Patient preferences for pilonidal sinus treatments: A discrete choice experiment survey

A range of treatments are available for pilonidal sinus disease (PSD), each of which has a different risk/benefit profile. The aim of this study was to collect patient views on which interventions they would rather avoid and which outcomes they most value for PSD.


INTRODUC TI ON
Pilonidal sinus disease (PSD) is a common condition that affects 26 in 100 000 of the population [1]. Patients-predominantly of working ageacquire an abscess or sinus due to the obstruction and subsequent rupture of hair follicles in the natal cleft [2, 3]. As a chronic condition, patients present to primary or emergency care services experiencing a cycle of pain and infection significantly impacting psychosocial wellbeing [4,5].
The optimal treatment for pilonidal disease should result in rapid healing with minimal postoperative complications. However, it is not clear which if any surgical option meets these requisites, perhaps explaining why clinical practice varies across the UK [6,7]. Whilst there is a large body of literature to guide practice, this research mainly consists of single-centre cohort studies using different classification systems and outcome measures [6]. Surgical treatment typically involves different excision procedures of the affected area. The wound may be left open to heal by secondary intention or closed using techniques such as fibrin glue [8], midline or asymmetric closure with sutures, or rotational skin flaps. Some interventions (e.g., rotational flaps) may result in the need for longer hospital stay and recovery or prolonged nursing care with protracted periods of physical adaptations (e.g., leave open) [9]. Conversely others (e.g., Endoscopic Pilonidal Sinus Treatment, laser, Bascom I, pit picking and glue) may be minimally invasive, with rapid recovery but a higher chance of persistent disease and the need for further intervention. Whilst patients may be involved in making shared decisions about their treatment, many are often uninformed of and unprepared for the burden of potential postprocedural care on daily living or conversely the risk of persistent disease [9].
In 2018, the National Institute for Health Research commissioned the Pilonidal Sinus Treatment: Studying the Options (PITSTOP) study.
PITSTOP is a multicentre (n = 33) observational study exploring the effectiveness of excision and closure techniques in the UK [10]. To enhance our understanding of patient treatment preferences, we conducted a discrete choice experiment (DCE). This quantitative survey method is used to elicit preferences for treatment characteristics (also called attributes) [11]. The DCE task involves participants choosing the best treatment option when presented with a set of competing hypothetical treatment profiles. The DCE data can then be analysed to reveal the relative importance of treatment attributes and to understand what trade-offs respondents are willing to make between two competing attributes. DCEs have increasingly been used to identify patient preferences in health and healthcare [11][12][13][14]. However, the application of this method to pilonidal disease has not yet been undertaken. The aim of this study was to collect patient views on which interventions they would rather avoid and which outcomes they most value.

Development of the survey
The survey contained four sections: (1) patient characteristics and disease history, (2) treatment ranking exercise, (3) DCE tasks ( Figure 1) and (4) survey feedback questions. The DCE and ranking task were developed by conducting qualitative interviews with 20 patients to identify key attributes and important levels that they considered when choosing a treatment.
Further details of the qualitative interview process are reported elsewhere [9]. Thematic analysis of the qualitative interviews identified a list of factors that patients considered to be important when choosing a treatment for PSD (see Data S1). Patient representatives and clinicians in the team reviewed the initial list of themes and helped to select the most important attributes by first grouping similar themes into attributes for the ranking task and then selecting the two attributes that are most important for the DCE through iterative rounds of discussions. Nine attributes were included in the ranking task: (1) type of excision and closure, (2) type of anaesthetic, (3) length of hospital stay, (4) wound care, (5) pain medication requirement, (6) infection risk, (7) healing time, (8) risk of recurrence and (9) scarring. Clinicians in the team generated a treatment classification containing five groups of treatments. This classification development was important because there were approximately 12 different types of procedures used in the UK for PSD [6,15] and asking patients to rank a large number of treatments was too cognitively taxing. The final treatment descriptions for the ranking task based on the classification and the nine attributes were developed with clinical input and information from the literature and were piloted with patients prior to survey rollout (see Figure 2). design catalogue [16]. The questionnaire contained 16 hypothetical DCE tasks. In the DCE tasks patients were asked to choose between two combinations of outcomes with varying levels (see Figure 1). We presented forced unlabelled choices 'treatment A' or 'treatment B' to respondents avoiding the use of an 'opt out' alternative for the purposes of realism. The DCE also included dominant tasks where one treatment option was logically better to test if participants understood the DCE task.
The survey was piloted with three patients to gain feedback on comprehension, interpretation and complexity of the tasks. Upon receiving the pilot feedback, we included the ranking task information in a table (see Figure 2), changed a free text question to a multiple-choice question, and made some formatting changes to the text of the survey.

