Cost effectiveness of pre-operative gonadotrophin releasing analogues for women with uterine fibroids undergoing hysterectomy or myomectomy

Authors

  • Cynthia Farquhar,

    Corresponding author
    1. Cochrane Menstrual Disorders and Subfertility Group, Department of Obstetrics and Gynaecology, National Women's Hospital, University of Auckland, Auckland, New Zealand
      * Dr C. Farquhar, Department of Obstetrics and Gynaecology, National Women's Hospital, University of Auckland, Auckland, New Zealand.
    Search for more papers by this author
  • Paul M. Brown,

    1. Department of Community Health, University of Auckland, Auckland, New Zealand
    Search for more papers by this author
  • Sue Furness

    1. Effective Practice Institute, University of Auckland, Auckland, New Zealand
    Search for more papers by this author

* Dr C. Farquhar, Department of Obstetrics and Gynaecology, National Women's Hospital, University of Auckland, Auckland, New Zealand.

Abstract

Objective To conduct a cost effectiveness analysis of pre-operative gonadotrophin releasing hormone agonists (GnRHa) in women with uterine fibroids undergoing hysterectomy or myomectomy.

Design A cost effectiveness analysis using the effectiveness data from a systematic review of GnRHa.

Setting Secondary care.

Sample Women with uterine fibroids undergoing hysterectomy or myomectomy and women volunteers.

Methods Effectiveness data were used from a systematic review to construct a model and to calculate the cost per surgical outcome avoided. In order to evaluate the value women place on the outcome, a willingness to pay analysis of women volunteers was undertaken.

Main outcome measures (a) The cost of avoiding abdominal hysterectomy and the cost of avoiding a vertical incision at either hysterectomy or myomectomy; (b) The value that women place on avoiding abdominal hysterectomy and on avoiding a vertical incision at either hysterectomy or myomectomy. All costs are in NZ dollars.

Results For hysterectomy, the additional cost of treatment with GnRHa was NZ$1190 per case. The cost of avoiding one abdominal procedure was NZ$4577 per case and the cost of avoiding one vertical incision was NZ$6263. For a myomectomy, the additional cost of treatment with GnRHa was NZ$1535 per case. The cost of avoiding one vertical incision was NZ$4651 per case. These costs exceeded the benefit women placed on the outcomes.

Conclusion Although the pre-operative use of GnRHa results in benefits which include less frequent abdominal incisions in the case of hysterectomy and less frequent vertical incisions in the case of myomectomy, the benefits do not justify the costs. This study highlights the importance of considering both the benefits and costs in health care decisions.

Introduction

Uterine fibroids are a commonly occurring benign tumour of the uterus in women of reproductive age. They occur in one in four women over the age of 36 years and are the most common cause of major surgery in premenopausal women1–3. Two common operative procedures for women with uterine fibroids are hysterectomy and myomectomy. When hysterectomy is undertaken, a vaginal approach is preferred, as an abdominal approach is associated with increased complications4 and avoiding an incision is thought to be cosmetically desirable. An alternative procedure is a myomectomy, which may be the procedure of choice for some women as fertility is maintained. However, there is a risk of recurrence with myomectomy and the need for further surgery often arises3,5.

Previous studies have demonstrated that gonadotrophin releasing hormone agonists (GnRHa) treatment can result in fibroid shrinkage of 30% to 60%3,6,7 and gynaecologists have been encouraged to use the treatment8,9. A systematic review of 21 randomised controlled trials (RCTs) has reported that pre-operative treatment with GnRHa can lead to clinical benefits5. For myomectomy, the suppression of fibroid growth decreases blood loss during surgery, leads to faster post-operative recovery and maintains the chance of pregnancy. When a hysterectomy is performed, treatment with GnRHa increases the likelihood of having a vaginal rather than an abdominal approach5. For both myomectomy and abdominal hysterectomy, the use of GnRHa pre-operatively increases transverse incision rates5.

The purpose of this paper is to examine whether the benefits described in the systematic review justify the additional cost of treatment. That is, whether it is cost effective to use GnRHa as a pre-operative treatment prior to hysterectomy and myomectomy.

