Repair of oesophageal atresia by consultants and supervised trainees results in similar outcomes

Abstract

Introduction

Consultants and trainees require exposure to complex cases for maintaining and gaining operative experience. Oesophageal atresia (OA) repair is a neonatal surgical procedure with indicative numbers for completion of training. A conflict of interest may exist between adequate training, maintaining consultant experience and achieving good outcomes. We aimed to review outcomes of procedures performed primarily by trainees and those performed by consultants.

Methods

We carried out a retrospective case note review of all consecutive infants who underwent surgical repair of OA with distal tracheooesophageal fistula (TOF) between January 1994 and December 2014 at our institution. Only cases that underwent primary oesophageal anastomosis were included. Surgical outcomes were compared between cases that had a trainee and those that had a consultant listed as the primary operator.

Results

One hundred and twenty-two cases were included. A total of 52 procedures were performed by trainees, and 68 by consultants. Two cases were undeterminable and excluded. Infant demographics, clinical characteristics and duration of follow-up were similar between groups. All infants survived to discharge. Procedures performed by trainees and those performed by consultants as primary operators had a similar incidence of postoperative pneumothorax (trainees 4, consultants 3; p=0.46), anastomotic leak (trainees 5, consultants 3; p=0.29) and recurrent TOF (trainees 0, consultants 2; p=0.5). Overall 52% of cases had an anastomotic dilatation during follow-up, with no difference between the trainee and consultant groups (50% vs 53%; p=0.85).

Conclusions

Surgical outcomes for repair of OA/TOF are not adversely affected by trainee operating. Trainees with appropriate skills should perform supervised OA/TOF repair. These data are important for understanding the interrelationship between provision of training and surgical outcomes.

Introduction

All surgical training schemes require a level of operative independence to be reached over time through observation, assisting and performing procedures, with reducing degrees of supervision as experience safely allows.^1,2 Although this varies between specialty, and even between departments, the development and training of future surgeons is essential to continue high standards of care across surgical specialties and to provide safe and competent surgeons.

In the UK, specialist training in paediatric surgery consists of six years of higher specialty training, during which the junior surgeon is exposed to a wide range of conditions, many of which are rare. The syllabus and standards to be achieved are set by the Joint Committee for Surgical Training (JCST).³ To complete training, a trainee must demonstrate competence in performing a range of procedures with full competence defined as being ‘able to perform the procedure fluently ± be able to manage the complications’.⁴ To maximise training opportunities, the JCST recommends that ‘to enable minimum training requirements, trainees should be involved in every case possible and perform where suitable.’⁵

Oesophageal atresia (OA) is a rare congenital anomaly affecting between 1 in 5000 and 1 in 2500 live births worldwide,⁶ with 150–200 cases per year in the UK.⁷ Repair of OA and tracheooesophageal fistula (TOF) is an index case in the UK trainees’ logbook, with a recommendation for involvement in a minimum of ten cases during training, and the trainee performing at least six repairs,³ meaning that opportunities need to be maximised for trainees to gain sufficient experience and competence.

Within contemporary surgical practice, there are three needs: the need to provide high-quality care and best short- and long-term outcomes for patients; the need for adequate, if not excellent, training for surgeons of the future; and the maintenance of consultants’ own surgical skills and expertise (Fig 1). Any conflict of interest may cause one of these domains to suffer. Specifically, training opportunities may be limited if there is a belief that complex procedures performed by trainees result in worse outcomes than those performed by consultants.

Figure 1.

Potential conflicts in surgical training

The aim of this study was to determine whether, following repair of OA with distal TOF, there are differences in important clinical outcomes between cases performed primarily by trainees or consultants.

Methods

We performed a retrospective case note review of all consecutive infants who underwent surgical repair of OA between January 1994 and December 2014 at our institution. Cases were identified via a prospectively maintained database of neonatal surgical admissions over this time period. For the purposes of this study, we included only cases of type C OA (OA with distal TOF) that underwent primary oesophageal anastomosis at the initial surgical procedure. Other anatomical arrangements of OA/TOF and cases that did not achieve oesophageal continuity at the initial surgical procedure (eg cases of TOF ligation followed by delayed primary anastomosis) were excluded.

Demographic and clinical data were recorded, including gestational age, birth weight, presence of congenital heart disease, Spitz classification,⁸ anastomotic tension and preoperative ventilation. These factors were selected as being generalized markers of urgency and difficulty of caseload.

