Cost-effectiveness plane: Mohs micrographic surgery vs surgical excision for primary basal cell carcinoma.
Cost-effectiveness plane: Mohs micrographic surgery vs surgical excision for recurrent basal cell carcinoma.
Acceptability curve of Mohs micrographic surgery (MMS) vs surgical excision for primary basal cell carcinoma.
Acceptability curve of Mohs micrographic surgery (MMS) vs surgical excision for recurrent basal cell carcinoma.
Essers BAB, Dirksen CD, Nieman FHM, Smeets NWJ, Krekels GAM, Prins MH, Neumann HAM. Cost-effectiveness of Mohs Micrographic Surgery vs Surgical Excision for Basal Cell Carcinoma of the Face. Arch Dermatol. 2006;142(2):187-194. doi:10.1001/archderm.142.2.187
Copyright 2006 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2006
To assess the cost-effectiveness of Mohs micrographic surgery (MMS) compared with the surgical excision for both primary and recurrent basal cell carcinoma (BCC).
A cost-effectiveness study performed alongside a prospective randomized clinical trial in which MMS was compared with surgical excision.
The study was carried out from 1999 to 2002 at the dermatology outpatient clinic of the University Hospital Maastricht, Maastricht, the Netherlands.
A total of 408 primary (374 patients) and 204 recurrent (191 patients) cases of facial BCC were included.
Main Outcome Measures
The mean total treatment costs of MMS and surgical excision for both primary and recurrent BCC and the incremental cost-effectiveness ratio, calculated as the difference in costs between MMS and surgical excision divided by their difference in effectiveness. The resulting ratio is defined as the incremental costs of MMS compared with surgical excision to prevent 1 additional recurrence.
Compared with surgical excision, the total treatment costs of MMS are significantly higher (cost difference: primary BCC, €254; 95% confidence interval, €181-€324; recurrent BCC, €249; 95% confidence interval, €175-€323). For primary BCC, the incremental cost-effectiveness ratio was €29 231, while the ratio for recurrent BCC amounted to €8094. The acceptability curves showed that for these ratios, the probability of MMS being more cost-effective than surgical excision never reached 50%.
At present, it does not seem cost-effective to introduce MMS on a large scale for both primary and recurrent BCC. However, because a 5-year period is normally required to determine definite recurrence rates, it is possible that MMS may become a cost-effective treatment for recurrent BCC.
Basal cell carcinoma (BCC) is the most common form of nonmelanoma skin cancer worldwide. In the Netherlands, an estimated 30 000 new cases of BCC are diagnosed annually.1 Although it is not a life-threatening disease, recurrence of this predominantly facial tumor may cause considerable morbidity related to functional and aesthetic problems. Prevention of a recurrence, particularly within the facial area, is therefore an important goal in the treatment of BCC because it may have a considerable impact on a patient's quality of life.
The main conclusion of a systematic review on recurrence rates after treatment for primary BCC was that recurrence rates for different therapies could not be compared because of a lack of uniformity in the method of reporting.2 However, Mohs micrographic surgery (MMS) appeared to have the lowest recurrence rate, followed by surgical excision. Still, MMS is also a more labor-intensive and thus a more costly procedure. To assess whether MMS is a cost-effective treatment for BCC, it is necessary to relate costs to effectiveness. We report herein the results of our cost-effectiveness study, performed alongside what is to our knowledge the first prospective randomized clinical trial, in which MMS was compared with surgical excision for both primary and recurrent BCC.
