Int Urogynecol J DOI 10.1007/s00192-014-2412-z

ORIGINAL ARTICLE

Analysis of the learning process for laparoscopic sacrocolpopexy: identification of challenging steps Filip Claerhout & Jasper Verguts & Erika Werbrouck & Joan Veldman & Paul Lewi & Jan Deprest

Received: 16 September 2013 / Accepted: 16 April 2014 # The International Urogynecological Association 2014

Abstract Introduction and hypothesis We earlier demonstrated that the operation time of laparoscopic sacrocolpopexy (LSCP) by an experienced surgeon drops significantly after 30 cases to reach a steady state after 90. We now aimed to define the learning curve and to identify the most challenging steps for a trainee learning LSCP. Methods Prospective consecutive series of 60 patients undergoing LSCP performed by a trainee experienced in operative laparoscopy but not LSCP. Prior to the first case, the trainee primed his endoscopic suturing skills on an endotrainer for 15 h. His operation time and performance score were analysed using moving average analysis (MOA). The former and the occurrence of complications or short-term failures were compared with those of a concurrent control group consisting of patients operated on by a surgeon experienced in LSCP (teacher). The procedure was empirically divided into five consecutive steps (dissection of the promontory, the paracolic gutter and vagina, suturing of the mesh to the vault, stapling to the promontory, and peritonealisation). Results The MOA of the operation time demonstrated a learning curve for all steps, except for the dissection of and fixation to the promontory. The most time-consuming step is the J. Verguts : E. Werbrouck : J. Veldman : J. Deprest Pelvic Floor Unit, University Hospitals, KU Leuven, Leuven, Belgium J. Verguts : P. Lewi : J. Deprest Department of Development and Regeneration, KU Leuven, Leuven, Belgium J. Deprest (*) Department of Obstetrics and Gynaecology, UZ Leuven, Herestraat 49, 3000 Leuven, Belgium e-mail: [email protected] F. Claerhout Gynaecologie en Verloskunde, Sint Lucas Ziekenhuis, Brugge, Belgium

dissection of the vault, for which it took the trainee 31 procedures to achieve an operation time comparable to that of the teacher. Also, the quality of the dissection improved over time. Suturing of the implant to the vault and peritonealisation took only 10 and 6 procedures respectively. There was no difference in the occurrence of major complications and in one case the trainee asked for assistance. Conclusion Quality of LSCP improves with experience. Operation time falls as well, and the most time-consuming step is the dissection of the paracolic and perivaginal spaces. Prior training in laparoscopic suturing coincided with a short learning process for the phases requiring suturing. Keywords Sacrocolpopexy . Laparoscopy . Vault prolapse . Learning curve . Moving average analysis

Introduction Sacrocolpopexy has been shown to be very effective for treating apical vaginal defects [1]. Randomized trials have shown that abdominal sacrocolpopexy offers lower recurrence rates and less dyspareunia than sacrospinous fixation, but at the expense of a longer recovery time [1]. A laparoscopic approach may overcome this limitation, reducing blood loss and the need for excessive abdominal packing, which may all lead to lower morbidity. A recent randomised controlled trial, meanwhile, confirmed that laparoscopic sacrocolpopexy is equally effective as “open” abdominal sacrocolpopexy [2]. There are, however, a number of hurdles that may hinder the wider implementation of laparoscopic sacrocolpopexy (LSCP). First, the relatively long operating times may be a significant burden on the operation schedule. Second, the use of specific instruments and consumables may cause an additional cost. Third, the rarity of the operation is a limitation; thus, only a few surgeons may eventually acquire wide

Int Urogynecol J

experience with this operation. Last but not least, LSCP requires advanced endoscopic dissection and suturing skills, causing a lengthy and difficult learning process to obtain reasonable operation times and low conversion rates [3]. In a series of 206 consecutive patients operated on by a single surgeon, the operation time declined rapidly over the first 30 procedures, to drop down more slowly, thereafter, reaching a steady state after 90 cases. Obviously, this is a high number and we were pleased to obtain a better insight into the limiting factors or challenging steps of the procedure causing this. Identification of these limitations may allow for proactive steps in training, prompt novel instrument developments or the use of alternative techniques. For that purpose we embarked on a study quantifying the learning process of fellows for this procedure in order to identify potential bottle necks in the operation. This may be helpful in the design or comparison of specific training programmes or steering instrument development. Such information cannot be retrospectively gathered from the published data available or from our previous implementation study [3], because operation times and step-specific complications were purposefully not documented.

