Upload
andrea-manca
View
213
Download
1
Embed Size (px)
Citation preview
A cost–utility analysis of tension-free vaginal tape versuscolposuspension for primary urodynamic stress incontinence
Andrea Mancaa,*, Mark J. Sculphera, Karen Wardb,c, Paul Hiltonb,c
Objective To assess the cost effectiveness of tension-free vaginal tape compared with open Burchcolposuspension as a primary treatment for urodynamic stress incontinence.
Design Cost–utility analysis alongside a multicentre randomised comparative trial.
Setting Gynaecology or Urology departments in 14 centres in the UK and Ireland, including University-associated teaching hospitals and district general hospitals.
Population Women with urodynamic stress incontinence. Exclusion criteria were: (1) detrusor overactivity;(2) major voiding problems; (3) prolapse; (4) previous surgery for incontinence or prolapse.
Methods Resource use data were collected on all 344 patients in the trial, including length of hospital stay,time in theatre and management of complications; resource use was costed using UK unit costs at 1999–2000 prices.
Main outcome measures Health outcomes were expressed in terms of quality-adjusted life years (QALYs)between baseline and six months follow up, based on women’s responses to the EQ-5D health questionnaire.
Results Tension-free vaginal tape resulted in a mean cost saving of £243 (95% CI £341 to £201) comparedwith colposuspension. Differential mean QALYs per patient (tension-free vaginal tape � colposuspension)was 0.01 (95% CI �0.01 to 0.03). The probability of tension-free vaginal tape being, on average, less costlythan colposuspension, was 100%, and the probability of tension-free vaginal tape being more cost effectivethan colposuspension was 94.6% if the decision-maker was willing to pay £30,000 per additional QALY.
Conclusion The results from this trial suggest that, over a post-operative period of six months, tension-freevaginal tape is a cost effective alternative to colposuspension. The results will need to be reassessed on thebasis of longer follow up.
INTRODUCTION
Urinary incontinence is a significant health problem,
with 14% of women reporting the symptoms1. Urodynamic
stress incontinence accounts for approximately 50% of
urinary incontinence in women presenting to hospital2.
Physiotherapy has a cure rate of about 50%3 – 5, and surgery
is recommended for those women who fail to respond.
Colposuspension is the most frequently used form of
primary therapy for urodynamic stress incontinence, but
the morbidity and resource use associated with the proce-
dure are considerable.
A recent development in the treatment of urodynamic
stress incontinence is the use of a prolene tape (tension-free
vaginal tape) inserted using a tunnelling approach rather
than open dissection of the retropubic space. This can be
done under local anaesthesia and can be undertaken using
day-case admission6, allowing more rapid return to normal
activity. Given that the use of tension-free vaginal tape
results in changes in a range of health care resources as
well as health outcomes, it is important to assess its cost
effectiveness relative to standard surgical management with
colposuspension.
A multicentre prospective randomised controlled trial
comparing tension-free vaginal tape and Burch colposus-
pension for ‘primary’ urodynamic stress incontinence has
been undertaken in the UK7. As part of the study, patient-
specific resource use and health-related benefits were
prospectively collected during hospitalisation and for a
period of six months after discharge from hospital. This
paper reports the results of a cost–utility analysis under-
taken using these data, adopting a UK National Health
Service perspective.
METHODS
Full details of the design of the trial have been published
elsewhere7. Briefly, the clinical study was a prospective
multicentre randomised controlled trial enrolling 344
BJOG: an International Journal of Obstetrics and GynaecologyMarch 2003, Vol. 110, pp. 255–262
D RCOG 2003 BJOG: an International Journal of Obstetrics and Gynaecology
doi:10.1016/S1470-0328(03)02915-X www.bjog-elsevier.com
aCentre for Health Economics, University of York, UKbDirectorate of Women’s Services, Royal Victoria
Infirmary, Newcastle upon Tyne, UKcDepartment of Obstetrics and Gynaecology, University of
Newcastle upon Tyne, UK
* Correspondence: Dr A. Manca, Centre for Health Economics,
University of York, Heslington, York, YO 10 5DD, UK.