The following inclusion criteria for patients with a primary facial BCC were chosen: a histologically proven tumor larger than 1 cm, located in the H-zone of the face or of an aggressive histopathologic subtype. For patients with recurrent facial BCC, first- and second-time cases were included. The exclusion criterion was a life expectancy of less than 3 years, based on expert opinion of the dermatologist. After patients signed the informed consent, separate randomization procedures for both primary and recurrent BCC were performed. Because inclusion and randomization were performed per BCC case, patients with more than 1 BCC could be included multiple times. The clinical trial was designed to have a 90% power to detect a 6.5% difference in the recurrence rate between MMS and surgical excision for primary BCC and a 13.5% difference for recurrent BCC at the 5% level. Full details and results of the clinical trial are described elsewhere.3
All surgical procedures were performed under local anesthesia at the outpatient clinic. Mohs micrographic surgery is a specialized technique by which the tumor is removed layer by layer. During the first MMS stage, a bowl-shaped specimen is obtained and processed into horizontal frozen sections. These sections are histopathologically examined, and if tumor cells are found, a second MMS stage takes place. The procedure will be repeated until the entire area is tumor free. Surgical excision is a technique by which the obtained specimen is histopathologically examined by the bread-loafing (vertical sections) technique or by the quadrant method. In case of positive margins, a reexcision is performed.3
The primary clinical outcome of both studies was recurrence of a BCC. Secondary outcomes were quality of life, anxiety, and cost-effectiveness. Quality of life and anxiety were assessed at baseline and 6 months postoperatively by means of a generic instrument (the Dutch version of the Nottingham Health Profile [NHP]) and a domain-specific questionnaire (the Dutch version of the State-Trait Anxiety Inventory [STAI]).4,5 The NHP is a 38-item questionnaire that produces binary responses to a series of items that are related to the following 6 dimensions: physical mobility, sleep, bodily pain, energy level, emotional reactions, and social isolation. Each dimension of the NHP gives a score between 0 and 100, with 0 being the best and 100 being the worst quality-of-life outcome. The assumption was that primarily the emotional reactions domain would show a difference between the MMS and surgical excision.
The Spielberger STAI consists of a 20-question assessment scale that measures to what extent personality traits are susceptible to anxiety and a 20-question assessment scale to measure anxiety in a certain state. The STAI gives a score between 20 and 80, with 20 indicating no anxiety and 80, a high anxiety level. In our study, only the STAI was used to determine whether patients undergoing MMS showed a significant difference in anxiety level at baseline and 6 months postoperatively compared with the group undergoing surgical excision.
The economic evaluation was performed from a hospital perspective because no difference in out-of-pocket costs or use of other health care and non–health care services between the 2 surgical treatments was expected. In addition, the mean age for patients with this type of skin cancer is approximately between 60 and 65 years and therefore productivity losses were expected to be minimal. The costs and effects for primary BCC were collected during a period of 30 months while the follow-up period for recurrent BCC lasted 18 months. For both primary and recurrent BCC, costs and effects of MMS were compared with the costs and effects of surgical excision. Given the time horizon of both studies, costs and effects occurring after 1 year were discounted at 4%.6 The incremental cost-effectiveness ratio (ICER) was calculated as the difference in costs divided by the difference in effectiveness between MMS and surgical excision. The result of this calculation was defined as the incremental costs of MMS to prevent an additional recurrence. For the cost analysis, real resource consumption was measured and related to their costs, as recommended by the US panel on cost-effectiveness in health and medicine. This approach is also known as “microcosting,” which is a detailed inventory and measurement of resources consumed. A major advantage of microcosting is that it allows others to see how well the analysis matches their own situation in which patterns of care may differ.7
All volumes of use were based on the hospital information system and empirical time registrations. Unit costs were derived from the hospital financial department. Costs were calculated by multiplying volumes of use with the costs per unit. Direct costs of both treatments contained the personnel and material costs of all diagnostic procedures, surgery, and all outpatient visits. Indirect costs included the general hospital overhead, which was allocated to the direct costs as an overall percentage of 35%, according to the Dutch guidelines for cost calculation.6 All costs are presented in 2001 euros (€1 = US $0.89; €1 = £0.69). Table 1 provides an overview of the costs per unit.
The preoperative costs consisted of a diagnostic procedure and a preoperative outpatient visit. The personnel costs for surgery were determined by using empirical time registrations, whereby for every procedure (MMS or surgical excision), the dermatologist who performed the surgery registered the exact starting and ending time. This allowed us to calculate the mean personnel costs based on a substantial number of procedures. Total operative costs included not only the costs of the surgical procedure but also the costs of a possible reexcision as well as processing and examination of the histopathologic slides. Based on expert opinion, it was assumed that complications would lengthen the usual duration of a control visit by 50%, and thus the costs of complications were calculated as half the costs of a postoperative control visit. The costs of any additional cosmetic treatments were already included in the personnel and material theater costs if they were performed during the surgical procedure. However, if a cosmetic treatment, such as laser therapy, was performed after surgery, those costs were calculated separately based on cost information as provided by the hospital financial department. The costs of postoperative control visits consisted of 5 regular outpatient visits and any additional control visits.
