Long-acting reversible contraceptive (LARC) methods are highly effective in preventing unintended pregnancies. However, their uptake is low in much of the developed world. This study aimed at assessing the cost-effectiveness of LARC methods from the British National Health Service (NHS) perspective.
A decision-analytic model was constructed to estimate the relative cost-effectiveness of the copper intrauterine device (IUD), the levonorgestrel intrauterine system (LNG-IUS), the etonogestrel subdermal implant and the depot medroxyprogesterone acetate injection (DMPA). Comparisons with the combined oral contraceptive pill (COC) and female sterilization were also performed. Effectiveness data were derived from a systematic literature review. Costs were based on UK national sources and expert opinion.
LARC methods dominated COC (i.e. they were more effective and less costly). Female sterilization dominated LARC methods beyond 5 years of contraceptive protection. DMPA and LNG-IUS were the least cost-effective LARC methods. The incremental cost-effectiveness ratio of implant (most effective LARC method) versus IUD (cheapest LARC method) was £13 206 per unintended pregnancy averted for 1 year of use and decreased until implant dominated IUD in 15 years. Discontinuation was a key determinant of the cost-effectiveness of LARC methods.
LARC methods are cost-effective from the British NHS perspective. Practices improving user satisfaction and continuation of LARC method use should be identified and promoted.
Worldwide, 61% of women aged 15–49 years who were married or in a consensual union (∼635 million women) used some form of contraception in 2003. In developed countries, women relied mostly on oral contraceptives (16%), female or male sterilization (15%) and condoms (13%); only 9% of women used long-acting reversible contraceptive (LARC) methods. The respective percentages in developing countries were 6, 25, 3 and 18% (United Nations, 2003).
Prevention of unintended pregnancies by provision of contraception has been suggested to be cost-effective. McGuire and Hughes (1995) estimated that public provision of family planning services saved the British National Health Service (NHS) over £2.5 billion in 1991, and overall the public purse £25 billion by additional savings from income maintenance payments (e.g. child benefit and single parent allowance). Other US-based studies have demonstrated that all types of contraception provide great cost-savings to the health system (Sonnenberg et al., 2004; Trussell et al., 1995) and the broader public sector (Koenig et al., 1996).
In the UK, 53% of women of reproductive age use some form of reversible contraception, while another 22% have undergone permanent sterilization or have a partner who has been sterilized. Almost half (47%) of women under reversible contraception use oral contraceptives, 36% use barrier methods (e.g. male condom, caps and diaphragms) or behavioural methods (periodic abstinence and withdrawal) and only 17% choose LARC methods for contraceptive protection (O'Sullivan et al., 2005).
Although oral contraceptives can be very effective in preventing unintended pregnancies, they have been associated with poor compliance (Rosenberg et al., 1998) which often results in contraceptive failure (Ingelhammar et al., 1994; Bianchi-Demicheli et al., 2003). In contrast, female sterilization does not depend on users' adherence, is highly effective, but it has a permanent contraceptive effect. Notably, LARC methods combine reversibility with particularly high effectiveness, which does not rely (or relies at a small degree only) on users' compliance or correct use (Trussell, 2004).
Despite these advantages, the uptake of LARC methods in Britain and most other developed countries is rather low. Currently four LARC methods are licensed in the UK: the copper intrauterine devices (IUDs), licensed for 5 or 10 years of continuous use, the levonorgestrel intrauterine system (LNG-IUS), licensed for 5 years, the etonogestrel subdermal implant (referred to as ‘implant’ from this point onwards), licensed for 3 years and two progestogen injectable contraceptives, of which the depot medroxyprogesterone acetate injection (DMPA) is licensed for first-line use. IUDs, LNG-IUS and the implant are highly effective without requiring users' compliance. Beyond their licensed duration of use, old devices generally need to be removed and replaced by a new device in order to retain highly effective contraceptive protection. DMPA requires some degree of user compliance, comprising one visit every 12 weeks to a health professional for administration of an injection.
