Study design

This study follows a longitudinal, prospective observational design. The registry has no anticipated completion date; patients are recruited on a continuous basis and followed until they reach 18 years of age. Fig 1 summarises the registry workflow, from identification of eligible patients through consent and baseline data capture to the schedule of longitudinal follow-up assessments and the definitions applied at censoring.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. COCOE Registry workflow.

Identification of eligible patients, consent and assent, baseline and annual data capture, patient-reported outcome administration schedule, and censoring definitions.

https://doi.org/10.1371/journal.pone.0356813.g001

Study population and setting

Eligible paediatric patients (aged 0–17 years) with documented diagnoses of HD or ARM are invited to participate at the COCOE clinic, Montreal Children's Hospital. Prospective patients are identified through weekly reviews of colorectal clinic schedules.

Although HD and ARM are embryologically distinct conditions with different surgical approaches and long-term outcome profiles, both require lifelong multidisciplinary follow-up and are managed by the same specialist team at COCOE. Enrolling both conditions within a single registry infrastructure reflects this shared clinical pathway and the impracticality of maintaining two separate registries at a single centre. To respect these clinical and methodological distinctions, primary and multivariable analyses will be conducted separately for each disease cohort rather than pooled, with disease-specific instruments and predictors as detailed in the Statistical Analysis section.

Catchment and representativeness

The Montreal Children’s Hospital is one of two paediatric tertiary care centres in Québec and receives colorectal referrals from across the province, including from Nunavik and the Cree territory of James Bay, as well as selected complex referrals from other provinces. The registry therefore captures a substantial proportion, but not the entirety, of the incident paediatric HD and ARM population in Québec. To quantify representativeness rather than assert it, the age, sex, diagnosis and severity distribution of enrolled participants will be compared at each annual analysis against two external benchmarks: (i) expected provincial case counts, derived by applying published Canadian incidence estimates for HD (2.05 per 10,000 live births [1]) and ARM (approximately 1 in 2,162 births [3,4]) to Québec live-birth denominators; and (ii) the published case-mix of the ARM-Net and PCPLC registries [11,13]. Aggregate, non-identifying information on patients who are approached but decline enrollment (age band, diagnosis, and referral region) will also be recorded so that a non-participation comparison can be reported. Any material divergence, for example under-representation of remote communities or of the most severe Krickenbeck categories, will be reported explicitly alongside registry outputs and taken into account when interpreting external validity.

Eligibility criteria

Inclusion and Exclusion Criteria are described in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0356813.t001

Outcome measures

The primary outcome and its definition are described in detail below. Pre-specified secondary outcomes are presented in Table 2 below, each with its definition, instrument or data source, timing of ascertainment, and analytic role.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Pre-specified primary and secondary outcomes.

https://doi.org/10.1371/journal.pone.0356813.t002

The primary outcome is time to achievement of fecal continence from the date of definitive surgical repair, analysed as a time-to-event outcome. For participants with ARM, continence is assessed using the Baylor Continence Scale (BCS) [16], with social continence defined as no soiling episodes in the preceding 4 weeks per published BCS criteria, completed by the parent or caregiver proxy up to age 11 and by self-report from age 12. For participants with HD, continence is assessed using the Bowel Function Score (BFS) [17], with social continence defined as a BFS score ≥ 17, administered to the parent or caregiver proxy up to age 11 and by self-report from age 12. Both instruments are administered at every annual clinic visit beginning at age 3 years. The earliest clinically meaningful assessment for continence is age 3, consistent with normal developmental milestones for voluntary bowel control. Participants will be censored at age 18, death, transition to adult care, or loss to follow-up, whichever occurs first.

Pre-specified secondary outcomes, listed in priority order, are: (1) Hirschsprung-associated enterocolitis (HAEC), defined as at least one physician-confirmed episode meeting the Ruttenstock diagnostic criteria and requiring medical treatment, assessed at each annual visit [20]; (2) reoperation rate, defined as any unplanned surgical return to the operating room following index repair, assessed cumulatively; (3) disease-specific quality of life, measured using the HAQL [18] at ages 6, 10, and 14 years using age-appropriate versions; (4) transition readiness, assessed at ages 14 and 17 years using the validated Transition Readiness Assessment Questionnaire (TRAQ) [19], and (5) loss to follow-up rate, defined operationally in the Loss to Follow-Up section. Each secondary outcome will be reported descriptively and, where the events-per-variable threshold is met (see Sample Size section), included as an endpoint in pre-specified multivariable analyses with the analytic role specified in the Statistical Analysis section.