Univariate sensitivity analyses were undertaken by varying a cost or an effectiveness parameter one at a time to investigate the impact on the ICER. First, the personnel costs were varied by taking the upper and lower bound of the 95% confidence interval (CI) of the procedure time. Second, the effect difference between MMS and surgical excision as calculated for the power analysis at the start of the study was used. Third, the actual recurrence rates after a follow-up of 1.5 and 2.5 years for recurrent and primary BCC, respectively, were extrapolated to a 5-year period to estimate the effect difference between the 2 treatments and subsequently the 5-year ICER. The assumption was that, within the observed follow-up period, 70% of the recurrences for the primary and 50% of the recurrences of the recurrent carcinomas had occurred.8,9
For statistical analysis, SPSS-PC for Windows version 11.0 (SPSS Inc, Chicago, Ill) was used. P<.05 was considered to indicate statistical significance. Data of the cost analysis were analyzed by the intention-to-treat principle in which mean substitution for the missing cases was used. Data that combined both costs and effectiveness were, similar to the clinical results, analyzed by modified intention to treat in which the missing cases were excluded.3 The χ2 test was used to calculate log-likelihood. Regression analysis was used to investigate the effect of location and histopathologic subtype on the cost difference between MMS and surgical excision. If nonnormal distributions were present, log-transformation of the costs was applied, after which a Kolmogorov-Smirnov test was used to test for normality. Confidence intervals for the mean differential costs were obtained by the bootstrap method. This method estimates the sampling distribution of a statistic through a large number of simulations, based on sampling with replacement from original data.10 To account for the uncertainty surrounding the ICERs, bootstrapping was also used (1000 replications).11 Results of the bootstrap analysis for the ICERs were presented in cost-effectiveness planes. A cost-effectiveness plane is a graphical presentation of 4 situations or quadrants in which additional costs and additional health outcome effects of a new therapy are compared with the standard therapy. Furthermore, acceptability curves were calculated that show the probability of a new therapy being more cost-effective than the usual treatment for various threshold values, in this case being the maximum amount a hospital would be willing to pay extra for MMS to prevent 1 additional recurrence.
Quality-of-life scores were tested for normality by the Kolmogorov-Smirnov test. Differences in time of quality-of-life scores between MMS and surgical excision were analyzed by either a Mann-Whitney test or the independent group unpaired, 2-tailed t test.
There were 408 cases of primary BCCs among 374 patients. The mean (SD) age of the patients was 67.7 (12.7) years, and 60% were male. There were 190 patients with 204 recurrent BCCs. The mean (SD) age of this group was 67.9 (11.7) years and 58% were male.
A total of 408 cases of primary BCC were included, of which 204 were randomized to surgical excision and 204 to MMS. For the 204 recurrent BCCs, 102 were randomized to surgical excision and 102 to MMS. At 30 months' follow-up within the primary BCC group, 5 recurrences were found after surgical excision and 3 after MMS. Thirty-three patients (16%) from the surgical excision group and 44 patients (22%) from the MMS group were lost to follow-up.
For the recurrent BCC group, at 18 months' follow-up no recurrences were found after MMS, whereas 3 were found recurrences after surgical excision. Nine patients (9%) from the surgical excision group and 7 patients (7%) from the MMS group were lost to follow-up.
For the primary BCCs, no difference in postoperative complications between surgical excision and MMS was seen (28 [14%] vs 24 [12%]), whereas for recurrent BCCs more complications occurred after surgical excision than after MMS (19 [19%] vs 8 [8%]; P = .02)
As given in Table 2, the number of patients included in the quality-of-life study is based on a subset of the total patient population owing to a number of reasons. First, because inclusion was done per BCC case, a patient with multiple BCCs could be randomized to both surgical excision and MMS. However, to determine whether there was a difference in quality of life between patients undergoing MMS or surgical excision, only patients with a single BCC were interviewed. Second, a number of patients were lost to follow-up owing to administrative restrictions. Given that only 1 researcher was available to administer the questionnaires after the visit to the dermatologist, patients often had to wait in turn before they could be interviewed. A number of patients refused to do so, because they either had commitments elsewhere or were not willing to wait.
The mean scores preoperatively and postoperatively for the NHP and STAI (Table 2) show that patients reported a good health-related quality of life and a minimum level of anxiety. There was no statistically significant difference in NHP scores at 6 months after surgery compared with baseline between MMS and surgical excision for both the primary and recurrent BCC groups. In addition, there was no statistically significant difference in anxiety level at 6 months compared with baseline between the 2 treatments.