Methods

The costs of medical care to both the health system (e.g. consultation, treatment and hospitalisation) and the patient (lost wages due to treatment) were priced at current prices (NZ$). Other indirect outcomes from the treatment, such as pain and suffering, are not considered due to lack of data.

The results describe the cost per abdominal entry avoided (for hysterectomies) and cost per vertical incision avoided (when an abdominal incision is necessary). The probabilities for these outcomes were taken from a systematic review of 21 RCTs5. The use of pre-operative GnRHa for hysteroscopic or laparoscopic surgery was not considered in this study as no RCTs of pre-operative GnRHa were found. The systematic review provided no evidence on the value women place on these outcomes5. Such information is necessary in order to determine whether the additional cost of treatment is justified. This study therefore augments the systematic review by providing evidence on the extent women value each outcome. A willingness to pay survey administered to 30 women in New Zealand was used to indicate the value placed on having a transverse instead of a vertical incision (for both a hysterectomy and a myomectomy) and for a vaginal rather than an abdominal approach for hysterectomy.

Costs associated with three stages of the treatment were examined: pre-operative phase (commencing three months before surgery), the hospitalisation phase and the post-operative period (three months). Thus, the time period for the study was six months. Only direct costs to the health sector and to the patients were examined. The health care resources used are summarised in Table 1.

Table 1.  Health care resource usage.
 Pre-operative phaseIntra-operative phasePost-operative phase
Hysterectomy
Treatment with GnRHathree consultations for treatment plus 10% chance of additional consultationSurgery: 54 minutesone consultation
Hospital stay: two days for vaginal approach, four days for an abdominal approach
No treatment Surgery: 60 minutesone consultation
Hospital stay: two days for a vaginal approach, four days for an abdominal approach
Myomectomy
Treatment with GnRHathree consultations for treatment plus 10% chance of additional consultationSurgery: 54 minutesone consultation plus 65% chance of recurrence
Hospital stay: three days
No treatment Surgery: 60 minutesone consultation plus 28% chance of recurrence
Hospital stay: four days

Treatment includes monthly doses of GnRHa analogue. This medication is not subsidised in New Zealand. The price used in the analysis was the market (unsubsidised) price.

Three injections were required (at one-month intervals). Injections were administered at a gynaecology outpatient clinic.

Previous studies indicate that approximately 10% of the women will require a consultation due to side effects from the medication5. The cost of the consultation is included.

The average time needed for a consultation is assumed to be one hour (including travel time). Value of time spent away from work was assessed using the average hourly wage rate for women in New Zealand.

As mentioned above, the benefits of the GnRHa treatment include a lower risk of complications from surgery and postsurgical infections. These benefits are reflected in quicker surgery time and a shortened hospital stay.

Total theatre costs for surgery were determined using the average cost per minute of theatre time and the number of minutes per surgery. The theatre team was assumed to consist of a consultant gynaecologist, a consultant anaesthetist and three surgical nurses. Average theatre time was available from the systematic review.

The average cost per minute was based on the market rate charged by National Women's Hospital Auckland to private patients. As such, it includes both the variable (staff and supplies) and the fixed (overhead) costs associated with running the hospital.

The participant's stay in the hospital was costed using the number of days and the average cost per day in the general medical ward. As with the theatre cost, ward cost per day uses market prices charged by the National Women's Hospital to private patients.

Note that the cost per ward day includes the normal use of resources (such as nurse time and standard medication). Because the hospital includes the use of antibiotics in these estimates, it is not necessary to include the cost of adverse events and infections separately. The additional cost will be reflected in the increased hospital stay (although not the pain and suffering incurred by the patient). The systematic review also indicated that treatment with GnRHa analogues did result in less blood loss and pre- and post-operative haemoglobin. However, although the difference was statistically significant, the reductions (58 mL—one unit of blood is 500 mL) were neither clinically nor economically significant. Overall, blood transfusions were not significantly reduced.

One routine consultation at two to three months after surgery was priced at the market rate. One hour lost wages was used for patient's time.

