SEARCH

SEARCH BY CITATION

Keywords:

  • Cost-effectiveness;
  • endometrial ablation;
  • heavy menstrual bleeding;
  • levonorgestrel intrauterine devices;
  • thermachoice

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objective  To compare the cost-effectiveness of levonorgestrel intrauterine system (LNG-IUS) (Mirena®; Schering Co., Turku, Finland) and thermal balloon ablation (Thermachoice™; Gynecare Inc., Menlo Park, CA, USA) for the treatment of heavy menstrual bleeding.

Design  An open, pragmatic, prospective randomised trial.

Setting  A menstrual disorders clinic at National Women’s Hospital, Auckland, New Zealand.

Population  Seventy-nine women with self-defined heavy menstrual bleeding randomised to the LNG-IUS (40 women) or the thermal balloon ablation (39 women).

Methods  Decision tree modelling using primary source data was used to identify the incremental cost-effectiveness of the two treatments.

Main outcome measures  Direct and indirect costs of medical treatment, including treatment costs, subsequent medical procedures, lost income and medical treatment for failed procedures. The change in quality of life as assessed by the Short Form-36 (SF-36) measured between time of treatment and 24 months was the primary outcome measure. Economic modelling examined the expected cost and outcome for a woman entering each treatment. Sensitivity analysis explored the robustness of the results.

Results  The expected cost of treatment was $NZ1241 ($US869) for the LNG-IUS and $NZ2418 ($US1693) for the thermal balloon ablation. The LNG-IUS was associated with an increase of 15 points on the SF-36 scale, compared with 12 points for the thermal balloon ablation. Sensitivity analysis indicates that the results are robust to a 25% decrease in the price of the primary cost drivers and to variations in the rates of failed treatment between the conditions.

Conclusion  The LNG-IUS would appear to be cost-effective when compared with the thermal balloon ablation for treatment of heavy menstrual bleeding.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

In recent years, second-generation endometrial ablation devices have tended to replace hysteroscopic ablation techniques for the treatment of menorrhagia. These devices include thermal balloon ablation, microwave ablation, diffused laser energy ablation, bipolar impedance controlled ablation, cryoablation and hot saline instillation.1–3 In a study, we compared the efficacy of a thermal balloon endometrial ablation device with a commonly used alternative treatment, the levonorgestrel intrauterine system (LNG-IUS).4 The results suggest that at 12 and 24 months of follow up, women with heavy menstrual bleeding treated with the LNG-IUS had significantly lower pictorial blood assessment chart (PBAC) scores than those treated with thermal balloon ablation. Both the treatments resulted in a significant increase in overall quality of life at 24 months follow up, but there were no significant differences between the treatments in overall quality of life, patient satisfaction or the number of women requesting an alternative treatment.

The purpose of this study was to compare the cost-effectiveness of the thermal balloon endometrial ablation device with the LNG-IUS. Economic modelling and primary source data were used to calculate the direct medical cost of medical care (treatment and use of other medical practitioners), indirect cost associated with income lost due to menorrhagia and costs anticipated due to failed treatment. A decision tree identifying the outcomes resulting from the treatments (completed or failed treatment) and the costs, changes in quality of life and probabilities for each outcome was used to determine the expected cost and outcome from each treatment. The incremental cost-effectiveness of the treatments was then determined.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Women recruited from March 1999 to July 2001 at National Women’s Hospital (Auckland, New Zealand) were randomly assigned to either LNG-IUS or thermal balloon ablation, preceded by a diagnostic hysteroscopy under sedation in the outpatient clinic. As described in more detail elsewhere,4 treatments were performed in an outpatient setting during the first 10 days of the participant’s menstrual cycle. All the women underwent a diagnostic hysteroscopy prior to the insertion of either the LNG-IUS or the thermal balloon ablation. Women who could not tolerate hysteroscopy or treatment under local anaesthetic were rescheduled to have their hysteroscopy or treatment under a general anaesthetic on the next available theatre list. The LNG-IUS (Mirena®; Schering Co., Turku, Finland) was inserted as per manufacturer’s instructions. Thermal balloon ablation was undertaken using the Thermachoice™ device (Gynecare Inc., Menlo Park, CA, USA) as per manufacturer’s instructions.

