such as the relief of climacteric symptoms, which could justify treatment with tibolone for longer than 5 years.

It could not be inferred whether all the relevant costs for the Swedish health service perspective were included, as the authors did not detail the costs considered at analysis. They did, however, report the costs related to the different health states and the degree of recovery. The costs and the quantities were not reported separately. The cost data were obtained from published studies. Sensitivity analyses on some of the costs were performed. Discounting was performed at a rate of 3%, and a 5% rate was considered in the sensitivity analyses in order to increase comparability with other studies. This was appropriate since a 25-year-period was considered at analysis. The price year was stated.

The authors compared the incremental cost per QALY gained to the results of other studies by extending tibolone administration from 5 to 10 years. They stated that these results were similar. No further comparisons appear to have been made. The issue of the generalisability of the results was addressed in the sense that it was clear to whom the study results would be applicable. The results were adequately presented and the authors' conclusions reflected the scope of the analysis.

The authors did not make any recommendations nor highlight any need for further research. However, the study emphasised the lack of data about some effectiveness estimators related to the intervention, such as the decline of the efficacy rate of tibolone after discontinuation of treatment, and the mortality rates associated with hip fractures in osteopenic and osteoporotic women. For this reason there seems to be a need for further research.

Supported by Organon AB.

Cooper C. The crippling consequences of fractures and their impact on quality of life. JAMA 1997;103:12S-19S.

Hammar M, Christau S, Nathorst-Boos J, et al. A double-blind randomised trial comparing the effects of tibolone and the continuous combined hormone replacement therapy in postmenopausal women with menopausal symptoms. British Journal of Obstetrics and Gynaecology 1998;105:904-11.

Studd J, Arnala I, Kicovic P, et al. A randomised study of tibolone on bone mineral density in osteoporotic postmenopausal women with previous fracture. Obstetrics and Gynecology 1998;92:574-9.

World Health Organization. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. WHO Tech Rep Ser 843, 1994.

Whittington R, Faulds D. Hormone replacement therapy II. A pharmacoeconomic appraisal of its role in the prevention of postmenopausal osteoporosis and ischaemic heart disease. Pharmacoeconomics 1994;5:513-54.

Jonsson B, Christiansen C, Johnell O, et al. Cost-effectiveness of fracture prevention in established osteoporosis. Osteoporosis International 1995;5:136-42.