Hum. Reprod. Advance Access published April 7, 2004

Human Reproduction Page 1 of 6 Hum. Reprod. Advance Access published April 7, 2004 DOI: 10.1093/humrep/deh219 An economic evaluation of laparoscopic ovarian diathermy versus gonadotrophin therapy for women with clomiphene citrate resistant polycystic ovary syndrome Cynthia M. Farquhar 1,2,3,5, Karen Williamson 1, Paul M. Brown 3,4 and Jules Garland 1 1 Department of Obstetrics & Gynaecology, 2 Fertility Plus, 3 Effective Practice, Informatics and Quality Institute, University of Auckland, Auckland, New Zealand and 4 Centre for Health Services Research and Policy 5 To whom correspondence should be addressed at: Department of Obstetrics & Gynaecology, University of Auckland, National Womens Hospital, Auckland, New Zealand. E-mail: c.farquhar@auckland.ac.nz BACKGROUND: Laparoscopic ovarian diathermy and gonadotrophin ovulation induction for women with clomiphene citrate resistant polycystic ovary syndrome have been shown to result in similar pregnancy rates, but their relative cost-effectiveness has not been evaluated. METHODS: A cost-minimization study was undertaken alongside a randomized controlled trial in women with anovulatory infertility secondary to clomiphene resistant polycystic ovary syndrome. Inclusion criteria were age less than 39 years, body mass index less than 35 kg/m 2, failure to ovulate with 150 mg of clomiphene citrate for 5 days in the early follicular phase, more than 12 months of infertility and no other causes of infertility. Laparoscopic ovarian diathermy was compared with three cycles of urinary or recombinant gonadotrophins. Direct and indirect costs were based on the results of a randomized trial. Results: The cost of a live birth was one third lower in the group that underwent laparoscopic ovarian diathermy compared to those women who received gonadotrophins (NZ$19 640 and NZ$29 836, respectively). Conclusions: This economic evaluation shows that treating women with clomiphene-resistant polycystic ovarian syndrome with laparoscopic ovarian diathermy results in a signi cant reduction in both direct and indirect costs. Key words: clomiphene citrate resistance/cost-effectiveness analysis/gonadotrophins/laparoscopic ovarian diathermy/polycystic ovary syndrome Introduction Anovulatory infertility secondary to polycystic ovary syndrome is a common ailment among women who attend fertility clinics (Balen et al., 1993). The preferred treatment has been ovulation induction with clomiphene citrate, with rates of ovulation reported at 70% after rst treatment (Franks et al., 1995). Women who do not ovulate after treatment with increasing doses of clomiphene citrate are described as being clomiphene citrate resistant. In the past, wedge resections to induce ovulation are reported being successful in up to 90% of women (Adashi et al., 1981). The introduction of urinary gonadotrophins and oral anti-estrogen agents (such as clomiphene citrate) in the early 1980s resulted in wedge resections of the ovary being largely abandoned. While often successful, treatment with gonadotrophins requires a considerable investment of patient time and can result in ovarian hyperstimulation and thus increase the likelihood of multiple pregnancies. The advent of laparoscopic surgery has raised the possibility of inducing ovulation by laparoscopic ovarian diathermy (Armar et al., 1993; Gjonaess et al., 1994; Donesky et al., 1995; Udoff et al., 1995). The surgical approach is not associated with ovarian hyperstimulation and can lead to consecutive ovulations without the need for further treatment. However, compared with hormonal treatment, the surgical procedure is more invasive, can result in tubo-ovarian adhesions and carries the risk of adverse longterm effects on ovarian function. We have recently published a randomized controlled trial comparing laparoscopic ovarian diathermy with gonadotrophins and reported similar pregnancy rates between the two treatments (Farquhar et al., 2002). Although the clinical outcomes may be similar, the treatments may differ in the cost of care. Finding the lower cost option is an important consideration for funders and providers of health services as choosing a treatment with similar outcomes and lower costs will free up health care resources. The purpose of this report is to compare the costs and outcomes of this trial of women with clomiphene-resistant polycystic ovarian syndrome who received either a surgical laparoscopic ovarian diathermy or up to three cycles of the follicle stimulating hormone (gonadotrophins) within the context of a randomized controlled trial. Materials and methods The clinical trial upon which this economic evaluation is based was an open pragmatic randomized controlled trial comparing a laparoscopic Human Reproduction ã European Society of Human Reproduction and Embryology 2004; all rights reserved Page 1 of 6