The primary outcomes for the cost-effectiveness analysis were changes in quality of life and (direct and indirect) costs associated with the treatments. Quality of life was measured using the Short Form-36 (SF-36) at pre-treatment and at 24 months. Information on the resources used by women in each treatment arm was collected at the time of pre-treatment and at 3, 12 and 24 months, with the women asked to list the medical care usage over the previous 3-month period.

Costs

All resources are priced in New Zealand dollars (2004).

Pre-treatment costs

The analysis reported in this study does not include the costs of pre-treatment consultations and clinic visits since these are assumed to have been the same for both the assigned treatments.

Treatment costs

For the purposes of the study, all the women received a hysteroscopy and biopsy prior to the implantation of the device. Although this is a standard procedure when using a thermal balloon ablation device, it is not a standard treatment with the LNG-IUS. In order to avoid biasing the results, the time required for the hysteroscopy and biopsy was not included for women undergoing treatment with the LNG-IUS.

Data for the procedures were collected at the time of treatment and included:

  • • 
    number of minutes required for preparation and administration of the procedure, including sedation and hysteroscopy.
  • • 
    number of minutes required in a recovery area of the clinic.
  • • 
    dose of medications required by the woman (paracetamol, diclofenac, fentanyl, midazolam and metoclopramide).
  • • 
    equipment required for the procedure: thermal balloon ablation and the LNG-IUS.

Using the data listed above, costs were calculated from prices supplied by National Women’s Hospital:

  • • 
    Procedures—$16 per minute for theatre time, $1.13 per minute for specialist’s time.
  • • 
    Recovery—$1.66 per minute (based on an hourly charge of $100 per session).
  • • 
    Dose—ranged from $0.10 for 1000 mg paracetamol to $2.50 per 250 mg of fentanyl.
  • • 
    Equipment—$900 for a thermal balloon ablation device and $280 for a LNG-IUS device.
Post-treatment medical costs

Data were collected at 3, 12 and 24 months on direct and indirect costs associated with menstrual bleeding. This included the number of visits to other health professionals (e.g. GPs), the number of workdays impaired by menstrual bleeding and the average number of hours per workday missed over the previous 3 months.

The price of other medical treatments was taken to be the standard rate charged by GPs for a consultation in New Zealand ($35).5 The income lost due to menstrual bleeding was estimated by multiplying the total number of hours missed by the individual during the questionnaire period (e.g. 0–3 months) by the standard hourly rate for working adult women in the age range of women in the trial ($21).6

Failed treatment costs

Procedures can often lead to complications and adverse effects that require additional medical procedures to mitigate further harm. These ‘failed treatments’ can result in additional costs beyond what would have been incurred in the absence of the adverse effects. Women suffering a failure of the LNG-IUS or the thermal balloon endometrial ablation device may have to undergo a hysterectomy. Previous studies7 suggest that approximately 25% of women visiting gynaecological clinics ended up with hysterectomies. For the purposes of this analysis, this figure was taken as the likely outcome for failed treatments for both conditions.

In a previous study, the cost for a hysterectomy at National Women’s Hospital was found to be $3332 (adjusted for inflation).8 Other expenses (such as clinic visits) were assumed to be similar across the treatments.

Outcomes

The SF-36 is a general measure of quality of life, including subsections on eight domains (physical functioning, role limitations—physical, bodily pain, general health, vitality, social functions, role limitations—emotional, mental health and self-reported health). The analysis reports the change in each participant’s overall SF-36 score (averaged across the eight domains) between pre-treatment and 24 months. For women who successfully completed the treatment, the SF-36 score at 24 months was subtracted from the SF-36 score at pre-treatment. For participants who failed treatment prior to 24 months, the SF-36 scores administered at the time of failure were used as a proxy for the value that would have occurred at 24 months.