women diagnosed with ‘primary’ urodynamic stress incon-
tinence. Patients were recruited to the trial from urogynae-
cology, general gynaecology and urology outpatient clinics
between May 1998 and August 1999, and were randomised
to either colposuspension (n ¼ 169) or tension-free vaginal
tape (n ¼ 175). A total of 34 women dropped out from the
study, 28 before surgery (23 colposuspension, 5 tension-
free vaginal tape) and 6 afterwards (4 tension-free vaginal
tape, 2 colposuspension). Of the 23 women who did not
undergo surgery in the colposuspension group, 20 withdrew
their consent, 2 discontinued the study due to protocol
violation and 1 patient withdrew for other reasons. Of the
five women who did not have surgery in the tension-free
vaginal tape arm, two withdrew their consent, two violated
the protocol and one withdrew for other reasons. Of the four
women who dropped out from the tension-free vaginal tape
group after surgery, two had a treatment failure, one did not
return for the follow up visit and one withdrew for other
reasons. Finally, one of the two women who decided not to
continue the study after colposuspension withdrew her
consent, while the other did not return for the follow up visit.
Baseline characteristics and clinical results of the ran-
domised trial are detailed elsewhere7. Women who pre-
sented with urodynamically proven stress incontinence
were invited to participate in the trial. Exclusion criteria
were detrusor overactivity, vaginal prolapse requiring treat-
ment, previous surgery for prolapse or incontinence, a
major degree of voiding dysfunction, neurological disease
and allergy to local anaesthetic. In terms of clinical results,
there was no statistically significant difference between the
cure rate in the two groups: 115 (66%) women in the tension-
free vaginal tape group and 97 (57%) in the colposuspension
group were objectively cured. Subjective cure of stress
incontinence was reported by 103 (59%) and 90 (53%) of
women in the tension-free vaginal tape and colposuspension
arm, respectively. This economic evaluation is conducted on
an intention-to-treat basis and includes only the 316 women
who underwent surgery.
Using case report forms completed by clinical staff,
patient-specific data on resource use were prospectively
recorded from hospitalisation to six months from discharge.
Where resource use was not expected to vary between
patients, estimates based on clinical expert advice were
used. Data were collected on resource use during two study
periods: main hospitalisation and follow up at six months.
The former comprised resource use in theatre, inpatient stay
and post-operative complications. Theatre resource use
included staff present (in holding bay, anaesthetic room, op-
eration theatre and recovery area) based on clinical advice;
consumables (in theatre and anaesthetic room); and drug use
(i.e. thromboprophylactics, anaesthetics and antibiotics).
Theatre resource use included any extra tension-free vaginal
tapes. Drugs used for anaesthesia, thromboprophylaxis, se-
dation and the prophylactic antibiotic regimen were set by
the protocol. Other patient-specific theatre resources com-
prised those associated with intra-operative complications.
Inpatient stay was the number of nights spent in the general
ward by each woman. Post-operative resource use included
that related to management of complications (e.g. fever,
wound infection, urinary tract infection), additional consum-
ables (e.g. drains, catheters), post-operative analgesia and
any return to theatre during main hospitalisation. Finally, the
six-month post-operative visit to clinic facilitated the col-
lection of data on additional surgical procedures, use of
concomitant medications, re-hospitalisations, number of
outpatient and day-case visits and general practitioner con-
tacts since hospital discharge.
The differential cost of treating patients in the two arms
of the trial was estimated by valuing the resource use
measured in the study using UK unit costs estimated at
1999–2000 prices, including value-added tax where appro-
priate. Staff time was costed using mid-range salaries
uprated for employers’ costs8 – 10. Consumables were costed
using manufacturers’ list prices, and drug costs were based
on those reported in the British National Formulary11.
Inpatient stay was costed using an estimated average hotel
cost per day in a gynaecological ward obtained from a
survey of three UK hospitals undertaken in 1995 and
uprated for health service inflation12. We assumed the hotel
cost of a day-case visit to be equal to the cost of an
outpatient visit13. Finally, the cost of a general practitioner
consultation was obtained from published estimates8. The
main unit costs used in the analysis are reported in Table 1.
The health outcomes of treatment, over six months follow
up, were expressed in terms of quality-adjusted life years
(QALYs). In order to measure women’s health status at
various points in time, women were asked to complete the
EQ-5D health questionnaire14 at baseline, at six weeks and
six months after hospital discharge. This is a standardised
non-disease specific instrument designed to describe and to
Table 1. Key unit costs used in the analysis.