Cases were divided into two groups, depending on the grade of the primary operator for OA surgery, as obtained from the operation notes. For cases performed by a trainee, the number of years before completion of specialist training was determined for each case. Important postoperative outcomes (predominantly complications) were compared between groups. A priori we defined these as pneumothorax, anastomotic leak, recurrent TOF formation and anastomotic stricture. In addition, while reviewing case notes we recorded other significant postoperative complications that required additional treatment. In this study, anastomotic tension was defined as being recorded as such in the operating note; the complication of pneumothorax was defined as a pneumothorax within 48 hours of surgery requiring drainage or resolving spontaneously with no later evidence of anastomotic leak. Anastomotic leak was defined by radiological or clinical evidence (ie saliva in the chest drain) of a leak, although not all had radiological confirmation. A stricture was defined as the need for dilatation at any age. During this time period, elective calibrations/dilatations and routine postoperative contrast studies were not performed. Chest drains were not routinely placed postoperatively.⁹

In addition to comparing outcomes between groups for the time period as a whole, we repeated the analysis for the first 10-year period and the second 10-year period independently. This sensitivity analysis allowed us to determine any difference in trainee involvement in cases and any evidence of improved overall outcome over time.

Data are median (range) or number of cases. Between-group comparisons were performed using Fisher’s exact test or the Mann–Whitney U test as appropriate. This study was part of a retrospective review approved by the UK National Research Ethics Service (15/WA/0153).

Results

During the study period, 168 newborn infants with OA with or without TOF were treated at our institution. Of these, 122 met the inclusion criteria for this study as having OA with distal TOF undergoing primary oesophageal anastomosis at the initial surgical procedure. Two infants were excluded as the grade of the primary operator was not identifiable from the surgical notes. The remaining 120 infants were included. For these 120 infants, the median gestational age was 39 weeks (29–42 weeks) and the median birth weight 2.8kg (1.3–4.2kg).

Sixty-eight procedures (57%) were performed by a consultant as the lead operator, and 52 (43%) by a trainee as the lead operator. All cases had a trainee and a consultant scrubbed for the procedure. Demographic and clinical parameters were similar between the groups, with the exception that all infants with congenital heart disease were operated on by a consultant as the primary operator (Table 1). Trainees were one to nine years (median four years) before completion of specialist training at the time the procedures were performed.

Table 1.

Clinical and demographic parameters of infants operated on by trainees or consultants as primary surgeons*

	Trainees (n = 52)	Consultants (n = 68)	p-value *
Male	32 (62%)	50 (74%)	0.17
Birth weight (kg)	2.8 (1.3–4.2)	2.8 (1.2–4.1)	0.11
Gestational age (weeks)	39 (31–41)	39 (29–42)	0.80
Congenital heart disease	0 (0%)	7 (10%)	0.02
Spitz group 1	50 (96%)	59 (87%)	0.11
Preoperative ventilation	8 (15%)	9 (13%)	0.79
Anastomotic tension	9 (17%)	11 (16%)	1.0

Data are number of cases (%) or median (range)

^*Fisher’s exact test or Mann–Whitney U test

A postoperative pneumothorax was seen in seven cases, anastomotic leak in eight cases, and recurrent TOF in two cases. A total of 62 infants (52%) had a dilatation at some stage during follow-up, thereby meeting our definition of a stricture. All infants survived to hospital discharge. The median duration of follow-up was 56 months (3–211 months) and was similar between procedures performed by consultants and trainees.

Incidence of the individual predefined surgical outcomes and other complications requiring additional treatment was similar between procedures performed by consultants and trainees (Table 2). The proportion of infants in each group who developed at least one complication was similar between the groups.

Table 2.

Postoperative outcomes of infants operated on by trainees or consultants as primary surgeons

	Trainees (n=52)	Consultants (n=68)	p-value*
Postoperative pneumothorax	4 (8%)	3 (4%)	0.46
Anastomotic leak	5 (10%)	3 (4%)	0.29
Recurrent TOF	0 (0%)	2 (3%)	0.50
Stricture	26 (50%)	36 (53%)	0.85
Dilatations per patient	0.5 (0–12)	1 (0–7)	0.69
Other	5** (10%)	6*** (9%)	–
At least one complication	31 (59%)	38 (56%)	0.71
Follow-up (months)	60 (4–217)	48 (3–211)	0.34

Data are number of cases (%) or median (range)

^*Fisher’s exact test or Mann–Whitney U test

^**Wound infection (5)

^***Wound infection (4), chyle leak (1), laryngeal nerve injury (1)

When comparing the first and second decades within our study, we identified that trainee operating rates declined from 47% in 1994–2004 to 36% in 2004–2015. When comparing outcomes across these two time periods, there was no difference in mortality, since all infants survived in both time periods. The rate of anastomotic stricture was lower in the first time period (43% vs 63%), and the rate of anastomotic leak was lower in the second time period (11% vs 3%).