C.M. Farquhar et al. Table I. Laparoscopic ovarian diathermy Mean units Unit cost Mean cost ovarian diathermy (Group 1) and with up to three cycles of gonadotrophin therapy (Group 2) (for a description of the study design and treatment outcomes, see Farquhar et al., 2002). Women were recruited from publicly funded and private tertiary level fertility clinics in Auckland, New Zealand between 1996 and 1999. The inclusion criteria for the study were as follows: age 20 to 38 years old, clomiphene citrate resistance, infertility of greater than 12 months duration, polycystic ovaries on ultrasound scan according to accepted criteria (Adams et al., 1985), a body mass index (BMI) of less than 33 kg/m 2 for women of European descent and less than 35 kg/m 2 for women of Paci c Island or NZ Maori descent, and normal semen analysis. The exclusion criteria were other known causes of infertility, including male factor infertility or known tubal disease. A societal approach was taken when conducting the economic evaluation, in that we included both direct and indirect health care costs. Resource data were collected prospectively from entry to the trial to completion of follow up for each woman. Women receiving the laparoscopic ovarian diathermy (Group 1) were followed for a maximum of 6 months or until they became pregnant. Those receiving gonadotrophins (Group 2) were followed until the end of the 3rd cycle or until they became pregnant. Direct costs: treatment and medical care The cost of health service resources was based on 1998 unit costs for National Women's Hospital, Auckland, New Zealand. All costs are expressed in New Zealand dollars. Laparoscopic Ovarian Diathermy Hospital costs Pre admit clinic 1 $100.00 $100.00 Fixed theatre costs 1 $292.50 $292.50 Fixed anesthesia 1 $72.33 $72.33 Theatre time (min) 54 $15.16 $818.64 Time on ward (h) 11.28 $20.94 $236.32 Total Hopsital cost $1519.67 Clinic visits Hospital. Laparoscopic Ovarian Diathermy is a surgical procedure requiring inpatient surgery in a general hospital, followed by a recovery phase within the hospital. As shown in Table I, the cost of the surgery Page 2 of 6 Monitoring Blood tests 8.44 $17.00 $143.48 Ultrasound 1.44 $70.00 $100.80 Clinic visits 1.18 $144.00 $169.92 Total clinic costs $414.20 Indirect costs Lost production by patient $676.00 Personal medical expenditure $18.00 Medicines $11.00 Counselling $0.00 Travel $74.00 Total indirect costs $779.00 Miscarriage ERPOC 1.0 $1279.45 $1279.45 Ultrasound 3.3 $70.00 $233.10 Blood tests 9.6 $17.00 $163.20 Clinic appointments 0.3 $144.00 $47.52 Total cost per miscarriage $1716.85 $240.36* Total cost per patient $2953.23 *Assuming a miscarriage rate of 14% consisted of a pre-operative phase, a xed theatre cost and the cost of the anaesthetist and the surgeon. All units were priced by the minute using standard hospital charges. The cost of post-surgical recovery in the hospital was based on an hourly charge in the general medical ward. Clinic visits. Women receiving surgery were required to visit the clinic approximately 2 weeks after surgery. Additional visits where required if complications or side effects arose from the surgery. The cost of the clinic visits was calculated based on the number of minutes spent in consultation (at $144 per hour). Gonadotrophin ovulation induction Women receiving an ovulation induction require up to three cycles of treatments. Each treatment cycle required a clinic visit during which blood tests were conducted, an ultrasound performed, luteal support provided and the drug was administered. The cost of the clinic visit was based on the number of minutes spent in consultation (at $144 per hour). The ultrasound was estimated to cost $70 per visit, each blood test $17 per test, and the luteal support $15 per unit. For the drug treatment, the number of units of gonadotrphin administered varied across patients. The average units used are reported in Table I. Each unit of treatment was priced at $30. Clinic visit. The cost of additional clinic visits for treatment of complications or side effects was based on the number of minutes spent in consultation (at $144 per hour). Adverse outcomes. For women in both groups, the most common adverse outcome is for the woman to have a miscarriage. This requires additional clinic visits

