The changing face of urinary continence surgery in England: a perspective from the Hospital Episode Statistics database

Objective: To quantify changes in surgical practice in the treatment of stress urinary incontinence (SUI), urge urinary incontinence (UUI) and post‐prostatectomy stress incontinence (PPI) in England, using the Hospital Episode Statistics (HES) database. Patients and Methods: We used public domain information from the HES database, an administrative dataset recording all hospital admissions and procedures in England, to find evidence of change in the use of various surgical procedures for urinary incontinence from 2000 to 2012. Results: For the treatment of SUI, a general increase in the use of synthetic mid‐urethral tapes, such as tension‐free vaginal tape (TVTO) and transobturator tape (TOT), was observed, while there was a significant decrease in colposuspension procedures over the same period. The number of procedures to remove TVT and TOT has also increased in recent years. In the treatment of overactive bladder and UUI, there has been a significant increase in the use of botulinum toxin A and neuromodulation in recent years. This coincided with a steady decline in the recorded use of clam ileocystoplasty. A steady increase was observed in the insertion of artificial urinary sphincter (AUS) devices in men, related to PPI. Conclusions: Mid‐urethral synthetic tapes now represent the mainstream treatment of SUI in women, but tape‐related complications have led to an increase in procedures to remove these devices. The uptake of botulinum toxin A and sacral neuromodulation has led to fewer clam ileocystoplasty procedures being performed. The steady increase in insertions of AUSs in men is unsurprising and reflects the widespread uptake of radical prostatectomy in recent years. There are limitations to results sourced from the HES database, with potential inaccuracy of coding; however, these data support the trends observed by experts in this field.

Hospital Episodes Statistics; HES; urinary incontinence; stress urinary incontinence; overactive bladder; post‐prostatectomy incontinence

Abbreviations

SUI stress urinary incontinence

UUI urge urinary incontinence

PPI post‐prostatectomy stress incontinence

HES Hospital Episode Statistics

TVT tension‐free vaginal tape

TOT tension‐free vaginal tape obturator

OPCS‐4 Office of Population Censuses and Surveys Surgical Operations and Procedures, Fourth Edition

OAB overactive bladder

TOT transobturator tape

AUS artificial urinary sphincter

Urinary incontinence represents a major health burden. Results from the Epidemiology of LUTS study, a cross‐sectional, population‐representative survey conducted in the UK, USA and Sweden, indicate that 0.4% of men and 14.8% of women experienced stress urinary incontinence (SUI) on laughing, sneezing or coughing at least a few times each week [1] . In addition, 4.5% of men and 13.1% of women were found to experience ‘urgency incontinence’ at least a few times each week.

The European Prospective Investigation into Cancer and Nutrition study was a population‐based, cross‐sectional survey, conducted between April and December 2005 in Canada, Germany, Italy, Sweden and the UK, using computer‐assisted telephone interviews [2] . This study reported overall weighted prevalences of SUI to be 0.6% in men and 6.4% in women, urge urinary incontinence (UUI) to be 1.2% in men and 1.5% in women and mixed SUI and UUI to be 0.6% in men and 2.4% in women.

Incontinence after radical prostatectomy is known to be common. The Scandinavian Prostate Cancer Group‐4 trial followed men up for a median period of 12.2 years and reported that 41% of men undergoing radical prostatectomy, compared with 3% of controls, experienced urinary leakage [3] . The European Association of Urology guidelines on urinary incontinence report prevalence estimates of post‐prostatectomy stress incontinence (PPI) of 5–50% [4] ; therefore, with the increase in use of radical prostatectomy, concomitant increases in the overall incidence of PPI are to be expected.

A range of surgical and non‐surgical techniques has been adopted in the treatment of urinary incontinence in recent years and the principles of management have changed with the introduction of newer, less invasive techniques. To investigate these changes within the NHS in England, public domain data from the Hospital Episode Statistics (HES) database were interrogated [5] .

