|
Cost-effectiveness of hormone replacement therapy for fracture prevention in young postmenopausal women: an economic analysis based on a prospective cohort study |
Fleurence R, Torgerson D J, Reid 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 This study looked at the use of hormone replacement therapy (HRT) for preventing fractures in young postmenopausal women.
Economic study type Cost-effectiveness analysis.
Study population The population comprised women aged between 45 and 54 years who lived within 20 km of Aberdeen (Scotland), who had had their bone mineral density (BMD) measured at the Osteoporosis Research Unit between November 1990 and November 1994.
Setting The setting was the Osteoporosis Research Unit. The economic study was carried out in Scotland, UK.
Dates to which data relate The effectiveness and resource use data were collected between January 1997 and July 1999, and referred to the period from November 1990 to July 1999. With the exception of HRT (year 2000), the costs for all resources related to 1997.
Source of effectiveness data The effectiveness evidence was 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 study.
Study sample The sample size was not determined in the planning phase of the study to assure a certain power. In addition, power calculations were not performed retrospectively. The sample was selected by inviting patients for a repeat BMD measurement. The invitations were issued, on average, 6.17 (standard deviation, SD=0.81) years after the original measurement. Seventy-one per cent of the study population (3,646 out of 5,119) re-attended for a second measurement. The initial sample appears to have been appropriate for the study question. Those patients who were using HRT at the time of their first BMD measurement were excluded from the study. The authors felt that including these patients in the statistical analysis would underestimate the effect of HRT on fracture incidence because the efficacy of HRT on fractures is, at least in part, due to preserving bone mass. This meant that the group was made up of 732 patients who were current HRT users and 1,141 patients who had never received HRT.
Study design This was a cohort study that was carried out in a single centre. Information on the women's HRT use and self-reported fracture history were collected retrospectively for, on average, 6.2 years. Due to the study design there was no loss to follow-up.
Analysis of effectiveness All the patients included in the study were accounted for in the analysis. The primary health outcome was the number of fractures. The authors reported that they used odds ratios as the measure of association between HRT use and fracture incidence, simultaneously adjusting for BMD and other potential confounders using logistic regression analysis.
Effectiveness results HRT was strongly associated with baseline BMD status, with 27.3% (n=428) of postmenopausal women with high BMD (greater than the lowest quartile) at both sites (hip and spine) currently using HRT, 42.0% (n=185) of women with low BMD at the hip or spine using HRT, and 53.2% (n=133) of women with low BMD at the hip and spine using HRT, (p<0.001 for trend).
During the years of follow-up there were 146 incident fractures, of which 102 were among those who had never used HRT and 44 were among the current HRT users. Common sites of fracture were the wrist (36), ankle (25), digits (23), and foot or hand (20). The authors also report 3 femoral neck fractures.
A multivariate analysis of risk of fractures in premenopausal women and current HRT users, compared with postmenopausal women who had never used HRT, was conducted. This adjusted for baseline spine and femoral neck BMD, hysterectomy, age, body weight and time between scans. The results showed that women who had used HRT retained a similar reduced risk of fracture as premenopausal women (i.e. a 52% reduction in all fractures; p=0.0003).
A further multivariate analysis (adjusted as before) showed that HRT seemed to have the same relative effect whether women had a low or high BMD.
Clinical conclusions The authors reported that, after an average of 6.2 years of follow-up, HRT use significantly reduced incident fractures by 52% (95% confidence interval, CI: 18 - 67).
Measure of benefits used in the economic analysis The measure of benefits used was the number of fractures averted. This was measured directly from incidents reported by patients in questionnaires. The questionnaires appear to have asked for the number of fractures after initial BMD measurement, on average over the last 6.2 years.
Direct costs The resource quantities were reported separately from the costs. The costs included in the analysis appear to have been those of the health service. The cost of HRT was derived from the Monthly Index of Medical Specialties (May 2000). The cost of a general practitioner visit was taken from Netten and Dennet, while the costs for a mammography and cervical screening came from the National Office of NHS Cancer Screening. All other costs included in the analysis were taken from the Aberdeen Royal Hospital NHS Trust Price List. These costs were for the endocrine clinic, menopause clinic, hip fracture, leg fracture, wrist, elbow or arm fracture, ankle fracture, other fractures, hysterectomy, breast surgery (including follow-up visits), outpatient episodes and other episodes. The costs of the various types of fractures excluded follow-up visits. Outpatient episodes covered general surgery, gynaecology, cardiology clinic, osteoporosis clinic, medical oncology and other attendances.
A model was not used to extrapolate to a longer timeframe or another setting. The costs of HRT treatment, a visit to the general practitioner and mammography were derived from the literature. The authors reported that, for inpatient and outpatient episodes, expert opinion was sought to determine the types of interventions that might be related to the use of HRT. Expert medical opinion was also sought to verify the choice of accounting codes that corresponded to the interventions recorded from the questionnaires. Discounting was relevant, and the costs of treatment and clinic visits were discounted at a rate of 6%. No explicit justification was given for this rate. The study reported the mean costs. The cost data all referred to 1997, except for the cost of HRT which referred to 2000.
Statistical analysis of costs Mean values and SDs were reported for the costs. The authors reported that mean differences (with 95% CIs) in costs and effects were calculated using both Student's t-test and a test based on a nonparametric bootstrap test with 2000 replications to account for the possible non-normality of the distribution. The results were robust for the analysis on the whole sample, but differed for the analysis on the high-risk group (small sample). Hence the bootstrap results were employed for this group. The authors reported that Fieller's theorem was used to obtain CIs for cost-effectiveness ratios and that the cost-effectiveness acceptability curves were calculated using the bootstrap method.