Economic modelling

An economic model was developed to show the treatment outcomes over the trial period (Figure 1). The probabilities of a participant failing treatment or continuing to the next data collection period were estimated from the study data. The following assumptions were made in construction of the model:

image

Figure 1. Decision tree analysis with total costs and changes in SF-36 scores as endpoint nodes. TBA, thermal balloon ablation.

Download figure to PowerPoint

  • • 
    Participants were asked to report their resource usage (e.g. GP visits, missed workdays and medication usage) over the preceding 3 months. In order to extrapolate the resource use to the entire period between questionnaires (preceding 9 months for the 12-month questionnaire and preceding year for the 12-month questionnaire), an average cost per month was calculated for each subject using the most recent data. The total cost for the preceding sample period was calculated by multiplying the number of months by the average monthly usage.
  • • 
    For participants who withdrew prior to the conclusion of the study period due to failed treatment, the GP and the lost income costs they would have incurred had they remained in the study were extrapolated from data measured at the time of withdrawal. That is, an average cost per month from the most recent period was multiplied by the number of months remaining till the end of the study period (24 months) to form an estimate of their resource usage.
  • • 
    For participants who withdrew prior to the conclusion of the study, SF-36 score at the time of withdrawal was used as a proxy measure for their SF-36 score at 24 months.
  • • 
    Costs are discounted at 0.05 per year (benefits were not discounted since the change in SF-36 was used as the outcome measure).

The final expected cost per woman was calculated by multiplying the transition probabilities by the costs associated with each data collection point. Similarly, the final expected outcome was the probabilities at each node multiplied by the average change in SF-36 scores associated with each final node.

Sensitivity analysis

The results section describes sensitivity analysis designed to explore the effect of plausible changes in the key parameters and variables. The key variables were considered to be a) the price of medical treatment and b) the probabilities of failed treatment. In addition, the section also reports the total cost of the LNG-IUS under the assumption that a hysteroscopy was not performed before treatment.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Between March 1999 and July 2001, 177 women were assessed in the menstrual disorders clinic and 83 considered for inclusion. Of the 83 women randomised, 4 were excluded after randomisation when a hysteroscopic abnormality (submucous fibroids) was found. Forty women were treated with the LNG-IUS and 39 treated by thermal balloon ablation. There were no significant differences in patient characteristics following randomisation (see table 1 in Busfield et al., for details).

Table 1.  Average resources values per women and unit prices
Resource categoriesLNG-IUS, n= 40Thermal balloon ablation, n= 39Price per unit
MeanMedianMeanMedian
  1. Values in parentheses represents SD.

Procedure32.63 minutes (11.14)31.65 minutes67.68 minutes (14.73)67 minutes$16 per minute theatre time, $1.13 per minute specialist time
Recovery29.74 minutes (13.63)29.70 minutes39.42 minutes (18.91)39.4 minutes$1.66 per minute in hospital clinic
Medications1027 total mg (169.10)1050 total mg1090 total mg (218.31) Ranges from $0.10 per 1000 mg of paracetamol to $2.50 per 250 mg of fentanyl
Equipment1 unit1 unit$900 for thermal balloon and $280 for LNG-IUS device
GP0.05 visits per person (0.22)0 visits per person0.08 visits per person (0.27)0 visits per person$35 per 15 minute consultation
Lost income$147.52 per person (481.45)$0 per person$96.38 per person (244.50)$0 per person$21 per hour