Item of resource Unit Unit cost (£) Source
Ward ‘hotel’ cost Day 103.00 Ref. 12
Theatre
Staff (tension-free vaginal tape) Minute 2.24 Refs 8– 10
Staff (colposuspension) Minute 2.87 Refs 8– 10
Anaesthetic room
Staff Minute 0.85 Ref. 9
Recovery area
Staff Minute 0.20 Ref. 9
Overheads Minute 2.33 Ref. 12
Key consumables
Tension-free vaginal tape Item 359.45 Ethicon
Staple gun Item 113.42 Ethicon
Outpatient and day-case visits Visit 62.00 Ref. 13
General practitioner visit Visit 15.75 Ref. 8
256 A. MANCA ET AL.
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
value health status. Health status is defined in terms of five
dimensions: mobility, self-care, usual activities, pain or
discomfort and anxiety or depression. Each of these dimen-
sions has three levels of severity: no problems, moderate
problems or extreme problems. In completing the EQ-5D, a
patient defines their health state in terms of the five dimen-
sions, which is transformed into a weighted health state
index score or ‘utility’. Using values elicited from the UK
population15, the index typically ranges between 0 (equiva-
lent to death) and 1 (equivalent to good health), although a
small number of health states are valued as worse than death.
EQ-5D scores at baseline, six weeks and six months were
used to calculate patient-specific QALYs, which were esti-
mated using the area under the curve method16,17. This is
achieved, for each patient, by weighting the time between
the three EQ-5D responses using the health state index score.
Therefore, the QALY seeks to capture the impact of the
alternative treatments in terms of both morbidity and mor-
tality on a single dimension.
Statistical analysis was undertaken using STATA 6.018.
Estimates of mean costs and QALYs for the two treatment
arms were calculated over six months follow up. Costs have
been grouped under four headings: theatre cost, hospital
‘hotel’ (i.e. ward) cost, other post-operative complications
and follow up cost at six months. To indicate the pattern of
the health outcomes over the study period, EQ-5D scores at
baseline, six weeks and six months have been reported.
Given that the time horizon of the analysis was less than a
year, total costs and QALYs remain undiscounted.
To account for the skewed nature of the resource use
data, 95% confidence intervals for the differential costs and
QALYs have been calculated using non-parametric boot-
strap (based on the 2.5th and 97.5th centiles)19,20. In some
patients, resource use data and EQ-5D responses were
wholly or partially missing. Under the assumption that data
were missing completely at random21,22, those observations
where either length of stay in theatre or one of the EQ-5D
assessments was missing were excluded from the base case
analysis. As a result, 53 observations in the tension-free
vaginal tape group and 49 in the colposuspension group
were excluded from the initial analysis, and the base-case
analysis was conducted on a complete case data set of 214
Table 2. Main resource use measured during the trial based on the complete case analysis. Values are expressed as n (%), mean [SD] or median
{interquartile range}.
Item of resource Tension-free vaginal tape (n ¼ 117) Colposuspension (n ¼ 97)
Initial hospitalisation
Length of stay in hospital (days) 2.29 [1.9] 6.67 [1.78]
2 {1– 3} 6 {5–8}
Time in anaesthetic room (minutes) 15.7 [8.9] 18.6 [8.9]
15 {10–20} 17 {12– 25}
Time in theatre (minutes) 39.9 [15.4] 51.7 [22.6]
40 {29–48} 50 {35– 60}
Time in recovery area (minutes) 53.4 [41.8] 97.1 [41.2]
45 {31–60} 96 {69– 120}
Return to theatre* 1 (0.85) –
Minutes in theatre 60 –
Use of tension-free vaginal tapes
1 tape 110 (94.02) –
2 tapes 7 (5.98) –
Additional procedures
0 113 (96.59) –
1 4 (3.41) –
Follow up period
Day case visits
0 112 (95.73) 95 (97.94)
1 5 (4.27) 2 (2.06)
Outpatient visits
0 87 (74.36) 74 (76.29)
1– 2 28 (23.93) 21 (21.65)
3– 4 1 (0.85) 1 (1.03)
5– 6 1 (0.85) 1 (1.03)
General practitioner visits
0 77 (65.81) 57 (58.76)
1– 2 30 (25.64) 28 (28.87)
3– 4 6 (5.13) 9 (9.27)
5– 7 4 (3.41) 3 (3.09)
Re-admissions 2 (1.71) 12 (12.37)
Mean length of stay (min– max) 5 (3 –7) 2.6 (1 – 5)
* During initial hospitalisation.
