Cost Effectiveness of Laparoscopic Versus Mini-Incision Open Donor Nephrectomy: A Randomized Study

Abstract

Background. 

Cost-effectiveness remains an issue surrounding the introduction of laparoscopic donor nephrectomy (LDN).

Methods. 

In a randomized controlled trial the cost-effectiveness of LDN versus mini-incision open donor nephrectomy (ODN) was determined. Fifty donors were included in each group. All in-hospital costs were documented. Postoperatively, case record forms were sent to the donors during 1-year follow-up to record return-to-work and societal costs. To offset costs against quality of life, the Euroqol-5D questionnaire was administered preoperatively and 3, 7, 14, 28, 90, 180, and 365 days postoperatively.

Results. 

Mean total costs were €6,090 (US$7,308) after LDN and €4,818 ($5,782) after ODN (P<0.001). Disposables influenced the cost difference most. Mean productivity loss was 68 and 75 days after LDN and ODN respectively, corresponding to €783 ($940) gained per donor after LDN. The main gain in quality of life in the LDN group was realized within 4 weeks postoperatively. LDN resulted in a mean gain of 0.03 quality-adjusted life years at mean costs of €1,271 ($1,525) and €488 ($586) from a healthcare perspective and a societal perspective, respectively. This implies that one additional Quality-Adjusted Life Year after LDN costs about €16,000 ($19,200) from a societal point of view and about €41,000 ($49,200) from a health-care perspective. Activities other than work were resumed significantly earlier after LDN (66 vs. 91 days, P=0.01).

Conclusion. 

In addition to a clinically relevant donor-experienced benefit from LDN, this technique appeared, given a societal perspective, a cost-efficient procedure mainly due to less productivity losses.

Live kidney donor transplantation has become the most realistic option to expand kidney transplantation and has several benefits over transplantation of kidneys from deceased donors including shorter waiting lists, better initial graft function, and longer graft survival (¹). From the perspective of the donor, laparoscopic donor nephrectomy (LDN) has become the preferred technique to procure the kidney (^2–5).

In general, the widespread introduction of a novel surgical technique is only acceptable if the technique offers an equal or better treatment to patients, if the technique can be mastered relatively easily and if the technique is cost efficient. With regard to the first aspect, evidence has mounted that LDN provides clinically relevant benefits to the donor including less pain, earlier convalescence, and superior quality of life (^2–7). After early concern with regard to graft function after LDN, recent reports showed similar graft function after laparoscopic and open donor nephrectomy (^{4, 8}). With regard to the second aspect, in the United States more than 60% of the live donor kidneys is procured laparoscopically with acceptably low mortality and morbidity rates (⁹). In Western Europe, LDN has been introduced by more than 40% of the transplant centers (¹⁰).

The missing keystone impeding widespread acceptance of laparoscopic surgery for live kidney donation is the cost-efficiency issue. The assumption that LDN is associated with higher costs and the idea that the additional time and money spent on LDN are not rewarding are common reasons to stick to open techniques (¹⁰). We evaluated the cost effectiveness of LDN versus mini-incision open donor nephrectomy (ODN) alongside a randomized, blinded clinical trial.

PATIENTS AND METHODS

Patients

Live kidney donors at the University Medical Centers in Rotterdam and Nijmegen in the Netherlands were considered eligible for participation in the Living Donors trial. The medical ethics committees of both hospitals approved the study protocol. In this study, a complete laparoscopic approach for live kidney donation was applied to 50 donors, whereas 50 others underwent a mini-incision muscle-splitting open approach. Four experienced surgeons supervised all procedures. Details of donors, recipients, screening, surgical procedures and perioperative analgesia have been published elsewhere (³).

Briefly, all donors scheduled for donor nephrectomy on the following day were randomized using sealed opaque envelopes. To further guarantee opacity, the papers that contained the procedure to be performed and the trial number were double-folded twice. The trial statistician had created a computer-generated randomization list. There was no stratification per center. Donors had to have adequate knowledge of the Dutch language and had to give informed consent. All personnel outside the operation team was kept unaware of the surgical approach using a standard pattern of blood stained dressings. These dressings were removed at discharge only. Donors were only discharged from the hospital if they resumed a normal diet and were able to walk stairs.

To harvest the kidney during LDN an endobag was inserted via a Pfannenstiel incision. To harvest the kidney in ODN, a 10- to 12-cm skin incision was made anterior to the 11th rib with subsequent splitting of the oblique and transverse abdominal muscles. A mechanical retractor (Omnitract Surgical, St. Paul, MN) provided the required working space.

