Overview
The Idea, Development, Exploration, Assessment, Long-term follow-up (IDEAL) framework143 was designed to improve the quality of surgical research. Several large, multicentre, prospective longitudinal cohorts have followed this framework, allowing some understanding of variations in practice and their effects on outcomes of other surgical techniques.144–146 PITSTOP also followed this framework, with the aim of answering some key questions in the surgical field of pilonidal disease. The need for this approach to pilonidal disease is clear. While the literature on this subject is vast, the quality of this literature is poor. An initial mapping exercise found that only 12% of the 983 identified primary research articles were randomised trials, and our current understanding of treatment relies mainly on poorly designed cohort studies which cannot provide us with reliable and reproducible estimates of treatment effects.119 There is an absence of clear front-running surgical interventions.13 Interventions are numerous and there are issues with heterogeneity of definitions and measurements of outcome.13
Given the multitude of management options available, the initial work stream focused on surgeon preferences. A survey of over 100 UK surgeons was considered representative of real-world UK practice. Even with evidence to the contrary, a substantial proportion of surgeons who answered the survey perceived very high failure rates regardless of intervention. Again, regardless of the evidence, many practised non-surgical interventions and carried out procedures that experts in the field would consider obsolete (namely excision and leave open or midline closure). Even when minimally invasive procedures were perhaps appropriate, they were not considered as options by a substantial proportion.
This apparent disregard for the evidence could relate to the recognition of a low-quality evidence base and dismissal of the literature.13 It could also relate to the unglamorous nature of pilonidal surgery. The specialist colorectal surgeon who is often tasked with managing the condition in the UK may tend to focus on more challenging conditions. ‘Pilonidal sinus cases are often just fillers on all day lists’ is a quote from an involved member of the core clinical team. The apprentice style of UK surgical training for pilonidal disease may also offer an explanation for the perpetuation of outdated techniques or even newer techniques done incorrectly. A mentor surgeon who has ‘always done it this way’ and a training syllabus that fails to keep pace with modern developments may be contributors.
Although a survey hints at the real-world experience of pilonidal sinus surgery in the UK, a more robust confirmation of this experience is required. The main work package of the PITSTOP trial was therefore a prospective cohort study. Involving over 30 centres throughout the UK, this again was considered reflective of current practice. The multitude of procedures utilised, and the most performed procedures, were consistent with the survey findings. Indeed, the continued use of potentially obsolete procedures (excise and leave open and midline closure) was also confirmed.
While 40% of procedures were classed as minimally invasive, the disease characteristics of the cohort suggest that more patients would have been eligible for such techniques. This is pertinent because such procedures were shown to result in less pain, less complication risk and more rapid return to normal activities, even after accounting for case mix. While the chance of cure was increased with the more major excisions, it may be that some patients would prefer to trade more rapid and less complex recovery for a moderate decrease in efficacy.
Perhaps the most surprising result was the protracted time a significant proportion of patients took to heal and return to normal activities. It is possible that one-quarter of patients had not healed or had persistent disease at 6 months regardless of type of intervention. In an essentially active working population, possibly one in eight patients had not resumed normal activities. This has important implications in terms of the impact on health resource as well as the ability to counsel patients accurately before surgery. It contradicts most of the literature on the subject, with many studies reporting extremely low failure rates.39,40,147
One explanation for this difference between our findings and the reported literature is the definition of ‘failure’ compared with ‘recurrence’. This is a major issue with the existing evidence. Very few studies attempt to define what is meant by ‘recurrence’. A review of the relevant literature taken from the most recent guidelines148 indicates that over 85% of studies that investigate recurrence as an outcome fail to define it at all.