TENSION-FREE VAGINAL TAPE VERSUS COLPOSUSPENSION 257
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
patients (117 tension-free vaginal tape, 97 colposuspen-
sion). In the sensitivity analysis, the implications of the
missing completely at random assumption for the results of
the analysis were assessed through the use of an alternative
assumption that data were missing at random. This is
equivalent to saying that cases with incomplete data differ
from cases with complete data, but the missing data pattern
is fully predictable from other variables in the data set. On
this basis, we imputed the incomplete values using a
multivariate multiple imputation procedure21 – 23 and con-
ducted a new analysis on the entire data set of 316 patients.
The cost effectiveness of tension-free vaginal tape versus
colposuspension was assessed by relating the mean differ-
ential costs per patient of the two forms of surgery, to their
differential effectiveness in terms of mean QALYs per
patient measured over the six-month follow up period. One
treatment can be defined as more cost effective than its
comparator if one of the following conditions apply: (a) it is
less costly and more effective (i.e. it dominates its compar-
ator); (b) it is more costly and more effective, but its ad-
ditional cost per extra QALY is considered worth paying by
decision-makers; and (c) it is less costly and less effective,
but the additional cost per extra QALY of its comparator is
not considered worth paying by decision-makers. The point
estimates of mean costs and effects can be used to identify
which of these three conditions applies. However, these
means are estimated with uncertainty. Therefore, to account
for uncertainty due to sampling variation, we plotted a cost
effectiveness acceptability curve24 – 27. Given the data col-
lected within the trial, this curve shows the probability of the
tension-free vaginal tape being more cost effective than colpo-
suspension for different levels that the decision-maker may
be willing to pay for an additional QALY28. This is a Bayesian
approach to the presentation of cost effectiveness data29,
although a full Bayesian analysis has not been undertaken.
RESULTS
Table 2 details the main elements of resource use in the
trial. The mean length of stay in hospital with tension-free
vaginal tape was 2.29 days (interquartile range: 1 to 3), as
opposed to 6.67 days with colposuspension (interquartile
range: 5 to 8). Mean time in theatre with tension-free
vaginal tape was 40 minutes (interquartile range: 29 to
48), against 52 minutes (interquartile range: 35 to 60) with
colposuspension. Tension-free vaginal tape patients re-
quired a larger number of resources in terms of additional
surgical procedures and outpatient visits. In the colposus-
pension arm, 12 women (12.4%) were readmitted to hos-
pital by six months follow up, whereas two (1.7%) were
readmitted in the tension-free vaginal tape group.
Table 3. Estimates of mean cost of tension-free vaginal tape compared with colposuspension (UK£). Unless stated, all costs are based on complete cases.
Values are expressed as mean (median) [interquartile range].
Tension-free vaginal tape (n ¼ 117) Colposuspension (n ¼ 97)
Costs
Theatre cost 720 (690) [634 to 754] 437 (422) [343 to 518]
Hospital ‘hotel’ cost 236 (206) [103 to 309] 687 (618) [567 to 824]
Other post-operative cost 28 (0.75) [0 to 4.22] 76 (27) [25 to 30]
Follow up cost at 6 months 74 (46) [0 to 93] 101 (46) [0 to 116]
Total cost per patient 1058 (935) [839 to 1100] 1301 (1195) [1050 to 1449]
Differential costs (tension-free vaginal tape minus colposuspension) �243 [�341 to �201]*
Differential costsy (tension-free vaginal tape minus colposuspension) �242 [�340 to �183]*
* 95% non-parametric confidence interval based on 1000 bootstrap replications: lower band ¼ 2.5 centile; upper band ¼ 97.5 centile.y Based on 316 patients (tension-free vaginal tape ¼ 170; colposuspension ¼ 146), after multivariate multiple imputation.
Table 4. Estimates of mean QALYs of tension-free vaginal tape compared with colposuspension. Unless stated, all results are based on complete cases.
Values are expressed as mean (median) [interquartile range].
