Cost-effectiveness analysis of open colposuspension versus laparoscopic colposuspension in the treatment of urodynamic stress incontinence

Authors


Dr A Manca, Centre for Health Economics, Alcuin A, University of York, York YO10 5DD, UK. Email am126@york.ac.uk

Abstract

Objectives  To compare the cost effectiveness of laparoscopic versus open colposuspension for the treatment of female urinary stress incontinence.

Design  Cost utility analysis alongside a randomised controlled trial.

Setting  Six gynaecological surgical centres within the UK.

Population/Sample  Women with proven stress urinary incontinence requiring surgery.

Methods  Open abdominal retropubic colposuspension or laparoscopic colposuspension carried out by experienced surgeons.

Main outcome measures  Cost, measured in pounds sterling and generic health-related quality of life, measured using the EQ-5D. The latter was used to estimate patient-specific quality-adjusted life years (QALYs).

Results  Healthcare resource use over 6-month follow up translated into costs of £1805 for the laparoscopic arm and £1433 for the open arm (differential mean cost £372; 95% credibility interval [CrI]: 274–471). At 6 months, QALYs were slightly higher in the laparoscopic arm relative to the open arm (0.005; 95% CrI: −0.012 to 0.023). Therefore, the cost of each extra QALY in the laparoscopic group (the incremental cost-effectiveness ratio [ICER]) was £74,400 at 6 months. At 24 months, the laparoscopic arm again had a higher mean QALY score compared to the open surgery group. Thus, assuming that beyond 6 months the laparoscopic colposuspension would not lead to any significant additional costs compared with open colposuspension, the ICER was reduced to £9300 at 24 months. Extensive sensitivity analyses were carried out to test assumptions made in the base case scenario.

Conclusions  Laparoscopic colposuspension is not cost effective when compared with open colposuspension during the first 6 months following surgery, but it may be cost effective over 24 months.

Introduction

As highlighted in the accompanying clinical article,1 female urinary incontinence is a common clinical problem for which Burch colposuspension is a standard surgical approach. Colposuspension can be carried out as an open procedure or as a minimally invasive laparoscopic procedure. However, there is little robust evidence regarding the comparable cost effectiveness of the two alternative methods. As well as investigating the clinical effectiveness of alternative treatment strategies, in a limited budget environment like that in which many publicly funded healthcare systems operate, it is vital to assess whether any additional costs incurred by adopting the more expensive treatment strategy are counterbalanced by additional health benefits for the patients. Cost-effectiveness analysis (CEA) provides a formal analytical framework within which to assess whether a particular health technology provides ‘good value for money’. In principle, laparoscopic colposuspension may be expected to confer potential economic advantages when compared with the open procedure since women receiving laparoscopic surgery may have reduced length of hospital stay, better health-related quality of life in the postsurgery period and a quicker recovery.2 However, previous economic analyses of these two colposuspension techniques are methodologically weak and have not been conducted as part of published randomised controlled trials.3–5 Furthermore, since most economic analyses were carried out for other healthcare systems, the relevance of their results to the UK NHS decision-making process is unclear.

The Colposuspension Laparoscopic or Open (COLPO) trial funded by the UK Medical Research Council (MRC) is the largest randomised controlled trial to simultaneously assess the clinical and cost effectiveness of open versus laparoscopic colposuspension in patients in UK. This article reports the CEA of this trial.

Methods

Study design

Full details of the trial are reported in the companion article.1 Briefly, the study was a multicentre randomised controlled trial, taking place in six gynaecology units in the UK.

In total, 291 women were randomised to the trial—144 to the laparoscopic group and 147 to the open surgery group. Among the study participants, five were randomised but did not undergo surgery (two in the laparoscopic and three in the open surgery group) and were therefore excluded from this analysis, leaving a total sample of 286, following a modified intention-to-treat analysis approach. There were no serious complications relating to the surgery during the trial. One death occurred in the laparoscopic arm from causes not related to the urodynamic condition or to the colposuspension surgery received.

Measurement of resource use

Healthcare resource use relating to the surgery, associated hospital stay and first 6 months after hospital discharge was collected prospectively for each participant. Data collected are described below under the relevant headings.