The high start-up costs associated with LARC methods (relating to the cost of the devices and fitting equipment as well as staff time) might be an obstacle to their wider use. Thus the objective of this study was to assess, from the perspective of the British NHS, the cost-effectiveness of LARC methods available in the UK in comparison to other contraceptive methods for women. The analysis is based on an economic model constructed for a clinical guideline on long-acting reversible contraception, developed to provide guidance to healthcare professionals, managers and women seeking contraceptive advice in England and Wales (National Collaborating Centre for Women's and Children's Health, 2005). The guideline was commissioned by the National Institute for Health and Clinical Excellence (NICE). LARC methods considered in this study include the IUD licensed for 10 years, LNG-IUS, implant and DMPA. The comparators are the combined oral contraceptive pill (COC) and female sterilization. Other popular contraceptive methods, such as the male condom and male sterilization, which were included in the initial analysis conducted for the NICE guideline, are not examined in this paper, as their use does not depend exclusively on women's choice and compliance.
Materials and Methods
Economic model structure
A decision-analytic Markov model was constructed to assess costs and benefits associated with use of LARC methods, COC and female sterilization. The model took into account the different licensed durations of use of IUD, LNG-IUS and implant, which were expected to affect their relative cost-effectiveness over time. It also considered the discontinuation associated with LARC methods, which is likely to strongly affect their cost-effectiveness, owing to the subsequent use of other, overall less effective, contraceptive methods, thus increasing the risk of contraceptive failure. Moreover, discontinuation incurs extra costs of further visits to health professionals for premature (i.e. sooner than the initially intended) removal of a device and initiation of another contraceptive method.
The economic analysis considered consecutive time frames from 1 to 15 years of intended contraceptive use to explore the changes in the relative cost-effectiveness of LARC methods over time. Intended contraceptive use means that women were assumed to intend, at method initiation, to use the method for the full time frame examined, but were free to discontinue at any point during the simulation. The maximum time frame of 15 years was chosen to reflect the full duration of contraceptive effect of female sterilization on women, i.e. from 35 years of age, average age of women undergoing sterilization in the UK (Rowlands and Hannaford, 2003), to 50 years of age, average age of menopause in British women (Lawlor et al., 2003).
According to the model structure, six hypothetical cohorts of 1000 sexually active women of reproductive age adopted one of the six contraceptive methods assessed, respectively, and were subsequently followed up from 1 and up to 15 years. The model was run in yearly cycles. Every year a proportion of women in each cohort discontinued the chosen method and switched to another contraceptive method or no method. Discontinuation in the form of surgical reversal was not considered for women having undergone sterilization.
The range of contraceptive methods (including no method) adopted by women who discontinued the initiated method were summarized in the concept of the ‘average contraceptive method’. This concept was developed in order to consider the impact on cost-effectiveness of discontinuation per se rather than of specific patterns relating to contraceptive method switching. Moreover, assuming different switching patterns for the various cohorts of women initiating each of the assessed methods would imply populations of women with different profiles, and this would make comparisons across cohorts inappropriate. The average contraceptive method consisted of all contraceptive methods available in the UK, weighted (in terms of cost and contraceptive efficacy) according to reported contraceptive usage rates for women of reproductive age requiring contraception because of the risk of unintended pregnancy (O'Sullivan et al., 2005). The proportions of women switching to female or male sterilization following discontinuation of the initiated method were estimated based on annual incidence rates rather than prevalence (Rowlands and Hannaford, 2003).
Within each year, women either received the contraceptive benefits of the method used during this year (initiated method or average contraceptive method), or faced a contraceptive failure and unintended pregnancy. Four possible outcomes of unintended pregnancy were modelled: continuation of pregnancy and birth, miscarriage, abortion and ectopic pregnancy. In years within which discontinuation occurred, contraceptive effectiveness and relevant costs were attributed by 50% to the initiated method and 50% to the average contraceptive method. In the case of contraceptive failure following sterilization, repeat of the procedure was assumed.
A schematic diagram of the Markov model structure developed for the economic evaluation is presented in Fig. 1.
Costs and outcomes considered in the analysis
Costs were estimated from the perspective of the NHS. They included the cost of contraceptive provision and costs associated with outcomes of unintended pregnancy. Contraceptive provision costs included ingredients, health professionals' time and equipment for insertion and removal of IUD, LNG-IUS and implant. Initial counselling and follow-up visits for all methods as well as additional time required for insertion and removal of a device or for injection of DMPA were taken into consideration.
Outcomes were expressed as the total number of unintended pregnancies resulting from contraceptive failure of the initiated method plus the average contraceptive method adopted after discontinuation. Therefore, the overall effect of each evaluated method was determined not only by its clinical effectiveness, but also by its discontinuation rate.