Table 3 describes the mean (SD) costs per BCC. When comparing MMS with surgical excision, the statistically significant differences were first related to the personnel costs of surgery because of longer theater time (mean [SD] theater time, MMS vs surgical excision: primary BCC, 155  minutes vs 60  minutes; recurrent BCC, 190  minutes vs 91  minutes) and second to costs of pathologic examination. Both cost components were significantly higher for MMS, which resulted in a mean difference in total costs of €254 per primary BCC and €249 per recurrent BCC.
Additional analyses (Table 4 and Table 5) were performed to test whether the cost difference between MMS and surgical excision was different for specific locations or nonaggressive or aggressive histologic types. Dermatologists (New York University Medical Center, New York) distinguished 7 facial locations: (1) forehead/temporal area, (2) cheek/chin, (3) nose/perinasal area, (4) lips, (5) periocular area, (6) ears, and (7) periauricular area.3 Results for the primary BCC indicated that interactions between treatment and location (location × therapy, F6,324 = 0.478; P = .82) or histologic subtype (aggressive subtype × therapy, F1,324 = 0.051; ; P = .82) or both (location × aggressive subtype × therapy, F6,324 = 1.072; P = .38) were not statistically significant. This means that the difference in costs between MMS and surgical excision applies for all locations and persists regardless of whether histologic type is aggressive or not. Results for the recurrent BCC also demonstrated that none of the interactions was statistically significant (location × therapy, F6,181 = 0.441; P = .85; aggressive subtype × therapy, F1,181 = 0.05; P = .82; and location × aggressive subtype × therapy, F3,181 = 0.31; P = .82).
As given in Table 6, the difference in recurrence rate between MMS and surgical excision for primary BCC at 30 months is 0.0091. When the mean difference in costs is divided by this effect difference, the incremental cost is €29 231 per recurrence avoided. For recurrent BCC, the effect difference after 18 months' follow-up is 0.032, which gives an ICER of €8094 per recurrence avoided. To estimate the uncertainty surrounding the ICERs, cost-effectiveness planes and acceptability curves were calculated.
The cost-effectiveness plane for primary BCC (Figure 1) shows that 69% of all ICERs are located in the quadrant where MMS is more effective but also more costly, while 31% is in the quadrant where MMS is inferior (ie, has higher costs and lower effectiveness). For recurrent BCC (Figure 2), all cost-effectiveness pairs are situated within the quadrant where MMS is more effective but also more costly.
The acceptability curve for primary BCC (Figure 3) shows that, for the calculated ICER of €29 231, the probability of MMS being cost-effective is 40%. For recurrent BCC (Figure 4), the acceptability curve demonstrates that based on the ICER of €8094, the probability that MMS is cost-effective amounts to 30%.
The results of the sensitivity analysis (Table 7) demonstrate that taking the upper or lower bound of the 95% CI of the surgical procedure time has a minor effect on the incremental costs for both primary and recurrent BCC. The difference in effectiveness between both treatments has a significant impact on the ICERs. Based on the effect difference as calculated for the power analysis of the study, the ICER for primary BCC decreases to €3877 per recurrence avoided, while the ratio for recurrent BCC falls to €1844. Extrapolating the data of the actual effect difference to a 5-year follow-up period shows that, in particular, the ratio for the recurrent carcinomas is reduced substantially to €4047.
To our knowledge, this is the first study investigating the cost-effectiveness of MMS compared with surgical excision for primary and recurrent BCC. Although retrospective research by Cook and Zitelli12 and a recent prospective study by Bialy et al13 compared the costs of MMS and surgical excision, they did not relate the effectiveness of both treatments to their costs. As such, they describe a cost-comparison analysis. Moreover, both include various types of nonmelanoma skin cancer, whereas our study focuses exclusively on primary and recurrent BCC. Another study reported total costs of nonmelanoma skin cancer treatment in general but did not calculate costs of the separate treatment modalities.14
Results of the clinical trial showed that the effect difference between the 2 surgical treatments was small for both groups. A possible explanation for the lower recurrence rate after surgical excision is the fact that all incompletely excised BCCs were re-treated immediately.3
The quality-of-life results demonstrated that there was no statistically significant difference at 6 months compared with baseline between patients undergoing MMS or surgical excision. However, it is possible that a generic measure such as the NHP or even a domain-specific measure such as the STAI is not sensitive enough to detect a difference in treatment effect for patients with this type of skin cancer. Because BCC often is localized in the facial area, it is likely that, for instance, psychosocial or functional issues resulting from the skin cancer and its treatment are more relevant. Therefore, we developed, parallel to the clinical trial, a disease-specific questionnaire focused on the perceptions of patients regarding the health and aesthetics of their facial skin. The outcomes of this study are described in a separate article.15
The results of the cost analysis showed that the mean costs of MMS were significantly higher, mainly due to longer surgery time and higher costs of pathologic examination. More detailed analyses revealed that this cost difference persisted, irrespective of histologic subtype and location.