The HES database captures routine administrative data from all hospital admissions and procedures in England. Diagnoses are coded using the International Classification of Diseases, 10th Revision and operations are coded using the Office of Population Censuses and Surveys Surgical Operations and Procedures, Fourth Edition (OPCS‐4) [6] , [7] . Although the HES database was originally developed for the purposes of commissioning and reimbursement, it is increasingly being used for audit, quality improvement and policy decision‐making.

We present the first comprehensive analysis of trends in the surgical management of UUI, SUI and PPI in the NHS in England, from 2000 to 2012.

Public domain information from the HES database, selected using specific four‐character OPCS‐4 codes referring to procedures undertaken for urinary incontinence, was retrieved from the HES website [5] . The full list of procedures included is presented in Table [NaN] [5] . Summary tables for each procedure, presented on the HES website, record the number of procedures undertaken each year. These numbers were extracted and then plotted by year, to show trends in surgical practice, from 2000 to 2012 (Figs [NaN] , [NaN] , [NaN] , [NaN] , [NaN] , [NaN] ). The same codes were used for each year of the study period.

Surgical procedures for urinary incontinence and their four character codes extracted from the HES database 5

1 TVT, tension‐free vaginal tape; TOT, transobturator tape; AUS, artificial urinary sphincter; OPCS, Office of Population Censuses and Surveys Surgical Operations and Procedures.

Table [NaN] shows the number of procedures undertaken for SUI and for urge incontinence and related indications, between 2000 and 2012. Within SUI surgery, a general increase in the use of synthetic mid‐urethral tapes was observed. Admissions for transobturator tape (TOT) increased markedly after the introduction of their OPCS‐4 code in 2006–2007, when 4884 procedures were carried out, to peak at 7010 cases per year in 2007–2008 (Fig. [NaN] ). Similarly, by 2008–2009, admissions for tension‐free vaginal tape (TVT) insertion reached 4505. A substantial fall in the number of admissions for colposuspension procedures was observed, from 3713 in 2000–2001, to 192 in 2011–2012. Notably, coded admissions for partial or complete removal of TVT and for removal of TOT each increased rapidly from 2005–2006 to 2011–2012. The mean age of patients undergoing tape insertions in general was 53.1 years and the mean age of patients undergoing all forms of tape removal was 53.3 years. A detailed record of the age distribution of operations involving mid‐urethral tapes is shown in Tables [NaN] , [NaN] , [NaN] , [NaN] .

Procedures, in numbers of admissions, by year (2000–2012)

2 TVT, tension‐free vaginal tape; TOT, transobturator tape.

Age distribution of admissions for introduction of tension‐free vaginal tape, 2006–2012

3 TVT, tension‐free vaginal tape.

Age distribution of admissions for introduction of transobturator tape, 2006–2012

4 TOT, transobturator tape.

A , Age distribution of admissions for total removal of tension‐free vaginal tape ( TVT ), 2006–2012. B , Age distribution of admissions for partial removal of TVT , 2006–2012

Age distribution of admissions for removal of transobturator tape, 2006–2012

5 TOT, transobturator tape.

Admissions for bladder botulinum toxin A injection have risen to 8000 per year since the introduction of the code in 2005–2006 (Fig. [NaN] ). Percutaneous sacral nerve modulation coding has also increased since the code was introduced in 2000–2001, to 249 recorded cases in 2011–2012 (Fig. [NaN] ). By contrast, the number of ileocystoplasty admissions fell from 155 in 2000–2001 to 89 in 2011–2012 (Fig. [NaN] ). Table [NaN] shows the number of procedures performed in three different age ranges, and the mean age of patients undergoing ileocystoplasty each year, since 2000–2001.

An approximately fivefold increase (from 53 to 261) was observed between 2000 and 2012 in the number of artificial urinary sphincters (AUSs) inserted in men (Fig. [NaN] ).

Analysis of HES data between 2000 and 2012 shows substantial changes in surgical practice for the management of UUI, SUI and PPI in England.