Indirect Costs The indirect costs were not reported.
Sensitivity analysis A simple one-way sensitivity analysis around the cost of HRT was carried out. The method used to determine the range over which the cost of HRT was varied was not reported. The area of uncertainty investigated was generalisability of the results or costs.
Estimated benefits used in the economic analysis See the 'Effectiveness Results' section.
Cost results The total mean unadjusted cost was 790 (SD=700) for the HRT group and 491 (SD=564) for the no HRT group. The mean difference was 299 (95% BCa CI: 241 - 364) when calculated using 2000 bootstrap replications to obtain the bias-corrected and accelerated intervals, and 295 (95% CI: 238 - 359) when assuming a normal distribution of the mean. The corresponding discounted cost was 711 (SD=621) for the HRT group and 440 (SD=502) for the no HRT group. The mean difference was 275 (95% BCa CI: 228 - 332) when using the bootstrap approach, and 275 (95% CI: 221 - 330) when assuming a normal distribution of the mean.
For women who had had a hysterectomy, the total mean unadjusted cost was 785 (SD=631) for the HRT group and 632 (SD=778) for the no HRT group. The mean difference was 153 (95% BCa CI: 3 - 300) when calculated using the bootstrap approach, and 152 (95% CI: 1 - 304) when assuming a normal distribution of the mean. The corresponding discounted cost was 705 (SD=562) for the HRT group and 566 (SD=692) for the no HRT group. The mean difference was 138 (95% BCa CI: 6 - 275) when using the bootstrap approach, and 140 (95% CI: 5 - 274) when assuming a normal distribution of the mean.
The average discounted costs per woman using HRT were 275 higher per woman in the HRT group than in the non HRT group. In the group of women with prior hysterectomy, the use of HRT increased costs by an average of 138 per patient.
The costs of adverse events and knock-on effects were dealt with in the costing.
Synthesis of costs and benefits The estimated benefits and costs were combined through the cost per fracture averted. Cost-effectiveness acceptability curves were also displayed.
An incremental analysis was performed. The cost per fracture averted, using discounted costs, was 11,020 (95% CI: 8,625 - 13,872) in the whole sample and 1,784 (95% CI: 59 - 3,532) in the high-risk sample (hysterectomised women).
The authors noted that, by using the cheaper HRT treatment, the cost per fracture averted decreased dramatically. For example, with costs discounted, the incremental cost-effectiveness ratio decreased to 5,503 (95% CI: 3,391 - 7,807). In the hysterectomy group it decreased to 368 (95% CI: cost-saving - 2,078).
Authors' conclusions This study showed that hormone replacement therapy (HRT) was more cost-effective, in terms of cost per fracture averted, in high-risk populations who were encouraged to take HRT.
CRD COMMENTARY - Selection of comparators Only one technology was used in this study, but it was administered to two different patient samples. The choice of the two samples was based on the wish to exclude patients who were taking HRT at the time of their first BMD, as it was felt that including the baseline BMD of this group in any statistical analysis would underestimate the effect of HRT on fracture incidence. You should decide if this is applicable in your own setting.
Validity of estimate of measure of effectiveness The analysis was based on a cohort study, which was appropriate for the study question. There is no strong reason to believe that the study sample was not representative of the study population. The authors reported that, at analysis, the group who were taking HRT were slightly younger and lighter than the group who had never used HRT. They were also more likely to have had a hysterectomy and to have a lower BMD. Statistical analyses were undertaken to account for potential biases and confounding factors.
Validity of estimate of measure of benefit The estimation of benefits was obtained directly from the effectiveness analysis. The choice of estimate is implicit from the justification the authors gave for carrying out this study.
Validity of estimate of costs All the categories of cost relevant to the perspective adopted appear to have been included in the analysis. The costs and the quantities were reported separately. No statistical analysis of the quantities was performed. The only sensitivity analysis undertaken was around the cost of HRT. Charges were not used as a proxy for prices. The date to which the costs referred was reported.
Other issues The authors made appropriate comparisons of their findings with those from other studies. The issue of generalisability to other settings was not addressed. The authors do not appear to have presented their results selectively and their conclusions reflected the scope of the study. The authors reported that there might have been some selection bias, because not all women who were invited to attend for a second BMD measurement actually did so. The authors felt that this may not necessarily be a significant limitation to their study since the attendance rate was high (between 70 and 80%). In addition, they had already shown that women who did not attend for a bone mass measurement tended to be at lower risk of osteoporosis compared with those women who did attend (see Other Publications of Related Interest).
Implications of the study The authors did not report any implications.
Source of funding Funded by the Rheumatology Department Trust Funds, Derbyshire Royal Infirmary.
Bibliographic details Fleurence R, Torgerson D J, Reid D M. Cost-effectiveness of hormone replacement therapy for fracture prevention in young postmenopausal women: an economic analysis based on a prospective cohort study. Osteoporosis International 2002; 13(8): 637-643 Other publications of related interest Torgerson DJ, Garton MJ, Donaldson C, Reid DM, Russell IT. Recruitment methods for screening programmes: trial of an improved method with a regional osteoporosis study. BMJ 1992;307:99.
Indexing Status Subject indexing assigned by NLM MeSH Bone Density /physiology; Cost-Benefit Analysis; Epidemiologic Studies; Estrogen Replacement Therapy /economics /statistics & Female; Fractures, Bone /epidemiology /prevention & Humans; Menopause /physiology; Middle Aged; Scotland /epidemiology; control; numerical data AccessionNumber 22002001531 Date bibliographic record published 28/02/2005 Date abstract record published 28/02/2005 |
|
|
|