During the study period, 27% (21) of the women experienced treatment failure. In the LNG-IUS group, the 11 treatment failures included the following: by 3 months, one LNG-IUS expelled and two removed because of pain; by 12 months, one woman with menorrhagia, two LNG-IUS removed because of adverse effects and two LNG-IUS expelled; by 24 months, one woman with menorrhagia, one woman with actinomycoses and one LNG-IUS expelled. In the thermal balloon ablation group, the ten treatment failures included the following: at 3 months, no treatment failures; by 12 months, four women with menorrhagia and at 24 months, five women with menorrhagia and one woman with dysmenorrhoea. There were no serious complications in either treatment group. As shown in Figure 1, the probabilities at each stage are as follows:

  • • 
    LNG-IUS—0.925 (37/40) of continuing treatment at 3 months, 0.865 (32/37) of continuing treatment at 12 months and 0.906 (29/32) of completing the study period (24 months).
  • • 
    Thermal balloon ablation—1.00 (39/39) of continuing treatment at 3 months, 0.897 (35/39) of continuing treatment at 12 months and 0.829 (29/35) of completing the study period (24 months).

Table 1 shows the average and medical values for the resources associated with the delivery of the treatments (including diagnostic hysteroscopy under sedation for women receiving treatment with a thermal balloon) and the prices/costs applied to each unit. The procedure for the thermal balloon ablation was associated with longer procedure times (average time of 68 versus 33 minutes) and recovery time (average time of 39 versus 30 minutes) than for the LNG-IUS. Neither group required a high number of visits to a GP for menstrual bleeding during the 24-month follow up (0.08 and 0.05 probability of a visit). The LNG-IUS was associated with greater lost income than the thermal balloon ablation ($148 versus $96), although it was primarily due to one participant receiving the LNG-IUS, who incurred significant lost income over the trial period ($2551).

The costs associated with the timing of failed treatments are shown in Table 2 and Figure 1. For each time period, the thermal balloon ablation was associated with higher overall costs than the LNG-IUS; for women who ceased treatment between 3 and 12 months ($3088 versus $1847), between 12 and 24 months ($3452 versus $1683) and for women who completed treatment ($2113 versus $1059). Of the resource categories, the cost of the procedure constituted the most significant source of overall cost (37% for LNG-IUS and 43% for thermal balloon ablation), followed by the cost of equipment (18 and 39%, respectively). As shown in Table 2, when combined with the probabilities of being at each associated node, the LNG-IUS was associated with a lower expected cost ($1241) than the thermal balloon ablation ($2418), a difference of $1177.

Table 2.  Average costs for participants based on when they ceased treatment: by treatment
CostsLNG-IUSThermal balloon ablation
Failed treatmentCompleted treatmentFailed treatmentCompleted treatment
0–3 months3–12 months12–24 months0–3 months3–12 months12–24 months
  1. Values in parentheses represents SD.

Procedure ($)416.83 (19.78)565.29 (93.04)673.78 (272.11)560.71 (305.73)1203.46 (199.46)1330.43 (340.48)1118.18 (230.59)
Recovery ($)52.57 (42.59)48.14 (11.80)50.35 (2.54)49.14 (23.70)109.47 (67.31)57.82 (14.37)60.93 (22.47)
Medications ($)1.20 (0.08)1.69 (1.04)1.10 (0.09)1.23 (0.57)2.71 (1.20)4.29 (1.74)3.58 (1.44)
Equipment ($)280 (0)280.00 (0)280.00 (0)280 (0)900 (0)900 (0)900 (0)
GP ($)0 (0)0 (0)0 (0)2.41 (9.03)0 (0)5.83 (14.29)2.42 (9.02)
Lost income ($)0 (0)184.80 (413.25)56.00 (96.99)165.83 (542.19)105.00 (210.00)423.60 (427.58)27.51 (126.14)
Failed treatment ($)805.50 (0)767.14 (0)730.61 (0)0 (0)767.14 (0)730.61 (0)0 (0)
Probabilities7.50% (n= 3)12.50% (n= 5)7.50% (n= 3)72.50% (n= 29) 10.26% (n= 4)15.38% (n= 6)74.36% (n= 29)
Total cost ($)1556.10 (9.89)1847.06 (381.89)1683.35 (143.29)1059.32 (585.78)3087.71 (468.93)3452.49 (458.56)2112.62 (248.19)
Expected cost ($)1241.852418.76

