Ovarian cancer affects approximately 225,500 women and accounts for 140,000 deaths worldwide each year.1 More than 90% of the patients’ conditions are diagnosed with epithelial ovarian cancer (EOC).2 Epithelial ovarian cancer is staged by standards defined by the International Federation of Gynecology and Obstetrics (FIGO).3 The overall 5-year survival rate is approximately 40%, varying from 25% for advanced-stage disease (FIGO stages IIB-IV) to 90% for early-stage disease (FIGO stages I-IIA).4 Early diagnosis is difficult because of the nonspecific symptoms of EOC. Therefore, 75% of the women present with advanced-stage disease.4 The present study focuses on the treatment of patients with advanced-stage EOC.
Between diagnosis, surgery, and the initiation of chemotherapy, gaps of several weeks exist. Reducing these time intervals may benefit the patient and may lead to a reduction of costs. Costs of cancer care are expected to escalate more rapidly in the near future mainly because of more expensive treatments and improved survival.5 The challenge is to keep cancer care affordable to individuals and society. In this light, it is important to assess whether innovations, such as early-initiated treatment, represent value for money. Such an evaluation is ideally performed at an early stage because its results may be helpful in designing an optimal treatment strategy and in informing policy and research decisions.6 Decision-analytic models may be used to balance timeliness and accuracy. Although a model represents a simplification of reality, it is a pragmatic tool to synthesize available evidence from a range of sources to inform decision makers on relatively short notice.
The aim of the present study was to review and synthesize all available evidences to explore the cost-effectiveness of early-initiated treatment of patients with suspected advanced-stage EOC compared with that of current treatment. Overall survival, progression-free survival, health-related quality of life (HRQOL), and costs of the separate events were assumed to remain equal.
MATERIALS AND METHODS
Discrete Event Simulation
A discrete event simulation (DES) was developed using commercially available software (Arena; Rockwell Automation Inc, Milwaukee, WI). Discrete event simulation was the preferred modeling technique because it represents the course of events for individual patients, taking into account patient-specific characteristics. Furthermore, DES makes it possible to represent time as a continuous variable rather than in discrete cycles (eg, cycles of 1 year). This makes DES flexible and particularly useful in this study because a reduction of treatment time was 1 of the primary aims of early-initiated treatment.7
The target population consists of patients with suspected advanced-stage EOC who are eligible for primary surgery or surgery after neoadjuvant chemotherapy. We chose to include only those patients with suspected advanced-stage EOC because of the complexity of treatment and the low survival rates. The low survival rates of advanced-stage EOC patients emphasize the importance of maintaining HRQOL, so these women are expected to benefit most from early-initiated treatment. In addition, more evidences are available on treatment of patients with advanced-stage EOC compared with that in early-stage disease.
The current treatment of a patient with suspected advanced-stage EOC was modeled following the flow chart shown in Figure 1. Based on clinical, biochemical, and radiological evaluation, a decision is made whether to perform primary cytoreductive surgery or to administer neoadjuvant chemotherapy. In case of doubt, an exploratory laparoscopy may be performed to assess operability. If eligible for surgery, the patient undergoes primary cytoreduction followed by 6 cycles of platinum-based chemotherapy. Cases with no residual mass larger than 1.0 cm that did not have bowel surgery are eligible for intraperitoneal (IP) chemotherapy. If not, the patient is treated with intravenous (IV) chemotherapy only.
Patients not suitable for primary surgery receive 3 cycles of (neoadjuvant) IV chemotherapy followed by interval cytoreduction and again 3 cycles of IV chemotherapy (Table, Supplemental Digital Content 1, http://links.lww.com/IGC/A191).
Early-initiated treatment is similar to current treatment with 2 main differences. First, the time until exploratory laparoscopy is shortened (21 vs 2 days). Second, the time until primary treatment (cytoreduction or neoadjuvant chemotherapy) is shortened (7 vs 2 days). These time intervals were derived from expert opinion and are considered the minimal reasonable time intervals. As mentioned previously, we assumed that overall survival, progression-free survival, HRQOL, and costs of the separate events do not differ between the 2 treatment strategies. Only the time to these events differed.
Figure 2 provides an overview of the model flow. The simulation was run for 1000 patients until death (lifetime). At the start of the simulation, individual patients were simulated and assigned an age and stage of disease. Hereafter, patients were twinned to produce 2 identical cohorts to make sure that there were no differences in baseline characteristics between the 2 arms and to make it possible to compare the twinned patients. One of the twinned patients received current treatment, whereas the other received early-initiated treatment.