Healthcare resource use and cost data utilized in the model
Healthcare resource use related to provision of contraceptive services and to outcomes of unintended pregnancy was based on expert opinion and UK national reports (Department of Health, 2006a,b; Macdowall et al., 2000; Scottish Programme for Clinical Effectiveness in Reproductive Health, 2003). Unit prices referring to healthcare staff time, ingredient costs and costs of medical procedures were based on British national sources of healthcare costs (British Medical Association and Royal Pharmaceutical Society of Great Britain, 2007; Department of Health, 2003 and 2006b). Costs were uplifted to 2005 prices using the Hospital and Community Health Services Pay and Prices inflation index (Curtis and Netten, 2005). Costs were discounted at an annual rate of 3.5%, recommended by NICE (National Institute for Health and Clinical Excellence, 2006). Details on the cost data utilized in the base-case analysis as well as on the methods adopted for their calculation are provided in Supplementary Data 1.
Clinical data considered in the economic analysis
Annual contraceptive failure rates for LARC methods were based on data derived from RCTs (Toppozada et al., 1983; World Health Organization, 1990; Andersson et al., 1994; Croxatto et al., 1999; Edwards and Moore, 1999). Annual failure rates for the other contraceptive methods were taken from a review on contraceptive effectiveness (Trussell, 2004) and a large cohort study (Peterson et al., 1996). For COC, typical use was assumed (i.e. routine use, not always consistent or correct).
Discontinuation rates of LARC methods and COC were based on observational studies as rates reported in clinical trials are usually lower than those observed in routine clinical practice and do not always represent real-life decisions. Annual discontinuation rates of IUD, LNG-IUS and implant were based on UK studies (Cox et al., 2002a,b; Lakha and Glasier, 2006a), to reflect routine patterns representative of British women. First year discontinuation rates for DMPA and COC were estimated based on observational non-UK data, owing to lack of UK figures (Westfall et al., 1996; Paul et al., 1997; Rosenberg and Waugh, 1998; Colli et al., 1999; Hubacher et al., 1999; Trussell and Vaughan, 1999; Grady et al., 2002). For longer time periods, where relevant data were not available, discontinuation rate estimates for all contraceptive methods were based on expert opinion.
For implant, DMPA, COC and the average contraceptive method, relative probability of ectopic pregnancy was assumed to equal that in the general female population (Tay and Walker, 2000). The relative probabilities of ectopic pregnancy for IUD (Sivin, 1991; Xiong et al., 1995; Furlong, 2002), LNG-IUS (Sivin, 1991; Furlong, 2002; Backman et al., 2004) and female sterilization (Peterson et al., 1997) were estimates specific to each method, so as to reflect the fact that, although use of contraception reduces the absolute risk of ectopic pregnancy, the relative proportion of ectopic pregnancies out of all pregnancies resulting from contraceptive failure following use of IUD, LNG-IUS or female sterilization is significantly higher than the respective proportion in the general female population. The relative probabilities of the remaining outcomes of unintended pregnancy for all contraceptive methods were estimated based on UK statistics (Government Statistical Service for the Department of Health, 2005; Office for National Statistics, 2005; Scottish Office, 2005), and a number of assumptions supported by published literature (Forrest, 1994; Henshaw, 1998; Lakha and Glasier, 2006b). Outcomes were discounted at an annual rate of 3.5% recommended by NICE (National Institute for Health and Clinical Excellence, 2006). Details on the clinical data utilized in the analysis and the assumptions used at their estimation are provided in Supplementary Data 1.
Sensitivity analysis explored the robustness of the results under the uncertainty characterizing input parameters and assumptions utilized in the model. Estimated 95% confidence intervals (CIs) of annual failure rates and discontinuation rates of all methods were utilized in one- and two-way sensitivity analyses. Method costs of all methods (including the average contraceptive method) were varied by ±10% of their base-case values to test whether base-case results would be affected. Additional scenarios investigated included:
Perfect use of COC, i.e. consistent and correct use resulting in higher effectiveness; a 0.3% annual failure rate was used (Trussell, 2004).
No LARC method discontinuation; this scenario explored to what degree discontinuation affects the cost-effectiveness of LARC methods
Replacement of IUD device every 5 years, according to licensed use for some devices in the UK
Consideration of the fact that a proportion of British women discontinue contraception because they are no longer at risk for unintended pregnancy (e.g. stopped being in a heterosexual relationship, intend to become pregnant etc.) (O'Sullivan et al., 2005).