Tension-free vaginal tape (n ¼ 117) Colposuspension (n ¼ 97)
EQ-5D values
Baseline 0.778 (0.81) [0.71 to 0.92] 0.785 (0.81) [0.71 to 0.92]
Six weeks 0.788 (0.85) [0.71 to 0.92] 0.754 (0.76) [0.69 to 0.88]
Six months 0.806 (0.85) [0.73 to 0.92] 0.794 (0.85) [0.73 to 0.92]
QALYs 0.397 (0.42) [0.35 to 0.45] 0.387 (0.40) [0.35 to 0.44]
Differential QALYs (tension-free vaginal tape minus colposuspension) 0.010 [�0.010 to 0.030]*
Differential QALYsy (tension-free vaginal tape minus colposuspension) 0.012 [�0.006 to 0.029]*
* 95% non-parametric confidence interval based on 1000 bootstrap replications: lower band ¼ 2.5 centile; upper band ¼ 97.5 centile.y Based on 316 patients (tension-free vaginal tape ¼ 170; colposuspension ¼ 146), after multivariate multiple imputation.
258 A. MANCA ET AL.
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
Estimated mean costs per patient in the two arms of the
trial are reported in Table 3. Although tension-free vaginal
tape patients had a shorter time in theatre, the associated
cost saving was offset by the cost of tension-free vaginal
tape-specific consumables, which resulted in total theatre
cost being higher in the tension-free vaginal tape arm.
However, when hospital ‘hotel’ costs are also considered,
tension-free vaginal tape is, on average, less costly, due to
the shorter length of hospital stay associated with its use.
This remains the case when other post-operative and follow
up costs are included. The mean total cost per patient was
estimated to be £1058 (interquartile range: £839 to £1100) in
Fig. 1. Bootstrap replications (n ¼ 1000) of the mean differences in costs and QALYs generated from the trial data.
Fig. 2. Cost effectiveness acceptability curve. CCA ¼ complete case analysis (tension-free vaginal tape ¼ 117; colposuspension ¼ 97); MI ¼ multivariate
multiple imputation (tension-free vaginal tape ¼ 170; colposuspension ¼ 146).
TENSION-FREE VAGINAL TAPE VERSUS COLPOSUSPENSION 259
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
the tension-free vaginal tape group and £1301 (interquartile
range: £1050 to £1449) in the colposuspension group. Mean
differential cost of tension-free vaginal tape minus colpo-
suspension was therefore �£243 (95% CI: �£340 to
�£201); that is, a cost saving from tension-free vaginal tape.
Table 4 reports the EQ-5D scores at baseline, six weeks
and six months in the two treatment arms. The mean
QALYs per patient were 0.397 (interquartile range: 0.35
to 0.45) for tension-free vaginal tape and 0.387 (interquar-
tile range: 0.35 to 0.44) for colposuspension. The differ-
ential mean QALYs between the two groups was 0.01 (95%
CI: �0.01 to 0.03).
On the basis of the point estimates of mean cost and
QALYs presented in Tables 3 and 4, tension-free vaginal
tape dominates colposuspension; that is, over six months, it
generates higher mean QALYs and results in lower mean
health service costs. However, mean costs and QALYs are
estimated with uncertainty. This is evident when looking at
Fig. 1, which provides a graphical representation of the
joint distribution of differential mean costs and QALYs
after 1000 bootstrap replications. A simple visual inspec-
tion shows that the simulations fall completely below the
horizontal dotted line on the cost axis. However, a consid-
erable proportion of the simulations falls to the left of the
vertical dotted line on the QALYs axis. In other words, all
the uncertainty in this comparison relates to the mean
difference in QALYs between the two forms of surgery.
Uncertainty in mean costs and outcomes is reflected in
the cost effectiveness acceptability curve shown in Fig. 2.
The continuous line shows the cost effectiveness accept-
ability curve for tension-free vaginal tape for the base-case
analysis. This indicates that the probability of tension-free
vaginal tape being, on average, less costly than colposus-
pension is 100%—this is the point on the curve where the
decision-maker is not willing to pay anything additional for
an extra QALY. The probability of tension-free vaginal
tape being more cost effective than colposuspension is
94.6% when the decision-maker is willing to pay at least
£30,000 per additional QALY. As Fig. 2 shows, even if the
decision-maker is willing to pay up to £100,000 per
additional QALY, the probability that tension-free vaginal
tape is cost effective remains above 85%.
A key determinant of the mean cost saving achieved by
using the tension-free vaginal tape is the difference
between the two procedures in terms of inpatient hospital
stay. A sensitivity analysis was undertaken to investigate the
role of mean differential inpatient stay on (1) the probability
for tension-free vaginal tape of being, on average, cost
saving, and (2) the probability of tension-free vaginal tape
being more cost effective than colposuspension when the
decision-maker is willing to pay £30,000 for an additional
QALY. A visual inspection of Fig. 3 suggests that, all else
being equal, the tension-free vaginal tape is more likely to be
cost saving compared with colposuspension, provided that
the difference in terms of inpatient hospital stay in favour of
the tension-free vaginal tape is not less than two days.