Pelvic floor dysfunction affects over half of middle‐aged women [8] . The present analysis shows that >10 000 women underwent surgery for SUI in 2011–2012. The vast majority of these procedures were either TOTs or TVTs. By contrast, in 2000–2001, before the introduction of codes for mid‐urethral tapes, colposuspension accounted for >3000 procedures annually in the NHS in England. Notably then, in addition to tapes replacing colposuspension, there has been a significant increase in the number of procedures coded for SUI overall.

Hilton [8] offers insight into the various factors underlying this apparent increase in intervention for SUI, including a change in women's attitudes to incontinence and incontinence surgery; previously, incontinence was either simply accepted, or required invasive surgery. The introduction of TVT and TOT made available less invasive alternatives and probably also encouraged surgeons to offer surgical management at an earlier point in the natural history of the condition.

It should also be noted that the rate of increase in coded TVT and TOT procedures is influenced both by the number of procedures performed and the number of procedures coded; newly introduced codes appear to be assimilated at varying rates at different hospitals, resulting in varying reliability of clinically coding. Both TVT and TOT were performed in the NHS in England before the introduction of separate codes for each, and the curves that appear to show the uptake of this procedure must be interpreted with some caution; again, they represent changes in coding practice as well as surgical practice.

In interpreting these data, we must also allow for the potential for clinical coders to confuse TVT and TOT, potentially rendering precise comparisons between the uptakes of the two procedures unreliable. Nevertheless, the relative proportions of each procedure represented by the coded data seem realistic.

Concerns regarding tape removal are increasing among patient groups [9] . The HES summary tables offer insight into the scale of this issue. Tables [NaN] and [NaN] show the rates of tape removals, calculated as a percentage of tape insertions in each year, for TOT and TVT, respectively. Notably, the rate of TVT removal is ∼10 times higher than the rate of TOT removal. Even allowing for differences in coding practice and reliability between the two kinds of tapes, this finding supports the supposition that rates of TVT removal are significantly higher than rates of TOT removal in the NHS in England. Indeed, the rate of TVT removal derived from the HES summary data would imply an even higher rate of TVT complication than the 4.6% reported from data incorporating 5‐year follow‐up of 176 patients undergoing TVT in the UK and Ireland [10] .

Nilsson et al. [11] recently reported outcomes from a cohort of 90 patients, followed up for 17 years. From their single‐centre series, only one case of a minimal, symptom‐free tape extrusion and no other tape complications were reported. Over 90% of the women were objectively continent.

A recent meta‐analysis conducted for the European Association of Urology incontinence guidelines panel incorporated data from 34 randomised controlled trials, including 5786 female patients, and compared retropubic tape with TOTs [12] . No differences in subjective or objective cure rates were reported, but rates of adverse effects including voiding dysfunction and de novo urgency and adverse events including bladder, urethral and vaginal perforation were significantly lower in patients undergoing TOT insertion compared with retropubic tape insertion.

A large, recently updated systematic review comparing outcomes from transurethral tapes and colposuspension concluded that retropubic tapes achieve higher rates of continence than colposuspension but have a far higher risk of intra‐operative complications [13] . Retropubic tapes achieved slightly higher objective cure rates than TOTs but subjective cure rates were similar, and the risk of bladder and vaginal perforations and storage LUTS was lower after TOT procedures. The same systematic review found no significant difference in the rate of delayed complications, including erosion and tape removal, between TVT and TOT procedures. This finding is at odds with the observation made from the HES summary data. The authors did, however, comment on the relative inadequacy of the trials included to evaluate outcomes of relatively low incidence and commented on the superiority, for this purpose, of surgical case series and registries. Arguably, HES data are even better suited to the task of evaluating delayed outcomes of lower incidence, as they enable large numbers of patients to be followed indefinitely; therefore, the higher rates of tape removal suggested within the present analysis may in fact be more reliable than those reported within existing trials.

Interestingly, it has been shown, in a randomised controlled trial of 537 women comparing retropubic and TOTs, that increasing age was an independent risk factor for failure of surgery in patients aged > 50 years [14] . Summary HES data are not optimised to confirm such a finding, but patient‐level HES data analysis would be well suited to examining this at the population level.