Sampling
Sample sizes for DCE studies vary substantially in the literature, with sizes ranging from 100 to 1000+ respondents [17]. We adopted the rule of thumb formula used by Orme [18]

Participants and recruitment
All participants aged 16 and above with symptomatic PSD, referred for elective surgical treatment and participating in the PITSTOP cohort, were invited to take part in the DCE survey. Participants were emailed a link via Qualtrics which included a participation information sheet including data protection information, consent form and the questionnaire. Patients with symptomatic PSD not participating in the PITSTOP cohort study could also take part in the questionnaire by accessing a QR code advertised on a study leaflet. The study leaflet was displayed in 33 National Health Service Foundation Trust colorectal outpatient clinics. The study leaflet was also shared on the PITSTOP study Twitter page.

Data analysis
Descriptive statistics were calculated for patient characteristics, disease history and survey feedback variables. DCE responses were modelled using conditional logistic regression where the dependent variable was the preferred treatment choice and independent variables were risk of infection/persistence and recovery time. Linearity F I G U R E 2 Treatment ranking exercise.
of the attributes was assessed ( Figure 3) before deciding to treat risk as a linear variable. Regression coefficients were used to estimate the relative importance of attributes. Maximum acceptable risk is the rate at which patients are willing to give up a benefit in one attribute in exchange for an improvement in another attribute. Maximum acceptable risks were calculated by taking the ratio of recovery time coefficients divided by the infection/persistence coefficient and 95% confidence intervals calculated using the delta method [19]. Furthermore, latent class models were used to analyse individual heterogeneity and to identify subsets of patients with varying preferences. The optimal number of classes was selected using the Bayesian information criterion (BIC) and the consistent Akaike information criterion (CAIC) and model parsimony. Data were analysed using STATA 17.

Participants
Between April 2020 and February 2022 there were 423 unique visitors to the online survey platform and of these 150 participants consented to take part in the survey and three people declined, giving a response rate of 35%. Of the 150 participants who consented, 111 participants completed the survey giving a completion rate of 74%.
The characteristics of the 111 participants who completed the survey are reported in Table 1. The majority of respondents were men (68%), age between 17 and 29 years (66%), employed (80%), white (87%) and educated to A-level or above (81%). The majority of participants had at least one PSD surgery with only six patients reporting not having surgery and one person having 10 surgeries.

Patient preferences
Regression modelling results are presented in Table 2;  show that the direction of effects is as expected where, on average, patients prefer treatments with lower risk of infection/persistence (see Figure 3). Also, compared to a treatment with a 12-week recovery period, participants prefer treatments with shorter (1, 2 or 6 weeks) recovery periods. However, preferences for recovery time attribute levels were not linear as participants preferred a 2-week recovery period over a 1-week recovery period relative to 12 weeks (Figure 3).