?Evaluation of laparoscopic ovarian diathermy? as well as blood tests, an ultrasound and evacuation of retained products of conception (ERPOC). Indirect costs: productivity losses, travel and other costs Women in both conditions were asked to record the number of days/ hours that they had been unable to work or perform normal home duties after treatment. The mean number of working days lost in each treatment group (due to factors such as recovery after treatment and time spent at the clinic) was used to calculate lost productivity. Allowance was not made for reduction in paid production due to absence from work for partners or other caregivers. Each clinic visit was estimated to be 30 min duration and each clinic appointment 1 h. The cost per hour of time spent in treatment and away from work was based on the lost productivity. The value of the productivity loss was determined using an average hourly rate derived from the total daily income ($113.30) for women in the Auckland area (Quarterly Employment Survey May 1998, Statistics New Zealand, Wellington). In addition, travel costs to and from the clinic were calculated using return distance from home address to NWH. A transport cost of $NZ0.81 per kilometre was used (New Zealand Automobile Association, 2001). Additional indirect costs such as visits to other medical practitioners, visits to counsellors and the purchase of over the counter medicines were also calculated. Outcome measures Two effectiveness ratios are presented: cost per live birth and cost per pregnancy. While the former re ects the ultimate purpose of the interventions, the latter is reported in recognition that there are many factors that can in uence the miscarriage rates. Statistical methods Analysis of all outcomes in the trial was on an intention-to-treat basis. Data on all randomized women were included for the economic evaluation. Statistical analysis was performed using SAS version 6.12 (Cary, North Carolina, USA). Sensitivity analysis For the laparoscopic condition, the primary determinant of cost was the surgery and subsequent recovery period. To examine the sensitivity of the results to changes in this factor, hospital costs were varied by 25% and by 50%. For the ovulation induction therapy, drug treatment costs were the major determinant of cost. The drug costs were varied by 25% and by 50% in order to examine the sensitivity of the results to changes in treatment costs. Results Twenty-nine women were randomized to Group 1 with laparoscopic ovarian diathermy (LOD) and twenty-one to Group 2 with gonadotrophin ovarian induction (OI). Two women who were randomized to OI withdrew prior to treatment. In the OI group, three women became pregnant after Cycle 1 (two miscarriages, one live birth), two became pregnant after Cycle 2 (one miscarriage, one live birth), and two who became pregnant after Cycle 3 subsequently had live births. None of the women required hospitalization for hyperstimulation. In the LOD group 28 women underwent surgery, one withdrew prior to treatment and eight women conceived. Information on the use of resources was not available for all women in the trial. In the LOD group, data on monitoring was missing for one patient who moved abroad after surgery. In the OI group three women lacked monitoring data. One of these women's data on gonadotrophin usage was also missing. These three women received treatment at centres out of Auckland. Of the 21 women in the OI condition, three (14%) suffered a miscarriage. Because previous studies have found similar rates of miscarriage between the treatment options, the failure to observe miscarriages in the LOD condition