Data Collection

Direct Treatment Costs

Costs prior to admission to the hospital including charges of screening and imaging were not taken into account as these were similar for both techniques. Direct treatment costs constituted of personnel, material, and capital costs such as: total operating time, hospital days (prime and recurrent), capital costs associated with the operating theater for both procedures (monitors, endoscopic tower, etc.), personnel costs such as surgical costs (surgeon, assistant), anesthetics costs (anesthesiologist, assistant), operating nurses and material costs (disposables). Furthermore outpatient visits, general practitioner consultations, and home care related costs were included. Hospital overhead costs were added as a fixed percentage (35%) to the costs of the personnel in the operation room. Unit resource prices were based on guideline prices according to the Dutch Insurance Board to improve generalization of the results and limit dependence on local negotiations between healthcare instances and the insurance companies (i.e., standard prices were available for the rate of a day in a university hospital, the rate for a visit to the general practitioner) (¹¹). If these prices were not available real costs prices were determined (i.e. the costs of the endobag to retrieve the kidney). The trial coordinator in either center attended all operations from the arrival of the donor until leave to the recovery room. Together with the scrub nurse use of all instruments and other items such as the sheets to cover the operation field were documented. Costs of sterilization of reusable instruments were calculated. Depreciation of hardware used during the procedures such as monitors, the endoscopic tower and the mechanical retractor was included. Salaries of anesthesiologist, surgeons, nurses and supporting personnel in the operation room were all included and expressed as a function of the time spent in the operation room.

At the surgical ward, a standard price was counted for every day spent in (an academic) hospital according to the national guideline (¹¹). All costs related to the readmission of single donor including administration of intravenous antibiotics were added to the total in-hospital costs.

Use of health care resources during follow-up were registered using case record forms that were administered preoperatively, 1, 2, 4, and 6 weeks postoperatively, and every 2 weeks thereafter until complete return to preoperative activity. Donors were asked whether or not they had housekeeping or home care, whether or not they visited their general practitioner or a doctor in the hospital, or the first aid department. Participants were also asked to quantify these items. For these resources, guideline prices were available. These data were cross-checked with data from the electronic patient files from the hospital. If donors did not reply to the case record forms, we sent a reminder. If they did not respond to the reminder, we called them on the phone. We did not attempt to incorporate charges related to the recipient including dialysis, renal transplantation, and immunosuppressive therapy of recipients.

Societal Costs

Additional burden to society was quantified by calculating productivity losses for donors who performed paid labor preoperatively using the friction costs method. The friction costs method assumes that an employee on sick leave is replaced if this sick leave takes too long. The friction costs period was set at 154 days on average based on 2003 figures. This implies that an employee on sick leave for more than 154 days would have been replaced. Productivity loss was fixed at an average of 80% of €40 ($48) per hour. With the aforementioned case-record forms data were gathered on preoperative occupation, including whether this occupation was physically demanding or not and whether the job was part-time or not. Up to every lost hour, productivity losses were documented, therewith also addressing those donors who started working part-time first and resumed regular working hours later. Because the donor’s physical condition is known not be the only factor determining return to work in a previous study the day that patients resumed 90% and 100% of daily activities was also determined (¹²). We did not give any advice to the donors when to resume work.

Cost-Utility Analysis

Quality of life (QOL) during follow-up, measured on the Euroqol-5D was defined as utility (¹³). QOL of both surgical modalities was evaluated preoperatively and 3, 7, 14, 28 days and 3, 6, 12 months postoperatively. Health status was described according to five attributes: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each attribute had three levels: no problems, some problems, or severe problems. From these EQ-5D scores, Quality- Adjusted Life Years (QALYs) were derived using the trapezium rule. One QALY means one year in good health. The trapezium rule is a way to approximately calculate the definite integral over a function f(x). The trapezium rule applied to QALYs is conducted by approximating the region under the function QALY = f(QALY*t) by a trapezium and calculating its area, In this equation t expresses the aforementioned Euroqol-5D measurement points at baseline. A 0–100 VAS scale measuring health (with 0 indicating very poor health and 100 indicating superior health) was also included to document health, but did not attribute to the cost-utility analysis.