Range of abortion rates following contraceptive failure between 0 and 80%
Range of annual discount rate between 0 and 6%, suggested by NICE (National Institute for Clinical Excellence, 2004).
Cost-effectiveness of LARC methods versus COC and female sterilization
All LARC methods dominated COC (i.e. they were more effective and less costly) between 2 and 15 years of use; DMPA and IUD dominated COC also at one year of use. Female sterilization was overall more effective than all LARC methods. It was also more costly for time frames up to 4 years, resulting in incremental cost-effectiveness ratios (ICERs) versus LARC methods that reached a maximum of £38 197 per pregnancy averted. However, incremental costs decreased as duration of effect increased, until female sterilization became dominant; this occurred at 5 years over DMPA and at 6 years over the other LARC methods. Thus, for 6 years of contraceptive protection and above, female sterilization dominated all LARC methods.
Cost-effectiveness between LARC methods
DMPA was dominated by the other LARC methods between 2 and 15 years of intended use. At one year of use DMPA was the cheapest but also the least effective LARC method. LNG-IUS was dominated by absolute or extended dominance across all time frames examined (note: absolute dominance occurs when one option is less effective and more costly than another. Extended dominance occurs when an option is less effective and more expensive than a linear combination of two other options). Implant was the most effective and IUD was the cheapest (except for one and 15 years of use) among LARC methods. The ICER of implant versus IUD was £13 206 per pregnancy averted at one year of use, but gradually decreased as years of intended use increased, with slight increases at 4, 7, 10 and 13 years (owing to high implant reinsertion costs incurred in those years, which increased the average annual cost of implant and hence its incremental average annual cost over IUD). At 15 years of contraceptive use, implant dominated IUD. The fluctuation of the ICER of implant versus IUD overtime is shown in Fig. 2. Fig. 3 provides results for all methods over 10 years of intended contraceptive use in the form of a cost-effectiveness plane. Results for selected years of intended contraceptive use are presented in Supplementary Data 2.
Results of sensitivity analysis
The cost-effectiveness of LARC methods relative to COC was affected by assuming perfect COC use: compared with DMPA, COC was the dominant option for up to 12 years of use. Beyond this period DMPA became more effective at an extra cost. The cost-effectiveness of the other LARC methods relative to COC was less affected; IUD and implant dominated COC at 2 and 6 years, respectively, and above. LNG-IUS dominated COC from 5 years and beyond, with the exception of 6 years (due to high re-insertion costs of LNG-IUS in year 6). This was the only scenario of those tested that affected the cost-effectiveness of LARC methods compared with COC.
The rate of discontinuation determined the cost-effectiveness of LARC methods relative to female sterilization: at the extreme hypothesis of no LARC method discontinuation, female sterilization was dominated by one or more LARC methods across all time frames examined. In all other scenarios, results of the comparisons between LARC methods and female sterilization remained unchanged. The strongest effect of some scenarios (such as varying female sterilization costs and parameters of the average contraceptive method) was a delay or shortening in the time frame at which female sterilization became dominant, by 1–2 years at the maximum.
Discontinuation was also the major driver of the relative cost-effectiveness between IUD, LNG-IUS and implant. In contrast, the cost-effectiveness of DMPA versus other LARC methods was mostly unaffected by changes in discontinuation rates. Varying failure rates of the assessed methods in two-way sensitivity analyses (i.e. using the lower 95% CI of the failure rate of one method versus the upper 95% CI of the failure rate of its comparator) had no significant impact on the base-case results. Of the remaining scenarios, only variations in LARC method provision costs altered moderately the cost-effectiveness between IUD, LNG-IUS and implant, predominately in the longer term, i.e. from 9 years of use and beyond. Results of sensitivity analysis testing the impact of discontinuation on cost-effectiveness of LARC methods are summarized in Supplementary Data 2. Full results of sensitivity analysis are available upon request.