Materials and methods Study design, cases and operative technique This is a prospective study analysing the learning process of a single clinical fellow in performing LSC. It included 60 unselected, consecutive LSCPs for symptomatic anatomical stage II prolapse of the vaginal apex (point C200 procedures), further referred to as the teacher. No proficiency level was required to move across these phases. The trainee was neither assisted by, nor received feed-back on his performance from the teacher during the time he was performing his part of the operation. However, supervision was present and an intervention would have been made if a dramatic complication had been anticipated. The trainee could also spontaneously ask for assistance from this supervisor if he felt uncomfortable or unsafe. The operative setting and technique have already been described in detail elsewhere [6]. In essence, the promontory, paracolic gutter and vaginal vault are dissected using monopolar scissors and atraumatic forceps. Suspension was done with two separate macroporous polypropylene meshes, one sutured to the anterior and one to the posterior aspect of the vagina. A minimum of three rows of three sutures were placed using an extracorporeal knot-tying technique using a reusable knot pusher (Ranfac KPL-4, the Surgical Company, Eindhoven, The Netherlands). The anterior implant was connected to the posterior mesh at several locations and the entire construct was suspended free of tension to the promontory with 3–4 staples (EMS hernia stapler; Ethicon, Groot Bijgaarden, Belgium). We closed the pelvic peritoneum with a running suture (Monocryl 0, Ethicon) and over the promontory with staples. Outcome variables The outcome variables included were:

1. Dissection of the promontory, which is the step where major haemorrhage may occur 2. Dissection of the right parasigmoidal and para-rectal gutter and the vaginal vault, where perforations may occur, or the extent to which it is dissected may contribute to anatomical outcome

1. Operation time needed to accomplish each step of the operation 2. A score for performance 3. The occurrence of perioperative complications 4. The occurrence of failures

Int Urogynecol J

The operation time was recorded in minutes and calculated from the first incision up to the moment the surgeon considered the step completed. The suturing time was recorded from the moment the first suture entered the abdominal cavity till the last suture was cut. We recorded the number of sutures and post hoc the time per suture was calculated by dividing the total suturing time by the number of sutures. If anatomy needed to be restored or adhesiolysis was required, the time required to do so was registered separately. At the end of each step the performance was scored by the teacher using a visual analogue scale (0–10), with 0 indicating worst performance and 10, best performance. To score the performance the teacher took the following characteristics into account: 1. For steps 1 and 2 the dissection planes together with extension 2. For steps 3 and 5 the number of sutures, the quality of the knots and the possible presence of folds within the mesh 3. For step 4 the number of staples and their location Complications were classified as either intraoperative or within the first 6 postoperative weeks. They were considered to be either major (perforation of the vagina, urinary tract or bowel, unintended additional surgery, haemorrhage requiring reintervention, thromboembolism, re-admission or re-operation within 6 weeks) or minor (signs of either local or systemic infections, haemorrhage requiring transfusion without the need for reintervention, mesh erosion). Failures were classified as intraoperative or postoperative. Intraoperative failure would be when the trainee was not able to accomplish a certain step himself and called for assistance from the teacher. Postoperative failure was defined as anatomical recurrence within 3 months of the operation, i.e. stage II prolapse at any compartment according to Bump et al. [7]. Statistical analysis Operative time is very often used to measure task efficiency, although it remains a crude proxy and may not be related to true proficiency. It is, however, easily collectable and has been used very often when studying minimally invasive procedures. Because of patient-to-patient variability it is difficult to find a pattern in a scattergram of serial data. First, we analysed the data in chronological blocks of 10 patients to determine changes in operation time, performance score, occurrence of complications and failures [8, 9]. Trends in operation time were additionally tested using the moving average method (MOA). This averages a number of successive observations centred on the time point of current interest. Moving averages smooth individual fluctuations in continuous variables and disclose relevant trends [10]. To control, we used the

data of 30 unselected concomitant controls, who were operated on by the teacher when the fellow was not available. Data were analysed using SPSS 15 (SPSS, Chicago, IL, USA) and for MOA, JMP was used (JMP 7.0; SAS Institute, Cary, NC, USA). Differences between groups were tested using ANOVA or Pearson’s Chi-squared test. Logistic regression was used to identify factors (age, BMI and case number) potentially associated with difficulty or the occurrence of complications or failures and linear regression analysis for operation time. Factors that showed up in a univariate analysis were included in a multivariate regression. The MOA was used to construct the learning curve for each part of the procedure (detailed in Claerhout et al. [3]). We used a smoothing window of consecutive data points to determine a trend. The operation times of the trainee were compared with those of the teacher in different ways. First, we determined the teacher’s mean operation time and SD for each step. Second, we normalised the operation times of the trainee (mean operation time by the trainee minus reference time [mean operation time teacher] divided by the SD of the teacher operating time) and made an MOA of these normalised values, so that the position of the trainee in respect of the teacher could be determined. A value between 0 and 2 indicates that there was no significant difference in operation time between the trainee and the teacher. Conversely, a value above (or under) 2 indicates a significant difference in operation time (p

Analysis of the learning process for laparoscopic sacrocolpopexy: identification of challenging steps.

We earlier demonstrated that the operation time of laparoscopic sacrocolpopexy (LSCP) by an experienced surgeon drops significantly after 30 cases to ...
367KB Sizes 2 Downloads 3 Views