|
Should postmenopausal women with rheumatoid arthritis who are starting corticosteroid treatment be screened for osteoporosis? A cost-effectiveness analysis |
Solomon D H, Kuntz K 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 This paper compared three strategies, namely waiting, screening and treating, for the prevention of corticosteroid-induced osteoporosis in postmenopausal women with rheumatoid arthritis (RA): 1. "Watchful waiting", which involved screening no one and treating only after an osteoporotic fracture occurred. 2. "Screen and treat", which involved treating selectively on a bone mineral density (BMD) score of less than -1.0. 3. "Treat all", an empirical treatment which involved treating all women without BMD testing.
Economic study type Cost-effectiveness analysis.
Study population The model was restricted to 55-year old postmenopausal women with RA. Two clinical groups were evaluated: those taking oestrogen replacement therapy (ORT) and those not receiving ORT due to refusal or contraindications. The study assumed that the use of ORT was a feature of the cohort under observation rather than a decision considered in the model. All women were assumed to be taking at least 1,000 mg of elemental calcium and 400 IU of vitamin D daily. The ORT cohort took 0.625 mg of conjugated oestrogen, or its equivalent, daily. The treatments for osteoporosis were assumed to be 400 mg etidronate daily, for 2 weeks out of every 3 months, and 10 mg alendronate daily. All women were assumed to be 100% compliant with their medication.
Setting The setting was community and primary care. The economic study was carried out in the USA.
Dates to which data relate The effectiveness evidence and resource use data were gathered between 1984 and 1998. The unit cost year was 1998.
Source of effectiveness data The effectiveness data were derived from a review and/or synthesis of previous studies.
Modelling A Markov (state-transition) model was used.
Outcomes assessed in the review The authors stated that in the model that calculated quality-adjusted life-years (QALYs), they used published estimates of utilities for the following categories of patients: well, without a fracture; hip fracture; other fracture; nursing home. The model also included estimates for the effect of BMD on the femoral head and transition probabilities for: repeat fracture, hip fracture death, other fracture death, nursing home post-hip fracture, and nursing home post-other fracture.
Study designs and other criteria for inclusion in the review The authors did not report the study designs included in the review. The analysis was restricted to 55-year old postmenopausal women with RA and the authors reported that the medical literature related to populations similar to the cohorts simulated in the model.
Sources searched to identify primary studies The authors did not report the sources used to identify the primary studies used to provide estimates for the model.
Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Number of primary studies included Eighteen references were used to provide estimates for the variables used in the model.
Methods of combining primary studies The methods used to combine the primary studies were unclear
Investigation of differences between primary studies Results of the review The model parameters were based on the following data, expressed as mean (range: biased towards treatment - biased towards observation):
The effect, i.e. percentage change, on femoral BMD: 0.8 (range: -0.9 - +2.4) for corticosteroids, calcium and vitamin D; 1.6 (range: 0.001 - 4.5) for ORT ; 0.001 (range: 1.7 - 0.001) for calcitonin; 1.0 (range: 2.0 - 0.001) for etidronate; 4.0 (range: 4.6 - 3.4) for alendronate.
Markov states, transition probabilities: 2.00 (range: 1.75 - 2.25) for repeat fracture; 0.20 (range: 0.30 - 0.10) for hip fracture death; 0.036 (range: 0.039 - 0.033) for other fracture death; 0.65 (range: 0.89 - 0.41) for nursing home post-hip fracture; 0.05 (range: 0.08 - 0.02) for nursing home post-other fracture.
Outcomes, utilities: 0.90 (range: 0.95 - 0.83) for well, without fracture; 0.60 (range: 0.50 - 0.75) for hip fracture; 0.80 (range: 0.67 - 0.92) for other fracture; 0.40 (range: 0.30 - 0.50) for nursing home.
Methods used to derive estimates of effectiveness The authors made several assumptions.
Estimates of effectiveness and key assumptions The sensitivity and specificity of DXA scanning was assumed to be 100%. One hundred percent treatment compliance was assumed.
Measure of benefits used in the economic analysis The measure of benefits was QALYs. The model used age-adjusted time trade-off estimates for utilities from a population unaffected by arthritis, for patients with RA without fractures. The utilities for fracture states were measured using a time trade-off method in a non-RA group who had sustained fractures. The nursing home utilities were based on estimates for patients with severe disabilities after osteoporotic fractures.
Direct costs The direct medical costs included the cost of RA, fracture states and nursing home care. The costs and outcomes were discounted at 3% per year. The medication costs were calculated from a published source. The costs were adjusted for inflation and standardised to 1998 using the medical care component of the Consumer Price Index. The study reported average costs. The cost data included in the model were: $2,524 (range: 2,481 - 6,165) for RA; $100 (range: 75 - 125) for DXA scan; $452 (range: 339 - 565) for calcitonin; $243 (range: 182 - 304) for etidronate; $543 (range: 407 - 679) for alendronate; $36,988 (range: 46,235 - 27,741) for nursing home; $13,449 (range: 18,582 - 6,577) for hip fracture; $921 (range: 1,069 - 600) for other fracture.
Indirect Costs The indirect costs were not included in the model and the authors did not justify their decision to exclude them.
Currency US dollars ($). The conversion rates were not reported.
Sensitivity analysis A one-way sensitivity analysis was carried out using the following parameters: treatment threshold BMD (T) score criteria in the screen and treat strategy (BMD score range: -1.0 - -4.0); range of cost and efficacy estimates for alendronate and etidronate; fracture rates; annual rate of BMD loss (range: 0.4% - 2.0%); percentage of fractures attributed to hip versus other sites; utility values for RA without fractures and for the post-fracture states; probability of death or permanent disability after fracture; and the discount rate (range: 0% - 5%).