Theatre

Patient-specific data were collected using case report forms, which were completed with the help of hospital staff. Individual-level information obtained as part of the trial included: time for anaesthetic induction, time in theatre and details of any intraoperative complications, including the need for blood transfusions. Consistently with other trial-based CEA6 data that were not expected to vary significantly between women within each treatment arm (i.e. type and quantity of equipment, consumables and drugs used intraoperatively) were collected at the hospital level to minimise the burden of data collection. Hospital-specific questionnaires (one for each arm in the trial) were sent to a suitably qualified member of clinical staff who was asked to fill in information relating to—among others—staff present during theatre and patient preparation, during the surgery itself and during theatre turn around; intraoperative equipment; intraoperative consumables and intraoperative drug use. This was undertaken at the end of the trial to capture up-to-date data on resource use.

Hospital stay

Case report forms were used to capture data on individual’s length of stay in hospital (including ward type), drug use and details of complications during admission, i.e. urinary tract infection, associated treatments and return to theatre for any reason. Medical advice (H.K., A.S., S.M.) was given regarding the additional resource use associated with complications that occurred postoperatively during the trial, which had not been routinely captured by our data collection forms.

Follow up

Healthcare resource use was collected at 6 weeks and 6 months after hospital discharge using case report forms completed by the trial participants. Information was collected on any complications and further treatment required as well as GP visits, hospital outpatient visits and re-admissions and return to usual activity.

Unit costs

The cost analysis was conducted from the perspective of the UK NHS and Personal Social Service, following the ‘reference case’ methodology indicated by the UK National Institute for Health and Clinical Excellence (NICE) Guidance for Health Technology Appraisal.7 Healthcare resource use was costed at 2002–03 prices in pounds sterling. Ward-specific NHS hospital costs were calculated from national averages.8 Ward ‘hotel’ costs were calculated as a percentage of the fully allocated cost per day of a urology surgical bed, adopting the methodology used in a previous laparoscopic surgical trial.9 Nursing and doctor’s staff costs were calculated using midscale salaries taken from published figures10,11 and up-rated for National Insurance and employers’ costs where required.11 Equipment and consumables were costed using manufacturers least prices (with value added tax of 17.5%). Drugs prices were taken from the British National Formulary.12Table 1 shows the unit costs of major healthcare resources.

Table 1.  Main unit costs used to value healthcare resource use
Healthcare resourceUnitCost per unit (£)Source
Inpatient NHS hospital ‘hotel’ costs
Normal wardDay132.40Royal Pharmaceutical Society of Great Britain12 and Kind13
Theatre
Staff—laparoscopicMinute3.49Kind et al.14,15
Staff—open surgeryMinute2.99
OverheadsMinute2.36Kind13
Selected laparoscopic consumables
Disposable trocar (12 mm)Item54.00Manufacturer
Disposable endoshearsItems80.00Manufacturer
Visits
Outpatient day visitVisit79.00Royal Pharmaceutical Society of Great Britain12
GP visitVisit17.00Kind et al.15

Health outcomes

The health outcome measure for the economic analysis was quality-adjusted life years (QALYs), a measure reflecting both patient’s health-related quality of life and mortality into a single index. The data required for the calculation of QALYs were collected prospectively using the EQ-5D questionnaire, which was completed by each participant in the trial at baseline, 6, 12 and 24 months. The EQ-5D is a widely recognised and validated generic measure of health-related quality of life.13 It has five questions, each relating to a different dimension: mobility, self-care, ability to undertake usual activity, pain and anxiety/depression. Each question has three possible response levels: no problems, moderate problems and severe problems. Based on their combined answers to the EQ-5D questionnaire, participants can be classified as being in one of the 243 possible health states (plus unconscious or dead). Each of these health states has an associated utility score, derived from interviews with a large sample of the UK population.14,15 Where appropriate, QALYs were discounted using 3.5% annual rate, following NICE guidance.7 Differences in QALYs were adjusted for differences in baseline EQ-5D scores.16

Cost-effectiveness analysis

Although the study was planned as a noninferiority trial for clinical outcomes, this is unlikely to be true for economic outcomes. CEA were carried out by relating differences in costs between treatments to difference in QALYs. The CEA can lead to one of the following scenarios:

  • (a) Laparoscopic colposuspension is more effective than (in terms of differential QALYs), or at least as effective as, open colposuspension while being less costly.
  • (b) Laparoscopic colposuspension is more (or at least as) costly and less effective than open colposuspension.
  • (c) Open colposuspension is both more (less) effective and more (less) costly compared with laparoscopic colposuspension.