Analogously, when the decision-maker is willing to pay
£30,000 for an additional QALY, the tension-free vaginal
tape is more likely to be cost effective as long as the dif-
ferential inpatient length of stay for women in the tension-
free vaginal tape group is no more than one day higher than
for those women undergoing colposuspension.
Fig. 3. Impact of differential inpatient hospital stay on the probability of the tension-free vaginal tape being, on average, cost saving and cost effective.
— Probability of the tension-free vaginal tape being cost saving as function of the difference in hospital inpatient stay; - - Probability of the tension-free
vaginal tape being cost effective, when the decision-maker is willing to pay £30,000 for an additional QALY, as function of the difference in hospital in
patient stay.
260 A. MANCA ET AL.
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
Some of the unit costs used in the analysis might not be
representative of those in other UK hospitals. In particular,
the ‘hotel’ cost per inpatient day in hospital will vary
between centres. In order to explore the robustness of the
results to this unit cost, a sensitivity analysis was conducted
varying this cost between £80 and £250 per day. Unsurpris-
ingly, this had little impact on the cost effectiveness results.
Unit costs greater than the base-case value (i.e. £103 per
day) resulted in an even greater mean cost saving for the
tension-free vaginal tape, being the intervention with a
shorter mean length of stay in hospital. Similarly, the
tension-free vaginal tape is still cost saving for a hospital
‘hotel’ cost per day as small as £80.
Finally, a third sensitivity analysis investigated the
impact of missing data on the results of the present study.
The base-case analysis was conducted on the complete case
data set assuming that data were missing completely at
random. An alternative assumption is that data were miss-
ing at random. Under this assumption, multivariate multiple
imputation can be applied to handle the incomplete data
problem. The dotted line in Fig. 2 shows the cost effective-
ness acceptability curve obtained from the analysis of the
entire data set of 316 patients after the application of
multiple imputation. The conclusion of the analysis,
namely, a high probability of the tension-free vaginal tape
being cost effective across a range of willingness to pay
values, remains unaffected.
DISCUSSION
On the basis of the findings presented in this paper over
six months follow up, the tension-free vaginal tape results
in an overall mean cost saving per patient of £243 while
generating a mean improvement in health outcomes of 0.01
QALYs per patient. Although the cost of a tension-free
vaginal tape is markedly higher than the theatre consum-
ables used during colposuspension, this was more than
offset by a reduction in the mean hotel cost of hospital
stay. In this paper, the value for money of the tension-free
vaginal tape was characterised using a cost effectiveness
acceptability curve, which plots the probability of the
tension-free vaginal tape being more cost effective than
colposuspension as a function of the decision-maker’s
willingness to pay for an additional QALY. If decision-
makers are only interested in costs and do not attach any
value to an improvement in patients’ health outcomes, the
probability of the tension-free vaginal tape being cost
effective (i.e. cost saving) is 100%. However, any health
service will value health gain and, although a formal value
for an additional QALY has not been stated explicitly in
this context, it is possible to infer a broad range for this
parameter based on decisions previously taken regarding
what interventions should be funded30,31. Using a wide
range of values for an additional QALY, the probability of
the tension-free vaginal tape being more cost effective than
colposuspension remains appreciably higher than 80%. The
health service will need to decide whether this level of
certainty is sufficient to justify the widespread use of the
tension-free vaginal tape.
Economic evaluation uses a large number of variables
collected over a period of follow up and, in most trials, a
proportion of data is missing. For the base-case analysis,
only observations with complete data were included (62%).
A sensitivity analysis was conducted to explore the impact
of this decision. This showed that using multiple imputa-
tion rather than complete case analysis had little effect on
the results.
The objective of the analysis was to use unit costs which
are representative of UK hospitals. However, there is con-
siderable variation in some costs, in particular, the hotel cost
of an inpatient day in hospital13. The sensitivity analysis
showed that variation in this cost between two extreme
values of £80 and £250 per day (compared with a base-case
of £103) had little impact on the cost effectiveness results.