The mean patient age at tape insertion is equal to that at tape removal, 53 years (Table [NaN] ). This is consistent with the likelihood that most tape removals are performed within a relatively short time after insertion, mainly because of retention. This has been shown to be a favoured management strategy among experts in female urology [15] . Beyond this, however, firm conclusions cannot be drawn from HES summary data and it is well known that tape removal can become necessary several years after insertion [9] . Investigation into the median time from insertion to removal of tapes would benefit from a more detailed analysis, using patient‐level HES data. Existing registry data may also be useful, but are subject to the bias of selective reporting of cases, submission by clinicians being voluntary [8] .

Augmentation cystoplasty was traditionally used for the small capacity, high‐pressure, poorly compliant or overactive bladder (OAB) and aims to provide urinary storage, protect the upper urinary tract, provide continence and resistance to infection, and offer a convenient method of voluntary and complete emptying [16] .

Bladder wall injection of botulinum toxin A was first described, by Schurch et al.[17] , in patients with spinal cord injury with neurogenic detrusor activity. Since then the technique has been modified and can be performed on an outpatient basis using flexible cystoscopy [11] . Its benefit in the treatment of idiopathic and neurogenic detrusor activity, in terms of symptoms, urodynamic variables and quality of life, has since been demonstrated in a number of randomised controlled trials [18] , [19] , [20] , [21] , [22] , [23] , [24] . More recently, phase III randomised controlled trials have shown botolinum toxin A to be efficacious in treating refractory OAB [25] , [26] .

The rapid uptake in the use of botulinum toxin A is clearly demonstrated by the gradient of the curve in Fig. [NaN] , which does not appear to be slowing down. In addition to its efficacy and a favourable side‐effect profile, the popularity of the procedure undoubtedly reflects its minimal invasiveness, including the facility to provide treatment in an outpatient setting. Repeated treatments are often required to achieve sustained benefit and when interpreting the HES summary data, it must be noted that the number of admissions does not necessarily equate to the number of individual patients treated. For instance, a proportion of the coded procedures recorded in 2011–2012 will correspond to patients treated in previous years and, within any year, a patient may be captured more than once, if receiving more than one instillation.

Furthermore, an assumption has been applied that the OPCS code specifically referring to ‘injection of other therapeutic substance into bladder wall’ refers to botulinum toxin A injection. This assumption, however reasonable, must be acknowledged as such.

A relatively modest increase in the number of coded sacral neuromodulation procedures is observed from the HES summary data, the number of procedures per year reaching 249 in 2011–2012 (Fig. [NaN] , Table [NaN] ). Clearly, botulinum toxin A has been favoured in the treatment of OAB refractory to pharmacological treatment. This is likely to reflect, in part, the comparative lack of level 1 evidence supporting the use of sacral neuromodulation [27] , [28] , [29] , but also the complexity of commissioning and provision of this type of service in various parts of the UK. Furthermore, with urologists increasingly opting to inject botulinum toxin A under local anaesthesia, sacral neuromodulation is often less favoured. Despite this, sacral neuromodulation is approved by the National Institute for Health and Care Excellence and offers an excellent therapeutic option for refractory UUI and urgency–frequency syndrome.

The HES summary data reported in the present study show a steady decline in the number of ileocystoplasty admissions annually (Fig. [NaN] ), which then appears to level off at 80–90 cases per year, since 2008. This plateau may reflect the role asserted for augmentation cystoplasty in treating OAB, notwithstanding a degree of displacement by newer, less invasive techniques that has been asserted by leading specialists in the field, for example, where intravesical therapy has failed.

A series of more specific indications for augmentation cystoplasty is also highlighted, including: OAB in the context of underlying neurological disorder; incontinence secondary to congenital bladder abnormalities; transplantation in the context of significant lower urinary tract dysfunction and infection; and inflammatory bladder pathologies resulting in reduced function capacity (e.g. tuberculous bladder) [16] . It remains undetermined what proportion of ileocystoplasty cases are performed for these specific indications and whether that proportion has changed since the adoption of botulinum toxin A for treating OAB more widely. Although the mean age of patients admitted for ileocystoplasty fell by 6 years between 2000 and 2012, arguably of more significance is the relatively constant number of patients in the 0–15 years age group treated with this operation, reflecting its continued role in patients with neurological disorders and congenital bladder abnormalities (Table [NaN] ).