As discussed above, some women receiving a myomectomy will redevelop fibroids postsurgery. The evidence suggests that the majority of women who have a recurrence of fibroids receive either another myomectomy or a hysterectomy5,10. However, no evidence is available on the relative probabilities of each outcome. The subsequent analysis will assume that 50% of all recurrences result in a hysterectomy within two years after surgery, with the cost of the hysterectomy included as an additional cost. It is assumed these women will have the hysterectomy an average of two years after surgery. This long term cost will be discounted at an annual rate of 6%11.

All the effectiveness data were provided by a systematic review of 21 RCTs5. The odds ratios were recalculated to show an increase in events not a decrease. The primary benefit for patients from GnRHa treatment is the increased probability of having a preferred surgical outcome. As shown in Fig. 1, for women having a hysterectomy, GnRHa treatment increases the probability of having the preferred vaginal approach from 12% to 38% (OR 4.7, 95% CI 3.0–7.4). For those who do have an abdominal approach, the probability of a preferred transverse incision increases from 67% without the treatment to 84% with the treatment (OR 2.8, 95% CI 1.8–4.3). The resulting final probabilities are shown in Table 2.

Figure 1.

Probabilities associated with treatment.

Table 2.  Probabilities of outcomes with hysterectomy and myomectomy with and without GnRHa6.
 No treatment (%)Treatment with GnRHa (%)
Hysterectomy
Vaginal approach1238
Abdominal
  Transverse5952
  Vertical2910
Myomectomy
Abdominal
  Transverse67100
  Vertical330
Recurrence2865

No evidence is available from the systematic review on the extent to which a vaginal approach is preferred to an abdominal transverse incision or an abdominal transverse incision to a vertical incision. Yet although a cardinal ranking is not possible, the outcomes can be ranked according to clinical preference (an ordinal ranking): The vaginal approach being the most preferred, followed by an abdominal transverse incision and finally a vertical incision.

Given this ranking, the results are presented in two ways: the increased probability of having the best outcome (vaginal approach) relative to the abdominal, and the increased probability of avoiding the worst outcome (vertical incision for the abdominal approach).

For women receiving a myomectomy, all women receiving the GnRHa treatment were able to have a transverse incision (compared with only 67% for those without the treatment) (OR 8.95, 95% CI 1.3–60.1). However, the treatment increases the likelihood of a recurrence of fibroids. Whereas only 28% of the women might expect a recurrence after treatment, 65% of those receiving GnRHa might expect a recurrence (OR 4.0, 95% CI 1.1–14.7). These probabilities are shown in Table 2.

The benefits of the GnRHa treatment discussed above are partially reflected in cost savings to the health system. However, outcomes such as less pain and suffering after the operation and a shorter hospital stay also benefit the patient. Alternatively, there are negative side effects associated with the treatment (e.g. adverse reactions to the medication) that are only partially reflected in increased cost (e.g. the additional time the patient spends going to a consultation). Ideally, these factors should be included when determining the overall benefit of the procedure.

Unfortunately, the RCTs discussed in the systematic review provide no evidence on the value women place neither on the decreased hospital stay nor upon the value of the reduced pain and suffering during recovery. Similarly, no evidence is available on the impact of adverse side effects experienced either as a result of taking the medication. Nor do the existing studies indicate how much women value the reduced scarring associated with a vaginal approach (hysterectomy) or a transverse incision (hysterectomy and myomectomy).

In the absence of this information, an alternative approach is to ask women to provide an estimate of the amount they would be willing to pay to obtain the favoured outcome. Other reports have used this approach in order to gain an idea of how women value the benefits of medical care12,13. For the purposes of this study, 30 women (without known fibroids or menstrual problems) in New Zealand were asked their willingness to pay for the various treatment options. After reading a description of the likely effects of treatment and the expected outcomes, women were asked how much they would be willing to pay to have (a) a vaginal rather than an abdominal approach, (b) a transverse rather than a vertical incision (should an abdominal approach be required) and (c) a myomectomy instead of a hysterectomy. Although this evidence is not a substitute for a thorough examination of quality of life differences, it does provide an estimate of the amount women value the various outcomes.