The difference in the mean length of hospital inpatient
stay between the tension-free vaginal tape and colposus-
pension seems to be a crucial variable in the present study,
and the extent to which the measurement of this variable in
the trial is representative of what would emerge in routine
practice needs to be considered. If the period of time
patients remained in hospital was more a reflection of
clinical expectations than patients’ needs, the estimate of
differential length of stay may be an under- or over-
estimate. Blinding patients and non-surgical clinical staff
to the procedure undergone, as in other surgical trials32,
was not considered feasible here, in view of the differences
in incision and anaesthetic technique. It is possible that
length of stay for both procedures in the trial will become
lower in routine practice over time. The effect of this de-
cline on clinical success rates is unclear. A key finding in
this analysis is that, as long as average length of stay is at
least two days longer following colposuspension, the tension-
free vaginal tape will remain the less costly procedure. Given
the slight QALY gain with the tension-free vaginal tape,
the threshold that would ensure that the tension-free vaginal
tape is the more cost effective (rather than less costly)
procedure will be still lower. Further observation of both
procedures in routine practice is necessary to illuminate this
issue.
One related issue is that, because the trial from which the
data for this economic analysis are obtained is a pragmatic
one, the colposuspension procedure was not standardised
between surgeons. This lack of standardisation among the
14 units involved with the trial will increase the genera-
lisability of the study findings as this permits variation in
clinical practice to be reflected in the cost effectiveness of
the results. The Burch colposuspension undertaken in this
study was open surgery. Some economic comparisons of
this procedure with laparoscopic colposuspension have
been reported33 – 35. However, none of these studies was
undertaken alongside a randomised controlled trial, nor were
TENSION-FREE VAGINAL TAPE VERSUS COLPOSUSPENSION 261
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262
they full cost–utility analyses. Further research is required to
assess whether laparoscopic colposuspension can represent a
more cost effective use of resources than the open form of the
procedure and the tension-free vaginal tape.
The costs and benefits estimated here are based on a
period of follow up of six months. As more time elapses
from the point of initial surgery, some women may require
additional treatments for their urodynamic stress inconti-
nence. Further follow up of patients in this trial is essential
to see whether these longer term events have substantive
effect on the results presented here.
In conclusion, the results from this economic evaluation
suggest that, over a post-operative period of six months, the
tension-free vaginal tape is a cost effective alternative to
colposuspension. The results will need to be reassessed on
the basis of longer follow up.
Acknowledgements
The authors would like to thank the UK and Ireland TVT
Trial Group for recruiting patients into this trial. The
authors also thank the patients, technical, secretarial,
nursing and medical staff of the participating hospitals
and the monitoring staff of Ethicon. Ethicon financially
supported this study. The investigators were given
complete freedom to analyse the data and report the results
as they saw fit. Mark Sculpher is also supported by a Career
Award in Public Health funded by the NHS Research and
Development Programme.
References
1. Brocklehurst JC. Urinary incontinence in the community —analysis
of a MORI poll. BMJ 1993;306:832– 834.
2. Keane DP, Eckford SD, Shepherd AM, Abrams P. Referral patterns
and diagnoses in women attending urodynamic unit. BMJ 1992;
305:808.
3. Bo K, Talseth T, Holme I. Single blind, randomised controlled trial of
pelvic floor exercises, electrical stimulation, vaginal cones, and no
treatment in management of genuine stress incontinence in women.
BMJ 1999;318:487– 493.
4. Cammu H, Van Nylen M. Pelvic floor muscle exercises: 5 years later.
Urology 1995;45:113 –117.
5. Klaskov P, Belving D, Bishoff N, et al. Pelvic floor exercise versus sur-
gery for female urinary stress incontinence. Urol Int 1986;41:129–132.
6. Ulmsten U, Hendriksson L, Johnson P, Varhos G. An ambulatory
surgical procedure under local anaesthesia for treatment of female
urinary incontinence. Int Urogynaecol J 1996;7:81– 86.
7. Ward K, Hilton P. Prospective multicentre randomised trial of tension-
free vaginal tape and colposuspension as primary treatment for stress
incontinence. BMJ 2002;325:67–70.
8. Netten A, Dennett J, Knight J. Unit Costs of Health and Social Care.
Canterbury: PSSRU, University of Kent at Canterbury, 2000.
9. Review Body for Nursing Staff Midwives, Health Visitors and Pro-
fessions Allied to Medicine. Report 2000. London: HMSO, 2000.
10. Review Body on Doctors’ and Dentists’ Remuneration. Report 2000.
London: HMSO, 2000.
