|
Relative cost effectiveness of Depo-provera, Implanon, and Mirena in reversible long-term hormonal contraception in the UK |
Varney S J, Guest J F |
|
|
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 study examined the use of long-term hormonal contraception for women.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised women aged 30 years or older. The population comprised a sample from the General Practice Research Database (GPRD), which consisted of 16,835 women who received either LIS (n=6,080), ESI (n=277), or MAI (n=10,478). There were no significant differences between the groups (more than 60% were aged 30 to 39). The sample size allowed the detection of 5% difference in resource usage and continuation rates with 95% power and a Type I error probability of 5%.
Setting The setting was primary care and included women from UK general practices that contributed information to the GPRD. The economic study was carried out in the UK.
Dates to which data relate The resource use data related to the years 1997 to 2002 and were costed at 2002/03 prices. Data on the effectiveness of LIS and MAI were derived from a paper published in 2002. The effectiveness of ESI was derived from a paper published in 1998.
Source of effectiveness data The effectiveness of LIS and MAI were both derived from a published systematic review. The effectiveness of ESI was derived from a report in a single paper that referred to evidence from other published studies.
Modelling A decision model, which combined resource data from the GPRD and clinical effectiveness data from the literature, was used. The model estimated the expected annualised direct health care costs and consequences of the provision of each type of contraception. The time horizon was 5 years with LIS, 3 years with ESI and 1 year with MAI.
Outcomes assessed in the review The outcomes assessed in order to populate the model were contraceptive use, continuation rates and unintended pregnancy rates. General practitioner (GP) and nurse visits were also derived as inputs for the model. No other adverse events were incorporated in the model.
Study designs and other criteria for inclusion in the review Sources searched to identify primary studies Not stated. However, full details of the search strategy (which included MEDLINE, POPLINE, EMBASE and BIOSIS Previews) used for the included systematic review were provided in the paper.
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 Two primary studies, plus the GPRD, were included.
Methods of combining primary studies The primary studies were not combined, but were used independently to derive the effectiveness estimates.
Investigation of differences between primary studies Results of the review The probability of pregnancy per year was reported to be 0 for ESI, 0.001 for LIS and 0.003 for MAI. It remained constant for the duration of treatment (5 years for LIS, 3 years for ESI and 1 year for MAI).
Continuation rates were reported to vary by year. These ranged from 89 to 85% for LIS, from 96 to 90% for ESI, and were 100% for MAI.
The number of GP visits per year ranged from 1.19 to 0.32 for LIS, from 1.54 to 0.86 for ESI, and were 0.68 for MAI.
The number of nurse visits per year ranged from 0.24 to 0.02 for LIS, from 0.14 to 0.00 for ESI, and were 0.34 for MAI.
Measure of benefits used in the economic analysis The measure of benefit was the number of additional pregnancies avoided.
Direct costs The health care costs were reported at 2002/03 prices. Future costs were discounted at a rate of 3.5%. Drug use and clinical activity were extracted from the GPRD for the years 1997 to 2002. Medications were costed using the Medical Index of Medical Specialties. Visits to GPs, nurses and outpatients were costed using information produced by the Personal Social Services Research Unit.
Statistical analysis of costs GP visits, nurse visits and referrals were reported together with their standard deviations. A chi-squared test was used to examine the statistical significance of differences between treatments in the number of GP visits. The unit costs were treated as point estimates.
Indirect Costs No indirect costs were reported.
Sensitivity analysis Multivariate probabilistic sensitivity analyses were undertaken using Monte Carlo simulations (1,000 iterations). Continuation rates, resource use values and pregnancy rates were varied. Resource use and pregnancy rates were varied according to a log-normal distribution between the upper and lower 95% confidence intervals. Extreme values for continuation rates were derived from the literature, while variation in resource use was estimated empirically from the GPRD.
Estimated benefits used in the economic analysis Benefits were reported using the cost-effectiveness plane, which showed that starting contraception with ESI rather than MAI would lead to the greatest number of additional pregnancies avoided (0.003). The lowest number of benefits would be found when comparing ESI with LIS (0.001).
Cost results The total annual intervention costs amounted to 61.95 for ESI, 41.00 for LIS and 107.16 for MAI, with the costs discounted at 3.5%.