The study examines the cost effectiveness of pre-operative treatment with GnRHa. For women having a myomectomy, the results present the additional cost incurred for avoiding a vertical abdominal incision. For a hysterectomy, the results are presented in two ways: the additional cost incurred for avoiding a vertical abdominal incision, and the additional cost of avoiding an abdominal incision in favour of a vaginal approach. The evidence from the willingness to pay survey is used to assess the value placed on these outcomes (including the increased risk of fibroids recurrence for women with a myomectomy).

Results

The 21 RCTs examined for the systematic review encompass a total of 1888 women from more than 10 countries. Current market prices (NZ$ in 2001) were used to estimate the cost per each input (summarised in Table 3). Using these prices and the resource usage given in Table 1, the total costs for using GnRHa as a pre-operative treatment were calculated (Table 4).

Table 3.  Cost estimates for all aspects of care.
 UnitUnit cost (NZ$)Alternative values used during sensitivity analysis
Pre-operative
GnRHaper treatment370Hi: NZ$555
Low: NZ$185
GnRHa treatmentper consultation120Hi: NZ$180
Low: NZ$60
Patient time off work for treatmenthourly wage rate20Wages and time off work:
  Hi: NZ$30 at 3 hours
  Low: NZ$10 at 1 hour
Extra consultation for side effectsper consultation120Consultation
  Hi: NZ$180
  Low: NZ$60
Time off workhourly wage rate20Wages and time off work:
  Hi: NZ$30 at 3 hours
  Low: NZ$10 at 1 hour
Hospitalisation phase
Theatre costsper minute20Hi: NZ$30 per minute
Low: NZ$10 per minute
Length of stayper day450Hi: NZ$600 per day
Low: NZ$300 per day
Post-operative phase
Consultationper consultation120Hi: NZ$180
Low: NZ$60
Time off workhourly wage rate20Wages and time off work:
  Hi: NZ$30 at 3 hours
  Low: NZ$10 at 1 hour
Table 4.  Analysis of costs for hysterectomy and myomectomy.
 Treatment with GnRHaNo treatment
UnitsCost (NZ$)UnitsCost (NZ$)
Hysterectomy
Pre-operative
  GnRHathree at NZ$3701110  
three at NZ$120   
  Consultthree at NZ$20420  
  Add. Consult10% of one at NZ$120 and one at NZ$2014  
Hospitalisation
  Surgical time (minutes)54 at NZ$20108060 at NZ$201200
  Length of stay, vaginal approach38% of two at NZ$45034212% of two at NZ$450108
  Length of stay, abdominal: transverse or vertical62% of four at NZ$450111688% of four at NZ$4501584
Post-operative
  Consultationone at NZ$120120one at NZ$120120
one at NZ$2020one at NZ$2020
Total cost 4222 3032
Myomectomy
Pre-operative
  GnRHathree at NZ$3701110  
three at NZ$120   
  Consultthree at NZ$20420  
  Add. Consult10% of one at NZ$120 and one at NZ$2014  
Intra-operative
  Surgical time (minutes)54 at NZ$20108060 at NZ$201200
  Length of stay (days)three at NZ$4501350four at NZ$4501800
Post-operative
  Consultone at NZ$120120one at NZ$120120
one at NZ$2020one at NZ$2020
Cost of reoccurrence
  Hysterectomy50%×65% of NZ$3032/(1.06)98550%×28% of NZ$3032/(1.06)424
Total cost 5099 3564

As shown in Table 4, the expected cost for the pre-operative treatment of a patient with a hysterectomy was NZ$4222. The pre-operative treatment was nearly a third of the total (NZ$1544) and the hospital stay being the single major component. The decrease in surgical time (6 minutes) resulted in a reduction of only NZ$120. Thus, using GnRHa as a pre-operative treatment resulted in a net cost of NZ$1190. (Note: the cost of the Diagnostic Related Group in New Zealand for hysterectomy in 2000 was NZ$2700.)

Because the treatment and surgical outcomes for women having a myomectomy are similar to the hysterectomy, the treatment costs are similar. The major additional expense associated with treating women with GnRHa is the increased risk of having a recurrence. Using the assumption that 50% of women with a recurrence have a hysterectomy within two years, the increased chance of recurrence associated with the treatment results in an average additional expense of NZ$561. The additional cost of treatment for myomectomy women (NZ$1535) reflects the increased risk of recurrence.