When the costs of both surgical treatments were related to their effectiveness, an ICER of €29 231 per recurrence avoided was found for primary BCC and the ICER for recurrent BCC was €8094. The acceptability curves for the primary and recurrent BCC demonstrated that the probability of MMS being more cost-effective then surgical excision never reached 50% unless the hospital would be willing to pay even more than the calculated ICERs. The cost-effectiveness planes showed that most ratios were located in the quadrant where an increased effectiveness of MMS is only achieved at higher costs. In such a situation, the decision to replace surgical excision by MMS depends first on the threshold value (in this case, the maximum amount the hospital would be willing to pay extra to prevent 1 additional recurrence) and second, whether the estimated ICER lies below this threshold value.16 However, a maximum amount with respect to the treatment of BCC has not yet been determined, which makes interpretation and comparison of the ICERs difficult. This is often a problem when applying a disease-specific outcome measure instead of a more common measure for effectiveness such as quality-adjusted life year.17 Quality-adjusted life year was not used as an outcome measure because BCC and its treatment have no effect on life expectancy. In addition, it was expected that treating BCC had no substantial effect on general quality of life. Nevertheless, the ICERs of this study can be put into perspective by considering the consequences of recurrent BCC from a hospital perspective. It is likely that a patient with recurrent BCC will undergo a reoperation in the future. Therefore, it can be argued that an acceptable threshold value should at least include the hospital costs of repairing a recurrence. Furthermore, the value should account for the fact that not all recurrences can be repaired in the long run and that some patients will develop a second or third recurrence. Finally, although we did not examine patient perspective, it is possible that a patient's willingness to pay to avoid a recurrence may provide important information with respect to a reasonable threshold value.
Still, even when a recurrent BCC is repaired at a cost 3 times that of MMS (ie, €3438), the ICERs for primary and recurrent BCC are still too high to recommend a broad implementation of MMS. When the threshold value of €3438 is subsequently compared with the estimated ICERs for each location, it appears that, except for primary BCC of the ears and recurrent BCC of the cheek, MMS is not likely to be considered a cost-effective treatment. Furthermore, although the ICERs for primary BCC of the ears and recurrent BCC of the cheek seem more favorable, they should be interpreted with caution because they are based on small subgroups. However, when the threshold value is compared with the ICERs per histologic subtype, MMS might well be a cost-effective treatment for the group of aggressive recurrent BCC.
As shown by the sensitivity analysis, the ICERs for both primary and recurrent BCC were largely affected by the difference in recurrence rate. Because a 5-year period is normally required to determine definite recurrence rates, it is possible that in the future the ICER of MMS for recurrent BCCs may become more acceptable.
Correspondence: Brigitte A. B. Essers, MHSc, Department of Clinical Epidemiology and Medical Technology Assessment, University Hospital Maastricht, P. Debyelaan 25, PO Box 5800, 6202 AZ Maastricht, the Netherlands (email@example.com).
Financial Disclosure: None.
Accepted for Publication: May 18, 2005.
Author Contributions:Study concept and design: Dirksen, Krekels, and Neumann. Acquisition of data: Smeets and Essers. Analysis and interpretation of data: Dirksen, Essers, Prins, and Nieman. Drafting of the manuscript: Essers. Statistical analysis: Nieman. Critical revision of the manuscript for important intellectual content: Dirksen, Essers, Prins, Nieman, Krekels, Smeets, and Neumann. Obtained funding: Dirksen, Krekels, Neumann. Ms Essers had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Funding/Support: This study was financed by the Dutch fund for Investigative Medicine, a governmental institution financing research to improve health care in the Netherlands.
Role of the Sponsor: The sponsor had no role in the design and conduct of the study, in the collection, analysis and interpretation of the data, or in the preparation of the manuscript, review or approval of the manuscript.