Each patient was modeled individually from the moment the condition was diagnosed with suspected advanced-stage EOC. Each patient followed the path through the simulation, determined by the patient’s specific characteristics, and events and treatments experienced. When the treatment was completed, every patient was assigned with a unique time until recurrence and time until death, depending on their characteristics and type of treatment within the arm. When the simulation arrived at a certain event for a certain patient, the costs were counted and the effects up to that point were accrued. If applicable, the patient was also assigned a new HRQOL value. When the patient died, the final costs and effects were calculated.
An overview of all parameters used in the model and their sources is shown in Table 1. The probability that a patient underwent an exploratory laparoscopy (0.175) was derived from expert opinion, because no accurate estimation could be made based on literature or internal data. The probabilities of a complete (no visible residual mass), optimal (residual mass ≤1 cm), or suboptimal (residual mass >1 cm) result of primary cytoreduction were extracted from data published by van Altena et al.8
The quality-adjusted life year (QALY) was used as an outcome measure because it combines the quantity and the HRQOL.9 Health-related quality of life was considered as a single index utility value on a scale from 0 (representing death) to 1 (representing perfect health). Quality-adjusted life years are obtained by multiplying the health utility value with time (in years). The HRQOL of patients with recurrent EOC was extracted from data published by Krasner et al.10 They used the EuroQol-5D (EQ-5D) questionnaire, which allows the calculation of health utility values. In all other relevant published literature, HRQOL was assessed using cancer-specific instruments that do not allow the calculation of health utility values. We used an algorithm reported by Crott and Briggs11 to approximate the associated health utility value using scores of the QLQ-C30 (Table, Supplemental Digital Content 2, http://links.lww.com/IGC/A191).
The HRQOL of patients who received IP chemotherapy was assessed using the FACT-Ovarian (FACT-O).12 To approximate the difference in health utility value between patients who received IP chemotherapy and patients who received IV chemotherapy, we used a method developed by Bristow et al,13 which was also used by Havrilesky et al.14 The associated health utility value was calculated by dividing the reported FACT-O score by the maximum FACT-O score of 156. Future effects were discounted to their present value by a rate of 1.5%.15
Times Until Events
Times until the different events during the treatment were derived from the health care pathway in our own hospital. Median times until events after the treatment were extracted from literature.16–18 For each patient, we modeled time until recurrence, time until death from EOC, and time until death from natural cause. Time until death from natural cause was estimated using the life expectancy of women in the Netherlands.19 The altered parameters of early-initiated treatment were derived from expert opinion.
A health care perspective was used. All costs are in Euros and are listed in Table 1. Costs of IP and IV chemotherapy were derived in dollars from a cost-effectiveness analysis published by Havrilesky et al.14 They were converted to Euros (US$1 = €0.79 in 2006), and consumer price indices were used to convert these costs to the 2012 price level.19 Costs of an exploratory laparoscopy and cytoreduction were extracted from the passers pricelist.20 Future costs were discounted to their present value by a rate of 4%.15
Base Case Analysis
The 2 treatment strategies were compared in treatment time, costs, effects (in QALYs), and cost-effectiveness. Treatment time was defined as the time between the diagnosis and the completion of all chemotherapy cycles. The incremental cost-effectiveness ratio (ICER) was calculated by dividing the difference in costs by the difference in QALYs. It thereby represents the costs needed to gain 1 QALY. Whether a treatment strategy is deemed cost-effective depends on how much society is willing to pay for a gain in effect, which is referred to as the ceiling ratio. In the Netherlands, an informal ceiling ratio of €80,000 per QALY exists for diseases with a high burden.21
Deterministic Sensitivity Analyses
We performed 1-way deterministic sensitivity analyses to assess the minimal reduction of treatment time necessary to cause a difference in effects between the 2 treatment strategies. We changed the parameters that differ between the strategies over a range of values to assess the potential consequences on the outcome of the model.
Probabilistic Sensitivity Analysis
Probabilistic sensitivity analysis was carried out to address the impact on outcomes resulting from varying multiple parameters simultaneously.22 We varied the health utility values and times until events during treatment. Parameter values were drawn at random from the assigned distributions using Monte Carlo simulation with 1000 iterations. Thereafter, a cost-effectiveness scatter plot was created, and a cost-effectiveness acceptability curve (CEAC) was calculated, which shows the probability that a treatment strategy is cost-effective for a range of ceiling ratios.23
Base Case Analysis
The results of the base case analysis are presented in Table 2. The treatment times of current and early-initiated treatment were 27 and 24 weeks, respectively. Early-initiated treatment yielded 3.42 QALYs (95% confidence interval [CI], 3.22–3.61) per patient, for a total expected health care cost of €25,654 (95% CI, 24,580–26,729) per patient. Current treatment yielded 3.40 QALYs (95% CI, 3.21–3.59) per patient, for a total expected health care cost of €25,607 (95% CI, 24,524–26,691) per patient. This resulted in an ICER of €2592 per QALY gained for early-initiated treatment compared with current treatment.