might have resulted from the small sample size. In the subsequent analysis, it is assumed that the miscarriage rate in the LOD is also 14% (so as not to bias the results in favour of the LOD treatment). Laparoscopic ovarian diathermy The direct and indirect costs per patient are presented in Table I. As expected, hospital costs accounted for nearly half ($1520) of the total cost per patient ($2953). The next most signi cant cost component was lost productivity, re ecting the cost to the patient and society from the surgical procedure. Gonadotrophin ovulation induction Nineteen women received the rst cycle at an average cost of $1701 (Table II). The most signi cant cost component was the cost of the drug, accounting for over 50% of the total cost. Of those receiving the treatment, three became pregnant after the 1st cycle. The remaining 16 (84%) went to the next cycle. The average cost for those who had a second cycle was $1614, meaning an overall average cost to all women of $1356 ($1614 3 0.84). Two women became pregnant after the 2nd cycle, leaving 14 (74%) to receive the 3rd cycle. The average cost of the 3rd cycle was $1485 for those completing this cycle, or $1099 when averaged across all women. The indirect cost of treatment was $1065. In addition, three patients suffered miscarriages during the treatment, at an average cost per miscarriage of $1723. Factoring in the probabilities of becoming pregnant, the total expected cost for a women starting treatment was $5461 ($1701 + 0.84*$1614 + 0.74*$1485 + 0.14*$1716 + $1065). Cost per outcome measure Combining the direct and indirect cost of treatment, women receiving the laparoscopic treatment had an expected total cost of $2953, while women receiving the ovulation induction had an expected cost of $5461. Women who underwent ovarian diathermy had a 28% chance of becoming pregnant. Thus, the average cost for one pregnancy in the surgical group was NZ$10 938. The chance of pregnancy for women having ovulation induction was 33%, meaning that the average cost for one pregnancy in this group was NZ$16 549 (Table III). There were four live births in each arm of the trial. The average cost per live birth in the ovarian diathermy group was NZ$21 095 and in the ovulation induction group $28 744 (Table III). Sensitivity analysis The results suggest that treatment with laparoscopic ovarian diathermy is associated with a lower cost per pregnancy and cost per live birth. In order to test the sensitivity of the results to Page 3 of 6

C.M. Farquhar et al. Table II. Gonadotrophin ovulation induction Mean units Unit cost Mean cost Table III. Cost per outcome measure Cycle 1 Gonadotrophins 34.0 $30.00 $1020.00 Luteal support 3.3 $15.00 $49.50 Blood tests 15.3 $17.00 $260.10 Ultrasound 1.4 $70.00 $98.00 Clinic visits 1.9 $144.00 $273.60 Total cost cycle 1 $1701.20 Cycle 2 Gonadotrophins 31.1 $30.00 $933.00 Luteal support 2.9 $15.00 $43.50 Blood tests 16.1 $17.00 $273.70 Ultrasound 1.7 $70.00 $119.00 Clinic visits 1.7 $144.00 $244.80 Total cost cycle 2 $1614.00 $1355.76* Cycle 3 Gonadotrophins 28.7 $30.00 $861.00 Luteal support 2.6 $15.00 $39.00 Blood tests 14.3 $17.00 $243.10 Ultrasound 1.8 $70.00 $126.00 Clinic visits 1.5 $144.00 $216.00 Total cost cycle 3 $1485.10 $1098.97** Miscarriage ERPOC 1.0 $1279.45 $1279.45 Ultrasound 3.3 $70.00 $233.10 Blood tests 9.6 $17.00 $163.20 Clinic appointments 0.3 $144.00 $43.20 $1716.85 $240.36*** Indirect costs Lost production by patient $714.00 Personal medical expenditure $22.00 Medicines $16.00 Counselling $118.00 Travel $195.00 Total indirect costs $1065.00 Total cost per patient $5461.29 *Final cost of stage 2 is total cost 3 0.84, probability of having Cycle 2. **Final cost of Cycle 3 is total cost 3 0.74, probability of having Cycle 3. ***Final cost of miscarriage is total cost 3 14, probability of a miscarriage. Laparoscopic Ovarian Diathermy Ovulation induction Chance of pregnancy* 27% 33% Chance of live birth* 14% 19% Cost per patient $2953 $5461 Cost per pregnancy $10 938 $16 549 Cost per live birth $21 095 $28 744 *Differences between treatment groups are not statistically signi cant. changes in the primary cost variables, sensitivity analysis examined the impact of increasing hospital cost by 25% and by 50% and decreasing drug costs