The cost-effectiveness analysis was conducted from a health care perspective as well as a societal perspective, the latter including lost productivity. Patient costs and QALYs were combined and subjected to bootstrap analysis (¹⁴). Bootstrap analysis is a method to deal with nonnormality of two combined variables. By resampling with replacement from the original sample the chance that outliers significantly influence the cost-utility analysis becomes smaller (i.e., the readmission in the open group led to significant additional costs and may be considered an outlier). One thousand incremental cost-effectiveness ratios (ICERs) were generated. These ICERs were plotted in a cost-effectiveness plane from which an ICER acceptability curve was derived. In general, insurance companies and governmental organizations are willing to pay up to €50,000 per QALY (¹⁵). The maximum amount of money people are willing to pay per QALY is called the ceiling ratio or threshold.

Data Analysis and Statistical Considerations

Physical fatigue and physical function have been the primary and primary secondary endpoint of this trial (³). These endpoints were chosen to quantify surgical recovery and to focus on the physical burden on the donor after live kidney donation. Power calculations led to 50 donors in either group to establish a moderate significant difference in physical fatigue with a power of 80% and an alpha of 0.05. Return to work, return to daily activity and cost-effectiveness were in the original protocol defined as important secondary outcomes. To express costs in U.S. dollars ($), we applied a currency exchange ratio of $US1.2 per Euro. Categorical variables were compared with the chi-square test. Continuous variables were compared with the Mann Whitney U test. Repeated continuous variables were compared with repeated measurement analysis of variance. Repeated measures were adjusted for baseline values, donor’s sex and age. Analyses were conducted using SPSS (version 11.5, SPSS Inc., Chicago, USA). Data were analyzed according to the intention to treat principle. A value <0.05 (two-sided) was considered statistically significant.

Funding

This study was financially supported by unrestricted grants of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES) and the Dutch Kidney Foundation.

RESULTS

Between November 2001 and February 2004, 100 donors underwent surgery according to the randomization. Conversions or crossovers did not occur. During one-year follow-up response rates for quality of life ranged from 97% to 88%. Baseline characteristics are shown in Table 1. Thirty-four donors in either group were employed preoperatively. Eighteen donors in either group performed physically demanding work. Donors worked 34 hr on average in either group.

TABLE 1: Baseline characteristics

Intra- and Direct Postoperative Outcomes

Most of these outcomes have been published elsewhere (³). Presence of the donor in the operation room (induction of anesthesia, operation, extraction and perfusion of the transplant, and extubation of the donor) lasted 293 min on average during LDN and 234 min during ODN (P<0.001). Mean postoperative hospital stay was 3 days (range 1–6 days) in the LDN group and 4 days (range 2–8 days) in the ODN group (P=0.003). Three postoperative complications occurred in either group. The influence of five of these complications was limited (a transfusion with two packed red blood cells and two oral antibiotic cycles to treat wound infections, an untreated pulmonary infiltrate and an untreated urinary tract infection). However, one donor experienced an infected hematoma requiring readmission to the hospital and administration of intravenous antibiotics for 14 days.

Direct Treatment Costs

Direct treatment costs are categorized in Table 2. Higher costs in the operation room are the main reason why total charges for LDN are significantly higher. Fixed costs are higher in the LDN group because of significant higher charges for maintenance (€50 vs. €10, $60 vs. $12) and depreciation of the endoscopic tower and the monitors (€126 and €47, respectively, per laparoscopic operation, $151 vs. $56). Personnel costs were higher for laparoscopy, which is mainly caused by longer operation times. Disposables were significantly more expensive in the LDN group. In particular the use of ultrasonic shears, the endobag and endostapler were cost-drivers.

TABLE 2: Costs, return to work, and activity level

Charges for hospitalization were lower in the LDN group, because the average hospital stay was about one day shorter. In addition, expenses after discharge were lower as a consequence of fewer visits to the outpatient clinic on average and the aforementioned readmission in the ODN group.

Return to Work and Resumption of Daily Activities

In Table 2, data are presented on return to work and daily activities. After 55 and 58 days on average, donors in the LDN and ODN group respectively returned to work. Remarkably, 18% of the donors in either group had not resumed work by 3 months. Neither the operative approach, gender, nor the relation to the recipient influenced return to work. Donors who performed physically demanding work returned to work later than those who did not have a physically demanding job (67 vs. 45 days, P=0.006). The mean number of days of productivity loss was also calculated. This number is a summary measure for the total number of days lost. On average, productivity losses were 68 days in the LDN group and 75 days in the ODN group. This corresponded with 36 hr of additional lost productivity in the ODN group.