Unlike scenarios (a) and (b) where laparoscopic colposuspension is clearly superior (i.e. costs less and is more beneficial) in the first case but not in the second, scenario (c) above requires a ‘decision rule’, which in CEA is fairly simple. The decision rule requires the calculation of the incremental cost-effectiveness ratio (ICER), which is the ratio of the differential mean costs (between the two trial arms) and the differential mean QALYs (between the two trial arms). Once calculated, the ICER needs to be compared against the maximum amount the decision maker (e.g. NICE in the UK) is (on average) willing to pay to achieve an additional QALY. A treatment strategy is considered to be cost effective if the decision maker’s willingness to pay for an additional QALY is at least equal (or greater) to the ICER.

The economic evaluations were conducted from a Bayesian perspective. This facilitates the interpretation of CEA results in a way that can feed directly into the decision-making process regarding the value for money of healthcare technologies.17 Mean differential costs and QALYs are reported with associated 95% credibility intervals (CrI)—the Bayesian equivalent of confidence intervals—to reflect the sampling uncertainty surrounding their estimates. This approach is increasingly being adopted in published economic evaluations.16,18–21 The analyses were carried out using WinBUGS,22 a freely available statistical software for Bayesian analysis. Cost-effectiveness acceptability curves (CEAC)23–26 were plotted to represent the decision uncertainty surrounding the cost effectiveness of laparoscopic colposuspension. Given the data in the COLPO trial, the CEAC shows the probability that laparoscopic colposuspension is more cost effective than open colposuspension for different amounts the decision-maker may be willing to pay for an additional QALY.

Sensitivity analyses

All women in the trial were followed up for a period of 2 years. However, healthcare resource use data during the trial were only collected for the first 6 months of follow up, unlike EQ-5D data, which were available for the entire duration of the study (i.e. 24 months). CEA was therefore conducted for both time horizons: up to 6 months and up to 24 months, with the latter analysis making full use of available EQ-5D data. The base case analysis at 24 months assumed that no additional cost had occurred beyond 6 months. This hypothesis was explored in a sensitivity analysis to ascertain the amount of additional cost that would need to occur during the interval of 6–24 months in order to alter the conclusions of the base case CEA at 2 years.

Most resource use variables at 6 weeks, such as time in theatre and GP and hospital visits, had a very low level of missing data (<5%). Thus for the base case analysis, missing data for continuous variables were replaced with the group’s sample mean value, while missing hospital and GP visits were imputed using the group mode. At 6 months, there was a greater extent of missing hospital outpatient visit data (approximately 33%). Following clinical advice regarding the likelihood of a woman needing out-of-patient hospital visits at 6 months, the base case analysis assumed that no visits had taken place after 6-week follow up, where the information was missing. This assumption was also challenged in sensitivity analyses.

A further sensitivity analysis was carried out to assess the impact of using some reusable laparoscopic equipment (trocars, with cannulas, and endoshears) rather than the corresponding disposable items. This was carried out to expose the sensitivity of the cost-effectiveness results to variations in practice between hospital and surgeons.

In previous trials comparing open and laparoscopic colposuspension, the difference in length of stay between the two arms has ranged from approximately 0–4 days.27–31 We conducted a series of sensitivity analyses to investigate the robustness of the cost-effectiveness results when applied to differences in clinical practices concerning the duration of the inpatient hospital stay for women undergoing the two different procedures.

A final sensitivity analysis was conducted to address the potential concern that some of the unit costs used in the cost analysis might not be representative of the entire UK context. In particular, it could be argued that the ‘hotel’ cost of inpatient stay may vary between hospitals. To address this issue, we explored the impact of different ‘hotel’ costs for 1 day in hospital on the differential costs.

Results

Clinical outcomes

The clinical article accompanying this economic evaluation reports that laparoscopic colposuspension is not inferior to open colposuspension.1 This suggests that the choice of the optimal strategy in this patients’ population should be informed by cost-effectiveness considerations.