by 25% and by 50%. The results are shown in Table IV. The results suggest that increasing the hospital costs by 25% and reducing the drug costs by 25% does not substantially change the conclusions: laparoscopic ovarian diathermy remains the less costly option. Increasing the hospital cost by 50% and lowering the drug cost by 50% does equate the results, with the ovarian treatment being slightly less expensive than the laparoscopic surgery for both the cost per pregnancy ($12 851 and $13 752, respectively) and cost per live birth ($22 320 and $26 522, respectively). This Page 4 of 6 suggests not only that the laparoscopic treatment is more cost effective under the current pricing system, but that prices would have to change signi cantly (50%) in order for ovarian therapy to become less costly. Discussion This randomized controlled trial has demonstrated that laparoscopic ovarian diathermy has similar pregnancy outcomes to ovulation induction. Although the number of women was too few to prove that one strategy was superior to the other, the economic evaluation has shown that surgery is considerably less expensive than ovulation induction. Furthermore, as time goes on, the differences in cost are likely to be even greater as consecutive spontaneous ovulations result from laparoscopic ovarian diathermy allowing spontaneous pregnancies several years later. Ovulation induction may also be associated with more frequent multiple pregnancies, thus leading to even great costs of care (Callahan et al., 1994; Mugford et al., 1995). Finally, the other major advantage of laparoscopic ovarian diathermy is the ease and convenience of the treatment. The majority of women who underwent both procedures stated that they preferred the ovarian diathermy

?Evaluation of laparoscopic ovarian diathermy? Table IV. Sensitivity analysis Laparoscopic Ovarian Diathermy At 125% of hospital costs At 150% of hospital costs Hospital costs $1899.59 Hospital costs $2279.51 Clinic cost $414.20 Clinic cost $414.20 Indirect costs $779.00 Indirect costs $779.00 Miscarriage $240.36 Miscarriage $240.36 Total cost $3333.15 Total cost $3713.07 Gonadotrophin Ovulation Induction At 75% of drug costs At 50% of drug costs Cycle 1 $1446.20 Cycle 1 $1193.55 Cycle 2 $1355.76 Cycle 2 $963.90 Cycle 3 $1098.97 Cycle 3 $780.40 Miscarriage $240.36 Miscarriage $240.36 Indirect cost $1065.00 Indirect cost $1065.00 Total cost $5206.29 Total cost $4240.86 Cost per outcome Laparoscopic Ovarian Diathermy Ovulation induction At 125% of hospital costs At 150% of hospital costs At 75% of drug costs At 50% of drug costs Chance of pregnancy 27% 33% Chance of live birth 14% 19% Cost per patient $3333 $3713 $5206 $4241 Cost per pregnancy $12 345, $13 752 $15 777 $12 851 Cost per live birth $23 808 $26 522 $27 401 $22 320 procedure over ovulation induction with gonadotrophins (Farquhar et al., 2002). These ndings are consistent with the other randomized controlled trials although none of the other trials have published an economic evaluation (Farquhar et al., 2004). One recent economic evaluation using modeling also suggested that surgical management of this condition was the less expensive option. (Kovacs et al., 2002). The results from our study suggest that laparoscopic surgery is less expensive and will remain so unless hospital care costs increase by 50% and drug costs are lowered by 50%. There are four important caveats regarding this conclusion. First, the study involved a small number of women for whom there were no complications (e.g. hyperstimulation or serious adverse effects from surgery). The presence of one or more unfavorable events would have signi cantly altered the cost estimates. Care should therefore be taken in extrapolating these results to other situations. Secondly, the results do not incorporate the difference in value that women place on the two treatments in the trial. It is very possible that women might place a greater value on the convenience of one of the treatments than on the differences in the costs. Further study would be required to fully make this assessment. Thirdly, these data were collated in the New Zealand health care system, a publicly funded system with externally applied constraints on health care spending. Therefore, it is possible that these