The economic analysis demonstrated that, despite high initiation costs, LARC methods were more cost-effective than COC: they prevented a higher number of unintended pregnancies and produced net cost-savings. This finding was driven by the significantly lower failure rates of LARC methods, which are independent of (or, in the case of DMPA, only moderately dependent on) users' compliance, compared with those reflecting typical COC use. This feature makes them suitable for women with no established regular routine, who have been shown to poorly comply with common contraceptives (Rosenberg et al., 1995). According to the study findings, if 5% of British typical COC users of reproductive age (estimated at ∼177 000 women by combining data by O'Sullivan et al., 2005 and Office for National Statistics, 2006) switched from COC use to one of the LARC methods available in the UK for 5 years of intended use, the average annual reduction in unintended pregnancies would reach 7500, while the annual net cost-savings from the switch would approximate £9.5 million. It must be noted, though, that perfect (i.e. correct and consistent) COC use improves its relative cost-effectiveness.
Female sterilization was clearly more cost-effective than LARC methods when contraceptive effect exceeded six years, since beyond this period it was the dominant option. This finding was exclusively caused by high LARC method discontinuation rates. At the extreme scenario of no LARC method discontinuation, female sterilization became dominated by one or more LARC methods across all time frames examined. It is important to note that female sterilization is not a realistic option for women wishing to retain their fertility. Evidence suggests that women sterilized at a young age are likely to regret their decision to be sterilized, consider the possibility of a surgical reversal, and even obtain a reversal or undergo evaluation for IVF procedures at a later time (Curtis et al., 2006). For these women, LARC methods are a suitable form of contraception, since they combine high effectiveness with reversibility.
Among LARC methods, DMPA and LNG-IUS were dominated by absolute or extended dominance virtually across all time frames examined. Implant was the most effective and IUD was the cheapest option. With the exception of DMPA, the cost-effectiveness between the other LARC methods was vastly influenced by changes in discontinuation. Given the robustness of the results concerning DMPA, the latter should not be preferred solely on cost-effectiveness grounds when other LARC methods are indicated and available, especially when contraceptive protection is required for long time periods.
The findings of this analysis are overall consistent with those reported in previous studies conducted in the UK (French et al., 2000; Philips, 2000; Varney and Guest, 2004) and in the USA (Chiou et al., 2003). The analysis is, nonetheless, subject to a number of limitations. LARC method use is associated with side effects, predominantly menstrual disturbances (Affandi, 1998; Bhathena, 2001; Faculty of Family Planning and Reproductive Health Care Clinical Effectiveness Unit, 2004a,b). The presence of side effects affects users' quality of life and may require extra general practitioner consultations, medication and, less frequently, hospitalization. These factors were not addressed in the analysis. Nevertheless, menstrual disturbances are among the major causes of LARC method discontinuation (Paul et al., 1997; Cox et al., 2002a,b; Lakha and Glasier, 2006a); therefore, their impact on the cost-effectiveness of LARC methods has been partially captured in the model design. Besides, COC is characterized by similar side effects such as bleeding irregularities, which constitute a common reason for discontinuation too (Rosenberg and Waugh, 1998). Consequently, omission of common side effects of LARC methods and COC from the model structure is unlikely to have substantially affected the respective study results. Female sterilization is associated with a number of complications such as bowel injuries during the operation (Royal College of Obstetricians and Gynaecologists, 2004), which may require additional healthcare resource use; these were not dealt with in the analysis, but their impact on the relative cost-effectiveness of female sterilization is likely to be insignificant. On the other hand, events such as surgical reversal or IVF following regret of the procedure were also not modelled; consideration of such parameters might have altered the results in favour of LARC methods. Non-contraceptive benefits of LARC methods and subsequent expected cost-savings (such as management of heavy menstrual bleeding with LNG-IUS use) (Hurskainen et al., 2004) were also not considered. Inclusion of this parameter in the model might affect the results in favour of LNG-IUS.
While LARC methods are highly effective in preventing unintended pregnancies, they do not provide protection against sexually transmitted infections (STIs). The lack of protection against STIs does not affect their relative cost-effectiveness compared with COC or female sterilization, as none of these methods protects women from contracting STIs. In women at risk for STIs, LARC methods should always be used in conjunction with male or female condom. In an analysis carried out for the NICE clinical guideline (National Collaborating Centre for Women's and Children's Health, 2005), it was shown that use of LARC methods combined with male condom was more cost-effective than use of male condom alone, as it provided the same degree of protection against STIs, was far more effective in preventing unintended pregnancies, and, for this reason, was overall less expensive. It must be noted, though, that, although in theory the level of protection against STIs is the same between combination of LARC methods with male condom and male condom alone, in practice this may not be the case; women using LARC methods in conjunction with male condom may not use the latter consistently and correctly when they know that they are anyhow protected from unintended pregnancy by LARC method use.