Resource use and costs

Table 2 shows the results relating to the main healthcare resources use during the trial over the first 6 months of follow up. When compared with those undergoing open surgery, women who received laparoscopic surgery had a reduced mean length of stay in hospital (5.7 versus 6.4 days) and a reduced number of outpatient and GP appointments during follow up. However, the laparoscopic surgical procedure had much greater resource use implications than the open procedure, in terms of time in theatre (69 versus 54 minutes). Table 3 shows the cost estimates during the main hospitalisation and at follow up. Total theatre costs for the laparoscopic group were, as expected, markedly higher than the open surgery group theatre costs (£944 versus £464) due to longer theatre time and the extra equipment required in the laparoscopic surgery. Despite having a shorter length of stay in hospital, for the base case analysis, the total cost of the laparoscopic surgery at 6 months of follow up was subsequently higher than the total cost of open surgery (£1805 versus £1433), with a differential mean of £372 (95% CrI: 274–471).

Table 2.  Main healthcare resource use (total n= 286)
 Laparoscopic (n= 142)Open (n=144)
  1. IQR, interquartile range.

Surgery and hospital stay
Time of anaesthetic induction (minutes)
 Mean (SD)16 (6)17 (9)
 Median (IQR)15 (10–20)15 (10–20)
Time in theatre (minutes)
 Mean (SD)69 (21)54 (14)
 Median (IQR)65 (53–80)51 (44–63)
Days in hospital (all women were on normal ward)
 Mean (SD)5.7 (2.5)6.4 (2.3)
 Median (IQR)5.0 (4.0–7.0)6.0 (5.0–7.0)
6-week follow up, n (%)
No. of visits to GP (total)
 063 (44.4)53 (36.8)
 1–265 (45.8)83 (57.6)
 3–412 (8.5)7 (4.8)
 >42 (1.4)1 (0.7)
Outpatient visits
 0121 (85.2)109 (75.2)
 1–221 (14.7)34 (23.6)
 >201 (0.7)
Inpatient visits12 (8.5)11 (7.7)
Of those readmitted: number of days stayed (all on normal ward)
 Mean (SD)4.08 (3.0)2.66 (1.4)
 Median (IQR)3.5 (1.5–5.0)3.0 (1.0–4.0)
6-month follow up, n (%)
No. of visits to GP (total)
 096 (67.6)73 (50.7)
 1–245 (31.7)69 (47.9)
 3–41 (0.7)2 (1.4)
 >40 (0)0 (0)
Outpatient visits
 0131 (92.3)127 (88.2)
 1–210 (7.0)17 (11.8)
 >21 (0.7)0 (0)
Table 3.  Open and laparoscopic surgical costs mean £ (SD), using 2002–03 prices
 Laparoscopic, n=142Open, n=144
  • IQR, interquartile range.

  • *

    Laparoscopic minus open surgery.

  • **

    95% CrI obtained from Markov chain Monte Carlo estimation. The model was run for three different chains, with 7000 burn-in and 20 000 iterations.

Main hospitalisation
Theatre cost
 Mean (SD)944 (124)464 (76)
 Median (IQR)923 (851–1009)463 (410–516)
Hospital ‘hotel’ cost
 Mean (SD)749 (328)854 (301)
 Median (IQR)662 (530–927)794 (662–993)
Other postoperative cost
 Mean (SD)15 (31)17 (24)
 Median (IQR)12 (10–14)14 (11–16)
Follow up
6-week follow-up cost
 Mean (SD)77 (193)66 (115)
 Median (IQR)18 (0–56)19 (1–89)
6-month follow-up cost
 Mean (SD)20 (55)32 (95)
 Median (IQR)0 (0–18)18 (0–19)
Total cost
Mean (SD)1805 (471)1433 (362)
Median (IQR)1717 (1525–1930)1402 (1196–1595)
Differential mean cost (£)* (95% CrI)**372 (274–471)
Sensitivity analysis: 6-month outpatient visits
Missing in laparoscopic group = 1 visit
 Total cost1837 (471)1433 (362)
 Differential mean cost (£)* (95% CrI)**404 (305–502)
Missing in open surgery group = 1 visit
 Total cost1805 (471)1453 (361)
 Differential mean cost (£)* (95% CrI)**352 (251–447)
Sensitivity analysis: reusable equipment
Total cost1589 (471)1433 (362)
Differential mean cost (£)* (95% CrI)**156 (58–255)
Sensitivity analysis: impact of difference in inpatient hospital stay on difference in cost
No difference in hospital stay 473
Open surgery
 2 days longer inpatient stay 208
 3 days longer inpatient stay 75
 4 days longer inpatient stay −57