results will not apply to other health care systems. However, the aim of the sensitivity analysis was to take into consideration variations in costs. For example, the cost of hospital care in the US is known to be higher than in New Zealand although length of stay is usually shorter. However, the drug costs are also likely to be higher in the US or at least unchanged. Therefore the differences in treatments are likely to remain even in health care systems with high hospital and drug costs. Finally, the study period was only 6 months. It is possible that there are other, long term costs and adverse events associated with surgery or drug therapy. A longer term study is needed to identify these long term costs and consequences. The results suggest that cost per pregnancy for laparoscopic ovarian diathermy is approximately $5611 ($16 549 ± $10 938) less than with gonadotrophin ovulation induction. This translates into approximately $US3645 or 2132 pounds sterling. However, these estimates were derived using the cost of health care resources in New Zealand. In order to estimate the cost of the treatments in another health system, the price of health care resources in that market (e.g. cost per unit of gondotropin in the US) should be applied to the units of each resource shown above. In conclusion, this study has demonstrated that laparoscopic ovarian diathermy is a safe and cost-effective treatment for women with clomiphene-citrate resistant polycystic ovary syndrome. There are the obvious bene ts of avoiding multiple pregnancies and hyperstimulation, and the risks of the procedure can be balanced by the inconvenience and need for careful monitoring with gonadotrophins. Aknowledgements This work was supported by the Auckland Medical Research Foundation; grant number 81310. References Adams J, Franks S, Polson DW, Mason HD, Abdulwahid NA and Tucker M (1985) Multifollicular ovaries: clinical and endocrine features and response to pulsatile gonadotrophin releasing hormone. Lancet ii,1375±1378. Adashi EY, Rock JA, Guzick D, Wentz ASC, Jones GS and Jones HW (1981) Page 5 of 6

C.M. Farquhar et al. Fertility following bilateral ovarian wedge resection. Fertil Steril 36,320± 325. Armar NA and Lachelin GC (1993) Laparoscopic ovarian diathermy: an effective treatment for anti-oestrogen resistant anovulatory infertility in women with the polycystic ovary syndrome. Br J Obstet Gynaecol 100,161± 164. Balen AH, Tan S and Jacobs HS (1993) Hypersecretion of luteinising hormone: a signi cant cause of infertility and miscarriage. Brit J Obstet Gynaecol 100, 1082±1089. Callahan TL, Hall JE, Ettner SL., Christiansen CL, Greene MF and Crowley WF Jr (1994) The economic impact of multiple gestation pregnancies and the contribution of assisted-reproduction techniques to their incidence. New England Journal of Medicine 331,244±249. Donesky BW and Adashi EY (1995) Surgically induced ovulation in the polycystic ovary syndrome,: wedge resection revisited in the age of laparoscopy. Fertil Steril 63,439±463. Farquhar CM, Williamson K, Gudex G, Johnson NP, Garland J and Sadler L (2002) A randomized controlled trial of laparoscopic ovarian diathermy versus gonadotrophin therapy for women with clomiphene resistant polycystic ovarian syndrome. Fertil Steril 78,401±411. Farquhar C, Vandekerckhove P, Arnot M and Lilford R (2004) Laparoscopic drilling by diathermy or laser for ovulation induction in anovulatory polycystic ovary syndrome. Cochrane Library. Issue 1. Oxford: Update Software. Franks S, Adams J, Mason H and Polson D (1995) Ovulatory disorders in women with PCOS. Clin Obstet Gynaecol 12,605±633. Gjonaess H (1994) Polycystic ovarian syndrome treated by ovarian electrocautery through the laparoscope. Fertil Steril 41,205. Kovacs GT, Clarke S, Burger HG, Healy DL and Vollenhoven B (2002) Surgical or medical treatment of polycystic ovary syndrome: a cost-bene t analysis. Gynecol Endocrinol 16,53±55. Mugford M and Henderson J (1995) Resource implications of multiple births. In: Ward HR, Whittle M, editors. Multiple pregnancy. London: RCOG Press; 1995, 334±345. Udoff L and Adashi EY (1995) Polycystic ovarian disease: a new look at an old subject. Curr Opin Obstet Gynecol 7,340±343. Submitted on January 13, 2004; Accepted on February 18, 2004 Page 6 of 6