Maximum acceptable risk
According to the DCE results the optimum treatment that patients would prefer would have the lowest risk of infection/persistence with the fastest recovery times, but currently available treatments do not necessarily align with this scenario, so patients made tradeoffs and chose a treatment with some limitations but that on balance was acceptable to them. To measure this trade-off, the

Preference heterogeneity
Latent class analysis, a modelling approach to identify subgroups of respondents with distinct patterns of preferences for treatments, was used (Table 4). Two groups with distinct preferences (two classes) were identified based on model fit (BIC/CAIC scores) and model parsimony. Respondents in subgroup 1 strongly preferred treatments with lower risk of infection/persistence, so they were only willing to accept a small risk (1.51-2.15) in exchange for a treatment with faster recovery time. However, respondents in subgroup 2 strongly preferred treatments with shorter recovery time. Therefore, they were willing to accept higher risks of infection/persistence (22.35-34.67) to receive treatments with quicker recovery time. Only respondents' age could significantly differentiate whether they belonged to subgroup 1 or subgroup 2 and other sociodemographic characteristics were not significant (see Data S1).
Respondents in subgroup 1 were more likely to be in the age group 17-29 years and respondents in subgroup 2 were more likely to be in the age group of 30 years and above.

Ranking of treatments
The order of ranking for the best preferred treatments are complex

Participants understanding and engagement of the survey
The overall survey comprehension was high (see Table 5) with most participants reporting that they understood the DCE tasks (91%) and ranking task (86%). The median time to complete the survey was 12 min with a large range between 4 and 5388 min (as participants were able to stop and return to the survey). Most participants (84%) correctly answered the dominance DCE tests, where one treatment was logically superior to the other treatment.  choose the treatment profile on the right side of the questionnaire, exhibiting right side bias (see Table 2).

DISCUSS ION
This study assessed patient treatment preferences for PSD by con- This result is perhaps unsurprising given the likely protracted wound care in this group and its impact on wellbeing [5].
In addition, our results also showed preference heterogeneity, indicating the need for providing tailored treatments to suit subgroups of people with distinct preferences. People in the age group 30+ prioritized treatments with fast recovery and they were willing to accept a much higher risk of infection/persistence of around 35 percentage points. This suggests that people in the 30+ group would be willing to forgo leave open treatment and would much rather take the other treatments with shorter recovery. In our sample, people above 30+ were much more likely to report that they were either homemaking, retired or caring for someone else; these responsibilities may have influenced their preference for treatments with shorter recovery time. In contrast, people in the younger age group The literature on surgery for PSD is immense, perhaps reflecting the variety of surgical interventions that are advocated [20]. As there is no commonly used classification system [15], there is no stratification of disease and no universally accepted treatment algorithm. Judging by the plethora of papers detailing large single-centre cohorts of particular techniques [6], individual surgeons have their own preferred technique, and many patients have little choice in the decision. Outcomes are poorly reported, and some patients have not been asked what is important to them before surgery [9]. There is no core outcome set for pilonidal disease [21]. Assessing the literature in general, some broad assumptions can be made. For example, minimally invasive procedures usually result in rapid recovery but may not be as successful (in terms of reduced recurrence) than more invasive interventions (such as rotational flaps), particularly for more severe or recurrent disease [8,22]. Allowing a large wound to heal by secondary intention usually results in prolonged recovery [23]. Such implications should be discussed with the patient through a shared decision making process. There is a growing body of literature that supports this approach [9,24,25]

CON CLUS ION
Patients with PSD, particularly those above the age of 30, are willing to accept a higher chance of treatment failure in exchange for a more rapid recovery. This supports the use of minimally invasive procedures even if recurrent disease is more likely. Conversely, procedures that may result in a more protracted hospital stay and initial recovery such as rotational flaps may be preferred by some (especially the younger age group) if they offer a higher chance of cure.
The heterogeneity in preferences suggests that surgeons should offer a range of interventions and tailor treatments to individual patient preferences.

FU N D I N G I N FO R M ATI O N
This study was funded by the National Institute for Health Research (NIHR), Health Technology Assessment (HTA) programme (project number 17/17/02). The funder was not involved in the trial design, patient recruitment, data collection, analysis, interpretation or presentation, writing, editing of the report or decision to submit for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest. The study received approval from East of England Cambridge South Research Ethics Committee (REC reference 18/EE/0370).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

I N FO R M ED CO N S ENT
Informed consent was obtained from all individual participants included in the study.