|
The effect of alendronate on fracture-related healthcare utilization and costs: the fracture intervention trial |
Chrischilles E A, Dasbach E J, Rubenstein L M, Cook J R, Tabor H K, Black D M |
|
|
Record Status 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. Health technology The use of alendronate in the prevention of osteoporotic fractures.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised postmenopausal women with low bone mineral density (BMD), and pre-existing vertebral fracture, who were living in one of 11 metropolitan areas of the United States. Patients with major medical conditions or secondary osteoporosis were excluded.
Setting The setting was the community. The economic study was carried out in the USA.
Dates to which data relate The date when the effectiveness analysis was carried out was not stated. Full details of the study were published elsewhere (see Other Publications of Related Interest). The costs were reported in 1997 US dollars.
Source of effectiveness data The effectiveness data were derived from a single study.
Link between effectiveness and cost data The costing was undertaken retrospectively on the same patient sample as that used in the effectiveness analysis.
Study sample No power calculations or sample selection methods were reported. However, it should be noted that full details of the effectiveness study were published elsewhere (see Other Publications of Related Interest). The study sample appears to have been appropriate for the study question. The study enrolled 2,027 women with a history of vertebral fracture. Of these, 1,022 were randomised to the alendronate group and 1,005 to the placebo group.
Study design This was a multi-centre, double-blind, randomised controlled trial. The patients were randomised to receive either 5 mg alendronate for 2 years followed by 10 mg alendronate for 1 year, or 3 years of placebo. The patients were assessed at 3, 6, 12, 18, 24, 30 and 36 months. Full details of the study were published elsewhere (see Other Publications of Related Interest).
Analysis of effectiveness All the patients included in the study were accounted for in the effectiveness analysis. The outcome measure was a fracture event at the hip, wrist, vertebral column or 'other'. The two groups were broadly comparable at baseline.
Effectiveness results Over the 3 years of the study 13.6% (139) of the alendronate patients experienced any fracture, compared with 18.2% (183) of the placebo patient.
Of the 139 alendronate patients who experienced any fracture, 83 (59.7%) required fracture-related health services. Of the 183 placebo patients who experienced any fracture, 109 (59.6%) required fracture-related health services.
1.1% of the alendronate patients and 2.2% of the placebo patients experienced hip fractures. Of these, 100% (alendronate) and 95.5% (placebo), respectively, required health service use.
2.3% of the alendronate patients and 5.0% of the placebo patients experienced a vertebral fracture. Of these, 34.8% (alendronate) and 28.0% (placebo), respectively, required health service use.
2.2% of the alendronate patients and 4.1% of the placebo patients experienced a wrist fracture. Of these, 59.1% (alendronate) and 63.4% (placebo), respectively, required health service use.
10.4% of the alendronate patients and 10.7% of the placebo patients experienced an 'other' fracture. Of these, 42.5% (alendronate) and 55.6% (placebo), respectively, required health service use.
Clinical conclusions The authors concluded that alendronate reduces the incidence of fractures in women with low BMD and a history of vertebral fracture.
Measure of benefits used in the economic analysis No summary health benefit was used in this analysis. The study is therefore classified as a cost-consequences analysis.
Direct costs The authors showed an equivalent cost per patient for fractures averaged across the entire cohort. They did not show the total or unit costs separately. The costs included were relevant to a health service or health-care payer. The costs were calculated using diagnostic and resource codes (for example, ICD-9-CM and DRG). National fee schedules were then applied to these codes. The schedules used were the 1997 Medicare Fee Schedule, and the 1997 Medicare prospective payment system payment schedule. The authors also used a number of HCFA Financing Review Statistical Supplements from 1996 and 1997.
It was assumed that the costs were incurred over the 3-year period of the study, yet discounting was not employed. The price year was 1997.
Statistical analysis of costs Chi-squared tests were used to test for differences in the proportion of patients using hospital resources. The confidence intervals were estimated using resampling techniques. The per-patient relative risk of medical resource use was calculated using Mantel-Haenszel methods. Missing resource use data were imputed by resampling from pools of available cost data.
Sensitivity analysis No sensitivity analysis was carried out.
Estimated benefits used in the economic analysis See the 'Effectiveness Results' section.
Cost results The total cost of any fracture, averaged over the entire cohort, was $356 per alendronate patient and $547 per placebo patient, (not significant, p=0.114).
When analysed by fracture type, the only significant cost difference was between hip fractures, $132 per alendronate patient versus $312 per placebo patient, (p=0.036).