Ileocystoplasty admissions by age, 2000–2012

Rates of tranobturator tape removals, 2006–2012

6 TOT, transobturator tape.

Rates of tension‐free vaginal tape removals, 2006–2012

7 TVT, tension‐free vaginal tape.

The commonest indication for implantation of AUS in men is PPI. Estimates of the incidence of PPI vary widely, largely because of the wide variation in definitions of incontinence used throughout the literature [21] . The Scandinavian Prostate Cancer Group‐4 trial followed up a group of men for a median period of 12.2 years and reported that 41% of men undergoing radical prostatectomy, compared with 3% of controls, experienced urinary leakage [3] . The Prostate Cancer Outcomes Study, reporting 15 years’ follow‐up from a population‐based US study, reported incontinence in 9.6, 13.4 and 18.3% at 2, 5 and 15 years after surgery, respectively [30] . It is important to note that it is a minority of men who leak urine initially after radical prostatectomy that go on to require surgical treatment for their incontinence. Many are able to return to an acceptable degree of continence with conservative measures, such as pelvic floor exercises. Moreover, men with mild and intermediate severity incontinence, with a preference for a minimally invasive procedure may opt to undergo a male sling procedure if sphincter function is favourable. Nevertheless, as the number of radical prostatectomies increases, a concomitant increase in the overall incidence of PPI might be expected and indeed, this is observed in the summary data examined from HES (Table [NaN] ).

Interestingly, the observed increase in the number of AUS procedures in men, presumed to be secondary to PPI in the present analysis, appears to be at odds with existing single‐centre series that report a gradual decline in PPI with time [31] . It is important, however, to distinguish between studies such as these, which appear to document improvement in outcomes, as a result of increasing surgical experience, in epidemiological analyses of the use of this surgical procedure. Further monitoring of changes in the use of AUS insertions in men will be of interest for two reasons in particular. First, a combination of technical improvements in radical prostatectomy may lead to a slowing of the observed increase in the number of AUS procedures. Second, with a growing emphasis on minimal invasiveness, male sling procedures may increase in popularity, in favour of AUS.

The accuracy and reliability of HES data is dependent on the accuracy of clinical coding and there has been a tendency among clinicians to doubt that accuracy; however, there is growing evidence to attest to the accuracy of HES data. For example, a recent systematic review comparing HES data with medical case notes, found that accuracy was improving and concluded that routinely collected administrative data were sufficiently robust to support their use in clinical research [32] . HES data already have an established role in quantifying changes in the use of different surgical procedures [33] . Furthermore, it is likely that accuracy will improve further with increasing engagement of clinicians with the coding process [34] . Improvements in the reliability of coding procedures and thereby increased capture of procedures by the HES database present a potential issue when using HES to describe changes in practice; it can be impossible to differentiate actual changes in surgical practice from changes in coding practice. As general accuracy and reliability of coding improve, it might reasonably be expected that the number of procedures captured by HES would increase, even without an increased number of procedures actually being carried out.

The introduction of new codes adds complexity to the situation; these may refer to new procedures or may define existing procedures in new ways. In the present analysis, the same codes were used for each of the procedures throughout the study period. This is important, because when a new code is introduced, this may replace several codes, used in various combinations previously.

Although public domain HES data summary tables can describe, in broad terms, changes in the use of the various procedures used in treating SUI, more detailed patient‐level data have the potential to identify more intricate practice patterns. Access to these detailed extracts is restricted, requiring either permission from the National Information Governance Board, including justification of the potential benefit to patients of any proposed study, or encryption of the data to remove any patient‐identifiable data [35] .