The outcomes associated with each treatment are shown in Tables 1 and 3. For women who ceased receiving treatment between 3 and 12 months, the thermal balloon ablation was associated with a decrease in overall quality of life (2-point reduction) and the LNG-IUS with a 14-point increase. For women who ceased treatment between 12 and 24 months, the thermal balloon ablation was associated with an increase in overall quality of life by 4 points and the LNG-IUS with a 23-point increase. For women who completed treatment, the thermal balloon ablation was associated with an increase in overall quality of life by 14 points and the LNG-IUS with a 15-point increase. As shown in Table 2, when combined with the probabilities associated at each node, the LNG-IUS was associated with a 15-point increase in quality of life and the thermal balloon ablation with a 12-point increase.

Table 3.  Average SF-36 scores for participants based on when they ceased treatment: by treatment
SF-36LNG-IUSThermal balloon ablation
Failed treatmentCompleted treatmentFailed treatmentCompleted treatment
0–3 months3–12 months12–24 months0–3 months3–12 months12–24 months
  1. Values in parentheses represents SD.

SF-36 pre-treatment42.31 (32.64)53.31 (28.55)45.54 (32.05)68.38 (18.44)72.52 (7.62)59.25 (16.18)64.08 (14.51)
SF-36 24 months51.89 (28.06)67.09 (18.18)68.16 (42.37)82.96 (12.45)70.41 (15.75)63.71 (20.24)78.51 (16.85)
Probabilities7.50% (n= 3)12.50% (n= 5)7.50% (n= 3)72.50% (n= 29) 10.26% (n= 4)15.38% (n= 6)74.36% (n= 29)
Difference: SF-36 at 24 months—pre-treatment9.57 (16.58)13.78 (21.37)22.62 (27.84)14.58 (19.26)−2.10 (21.78)4.47 (20.42)14.43 (21.90)
Expected change in SF-3614.7012.09

In order to examine the robustness of the results, sensitivity analysis was performed by varying key variables. The sensitivity analysis focused on two factors; cost of treatment and probability of failed treatment.

Table 4 shows the results of decreasing the cost of the two largest components of total cost—equipment and procedure costs—by 25%. The results suggest that decreasing the costs did reduce the difference in overall expected cost difference from $1177 ($2418–$1241) to $857 ($1903–$1040), but the LNG-IUS remains less costly.

Table 4.  Sensitivity analysis: reducing procedure and treatment costs by 25% and reducing (increasing) probability of failures for thermal balloon ablation (LNG-IUS)
 LNG-IUSThermal balloon ablation
Failed treatmentCompleted treatmentFailed treatmentCompleted treatment
0–3 months3–12 months12–24 months3–12 months12–24 months
Costs
Procedure ($)312.62423.97505.35420.53902.60997.82838.64
Recovery ($)52.6748.1450.3549.14109.4757.8260.93
Medications ($)1.201.691.101.232.714.293.58
Equipment ($)210210210210675675675
GP ($)0.000.000.002.410.005.832.42
Lost income ($)0.00184.8056.00165.83105.00423.6027.51
Failed treatment ($)805.50767.14730.610.00767.14730.610.00
Total cost ($)1381.991635.741553.41849.142561.922894.971608.08
Expected values
Original probabilities (%)7.5012.507.5072.5010.2615.3874.36
 $1040.25$1903.89
Alternative probabilities (%)10.0010.0010.0070.000.000.00100.00
 $1051.51$1608.08

Due to the small sample size used to calculate the probabilities, sensitivity analysis examined variations in the probabilities of failed treatment. Table 4 shows the results under the assumption that the failure rate for LNG-IUS was 30% (rather than 27% as indicated in the sample) but was 0% for the thermal balloon ablation. The results suggest that the gap between the costs is reduced by the assumption from $626 to $557 ($1608–$1051). However, LNG-IUS remains lower in cost.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