Synthesis of costs and benefits Authors' conclusions Alendronate significantly reduces the proportion of patients experiencing fractures (139 out of 1022 versus 183 out of 1005), and hence fracture-related health care, by 25%. The cost-saving for all fractures is $190 per patient treated, (p=0.114). Alendronate significantly reduces the cost associated with hip-fracture related care by $181 per patient treated, (p=0.036).
CRD COMMENTARY - Selection of comparators The authors chose placebo as the comparator in this study. This allows the relative value of the treatment to be evaluated compared with no treatment. Care must be taken in interpreting potential cost-savings based on placebo controls. You must decide whether 'no treatment' represents current practice in your own setting.
Validity of estimate of measure of effectiveness The analysis used a multi-centre, double-blind, randomised controlled trial. This was appropriate for the study question.
The large study sample was representative of the study population, and the patient groups were shown to be comparable at analysis. The analysis of effectiveness appears to have been handled credibly: 14% of fracture events had missing data, but the authors stated that there was no significant difference in the percentage of incomplete data between alendronate and placebo patients.
Validity of estimate of measure of benefit The authors did not derive a summary measure of health benefit. Therefore, this was classified as a cost-consequences study.
Validity of estimate of costs Most of the relevant categories of costs from the perspective of a health-care payer were included, with the exception of 'office visits' (general practitioner visits) and 'free-standing urgent care centres'. The authors also excluded the cost of the drug itself. This will have underestimated the cost of alendronate patients, and may alter the financial conclusions of the study.
The authors used fee schedules based on diagnostic codes for hospital treatment costs. Unfortunately, they did not show which codes were applicable to which patients, nor did they show the unit costs. This limits the transparency and generalisability of the results to other settings.
The quantities of emergency room, hospital stays and nursing home/rehabilitation use were reported separately. Statistical analyses comparing alendronate with placebo were performed. Although the costs will have been incurred over a 3-year period, discounting was not undertaken. All the costs were reported in 1997 US dollars.
Other issues The authors made appropriate comparisons of their results with other studies. The issue of generalisability to other settings was addressed through comments on the use of one of the fee schedules (the Iowa State Ambulatory Patient Groups weights) being adopted throughout the United States. The authors did not appear to present their results selectively. However, the transparency of the costs section could have been enhanced by tabulating the unit costs and quantities incurred in the placebo and alendronate arms. The conclusions reflected the scope of the analysis, although it should be noted that, whilst alendronate does indeed reduce the incidence of all fractures by 25%, of those who do suffer a fracture there is no significant difference in the proportion requiring medical care. The implication is that there is a reduction in overall fracture risk, but not of severity.
The authors reported some limitations to their study. The first was the use of fee schedules. This has the advantage of being more generalisable to other settings. However, although the authors state that the use of Medicare schedules would be close to the true cost in elderly patients, it is likely to underestimate the cost for younger patients not covered by Medicare (who are most likely to suffer vertebral and wrist fractures rather than hip). The authors may also have underestimated the cost impact of alendronate by excluding 'office visits' (general practitioner visits) and 'free-standing urgent care centres'. The second limitation was the lack of power to detect significant differences in the cost. This is a common drawback from 'piggy-back studies', where the original trial is powered to detect clinical end points only. Cost differences usually require much larger sample sizes to detect significant differences.
The original trial suggested a possible reduction in non-fracture-related hospitalisations for patients taking alendronate. This could be due to fracture patients being more susceptible to further injury at a later stage. Thus, the potential cost-savings may be greater than indicated here. The study also excluded the cost of future fractures prevented. A modelling study would be required to establish the true long-term cost implications of alendronate.
Implications of the study A more complete economic evaluation, including the modelling of future fractures avoided, would be required to confirm the long-term potential cost-savings of alendronate.
Source of funding Supported by Merck Research Laboratories.
Bibliographic details Chrischilles E A, Dasbach E J, Rubenstein L M, Cook J R, Tabor H K, Black D M. The effect of alendronate on fracture-related healthcare utilization and costs: the fracture intervention trial. Osteoporosis International 2001; 12(8): 654-660 Other publications of related interest Black DM, Cummings SR, Karpf DB, et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fracture. Lancet 1996;348:1535-41.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Aged, 80 and over; Alendronate /economics /therapeutic use; Chi-Square Distribution; Confidence Intervals; Cost Savings; Double-Blind Method; Female; Fractures, Bone /economics /etiology /therapy; Health Care Costs; Humans; Middle Aged; Osteoporosis, Postmenopausal /drug therapy /economics; Patient Admission /economics /statistics & numerical data AccessionNumber 22001001830 Date bibliographic record published 31/08/2003 Date abstract record published 31/08/2003 |
|
|
|