Health outcomes

On average, women in both treatment arms had an improved EQ-5D score after 6 months (0.856 and 0.840, respectively for the laparoscopic and open colposuspension group) compared with baseline (Table 4). In the laparoscopic arm, the mean EQ-5D score was also improved at 12 months (0.865) compared with 6 months, whereas the open surgery group’s scores had fallen very slightly (0.837). At 24 months, both treatment groups had a mean EQ-5D score at a similar level to that observed at 6 months. Mean QALYs at 6 months were very similar (0.421 versus 0.416), with a differential mean QALYs in favour of the laparoscopic group of 0.005 (95% CrI: −0.012 to 0.023). At 24 months, those who underwent laparoscopic surgery had higher QALYs compared with those who underwent open surgery (1.677 versus 1.637) and a differential mean QALY at 2 years of 0.04 (95% CrI: −0.009 to 0.086).

Table 4.  Estimates of mean QALYs of laparoscopic versus open colposuspension (n= 286)
 Laparoscopic, n= 142Open, n= 144
  • *

    Laparoscopic minus open colposuspension.

  • **

    95% CrI obtained from Markov chain Monte Carlo estimation. The model was run for three different chains, with 7000 burn-in and 20 000 iterations.

Mean utility from EQ-5D values
Baseline0.8270.824
6 months0.8560.840
12 months0.8650.837
24 months0.8440.825
6-month follow-up analysis
Mean QALY0.4210.416
Differential QALYs* (95% CrI)**0.005 (−0.012 to 0.023)
24-month follow-up analysis
Mean QALY1.6771.637
Differential QALYs* (95% CrI)**0.04 (−0.009 to 0.086)

Cost-effectiveness analysis

The results from the 6-month analyses indicate that laparoscopic surgery produced, on average, greater QALYs (0.005) at an additional cost of £372 when compared with open surgery. This corresponds to scenario (c) illustrated in the Methods section. At 6-month follow up, the estimated ICER (i.e. differential cost/differential QALY) is £74,400. The CEAC for the 6-month analysis (Figure 1) indicates that laparoscopic colposuspension has a probability of 0% to be cost saving and a probability of only 20% to be relatively cost effective at £30,000. The latter is often cited as an unofficial implicit threshold for cost effectiveness, based on previous NICE decisions.32

Figure 1.

Cost-effectiveness acceptability curves.

Differential QALYs at 24 months showed that laparoscopic colposuspension was associated with an increase in QALYs of 0.04 when compared with the open surgery group. Assuming that additional costs beyond 6 months were not different between the two arms of the trial, these figures translate into an ICER of £9300. If NICE were willing to pay this amount for additional QALY, Figure 1 indicates that laparoscopic colposuspension would have a 50% probability to be cost effective. The probability that laparoscopic colposuspension is cost effective increases to 86% when the decision maker is willing to pay up to £30,000 for additional QALY. This result hinges on the crucial assumption that beyond 6 months laparoscopic colposuspension does not lead to any significant additional costs compared with open colposuspension.

Sensitivity analysis

In reality, it is possible that due to treatment failure, a proportion of women will need additional surgery. Hence, the key issue is whether there is a differential failure rate of laparoscopic colposuspension compared with open surgery that could lead to additional costs that would offset the increased QALYs. To reach an ICER at 24 months of £30,000 would require the mean cost of the laparoscopic group to increase by £828, giving a differential mean cost of £1200. Assuming that the next surgical option for women would be a tension-free vaginal tape (TVT™) procedure (estimated cost £1159 per women),33 our estimates suggest an increased cost of approximately £828 per women in the laparoscopic group as it would require 102 additional TVT procedures be carried out in this group between 6 and 24 months. This would require a 72% failure rate for the laparoscopic group in this study population at 24 months, which is unlikely to occur.34

Two sensitivity analyses, using 6-month data, were carried out to explore the impact of the assumption regarding the occurrence of hospital visits at 6 months on the results above. The first scenario analysis assumed that all women with missing hospital visits at 6 months in the laparoscopic group had attended one outpatient appointment at 6 months, while those with missing data who received the open surgery had not (Table 3). The alternate scenario reversed these assumptions, so all women with missing hospital visits at 6 months in the open surgery group were assumed to have had attended one outpatient appointment at 6 months. In the first scenario, the total cost for the laparoscopic group increased to £1837, with the differential mean cost reaching £404 (95% CrI: 305–502). This led to an ICER of £80,800. In the second scenario, the total cost for the open surgery group increased to £1453 and the differential mean cost was reduced to £352, leading to an ICER of £70,400.