Estimated benefits used in the economic analysis The benefits for each treatment were estimated for the lifetime of the population cohort: 15.661 QALYs for watchful waiting with etidronate; 15.668 QALYs for watchful waiting with alendronate; 15.741 QALYs for screen and treat with alendronate at T scores less than -1.0; and 15.745 QALYs for treat all with alendronate.
Cost results The estimated costs for each treatment, for the lifetime of the population cohort, were: $52,900 for watchful waiting with etidronate; $53,300 for watchful waiting with alendronate; $60,100 for screen and treat with alendronate at T scores less than -1.0; and $61,000 for treat all with alendronate.
Synthesis of costs and benefits The costs per QALY were reported as an incremental change in cost per change in QALYs, with each technology being compared to the next most expensive. They were: $66,100 per QALY for watchful waiting with alendronate; $92,600 per QALY for screen and treat with alendronate at T scores less than -1.0; and $224,300 per QALY for treat all with alendronate.
The strategies of screen and treat with etidronate, and treat all with etidronate were eliminated by dominance, i.e. they exhibited higher costs and lower effectiveness than the next most expensive technology. The costs and QALYs were discounted at 3% per year.
The sensitivity analysis revealed a lowest incremental cost per QALY of $24,300 for watchful waiting with etidronate versus alendronate, and $17,600 for watchful waiting with alendronate versus screen and treat with alendronate, when there was a bias toward treatment. The highest values were $172,100 and $426,100 respectively, when there was a bias toward watchful waiting.
Authors' conclusions The authors concluded that, compared with watchful waiting, the incremental cost-effectiveness ratio of a strategy of screening and treating postmenopausal female patients with RA with BMD T scores of less than -1.0 was greater than that of other well-accepted medical interventions. The cost-effectiveness ratios were more acceptable when a T score threshold treatment of less than -2.5 was used. The sensitivity analysis suggested that these estimates were significantly affected by the cost and efficacy of osteoporosis treatment, the fracture rates, and the proportion of fractures assumed to be non femoral.
CRD COMMENTARY - Selection of comparators The selection of comparators was not supported by published evidence. You, as a user of the database, should decide whether these comparators reflect the options available in current NHS clinical practice before the results can be generalised to the UK setting.
Validity of estimate of effectiveness:
This paper presented a Markov model to estimate the cost-effectiveness of a screen and treat strategy recommended by the ACR for the prevention of corticosteroid-induced osteoporosis. The paper was well written and clearly outlined the approach taken to develop the Markov model. A major limitation of this study was that the authors provided no details on how they conducted the review of the literature, and whether they took a systematic approach to identifying primary data sources. Furthermore, they did not explain what type of studies were used as primary data sources and how, if at all, they reviewed the quality of these studies. Thus, it is unclear whether the most appropriate studies were used to provide estimates to build the Markov model, and whether these were good quality studies.
Validity of estimate of benefits:
The authors used the terms QALYs and utilities interchangeably. The model used utility estimates, generated using the time-trade off approach, to calculate QALYs. The use of QALYs allows comparison of the results with other studies that have reported QALYs. The utility estimates were not directly available for all the possible outcomes used in the model. The authors state that further research is necessary to determine utilities for all the relevant outcomes for patients with RA.
Validity of estimate of costs Only direct medical costs were included in this model, which assumed a societal perspective. The authors did not explain why indirect costs or the costs of informal and social care, which would have been consistent with this perspective, were not incorporated into the model. The key variables identified from the sensitivity analysis suggest that more accurate information is required to improve the validity of the results generated by the model. However, the sensitivity analysis was only used to test the robustness of the data estimates. The authors do not appear to have tested the model structure or assumptions, for example, they assumed 100% compliance with treatment, which may be unrealistic. In addition, there is an implicit assumption that the overall health state of the women and treatment options is constant over time. Given the long time period from 55 years of age to death, it is likely that these factors would change. Finally, the authors did not specify if, and how, the model structure was validated to ensure it was a realistic representation of reality.
Other issues The issue of generalisability was addressed through a sensitivity analysis and in the discussion, and comparison was made with the results of other studies. Also, the authors cautioned that the model is only as good as the information used to complete it, and some estimates required were either completely missing or were inappropriate to the population cohort in question.
Implications of the study The authors state that the screen and treat strategy would be considered "moderately cost-effective" on the basis of its cost-effectiveness ratios, according to published guidelines (see Other Publications of Related Interest). The authors cautioned that the results from this model are limited by the lack of data on fracture and treatment efficacy in corticosteroid-treated patients.
Source of funding Supported in part by NIH, grant number AR-36308; Arthritis Foundation Investigator's Award.
Bibliographic details Solomon D H, Kuntz K M. Should postmenopausal women with rheumatoid arthritis who are starting corticosteroid treatment be screened for osteoporosis? A cost-effectiveness analysis. Arthritis and Rheumatism 2000; 43(9): 1967-1975 Other publications of related interest Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. CMAJ 1992;146:473-81.
Indexing Status Subject indexing assigned by NLM MeSH Arthritis, Rheumatoid /complications; Cost-Benefit Analysis; Estrogen Replacement Therapy /economics; Etidronic Acid /therapeutic use; Female; Humans; Mass Screening; Middle Aged; Osteoporosis, Postmenopausal /complications /diagnosis AccessionNumber 22000001585 Date bibliographic record published 28/02/2002 Date abstract record published 28/02/2002 |
|
|
|