In the true sense of the word, ‘recurrence’ refers to disease that has resolved after an intervention but then recurs after a sufficient period to indicate it is not simply the original disease. It should be differentiated from disease that never resolves, or symptoms that remain unresolved due to an unhealed wound. Therefore, if previous studies report on true recurrence, the incidence may be very low as this is likely to be rare. The most clinically valid outcome is a combination of true recurrence and failure of healing after a reasonable time point. We reported on this combination as being relevant clinically. However, the incomplete data on ‘recurrence’ specifically after at least 6 months and the fact that patients themselves reported on ‘recurrence’, adding an element of subjectivity, may limit our interpretation of the results, and further in-depth analysis is required to allow robust comparisons with other studies.
Another explanation of the difference between our data and the literature is the skill of the surgeon. The cohort study included multiple surgeons of varied expertise. It may be that certain experts can achieve the success portrayed in the literature.11 Analysis of individual surgeon data did show a difference in treatment failure rates between surgeons. Although numbers were small, this may justify that for optimal care patients, particularly those with complex disease, should be referred to specialist units. Alternatively, the skills of the more general surgeon should be enhanced.
Data from the consultant survey and the cohort study revealed a preference in favour of more aggressive interventions rather than minimally invasive procedures. This suggests that some surgeons may focus on cure rather than symptomatic improvement and believe that more major procedures result in a higher chance of cure even if minimally invasive procedures are possible. Our cohort data confirm a higher cure rate. However, patients may wish to trade this 10–15% increased chance of cure for significantly less pain, fewer complications and a more rapid return to normal activities. We explored this hypothesis utilising two qualitative methodologies: a mixed-methods substudy and a DCE.
The mixed-methods study suggested a lack of SDM for some patients, with many not being given a choice of procedure or informed fully about postoperative burden of care. This led to high levels of DR when procedures were not completely successful and protracted periods of recovery became necessary. Even when patients were involved in the decision-making, if not fully informed about postsurgical pathways, their expectations were often not met. Sufficient and accurate information about the risks of protracted postprocedural aftercare should be highlighted to address the false optimism many patients may have.99,101 If given a choice, some patients may elect for alternative procedures where outcomes may differ.
The element of patient choice was explored further in the DCE. While cure of the disease was considered a priority by many, some were prepared to trade the chance of complete cure for a less protracted recovery. This was particularly the case for older patients (> 30 years) where an acceptance of up to 35% increased risk of persistent disease was tolerated in exchange for a shorter recovery period. This again emphasises the need for SDM and tailoring treatment according to the individual patient and their treatment goals.
While these two workstreams suggested a need for improved decision-making and the potential for DR after surgery, the data from the cohort study looking at these parameters revealed conflicting results. The median CollaboRATE score, a tool for assessing the quality of SDM, was very high. In addition, 84% of those who were assessed for DR were either very satisfied or satisfied with the surgery. This is despite around 45% having complications of surgery and 25% having persistent symptoms 6 months after surgery. These contrasting data could be explained by social desirability bias – the tendency to report higher scores out of gratitude or deference.46
Of course, SDM becomes difficult if the surgeon only specialises in one technique.30 Such surgeons should consider expanding their armamentarium to provide an individualised recommendation and choice for the patient, or consider referring to a specialist who may be able to offer such a service.
The literature on pilonidal sinus surgery is confusing and misleading due not only to multiple interventions and no obvious gold standard comparator, the lack of definitions (particularly of recurrence), but also to the heterogeneity of disease severity. Many researchers make no attempt to classify or stratify disease. As such, it is difficult to draw meaningful conclusions about comparative studies. Attempts to classify disease for the purposes of improving the quality of research have been made, and these have been reviewed as part of the PITSTOP study.110 The Wysocki classification demonstrated moderate but acceptable agreement among the surgeons participating in this classification exercise. While there was only moderate agreement, the kappa exceeded previously reported inter-rater reliability of other surgical classifications such as grading of haemorrhoids120 and dysplastic colorectal adenomas.142 Agreement was similar among specialist, general and trainee colorectal surgeons, which offers reassurance that this system could be used across a range of surgeons. Only 2.5% of cases were considered not to fall in the four categories, indicating that the classification system incorporates reliably definable disease characteristics for almost all presentations of disease. Finally, there was a suggestion that the classification could be prognostically valid, with class 1 disease, in particular, being associated with more favourable outcomes.