For example, while it is of interest that the overall number of augmentation cystoplasty procedures performed annually has fallen, there remains a role for augmentation cystoplasty where less invasive options have failed. Patient‐level data may inform how frequently augmentation cystoplasty is required in patients who have previously received botulinum toxin A instillation or neuromodulation, the numbers of botulinum toxin A instillations individual patients have and the interval between failed non‐invasive treatment and definitive reconstructive surgery. Further detail such as lengths of hospital stay at each point in the pathway could also be derived.

Similarly, in the management of SUI, the increase in tape insertion, in favour of traditional colposuspension is clearly demonstrated within the summary data, but finer detail would require patient‐level HES data. Of particular benefit would be a more precise determination of intervals between tape insertion and removal. It may even be possible to derive risk factors for tape removal, using the full range of coded indicators contained within patient‐level HES extracts.

Likewise, the interval between radical prostatectomy and AUS insertion is not readily available within the public domain HES data summary tables, but may be determined using patient‐level HES data.

A further, more detailed analysis of changes in the surgical management of SUI would be valuable and informative. Patient‐level HES data may be able to answer some of the questions that the public access summaries cannot, yet an initial analysis of public access HES data provides access to information on thousands of patients, raising interesting and important questions concerning the contemporary surgical management of urinary incontinence.

A further limitation of the present analysis is our selection of a limited number of procedures of interest. Utilisation patterns with regard to bulking agents, AUS in women, male sling procedures and other surgical procedures used in the treatment of urinary incontinence have not been considered.

Notably, whereas HES data pertain solely to the NHS in England, similar data are available for the rest of the UK and indeed for many other countries worldwide. Formal consideration of summary data beyond the NHS in England was beyond the scope of the present study, but it has been shown that where individual procedures for continence are separately coded, such summary data can provide insight into changes in surgical practice and, potentially, patterns of complications from newer procedures. It would be of interest in future to consider variations in the rates of procedures undertaken for urinary incontinence between different healthcare settings and to consider the possible drivers for any such variation.

In conclusion, the present analysis quantifies shifts in practice toward less invasive surgical treatment options for SUI and UUI, with a decline in the practice of colposuspension and ileocystoplasty. Basic analysis of summary data on tape removals raises interesting questions concerning the reporting of complications, in the context of these shifts in practice. HES data may have an increased role therein, in the future. Meanwhile, AUS insertion in men has increased in incidence, with the widespread uptake of radical prostatectomy.

The importance of measuring trends in the use of different surgical procedures is to allow us to ensure that changes in practice are supported by an appropriate evidence base and, where newly introduced, are subject to appropriate scrutiny and outcome monitoring. Understanding shifts in surgical practice may also inform discussions about workforce planning, training and sub‐specialisation. This is highly relevant in the rapidly evolving field of urinary continence surgery.

Administrative data such as HES may ultimately yield more reliable information about surgical practice and outcomes than data from either clinical trials or registries. In particular, data such as these seem to be more sensitive to outcomes of lower incidence occuring at delayed intervals and are likely to assume an increasingly important role in health service research, clinical governance and research, both in England and internationally.

AS would like to acknowledge the MRC Centre for Transplantation: Medical Research Council (MRC) Centre for Transplantation, King's College London, UK – MRC grant no. MR/J006742/1. This research was supported by the National Institute for Health Research (NIHR) Biomedical Research Centre at Guy's and St Thomas' NHS Foundation Trust and King's College London. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.

None declared.

Graph: Procedures for stress urinary incontinence by year, 2000–2012.

Graph: image%5ft/bju12650-fig-0001-t.gif

Graph: Admissions for botulinum toxin A (BTXA) injection, 2000–2012.

Graph: image%5ft/bju12650-fig-0002-t.gif

Graph: Admissions for neuromodulation, 2000–2012.

Graph: image%5ft/bju12650-fig-0003-t.gif

Graph: Admissions for ileocystoplasty, 2000–2012.

Graph: image%5ft/bju12650-fig-0004-t.gif

Graph: Admissions for tape removal, 2000–2012.

Graph: image%5ft/bju12650-fig-0005-t.gif

Graph: Admissions for artificial urinary sphincter insertion in men, 2000–2012.

Graph: image%5ft/bju12650-fig-0006-t.gif