A third sensitivity analysis was carried out to assess the impact on differential mean costs at 6 months if reusable laparoscopic equipment (trocars and endoshears) were used rather than the disposable equivalents. Table 3 shows that the cost of the laparoscopic arm was reduced to £1589, the differential mean costs to £156 (95% CrI: 58–255), and thus, the ICER at 6 months was reduced to £31,200. The results at 24 months for these three sensitivity analyses indicate that the ICER would vary between £10,100 (when assuming all missing hospital visits in the laparoscopic arm equal to one and zero in the colposuspension group) to £3900 (when assuming reusable equipment was used in the laparoscopic arm).

All else being equal, assuming that open and laparoscopic surgery requires the same inpatient length of stay makes laparoscopic colposuspension more expensive than open colposuspension (differential cost: £473). In this case, laparoscopic surgery is not cost effective at 6 months (£473/0.005 =£94,525 per QALY gained) but has a favourable ICER at 24 months (£473/0.04 =£11,816 per QALY gained). However, in some countries it is more likely that clinical practice would require women undergoing open colposuspension to have additional inpatient days in hospital compared with those undergoing laparoscopic colposuspension. In this case, the latter becomes more attractive from the economic viewpoint, as the increased theatre costs associated with laparoscopic surgery are offset by the costs associated with higher inpatient hospital stay required by women having open colposuspension. The differential cost (i.e. laparoscopic − open colposuspension) decreases to £208 when open colposuspension requires 2 days longer inpatient stay compared with laparoscopic surgery (ICER: £5196 at 24 months and £41,546 at 6-month follow up) and £75 when the additional inpatient stay is 3 days (ICER: £1886 at 24 months and £15,085 at 6-month follow up). Finally, in those settings where women undergoing open colposuspension have to spend 4 or more additional days in hospitals, laparoscopic surgery would reduce the cost overall while being associated with a marginal benefit (Table 3).

A final sensitivity analysis was conducted to address the potential concern that ‘hotel’ cost for hospital inpatient stay used in the analysis might not be representative of the entire UK context. We varied this unit cost below and above the base case estimate of £132.40, from a minimum of £80 per day to a maximum of £250 per day. Unsurprisingly, the results of the CEA remained unchanged, given the mean differential hospital inpatient stay in the COLPO trial was 0.76 days.

Discussion

Implications of findings

Laparoscopic colposuspension surgery has a higher mean cost when compared with open colposuspension surgery over a 6-month follow-up period. Most of the excess cost in the laparoscopic arm is related to the increased theatre cost, including the cost of disposable consumables. This result agrees broadly with the findings of other studies in the area.3–5 Over the same period, increases in mean QALYs favour the laparoscopic group, but the differential mean is small, which is a similar finding to other comparisons of laparoscopic-assisted (or minimally invasive33) surgery with open surgery.9 Our findings suggest that, in the 6 months after surgery, laparoscopic colposuspension surgery is unlikely to be cost effective when compared with open surgery, even if use of reusable equipment is taken into account.

Analysis of 24-month data suggests that the laparoscopic procedure might be a cost-effective alternative in the medium term, provided that there are no major (more than a 72% differential re-operation rate for laparoscopic colposuspension) cost implications from treatment failure compared with the open group.

Statistical analysis from a recent systematic review estimates that those undergoing open colposuspension are less likely to have an objective (clinically assessed) failure during the first year when compared with laparoscopic surgery (relative risk = 0.63; 95% CI: 0.42–0.95).2 However, this review reports no difference in cure rate 1 year postsurgery. It has been reported that 5 years after receiving an open colposuspension, 80–85% of women will still be ‘dry’.2 There are limited corresponding data for laparoscopic outcomes after 18 months, although one trial has suggested that outcomes are poor.34 This trial’s surgeon was not experienced in laparoscopic colposuspension, so it is unclear whether long-term negative outcomes can be assigned to the surgical technique being investigated or to the surgeon’s lack of skill (or both).