Cost variability was incorporated into the Monte Carlo simulation.
The cost of an unintended pregnancy was excluded from the cost-effectiveness results.
Synthesis of costs and benefits Two-way incremental cost-effectiveness analyses were performed between each of the three treatment options. MAI was found to be both more expensive and less effective than either of the other two options. ESI resulted in fewer pregnancies than LIS, but at an additional cost of 20,953 per avoided pregnancy, a figure that was not considered cost-effective in comparison with the reported 912 threshold cost of an unintended pregnancy.
The Monte Carlo cost-effectiveness sensitivity analysis estimated that the probability of the incremental cost-effectiveness of LIS in comparison with MAI exceeding the threshold of 912 was greater than 90%. The same type of sensitivity analysis showed that the cost per averted pregnancy from using ESI in comparison with LIS would be less than the threshold value in 1 to 18% of cases.
Authors' conclusions Both the levonorgestral intrauterine system (LIS) and etonogestral subdermal implant (ESI) were dominant in comparison with the medroxyprogesterone acetate injection (MAI) from the perspective pf the UK National Health Service (NHS). ESI was not considered to be cost-effective in comparison with LIS. However, the choice of contraceptive should also consider consumer preferences, quality of life, clinical benefits, past medical history and the cost of an unintended pregnancy.
CRD COMMENTARY - Selection of comparators The three alternative medications appear to have been selected as representing current practice for long-term contraception. You should consider whether these are realistic options in your own setting.
Validity of estimate of measure of effectiveness The authors did not undertake a systematic review of the literature, but relied on other secondary sources for the estimates of effectiveness. However, one of the studies used was a well-conducted systematic review. The GPRD, which was stated to be the largest of its kind in the world (>30 million patient-years of data), was also used. The authors also undertook extensive sensitivity analyses to assess uncertainty in the model's estimates of effectiveness.
Validity of estimate of measure of benefit The estimate of avoided additional pregnancies was appropriate for the study objectives. However, it is condition-specific and does not permit cross-programme comparisons, as would be the case with a cost-utility analysis incorporating quality-adjusted life-years, for example.
Validity of estimate of costs Although the authors reported that the costs were estimated from the perspective of the NHS, they explicitly excluded the cost of treating unintended pregnancies from the cost analysis, preferring to use it as a threshold against which to judge the incremental cost-effectiveness of the alternative contraceptive techniques. The costs and the quantities were reported separately, which enhances the generalisability of the results. Chi-squared tests were carried out to test for differences between the contraceptive techniques in terms of the quantities of resources used. A sensitivity analysis of the prices was not conducted. The prices were appropriately discounted at the recommended UK Treasury rate.
Other issues The authors compared their results with those from other studies. The generalisability of their results to women aged under 30 years was felt to be irrelevant, as long-term reversible contraception is infrequently used by younger women. However, there was no discussion of the generalisability of the findings to other health care settings. The authors did not present their results selectively and their conclusions reflected the scope of the study. The authors reported a number of limitations to their study. For example, the exclusion from their model of some adverse effects, alternative contraception use following discontinuation of treatment, and the omission of non-prescription contraceptives (notably condoms).
Implications of the study The authors' findings indicated that there should be a variety of long-term contraceptive options available to consumers in order to provide patients with choice. In doing this, the maximum number of women can be protected and, therefore, the greatest savings to the NHS can be gained.
Source of funding Supported by Schering Health Care.
Bibliographic details Varney S J, Guest J F. Relative cost effectiveness of Depo-provera, Implanon, and Mirena in reversible long-term hormonal contraception in the UK. PharmacoEconomics 2004; 22(17): 1141-1151 Indexing Status Subject indexing assigned by NLM MeSH Adult; Contraceptive Agents, Female /administration & Contraceptives, Oral /economics; Cost-Benefit Analysis; Delayed-Action Preparations; Desogestrel /administration & Drug Administration Routes; Drug Implants; Female; Great Britain; Humans; Levonorgestrel /administration & Medroxyprogesterone Acetate /administration & Models, Economic; Prospective Studies; Treatment Failure; dosage /economics; dosage /economics; dosage /economics; dosage /economics AccessionNumber 22005008019 Date bibliographic record published 31/10/2005 Date abstract record published 31/10/2005 |
|
|
|