11. British National Formulary. Number 39 (March). London: British
Medical Association and the Royal Pharmaceutical Society of Great
Britain, 2000.
12. Sculpher MJ. Economic evaluation of minimal access surgery: the
case of surgical treatment for menorrhagia [PhD thesis]. Uxbridge:
Brunel University, 1996.
13. The Chartered Institute of Public Finance and Accountancy (CIPFA).
The Health Service Database. Croydon: CIPFA, 1999.
14. Kind P. The EuroQol instrument: an index of health-related quality of
life. In: Spilker B, editor. Quality of Life and Pharmacoeconomics in
Clinical Trials. Philadelphia, Pennsylvania: Lippincott-Rivera, 1996:
191– 201.
15. Kind P, Hardman G, Macran S. UK population norms for EQ-5D.
Centre for Health Economics Discussion Paper 172. York: Centre
for Health Economics, University of York, 1999.
16. Matthews JNS, Altman D, Campbell MJ. Analysis of serial measure-
ments in medical research. BMJ 1990;300:230 –235.
17. Orenstein D, Kaplan R. Measuring the quality of well-being in cystic
fibrosis and lung transplantation: the importance of the area under the
curve. Chest 1991;100:1016–1018.
18. Stata Statistical Software: Release 6.0 (program). College Station,
Texas: Stata, 1999.
19. Efron B, Tibshirani RJ. An Introduction to the Bootstrap. New York:
Chapman & Hall, 1993.
20. Chaudhary MA, Stearns SC. Estimating confidence intervals for cost-
effectiveness ratios: an example from a randomized trial. Stat Med
1996;15:1447–1458.
21. Rubin DB. Multiple Imputation for Non-Response in Surveys. New
York: Wiley, 1987.
22. Schafer JL. Analysis of Incomplete Multivariate Data. London: Chap-
man & Hall, 1997.
23. Little RJA, Rubin DB. Statistical Analysis with Missing Data. New
York: Wiley, 1987.
24. Van Hout BA, Al MJ, Gordon GS, Rutten FFH. Costs, effects and
C/E ratios alongside a clinical trial. Health Econ 1994;3:309– 319.
25. UK Prospective Diabetes Study Group. Cost effectiveness analysis of
improved blood pressure control in hypertensive patients with type 2
diabetes: UKPDS 40. BMJ 1998;317:720– 726.
26. Gray A, Raikou M, McGuire A, et al. Cost effectiveness of an inten-
sive blood glucose control policy in patients with type 2 diabetes:
economic analysis alongside randomised controlled trial (UKPDS
41). BMJ 2000;320:1373– 1378.
27. Delaney BC, Wilson S, Roalfe A, et al. Cost-effectiveness of an
initial endoscopy for dyspepsia in patients over age 50 years: a
randomised controlled trial in primary care. Lancet 2000;356:
1965– 1969.
28. Briggs AH. Confidence intervals or surfaces? Uncertainty on the cost-
effectiveness plane. Health Econ 1998;7:723– 740.
29. Briggs AH. A Bayesian approach to stochastic cost-effectiveness
analysis. Health Econ 1999;8:257– 261.
30. National Institute for Clinical Excellence, www.nice.org.uk.
31. Neumann PJ, Stone PW, Chapman RH, Sandberg EA, Bell CM. The
quality of reporting in published cost–utility analyses, 1976– 1997.
Ann Intern Med 2000;132:964– 972.
32. Majeed AW, Troy G, Nicholl JP, et al. Randomised, prospective, single-
blind comparison of laparoscopic versus small-incision cholecystectomy.
Lancet 1996;347:989–994.
33. Kung RC, Lie K, Lee P, Drutz HP. The cost-effectiveness of laparos-
copic versus abdominal Burch procedures in women with urinary
stress incontinence. J Am Assoc Gynecol Laparosc 1996;3:537 –544.
34. Loveridge K, Malouf A, Kennedy C, Edgington A, Lam A. Laparos-
copic colposuspension: is it cost-effective? Surg Endosc 1997;11:
762– 765.
35. Kohli N, Jacobs PA, Szw EHM, Roat TW, Karram MM. Open com-
pared with laparoscopic approach to Burch colposuspension: a cost
analysis. Obstet Gynecol 1997;90:411– 415.
Accepted 6 November 2002
262 A. MANCA ET AL.
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 255–262