The results of the analysis could be easily interpreted in cases of dominance. However, interpretation of a finding concerning one method being more effective than another at an additional cost would require attaching a monetary value on an unintended pregnancy averted by contraceptive use, beyond the financial cost of such an event, which has already been included in the analysis. However, the value of preventing an unintended pregnancy is difficult to determine; in order to make an estimate, one needs to consider the quality of life arising from contraceptive use, the psychological distress to the woman and her family caused by contraceptive failure, the value of a life foregone due to contraceptive use or resulting from contraceptive failure, and also the long-term costs and benefits to society associated with an unintended pregnancy (either occurring or averted). Currently, utility values, appropriate to inform a cost-utility study, are not available. Consequently, clear conclusions on cost-effectiveness in cases other than dominance cannot be reached.
Discontinuation is a key determinant of the cost-effectiveness of LARC methods. Discontinuation rates reflect to a certain extent women's acceptability of a method as well as their perceptions of associated risks and reactions to side effects. The discontinuation rates used in the analysis were based, where available, on observational UK studies; thus, the results are specific to the British population. Cost-effectiveness of LARC methods may differ in other settings and cultural contexts. Nevertheless, in all settings, reducing LARC method discontinuation is expected to improve their cost-effectiveness, with direct resource implications for the health service. It has been suggested that this can be achieved by providing women with detailed information on the efficacy, changes in bleeding patterns and side effects of LARC methods before initiation, combined with regular counselling and management of side effects during contraceptive use (Lei et al., 1996; Hubacher et al., 1999; Canto de Cetina et al., 2001; Flores et al., 2005).
The study results should be interpreted with caution in settings where clinical practice, healthcare resource use patterns and unit prices differ considerably from those in the UK. In order to generalize the results of this analysis in other settings, thorough consideration of their similarities and differences to the UK setting is required, especially in terms of patterns of discontinuation and healthcare resource use. Further research is required to identify ways to increase acceptability and reduce discontinuation of LARC methods, as this will result in reduction in the number of unintended pregnancies and potential abortions, improvement in the quality of women's life, and cost-savings for the family planning services. It must be emphasized that women's needs and preferences, lifestyle and any contra-indications should be carefully considered and discussed when helping women seeking contraception make an informed choice.
This study is based on a decision-analytic model developed for a clinical practice guideline on long-acting reversible contraception produced by the National Collaborating Centre for Women's and Children's Health, commissioned and funded by the National Institute for Health and Clinical Excellence in the UK.
Members of the LARC Guideline Development Group
Chris Wilkinson, Consultant in Sexual and Reproductive Health. Margaret Pyke and Mortimer Market Centres and Camden Primary Care Trust (Group Leader); Anna Glasier, Consultant in Sexual and Reproductive Health, University of Edinburgh and NHS Lothian (Clinical Advisor); Simon Barton, Consultant in Genitourinary Medicine, Chelsea and Westminster Healthcare NHS Trust; Alyson Elliman, Consultant in Family Planning and Reproductive Health, Croydon PCT; Sophie Mancey-Jones, General Practitioner, London; Shelley Mehigan, Specialist Nurse, Slough PCT; Sam Rowlands, General Practitioner, Specialist in Contraception and Reproductive Health, British Pregnancy Advisory Service (BPAS); Sue Ward, Specialist Nurse, South Downs Health NHS Trust; Stephanie Whitehead, Policy and Development Manager, Brook (patient representative); Joyce Howarth, University of Bristol (patient representative); Martin Dougherty, Executive Director, NCC-WCH; Moira Mugglestone, Deputy Director, NCC-WCH; Irene Kwan, Research Fellow, NCC-WCH; Michael Corkett, Senior Information Specialist, NCC-WCH; Anna Bancsi, Work Programme Coordinator, NCC-WCH; Hannah-Rose Douglas, Health Economist, NCC-WCH and London School of Hygiene and Topical Medicine; Ifigeneia Mavranezouli, Health Economist, NCC-WCH.
The Guideline Development Group defined the economic question, supported research at the development of the economic model and acted as an expert group where evidence was lacking. Special thanks to Hannah-Rose Douglas, Anna Glasier, and Joanne Lord. Part of this study was presented at the CES (Collège des Economistes de la Santé)—HESG (Health Economists' Study Group) Second Joint Meeting, 4–6th January 2006, City University, London.