The clinical analysis of the COLPO trial, where surgeons were experienced in performing laparoscopic colposuspensions, suggests that there is little difference in objective outcome between the two arms, and thus we assume failure rates, at 2 years.1 This may suggest that poor outcome during follow up reflects surgeons’ experience. This has been reported in trials of other laparoscopic techniques such as hernia repair where reoccurrence rates were lower in those surgeons who have performed more than 250 operations of this type.35 However, it is important to note that this was a post hoc evaluation and thus could be affected by confounding.

Even assuming that surgeons’ experience can affect the success rate of laparoscopic colposuspension, it is unlikely that failure rates in a laparoscopic group could ever occur to the extent required to increase the differential mean cost to that required to make laparoscopic surgery no longer cost effective at 24 months, and there is limited data to support this suggestion.34

The results of our sensitivity analyses indicate that the additional theatre cost related to laparoscopic colposuspension can be offset by a longer inpatient hospital stay required for women undergoing open colposuspension.

Limitations of study

The centres involved in this trial were all experienced in carrying out the laparoscopic procedure. This is reflected in the relatively small number of complications that occurred during the study. It is widely recognised that surgeons who are less experienced in laparoscopic techniques will have to undergo a learning curve on the first patients they operate on. These patients are arguably at a great risk of having intraoperative complications, and complications would add to the cost of the surgery. Thus, the cost of laparoscopic surgery carried out by nonexperienced surgeons may be greater than those reflected here for the period of time before they complete their learning curve. As discussed above, surgeon’s level of experience and skill could have implications for patients’ long-term failure rates.

The analysis could also have been carried out from the social perspective. In the COLPO trial, we have collected data regarding time of return to usual activities. The analysis of these data suggested that the proportion of women who returned to usual activity at 6 weeks was higher in the laparoscopic arm compared with the open surgery arm (54 versus 45%) and similar between the two arms of the trial at 6 months (e.g. 95 versus 98%).

It is also important to note that this study did not assess the quality-of-life impact related to any difference in recovery time between the two trial arms. Other trials of laparoscopic procedures35 report that laparoscopic surgery leads to shorter recovery times and therefore increased accrual of QALYs. If this were the case in this trial, differential QALYs could have been increased at the 6-month follow-up period, and this, in turn, could have potentially affected the cost effectiveness of laparoscopic surgery at 6 months.

TVT™ surgical technique has been widely adopted to treat urinary stress incontinence. A recent National Coordinating Centre for Health Technology Assessment report36 and associated article37 report that the TVT™ method is, in the short term, at least as effective as open colposuspension surgery and also cheaper. Two recent randomised controlled trial comparing TVT™ with laparoscopic colposuspension report that objective cure rates are better in the TVT™ group after 1 year.38,39 However, no corresponding CEA has been carried out, and further research is required to assess the potential difference in mean QALYs between laparoscopic colposuspension and TVT™ and the long-term failure rate of these techniques.

Conclusion

This study suggests that laparoscopic colposuspension may not be good value for money when compared with open colposuspension surgery in the short term (i.e. first 6 months following surgery), but it could be a cost-effective alternative over 24 months. In terms of application to clinical practice, taking this long-term view is important as in such health conditions as urinary incontinence, costs and benefits can manifest beyond the immediate period following surgery. Thus, this study suggests that laparoscopic colposuspension is a cost-effective approach, making the assumptions we have made in this analysis about re-operation rates.

It is also important to note that in countries where there may be a more marked difference in inpatient stay between open and laparoscopic surgery groups, the likelihood of laparoscopic surgery being cost effective would increase.

Acknowledgements

This trial was funded by the MRC. A.M. is funded by a Wellcome Trust Training Fellowship in Health Services Research. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the UK MRC and the Wellcome Trust. We would also like to thank Sara Moore for her clinical advice and the two anonymous reviewers for their useful comments.

Contribution to authorship

D.T., A.M. and H.K. were involved in planning the analysis. A.M. and J.D. were involved in developing and implementing the analysis. H.K. and L.N. coordinated the trial and A.S. provided clinical input. J.D. and A.M. wrote the original draft of the article and all authors reviewed and approved the final version.

COLPO study group

Graham Dunn, Vanessa Lawton, Fiona Reid, Nadia Ali, Mr Alfred Cutner, Dr John Osborne, Dr Tim Sayer, Mr Charles Cox, Dr Robert Hawthorn, Dr Ian Ramsay, Dr Christine Landon, Dr J.C.D., A.M., Prof H.K., Dr A.R.B.S., L.N. and Prof D.J.T.

Ancillary