Deterministic Sensitivity Analyses
Any reduction of the time until exploratory laparoscopy, time until primary surgery, and time until primary chemotherapy resulted in a difference in HRQOL in favor of early-initiated treatment—the smaller the reduction of the times until events, the smaller the differences in effects and costs. The ICERs did not differ from the aforementioned ICER.
Probabilistic Sensitivity Analysis
By performing a probabilistic sensitivity analysis, it was possible to determine the optimal treatment strategy for a range of ceiling ratios given the existing uncertainty. A cost-effectiveness scatter plot is presented in Figure 3. It shows that in all iterations, early-initiated treatment yielded more QALYs than current treatment. In 16% of the iterations, early-initiated treatment was less expensive. The treatment time of early-initiated treatment was shorter than the treatment time of current treatment in 100% of the iterations.
The CEAC is shown in Figure 4. For a ceiling ratio of €6000 per QALY, early-initiated treatment had a probability of 50% of being cost-effective compared with current treatment. For a ceiling ratio of more than €30,000 per QALY, early-initiated treatment had a 100% probability of being cost-effective compared with current treatment.
Resume of Results
Early-initiated treatment resulted in an expected reduction of treatment time of 3 weeks. The average expected difference in QALYs between the 2 treatment strategies was 0.02, in favor of early-initiated treatment. This equals approximately 1 full week in perfect health. Early-initiated treatment was €47 more expensive on average. The reported ICER of €2592 per QALY is far below the ceiling ratio of €80,000. These results show the potential cost-effectiveness of early-initiated treatment compared with current treatment, under the assumption that overall survival, progression-free survival, HRQOL, and costs of the separate events do not differ between the 2 treatment strategies.
Comparison With Literature
To the best of our knowledge, this is the first study to examine the cost-effectiveness of early-initiated treatment of patients with suspected advanced-stage EOC. Several studies have investigated whether the time interval between surgery and chemotherapy has prognostic impact on survival or progression-free survival; contradictory results were reported.24 However, no studies have been published that assessed the impact of shortening the time interval between the diagnosis and the primary treatment of EOC patients.
Strengths and Limitations
The major strength of this study is its timing in an early stage of the treatment strategy’s development. Its results are helpful in designing a strategy with the highest potential and informing policy and research decisions.6 The second major strength is the methodology. We used a decision-analytic model to synthesize all available evidences. If relevant additional evidence becomes available, the model can be easily updated. The third strength is the use of HRQOL (in QALYs) as an outcome measure. Of course, conventional end points such as progression-free and overall survival are very important. However, improvements in survival from ovarian cancer have not been due to cure but the ability of chemotherapy to slow the progression of disease, whereas the survival rates of patients with advanced-stage EOC remain low.25 Because most patients have no prospect of cure, emphasis should be laid on maintaining HRQOL of these patients.26 The advantage of QALYs is that both length and quality of life are combined.
Some potential limitations should also be discussed. First, most health utility values had to be estimated, because in most relevant published ovarian cancer trials, HRQOL was only assessed using cancer-specific instruments. These instruments do not allow the calculation of health utility values. A number of algorithms and estimation methods have come into use by which FACT or QLQ-C30 data can be used to approximate the associated health utility value.27,28 However, none of these methods have been validated for EOC patients. The algorithm to map QLQ-C30 scores was based on female patients with locally advanced breast cancer, whose QLQ-C30 scores were comparable to the scores reported by EOC patients. Further research into EOC-specific utilities is needed to accurately assess the impact of treatments. Currently, a multicenter trial is ongoing, in which health utility values are derived from EOC patients using the EQ-5D.29
Second, limited data were available, for example, about the effects of primary chemotherapy compared with primary surgery and the effects of IP chemotherapy.30,31 Furthermore, because the standard treatment of patients with recurrent EOC is poorly defined, we could not include the costs of additional treatment of these patients. These shortcomings are not specific for the present study. It is an acknowledged challenge of early evaluations to deal with lack of evidence.32 However, it is better to inform decisions with the available evidence under conditions of uncertainty than without any evidence at all. Since overall survival, progression-free survival, HRQOL, and costs of the separate events were assumed to be equal for the 2 treatment strategies, these shortcomings do not influence our conclusions. If additional evidence becomes available, the model can be easily updated.