Table 5 presents the additional cost of one additional patient receiving the desired outcome. For instance, because treatment increases the probability of a woman having a transverse incision for a myomectomy by a third (33%), we could expect one additional woman to receive the treatment for each three women treated. Thus, the total cost of the surgical outcome for the woman receiving the treatment must include the cost of treating all three. The result suggests it would cost approximately NZ$4577 for one woman to receive a vaginal rather than an abdominal hysterectomy. Avoiding a vertical incision over a transverse incision or a vaginal approach is even more costly, with NZ$6263 required for one vertical incision avoided. For a myomectomy, the expected cost of avoiding a vertical incision is NZ$4651.

Table 5.  Cost effectiveness of treatment with GnRHa.
 Cost difference per patient (NZ$)Increased probability per patient (%)Cost for one additional patient to receive the desired outcome (NZ$)
Hysterectomy
Vaginal rather than abdominal approach4222381190/0.26 = 4577
303212
119026
Avoiding a vertical incision1190291190/0.19 = 6263
10
19
Myomectomy
Avoiding a vertical incision50991002096/0.33 = 4651
−3564−67
153533

The willingness to pay responses provide evidence on the value women place on various outcomes. Comparing these estimates with the cost savings given in Table 5 provides evidence on whether the treatment provides value for money. Thirty women took part in the willingness to pay survey. Their average household income was NZ$68,667 (standard deviation NZ$28,885) and their average age was 38.3 years (standard deviation 7.5).

Table 6 shows the results from the willingness to pay questionnaire. When asked to imagine they were to have a hysterectomy, 83% (26 of 30) of the women indicated that they would prefer a vaginal to an abdominal approach. Of those 83%, the average amount they were willing to pay was NZ$644 (ranging from NZ$200 to NZ$2000). When asked to choose between a transverse and a vertical incision, 93% chose the former. The average willingness to pay for this outcome was NZ$594 (ranging again from NZ$200 to NZ$2000). These estimates are similar to the average responses by women when asked to imagine they were to have a myomectomy: 90% (27 of 30) indicated that they would prefer a transverse incision and were willing to pay an average of NZ$792 for that outcome (ranging from NZ$200 to NZ$2000).

Table 6.  Comparing cost of treatment with willingness to pay.
 Average from willingness to pay questionnaire, NZ$ (SD)Expected cost of treatment (NZ$, from Table 4)
Hysterectomy
Vaginal rather than abdominal approach644 (447)4577
Avoiding a vertical incision594 (407)6263
Myomectomy
Avoiding a vertical incision792 (516) 
Avoiding a hysterectomy4150 (5016) 
Net willingness to pay to avoid a vertical incision792
inline image
4651

One limitation of GnRHa pre-operative treatment for myomectomy was the increased risk of recurrence. Assuming that the recurrence leads to a 50% chance of a hysterectomy within two years, then the treatment is associated with the loss of possible fertility for a significant portion of the women. As an indication of the value women place on retaining the possibility of remaining fertile, the women were asked to indicate their willingness to pay for having a myomectomy rather than a hysterectomy. Of the 73% (11 of 15 respondents) of the women who preferred a myomectomy, their average willingness to pay was NZ$4150 (ranging from a low of NZ$300 to NZ$15,000). Thus, factoring the possibility of needing a hysterectomy in the future, the women would be willing to pay only a nominal amount (NZ$25) for the myomectomy (as shown in the last row of Table 6).

Comparing the willingness to pay estimates with the expected cost of treatment for each case reveals that in no case do the benefits from treatment justify the expense. Whereas the expected cost of having one woman receive a hysterectomy to have a vaginal approach rather than an abdominal approach was NZ$4577, the average willingness to pay was only NZ$644. For a woman receiving a myomectomy, the disparity is even greater, with the expected cost of NZ$4651 outweighing the net benefit of only NZ$25. Taken together, these results suggest that the benefits from treatment would not justify the expense.