Sample size and expected enrolment

The registry is designed for indefinite, continuous enrollment. We distinguish two enrollment streams: incident cases (children newly diagnosed with HD or ARM presenting to COCOE, estimated at a minimum of 5 per month) and previous cases (patients previously diagnosed but lost to follow-up and re-engaged, estimated at approximately 3 per month). Combined, the registry aims to enroll approximately 96 participants per year. Based on a prevalent backlog of an estimated 150–200 patients currently followed at COCOE, total enrollment over 5 years is projected at 480–680 participants.

While no minimum sample size is required for registry operation, we provide a feasibility projection. Based on published ARM cohort data reporting approximately 24–30% achieving full fecal continence [14], and HD cohort data reporting approximately 39% achieving a BFS ≥ 17 by adolescence, we project a combined continence event rate of approximately 30–35% within the first 5 years of registry operation [21]. This corresponds to approximately 145–240 continence events across the projected 480–680 enrolees, supporting inclusion of up to 14–24 covariates in the primary Cox model at an events-per-variable (EPV) threshold of 10 [22]. Importantly, this time-to-event framing includes all enrolled participants regardless of age at enrollment, resolving the limitation of the age-5 cross-sectional design by allowing older prevalent cases to contribute person-time and events from their enrolment date. Pre-specified multivariable analyses will be triggered only once this EPV threshold is confirmed at interim review; otherwise analyses will be limited to descriptive and unadjusted comparisons.

Study timeline and milestones

The registry opened to enrollment on 3 ^rd November 2025 and enrollment is ongoing. Formal analyses have yet to be conducted at the time of publication. Below is outlined the Projected Milestones and Study Timeline.

Projected milestones and Study Timeline are as follows:

Year 1: incident enrollment fully operational and backlog enrollment at least 25% complete;

Year 2: first annual data quality audit completed and first descriptive summary manuscript submitted;

Year 3: first interim analysis of the primary outcome conducted for participants who have reached age 5, if the EPV threshold is met;

Year 4 and 5: comprehensive registry report including all pre-specified multivariable analyses submitted for publication.

Milestone dates will be updated in subsequent protocol versions as the registry matures. As the study is a longitudinal prospective study, intermittent analyses of the data will be conducted throughout the study duration as is outlined in the Projected Milestones above.

Participant recruitment and consent

Written informed consent is obtained from parents or legally authorised representatives at enrollment. Age-appropriate assent is sought from all participants with developing decision-making capacity, consistent with TCPS2 Article 4.2 [23] and MUHC REB guidelines: verbal or picture-assisted assent is used for children aged approximately 7–12 years, and written assent is obtained from adolescents aged 13–17 years, consistent with the age of capacity for research participation under the Québec Civil Code [24]. Assent is documented in REDCap alongside the consent record.

Verbal consent is obtained where written consent is not feasible (e.g., telephone or virtual visits, or when translation is required) and is documented [23], with audio recording where permitted. Participation is entirely voluntary and has no bearing on the care received at Montreal Children's Hospital.

All registry data are de-identified: each participant is assigned a unique registry identifier, and names do not appear on any data collection forms. A master linking log is maintained separately on a password-protected encrypted drive accessible only to the PI. Data are not anonymised in the irreversible sense,  re-linkage to the participant remains possible during active follow-up for clinical and re-contact purposes. Full anonymisation, meaning permanent removal of all re-identification keys, will be applied only upon permanent withdrawal or at study closure.

When a participant turns 18 years of age, active data collection ceases. Data already collected will be retained in de-identified form pursuant to Section 12 of the Act Respecting the Protection of Personal Information in the Private Sector (Québec Law 25, as amended by Bill 64 [25]), which permits continued use of de-identified data for research purposes without renewed consent, subject to ongoing REB oversight (MUHC REB No. 2026–10997).

Loss to Follow-Up

A participant will be considered lost to follow-up if there has been no clinic encounter, completed survey, or successful contact (by telephone, email, or mail) for 18 or more consecutive months in a participant aged under 18 years who has not yet transitioned to adult care or withdrawn from the registry.

When a participant misses a scheduled visit without prior notification, the research assistant will initiate a standardised re-contact sequence: (1) two phone call attempts within 2 weeks; (2) email or written letter within 4 weeks, if contact information is on file; and (3) two further phone call attempts at 3 months. If no successful contact is made after this full sequence, the participant will be flagged as lost to follow-up in REDCap with the date of last confirmed contact recorded. All contact attempts and outcomes will be logged prospectively.

Rates and timing will be reported descriptively at each annual analysis. In survival and longitudinal analyses, the loss will be treated as censoring at the date of last confirmed contact. The potential impact of informative (non-random) censoring will be examined using inverse probability weighting (IPW) as a sensitivity analysis [26,27]. Factors independently associated with the loss will be examined using logistic regression (ever-lost vs retained) as a secondary analysis, with candidate predictors pre-specified to include: diagnosis type, age at enrollment, sex, aganglionosis length or ARM Krickenbeck category, area-level deprivation quintile, and distance from COCOE.