The difference regarding the mean day at which 90% of daily activities other than work were resumed approached statistical significance in favor of LDN. Complete resumption of activities occurred significantly earlier after LDN.

Quality of Life and Health Score

QOL as measured on the Euroqol 5-D is displayed in Figure 1. During 1-year follow-up the QOL score or so-called Tariff score was significantly higher after LDN (mean difference 0.035, P=0.047, 95% CI: 0.00 to 0.07) indicating better QOL. Converted to QALYs, LDN provided an additional 0.03 QALY, mainly generated within the first four weeks. In Figure 2 the results of the 0–100 VAS regarding health are presented. This graph is rather similar to Figure 1. During 1-year follow-up the health score has a tendency to be statistically superior after LDN (mean difference 3.3, P=0.054, 95% CI: –0.6 to 6.7) after adjusting for baseline differences, gender and age. Again the main difference is generated within the first 4 weeks.

FIGURE 1.: Tariff score of the Euroqol 5-D. Data are nonadjusted and displayed as mean±95% CI. Open squares: LDN, filled squares: ODN. Numbers underneath the x-axis represent the number of replies (left: LDN, right: ODN). Mean difference during 1-year follow-up: 0.035, P=0.047, 95% CI: 0.00 to 0.07.

FIGURE 2.: Health score of the 0–100 VAS. Data are nonadjusted and displayed as mean±95% CI. Open squares: LDN, filled squares: ODN. Numbers underneath the x-axis represent the number of replies (left: LDN, right: ODN). Mean difference during 1-year follow-up: 3.3 in favor of LDN, P=0.054, 95% CI –0.6 to 6.7.

Cost-Utility Analysis

Thirty-six additional working hours were lost per working donor in the ODN group. Thirty-six hours cost €1,152 ($1,382; Table 2; 36 hr×80%×€40 or $48 per hour). Therefore, additional productivity losses per donor were €783 ($940) after ODN (34/50×€1152 or $1382). Bootstrap analysis resulted in a mean gain of 0.03 QALY (95% CI: 0.00 to 0.07 QALY) at mean costs of €1,271 ($1,525; 95% CI: €853 to €1,690) and €488 ($586; 95% CI: €70 to €906) from a healthcare perspective and a societal perspective, respectively. Consequently the mean ICER was €41,278 and €15,844 ($50,000 and $19,000, respectively) per QALY. Acceptability curves are displayed in Figure 3. Given a societal perspective, there is an 80% probability that the ICER falls in a range of €3,000–50,000 or $3,600–$60,000 per QALY.

FIGURE 3.: Acceptability curves for laparoscopic donor nephrectomy (LDN) vs. open donor nephrectomy (ODN) calculated from a societal perspective (continuous line) and from a health care perspective (interrupted line). The ceiling ratio is the maximum amount of money governmental organizations or insurance companies are willing to pay per QALY. The probability represents the chance that LDN is cost-effective at a certain ceiling ratio. For example, if one is willing to pay up to €40,000 ($48,000) for one QALY, the probability that LDN is cost-effective can be determined by drawing a vertical line that crosses the x-axis at 40,000. The associated probabilities are the crossings between the vertical line and the two plotted lines; the probability that LDN is cost-effective is 0.43 from a health care perspective and 0.85 from a societal perspective.

DISCUSSION

This is the first paper that addresses cost effectiveness of laparoscopic versus mini-incision open donor nephrectomy alongside a randomized controlled clinical trial. The collection of data regarding costs was rather detailed and even included depreciation of hardware, home care, and visits to the general practitioner. As reported earlier the majority of costs were generated within the first 24 hr (¹⁶). We collected data on return to work prospectively, which is more reliable than telephone interviews at 1-year postoperatively as applied by investigators in a prominent previous study (²). Although we did not assess a difference in return to work in general, LDN resulted in less hours of lost productivity. This is in contrast with previous studies in which return to work differed significantly between laparoscopic and conventional open surgery (^{2, 12}). The disappearing difference between (partial) resumption of work may reflect improvement of the open technique. However, LDN rewards both employer and employee, because total productivity losses are lower and the donors experience better health and QOL. As described earlier by Andersen et al. sick leave is much longer in Europe than generally accepted in the Northern America (^{2, 17, 18}). This reflects both the social system in Europe, as well as the attitude towards work. Although we did not advice donors when to resume work, many donors got six weeks off for donation. Return to work contrasts with resumption of non-work related activities. We assume that donors in the laparoscopic group would have been able to start working earlier.