In order to examine the sensitivity of the results to the assumptions regarding the cost and effectiveness of treatment shown in Tables 1–5, two scenarios were examined. The first was a ‘worse case’ scenario in which actual costs were assumed to be 50% greater and effectiveness 50% worse (see ‘High estimates’ in Table 7). The second was a more favourable scenario in which the actual costs were assumed to be 50% less and the effectiveness twice as great (‘Low estimates’).

Table 7.  Sensitivity analysis for high and low cost estimates for hysterectomy and myomectomy.
 Cost difference per patient (NZ$)Increased probability per patient (%)Cost of one additional patient (NZ$)
Hysterectomy—vaginal approach
High estimates20101315,462
Low estimates566391425
Hysterectomy—avoiding vertical incision
High estimates20101118,273
Low estimates566272059
Myomectomy—avoiding vertical incision
High estimates25221615,746
Low estimates745491520

The results are shown in Table 7. Even under the most favourable assumptions (Low estimates), the total cost of a vaginal approach or a transverse incision outstrips the value placed on the outcome by women.

Discussion

This study examined whether it is cost effective to use GnRHa as a pre-operative treatment for women with fibroids. Using clinical outcomes and health care resource usage reported in a systematic review of 21 RCTs5 and estimates of the amount women value the outcomes from a willingness to pay survey, the paper concluded that the benefits from treatment (avoiding vertical incisions or avoiding an abdominal approach) do not justify the expense. Because the benefits were based on studies from a number of different settings (all appropriately in gynaecologic surgery) and a number of different countries, the evidence on outcomes is well founded. As outcomes in everyday practice are unlikely to be as good as the clinical trial, the evidence for the cost effectiveness of the treatment is likely to be even less than stated.

There are a number of limitations to the costings and analysis reported above. For instance, the cost of treatment may be overstated if the payers receiving the medication are at a lower cost. Furthermore, few studies considered women with very large fibroids and it is possible that women with fibroids greater than 14 weeks of size will have greater benefits that may justify treatment. Nor did the analysis include other societal costs (such as caregiver time and effort). These omissions reflect the limitations of dealing with data from systematic reviews (in which the review is dependent upon the detail provided in the published studies). Indeed, although there are many RCTs of hysterectomy, relatively few studies have examined myomectomy and only two small studies included recurrence of fibroids as an outcome.

The use of the willingness to pay estimates can also be questioned. The willingness to pay data reported here are open to a number of criticisms, including that the women may not have sufficient knowledge of the outcomes (being volunteers rather than patients), that the sample size (30) was too small to draw any conclusions and that insufficient consideration was given to individual differences (such as level of income). The data are included in order to provide some estimate of the value women place on the outcomes. A larger study among patients and across different settings would provide a more reliable information on the value that women place on these outcomes.

The need for willingness to pay estimates arises because none of the studies for the systematic review presented evidence of this kind. As shown above, the advantage of willingness to pay surveys and other similar approaches is to provide evidence on the value women place on the respective outcomes. This evidence is essential when determining whether the treatment benefits justify the costs. It has been used in both health and other contexts12,13. Although subject to debate, the values presented here suggest that the benefits from the pre-operative treatment are not highly valued by women, and (in the case of the myomectomy) the negative consequences result in no net gain as a result of the treatment. Thus, it is hoped that the inclusion of these results will highlight the need for more studies to include the values that women place on these outcomes.

How should clinicians use the information from this study? Increasingly, clinicians and patients are asked to consider cost as part of the decision making process. In the United Kingdom, the National Institute of Clinical Effectiveness considers cost as part of their process14. Health economics provides a methodology for comparing the benefits with the costs15. While some clinicians may view the field of health economics with suspicion, unless there is both clinician and patient input then the perspective of the economic evaluation may be narrow.

This economic evaluation grew out of a concern that although the evidence for pre-operative use GnRHa was consistent and statistically significant, neither the size of the benefit, the patients perspective nor the cost was considered in the studies. The economic perspective also suggests considering how the resources saved could be used in other ways. In this case, the results suggest that for the resources expended to avoid a vertical incision, one or two additional hysterectomies could be undertaken. Gynaecologists have long been encouraged to use GnRHa for women with fibroids undergoing hysterectomy and myomectomy, but as this economic evaluation alongside a systematic review of the research has shown, the benefits at this time do not justify the costs.

Ancillary