Third, known prognostic factors for survival such as the histologic diagnosis of the tumor, tumor grade, performance status, and the addition of bevacizumab have not been accounted for in this study because this was not necessary to the answer or research question.17,33–35 If needed in the future, these factors can be added to the model relatively easily.
Implications for Clinical Practice
The results of the present study show that a relatively simple organizational change in treatment strategy may lead to additional QALYs. We assumed that survival and HRQOL do not differ between the treatment strategies. However, early-initiated treatment may result in less worry and anxiety and may also result in treating patients in a better general condition, because the condition of advanced-stage EOC patients tends to decrease fast. All of which may lead to improved HRQOL. In addition, an earlier start of the treatment may result in better oncological outcomes. Most advanced-stage EOC patients present with a short history of feeling unwell (4–8 weeks in general), indicating rapid tumor growth. One may presume that installing treatment some weeks earlier may give better oncological results such as a higher percentage of complete cytoreduction, resulting in prolonged overall and progression-free survival. Therefore, we believe that the reported difference in QALYs may be an underestimation.
Early-initiated treatment may also increase patient satisfaction. Tools that measure patient satisfaction may even discriminate better between the 2 treatment strategies than measurements of HRQOL, such as the EQ-5D. Therefore, in addition to HRQOL, future studies on this subject should also assess patient satisfaction before, during, and after primary treatment because health utility values may not show all the benefits of early-initiated treatment.
The present study focused on advanced-stage EOC patients treated in the Netherlands. However, we have synthesized evidences from various sources and settings. Therefore, we believe that the results of this study are generalizable to patients in other hospitals in other countries where time gaps exist between the diagnosis, surgery, and the initiation of chemotherapy. It is questionable whether the results are also generalizable to a population of patients with suspected early-stage EOC. Because of the significant differences in survival and complexity of treatment, we also expect HRQOL to differ between the groups. Further studies are needed to investigate the possible effects of early-initiated treatment of patients with suspected early-stage EOC.
Given the current evidence, early-initiated treatment of patients with suspected advanced-stage EOC leads to additional QALYs and seems to be cost-effective compared with current treatment, under the main assumption that overall survival, progression-free survival, HRQOL, and costs of the separate events do not differ between the 2 treatment strategies.
1. Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin. 2011; 61: 69–90.
3. FIGO Committee on Gynecologic Oncology. Current FIGO staging for cancer of the vagina, fallopian tube, ovary, and gestational trophoblastic neoplasia. Int J Gynaecol Obstet. 2009; 105: 3–4.
4. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin. 2012; 62: 10–29.
5. Mariotto AB, Yabroff KR, Shao Y, et al. Projections of the cost of cancer care in the United States: 2010–2020. J Natl Cancer Inst. 2011; 103: 117–128.
6. Carlsson P, Jorgensen T. Scanning the horizon for emerging health technologies. Conclusions from a European workshop. Int J Technol Assess Health Care. 1998; 14: 695–704.
7. Caro JJ. Pharmacoeconomic analyses using discrete event simulation. Pharmacoeconomics. 2005; 23: 323–332.
8. van Altena AM, van den Akker PAJ, de Hullu JA, et al. The efficacy of a regional network for ovarian cancer
care: a population-based study. In: van Altena AM, ed. Improving Quality of Clinical Care in Epithelial Ovarian Cancer
Patients. Nijmegen, the Netherlands: IPSKamp; 2012.
9. Weinstein MC, Torrance G, McGuire A. QALYs: the basics. Value Health. 2009; 12 (suppl 1): S5–S9.
10. Krasner CN, Poveda A, Herzog TJ, et al. Patient-reported outcomes in relapsed ovarian cancer
: results from a randomized phase III study of trabectedin with pegylated liposomal doxorubicin (PLD) versus PLD alone. Gynecol Oncol. 2012; 127: 161–167.
11. Crott R, Briggs A. Mapping the QLQ-C30 quality of life
cancer questionnaire to EQ-5D patient preferences. Eur J Health Econ. 2010; 11: 427–434.
12. Wenzel LB, Huang HQ, Armstrong DK, et al. Health-related quality of life
during and after intraperitoneal versus intravenous chemotherapy for optimally debulked ovarian cancer
: a Gynecologic Oncology Group study. J Clin Oncol. 2007; 25: 437–443.
13. Bristow RE, Santillan A, Salani R, et al. Intraperitoneal cisplatin and paclitaxel versus intravenous carboplatin and paclitaxel chemotherapy for stage III ovarian cancer
: a cost-effectiveness
analysis. Gynecol Oncol. 2007; 106: 476–481.