The study compared the cost-effectiveness of the LNG-IUS with thermal balloon ablation for treatment of heavy menstrual bleeding. The results from primary data collection suggested a slight increase in overall increased quality of life for women with the LNG-IUS over the thermal balloon ablation. When considered with the results from the previous study,4 the results indicate that LNG-IUS is associated with better outcomes when measured by the PBAC and the SF-36 scores, although the differences are slight. The analysis presented here suggests that the LNG-IUS is less costly: $1241 ($US869) for the LNG-IUS compared with $2298 ($US1608) for the thermal balloon ablation. The differences are robust to changes in key drivers of costs, suggesting that the LNG-IUS is a cost-effective treatment when compared with the thermal balloon ablation.

Previous studies have examined the cost-effectiveness of other endometrial ablation devices9,10 and compared these devices with hysterectomy.11 This study is the first to directly compare the cost-effectiveness of a thermal balloon ablation device with the LNG-IUS, suggesting that the latter should be considered as an effective alternative for the treatment of heavy menstrual bleeding. Although the sample size is relatively small, the availability of primary source data is a significant improvement over economic models that must rely only upon previously published results.

There are a number of caveats regarding these conclusions. Because the sample size is relatively small, there is considerable uncertainty regarding the probabilities and cost estimates used in the analysis. The failure to find rare yet serious adverse effects might have also been an artefact of the small sample size. Although the time horizon of the study (2 years) was seen as sufficient to identify the immediate impacts of the procedures, additional long-term differences might emerge if the study period was extended. Although the sensitivity analysis explored the robustness of the results to changes in key parameters that would favour the thermal balloon ablation (no failed treatments and lower cost of procedure and equipment), the results should still be viewed with caution. This highlights the need for a study with a larger sample observed over a longer time horizon.

The relatively small difference in SF-36 scores (2-point difference overall) suggests that the treatments do not result in substantive differences in overall quality of life. However, the larger differences between treatments for women who fail are curious as it suggests that changes in the SF-36 scores may not be correlated with failed treatment. One explanation for these results might be the insensitivity of the SF-36. While the SF-36 is a robust and well-documented measure of global quality of life, there are aspects of heavy menstrual bleeding that are not captured by the SF-36. For instance, the SF-36 does not include specific questions on sexual functioning, and the pain and discomfort measures might be too global to detect differences in pain from menstrual bleeding. This might suggest that the relatively small differences in overall quality of life between treatments might understand the true differences. An alternative measure, such as a willingness-to-pay or conjoint analysis, might better reflect the impact of the treatments. Future cost-effectiveness studies should examine the appropriateness of these alternative measures for identifying the impact of heavy menstrual bleeding.

Decision making about treatments for heavy menstrual bleeding is complex; while costs may influence some decisions, the need to preserve fertility and the presence of painful symptoms would make balloon ablation inappropriate. Furthermore, the LNG-IUS can only be considered in women with a normal-sized and normal-shaped uterus and without fibroids greater than 3 cm in diameter. Some of the newer ablation techniques such as microwave endometrial ablation and water heating techniques might be able to be used in women with endometrial distortions and with fibroids up to 5 cm in diameter, although additional research on this topic is needed. In addition, some women do not wish to have an intrauterine device and also are averse to the concept of hormonal therapy albeit very small amounts.

Hysterectomy continues to be a useful option for women with heavy menstrual bleeding and it is likely that it always will be as there continues to be a failure rate with both LNG-IUS and endometrial ablation. Fortunately, the reported satisfaction with hysterectomy remains high for several years after the surgery in more than 90% of women who undergo the procedure. However, doing everything possible to avoid hysterectomy by offering the options of LNG-IUS or local ablative techniques or oral treatments such as antifibrinolytics and nonsteroidal anti-inflammatory agents is a wise and cost-effective approach to the management of women with heavy menstrual bleeding.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References