Discussion

In this study, we investigated trainee involvement in cases of repair of OA with distal TOF, and outcomes achieved when procedures were performed with either a trainee or a consultant as the primary operator. This was a single-centre study, and our centre is an established paediatric surgical training centre in the UK. Overall, just under half of all procedures (43%) were performed with a trainee as the primary operator, with a trainer scrubbed for all cases. We found that the grade of the primary operating surgeon did not affect the frequency of postoperative complications.

Adequate exposure to complex cases during training is essential if trainee surgeons are to successfully master operative techniques. The potential conflict of interest that exists between this training exposure and the need to achieve high-quality surgical outcomes is a daily challenge for all surgeons involved in training. In the field of paediatric general surgery, a small number of studies have reported no significant difference in outcome following procedures performed by either a consultant or a trainee for orchidopexy¹⁰ and pyloromyotomy.¹¹ Consultant presence in the operating room (but not necessarily as the primary operator or scrubbed assistant) was not significantly associated with improved 30-day adverse events during paediatric appendicectomy.¹² Of note, these procedures are not regarded as complex when compared with OA repair. To our knowledge, this is the first study to investigate the relationship between grade of trainee and surgical outcomes for a procedure as complex as OA repair. To remove any influence of anatomical arrangement of OA/TOF on surgical complexity, we restricted this cohort to cases of type C OA/TOF.

Data from other specialties also support trainee operating. Within paediatric cardiothoracic surgery (a highly complex specialty), data support trainee involvement in complex cases.^13,14 Centres that actively involve trainees in complex operating report outcomes that compare favourably with national standards, including reduced mortality and length of stay. A systematic review of 182 studies (including only 2 from paediatric surgery) that evaluated the relationship between complications and trainee involvement in cases similarly supports trainee involvement.¹⁵ Data revealed that when a trainee was the primary operator, there was a reduction in mortality and hospital length of stay. However, a significant bias in case selection was reported, resulting in consultants operating on more critical patients. The authors concluded that case selection is an important factor in providing best care for the patient. In the current study, the demographic and clinical factors that may make surgery more challenging (eg low birthweight, prematurity and preoperative ventilation) were similar in both groups, with the exception of infants with congenital heart disease, who were all operated on by a consultant as the primary operator. Although this likely demonstrates a level of case selection, there was evidence of trainees operating on preterm infants of low birthweight – cases that are also likely to be more technically demanding.

Study limitations

There are a number of potential limitations to our study. The outcomes we have reported were selected specifically for this study. We considered surgical complications to be the most appropriate outcomes for comparing procedures performed by trainees and consultants. Although it is possible that other outcomes, such as time to achieve full oral feeds or postoperative length of stay, may be different between procedures performed by trainees or consultants, there are other (non-surgical) factors that may influence these outcomes. We believe the outcomes we have reported are most closely related to the procedure itself and have the greatest potential to be influenced by the experience of the primary surgeon. It is possible that there are differences in outcomes that would be detected only in a larger study. Given that we have not detected any difference in the outcomes studied over a 20-year period, we would question the clinical relevance of such small differences, if indeed they do exist.

The overall number of cases and the number of procedures performed by trainees were distributed evenly over the time period, but the proportion of procedures performed by trainees was reduced in the second decade of the study.

The trainees were recorded at a range of experience between one and nine years before enrolment on the specialist register for paediatric surgery, but it was not possible to determine trainee experience in greater detail than this. As this is a retrospective review, it is not possible to determine the precise involvement of the trainee, even when named as the primary operating surgeon. In practical terms, the operations may be performed partly by trainee and consultant, with intraoperative decisions discussed between the two, depending on the ability and experience of the trainee and the consultant. It is unclear as to whether consultant-only operating would enhance the outcomes for patients, as within the consultant group there is no discernible difference in surgical outcomes.

We highlight the fact that a consultant was present and scrubbed at the operating table for all cases. We do not propose that trainees should necessarily be permitted to perform such complex surgery without direct consultant supervision. Equally, however, we do not believe it would be inappropriate for trainees deemed to have appropriate skill and experience to perform some complex surgery unsupervised. The ethos within our centre has been to develop trainees and encourage operating appropriate for their expertise and experience. Active supervision of trainees in this way is paramount to ensure safe operating, best care for patients, good surgical outcomes and good training opportunities. Discretion and appreciation of individual trainees’ ability and experience are required by trainers. Within this model, we have demonstrated that surgical outcomes can be achieved that are comparable or superior to national benchmarking data.^16,17

Conclusions

Surgical outcomes for repair of OA/TOF are not adversely affected by supervised trainee surgeons performing the procedure. Trainees with appropriate skill and experience should perform OA/TOF repair under supervision, and trainees operating on infants with OA/TOF should be encouraged with the aim of maximising experience in the training years to facilitate achieving competence. These data are important for understanding the interrelationship between provision of training and surgical outcomes and highlight the decline in trainee operating.