14. Havrilesky LJ, Secord AA, Darcy KM, et al. Cost-effectiveness
of intraperitoneal compared with intravenous chemotherapy for women with optimally resected stage III ovarian cancer
: a Gynecologic Oncology Group study. J Clin Oncol. 2008; 26: 4144–4150.
15. Health Care Insurance Board. Guidelines for Pharmaco-Economic Research (in Dutch). Diemen, the Netherlands: Health Care Insurance Board; 2011.
16. Armstrong DK, Bundy B, Wenzel L, et al. Intraperitoneal cisplatin and paclitaxel in ovarian cancer
. N Engl J Med. 2006; 354: 34–43.
17. du Bois A, Reuss A, Pujade-Lauraine E, et al. Role of surgical outcome as prognostic factor in advanced epithelial ovarian cancer
: a combined exploratory analysis of 3 prospectively randomized phase 3 multicenter trials: by the Arbeitsgemeinschaft Gynaekologische Onkologie Studiengruppe Ovarialkarzinom (AGO-OVAR) and the Groupe d’Investigateurs Nationaux Pour les Etudes des Cancers de l’Ovaire (GINECO). Cancer. 2009; 115: 1234–1244.
18. Vergote I, Trope CG, Amant F, et al. Neoadjuvant chemotherapy or primary surgery in stage IIIC or IV ovarian cancer
. N Engl J Med. 2010; 363: 943–953.
19. Statistics Netherlands. Available at: http://www.cbs.nl
. Accessed August 7, 2013.
21. Council for Public Health and Health Care. Sensible and Sustainable Care (in Dutch). Zoetermeer, the Netherlands: Council for Public Health and Health Care; 2006.
22. Briggs AH, Weinstein MC, Fenwick EA, et al. Model parameter estimation and uncertainty: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force–6. Value Health. 2012; 15: 835–842.
23. van Hout BA, Al MJ, Gordon GS, et al. Costs, effects and C/E-ratios alongside a clinical trial. Health Econ. 1994; 3: 309–319.
24. Mahner S, Eulenburg C, Staehle A, et al. Prognostic impact of the time interval between surgery and chemotherapy in advanced ovarian cancer
: analysis of prospective randomised phase III trials. Eur J Cancer. 2013; 49: 142–149.
25. Huang L, Cronin KA, Johnson KA, et al. Improved survival time: what can survival cure models tell us about population-based survival improvements in late-stage colorectal, ovarian, and testicular cancer? Cancer. 2008; 112: 2289–2300.
26. Hess LM, Stehman FB. State of the science in ovarian cancer quality of life
research: a systematic review. Int J Gynecol Cancer. 2012; 22: 1273–1280.
27. Hess LM, Brady WE, Havrilesky LJ, et al. Comparison of methods to estimate health state utilities for ovarian cancer
using quality of life
data: a Gynecologic Oncology Group study. Gynecol Oncol. 2012.
28. Pickard AS, Ray S, Ganguli A, et al. Comparison of FACT- and EQ-5D–based utility scores in cancer. Value Health. 2012; 15: 305–311.
29. Rutten MJ, Gaarenstroom KN, Van Gorp T, et al. Laparoscopy to predict the result of primary cytoreductive surgery in advanced ovarian cancer
patients (LapOvCa-trial): a multicentre randomized controlled study. BMC Cancer. 2012; 12: 31.
30. Morrison J, Haldar K, Kehoe S, et al. Chemotherapy versus surgery for initial treatment in advanced ovarian epithelial cancer. Cochrane Database Syst Rev. 2012; 8: CD005343.
31. Heintz AP, Odicino F, Maisonneuve P, et al. Carcinoma of the ovary. FIGO 26th Annual Report on the Results of Treatment in Gynecological Cancer. Value Health. 2006; 95 (suppl 1): S161–S192.
32. Greimel E, Daghofer F, Petru E. Prospective assessment of quality of life
in long-term ovarian cancer
survivors. International Journal of Cancer. 2011; 128: 3005–3011.
33. Greimel ER, Bjelic-Radisic V, Pfisterer J, et al. Randomized study of the Arbeitsgemeinschaft Gynaekologische Onkologie Ovarian Cancer
Study Group comparing quality of life
in patients with ovarian cancer
treated with cisplatin/paclitaxel versus carboplatin/paclitaxel. Journal of Clinical Oncology. 2006; 24: 579–86.
For the complete list of references, please contact J.Grutters@ebh.umcn.nl.