We concluded the PITSTOP study with a consensus exercise. This utilised a ‘so what, now what’ workshop incorporating data from the cohort study, an e-Delphi exercise and Qualtrics survey technology to consolidate patient and survey views as to the front-running policy and research statements. The policy statements highlighted some key outcomes from the other work packages and are included in the implications for practice and research discussed below.
Implications for practice
While minimally invasive procedures may not be suitable for all, they should form part of the armamentarium of each pilonidal sinus surgeon. Such interventions fit with the philosophy of not making the surgery worse than the disease itself.
The perceived high failure rate for pilonidal disease by many UK surgeons is a concern. Perpetuation of obsolete techniques by a substantial proportion, combined with newer techniques potentially done badly, emphasises the need for better guidance and training. National associations should take on this challenge by providing up-to-date guidance and influencing training through workshops and mentorship programmes.
Shared decision-making is essential, with patients offered an array of interventions allowing them to choose based on preferred outcomes. Many patients would be happy to trade a shorter recovery period for less chance of cure. They value the time to return to normal activities as an outcome, and this should be included in the decision-making process, aiding selection of interventions. If surgeons practise a ‘one fit for all’ intervention, they should consider learning a broader range of techniques or referring patients to a surgeon who can offer this service. An individualised approach based on the severity of disease and the wishes of the patient, combined with detailed information about interventions and potential aftercare, will improve patient expectations and reduce DR.
The Wysocki classification seems to provide a reliable tool and has some validity when it comes to prognostication. Further work is required to develop the tool to include some form of stratification of disease. This will help the surgeon in the choice of which interventions are appropriate.
Implications for research
The grouping of procedures into those that involve major excision and those that are minimally invasive could simplify both the process and the interpretation of future comparative trials. The impression from the core clinical team involved in PITSTOP was that this grouping was fair, although perhaps with the exclusion of excise-and-leave-open and midline closure techniques in the major excision group.
A classification system involving relevant disease characteristics is essential if future comparative trials are to be interpreted and meta-analysed in a meaningful way. Such a system should strive to include some form of stratification of disease severity. The Wysocki classification goes some way to meeting these requirements. Further application may allow development of a treatment algorithm or decision tree to aid surgical decision-making.
Future trials should include a robustly developed core outcome set which includes important PROs.
Equality, diversity and inclusion
No active steps were taken to make participation representative. Participants were representative of the disease population, with people of different disease severity included. White people were marginally overrepresented (see Table 5): 85% in PITSTOP versus 82% in the UK. We achieved a representative sample of Asian/Asian British people (9% in PITSTOP vs. 9% in the UK). We somewhat under-represented mixed/multiple ethnic groups (2% in PITSTOP vs. 3% in the UK) and black/African/Caribbean/Black British (1% in PITSTOP vs. 4% in the UK). This deficit could be addressed in future studies by opening more sites in London and the West Midlands and developing materials that are inclusive, accessible and encouraging to under-represented groups. Our core research team includes non-white members and a range of clinical and methodological expertise. Development opportunities were provided for entry-level researchers to present at conferences149 and act as first/corresponding authors30,110 on papers in peer-reviewed journals. The Associate PI scheme gave five junior clinicians, two of them non-white, the opportunity to contribute to the study.
Patient and public involvement
Patients informed the design of this study, ensured the methods selected were appropriate for patients, and reviewed and commented on questionnaires and other data collection methods. They advised on the appropriateness of the plain English summary and were named as co-applicants. Patient representatives steered the project through the research process, attending management group meetings. They assisted in the design of the protocol, patient information and consent forms. In the mixed-methods substudy, they assisted the analysts in developing themes from the data and contributed to the interpretation of data, with one person with lived experience acting as a co-author on the resulting publication. Expert patients were participants in the Delphi survey. Patients have helped us to design plain English summaries of findings.