The finding of favorable health scores and QOL in the LDN group is concordant with earlier published data of this trial (³). The Euroqol 5-D is a less sensitive instrument to measure QOL than the SF-36, but is primarily developed to express QOL in QALYs. This explains why the difference in QOL appears smaller than the differences between the groups with regard to the SF-36. The conversion of QOL in QALYs has been debated, but current guidelines in the Netherlands advise quantification of QOL in QALYs (¹⁹). The main advantage of this conversion into QALYs is that the gain of a therapy allows comparison to the gain of other therapies (i.e. heart transplantation or dialysis). Rapidly expanding costs of healthcare in Western countries force policy makers to draw decisions whether or not to allow new treatments. Cost-effectiveness may be decisive. The difference of 0.03 QALYs is smaller than the 0.06 QALYs found by Pace et al. (²⁰), which may reflect improvements of the open technique, but confirms the superiority of LDN from the perspective of the donor.

We applied rather conservative methods in the cost-utility analysis. An average salary of €40 ($48) per hour is moderate. Higher salaries would further offset the initial financial disadvantages of LDN. Furthermore, an advisory organ of the Dutch Ministry of Finance recently reported that treatments costing up to €80,000 (about $96,000) per QALY may be acceptable (¹⁹). In the United Kingdom, the acceptable threshold equals about €50,000 (¹⁵). For governmental organizations and policy makers, the societal perspective of cost effectiveness is the most interesting perspective because it offers the most detailed reflection of the treatment effect, including besides plane costs also the savings due to limited productivity losses. Regardless of a threshold chosen at €50,000 or €80,000, the incremental costs of laparoscopy are most likely justified.

Regardless this conservative perspective, more frequent use of endoscopic equipment, use of reusable material instead of the disposables and cheaper disposables, as well as less productivity losses will make LDN more attractive. Hand-assistance during laparoscopy or retroperitoneoscopy may be cost saving because the operation times may be reduced. However, studies providing a detailed cost-effectiveness analysis of laparoscopic kidney donation with and without hand assistance are lacking.

In this study, some aspects were not investigated. First, we did not combine data of donor and recipient. There has been some concern about slower graft-function of laparoscopically harvested grafts (²¹). With an increased number of graft-related complications and lower graft survival an initial benefit of LDN would vanish with a significant effect on cost effectiveness. Current data do not support earlier raised concerns (⁸). In our own study, we found a similar rate of graft-related complications after LDN and ODN and a similar 1-year graft survival. We focused on the donation procedure assuming that costs related to the preoperative work-up of the donor and the recipient, preoperative dialysis and transplantation were independent of the costs of the donation procedure. Inclusion of these costs in our analysis would have revealed that the costs of the live donor nephrectomy on its own and the incremental costs of laparoscopy were only a small fraction of the expenses for live kidney donor transplantation. Transplantation is cost saving (²²). This would be an argument to implement laparoscopic donor nephrectomy regardless of the value of the ceiling ratio provided that the donor experiences significant benefit and the safety to the recipient is comparable.

Second, we did not include data on the learning curve of LDN. It is well known that it takes several procedures to learn this technique. This probably implies longer operation times, more complications, and reoperations at the beginning of the learning curve. The number of significant postoperative complications was very small in the present study. Reoperations for bleeds, perforations, or later-presenting incisional hernias did not occur. Less experience would probably adversely affect the benefit of LDN and increase its costs. Finally, we did not investigate whether LDN led to more live kidney donors in Europe. Live kidney transplantation is the most cost-efficient treatment of end-stage renal disease (²²). In a recent survey, about half of the surgeons thought that LDN resulted in increased live kidney donation rates and the other half did not (¹⁰). If it were true that LDN leads to more donors, LDN would become a true dominant strategy because the financial benefits realized with more kidney donors easily outweigh slightly higher costs of the operative procedure.

In conclusion, these data confirm the donor-experienced benefit of laparoscopic surgery. Because LDN is also cost-efficient, at least given a societal perspective, LDN has truly become the standard of care for live kidney donation.

ACKNOWLEDGMENTS

The authors are very grateful to Dr. W. Hop, M.Sc., Ph.D., for statistical advice and Mrs. J. van Duuren-van Pelt for data management.