Clinical variables

Data collection is based on the standardised COCOE multidisciplinary clinical assessment form, adapted for input into REDCap [28,29]. This form captures all relevant clinical information encountered during routine care, including: patient demographics; referral details; primary diagnosis; involvement of subspecialists; current medications and allergies; vaccination status; detailed birth and surgical history; prior hospitalisations; relevant investigations; and multidisciplinary assessments. It also documents gastrointestinal, urological, gynaecological, dietary, developmental, social, and family history, as well as physical examination findings, growth parameters, and care plans. For research purposes, only de-identified data elements required to address the study objectives are abstracted, ensuring no additional procedures or burden beyond routine clinical care.

Data abstraction procedures and reliability

To ensure consistency of abstraction from the multidisciplinary assessment form, the registry operates under a written data dictionary and standard operating procedure that specify, for every REDCap field, the permitted values, the source document of record, and the rule to be applied when sources disagree. Operative reports and histopathology reports take precedence over clinic correspondence for surgical and diagnostic variables; the multidisciplinary assessment form takes precedence for functional and psycho-social variables. All abstractors complete a standardised training package and must reach at least 90% field-level agreement with the Study Coordinator across five benchmark charts before abstracting independently. REDCap instruments use closed-response fields, branching logic, and range and cross-field validation rules wherever possible, so that free-text entry is reserved for variables that cannot be pre-coded [28,29].

A random 10% sample of records is independently re-abstracted each quarter by a second trained abstractor blinded to the original entry. Agreement will be quantified using Cohen’s kappa for categorical variables and the intra-class correlation coefficient for continuous variables, against a pre-specified acceptability threshold of κ ≥ 0.80 or ICC ≥ 0.80. Fields falling below this threshold will trigger review of the field definition, retraining of abstractors, and re-abstraction of affected records. Reliability statistics will be reported in the first descriptive registry manuscript and at each subsequent annual audit.

Adherence to the protocol is itself treated as a monitored registry metric. The proportion of eligible patients approached, the proportion consented, the proportion of scheduled assessments completed within their pre-specified age window, and instrument-level completion rates are reported quarterly to the registry Director. Pre-specified action thresholds, for example an assessment-window adherence rate below 70% across two consecutive quarters, trigger review of clinic workflow and, where necessary, a protocol amendment. These adherence metrics will be published alongside outcome data so that readers can judge the completeness and fidelity of the underlying datasets.

Variables collected will include: demographic information (parent and child age, sex); Social Determinants of Health (gender, race/ethnicity, income, education, and first three digits of the postal code-based Material and Social Deprivation Index); health services utilisation (emergency room/outpatient visits and inpatient admissions over the past year); clinical data (diagnosis with severity, surgical procedures with dates, prematurity, comorbidities, and complications); and Patient-Reported Outcome and Experience Measure (PROM/PREM) data.

Social determinants of health

1. (a) Individual-level variables, collected directly from participants and caregivers using Statistics Canada standard wording [30]: sex assigned at birth (male, female, intersex); gender identity (man/boy, woman/girl, non-binary, another gender identity, prefer not to answer); race/ethnicity (self-identified, using Statistics Canada visible minority and Indigenous identity categories); household income before taxes (six categories: < $20,000; $20,000–$39,999; $40,000–$59,999; $60,000–$79,999; $80,000–$99,999; ≥ $100,000, aligned with Low Income Measure thresholds for Québec); highest level of education completed by the primary caregiver (less than high school diploma; high school diploma or equivalent; post-secondary certificate or diploma; bachelor's degree; graduate or professional degree); and employment status of the primary caregiver. For all individual-level SDOH variables, a “prefer not to answer” option will be provided and treated as missing in analyses, with missingness described and addressed using the plan detailed in the Statistical Analysis section.
2. (b) Area-level deprivation, derived from the Pampalon Material and Social Deprivation Index (MSDI), 2021 census-based [31], linked to participant residence via the first three characters of the postal code (forward sortation area, FSA) using the Statistics Canada Postal Code Conversion File. Material deprivation (a composite of employment, income, and education) and social deprivation (a composite of living alone, separated/widowed/divorced status, and lone-parent status) will be treated as separate dimensions and analysed as quintile scores (Q1 = least deprived to Q5 = most deprived). [32] The FSA will not appear in any publication output.

Severity and phenotype

For ARM, anatomical and surgical severity will be classified using the Krickenbeck International Classification [33], comprising the following major clinical groups: perineal fistula; rectourethral fistula (bulbar or prostatic); rectovesical fistula; vestibular fistula; cloaca; no fistula; and rectal atresia/stenosis. For HD, severity will be characterised by length of aganglionosis confirmed by operative report and histopathology, using the following categories: short-segment (rectosigmoid transition zone); long-segment (transition zone beyond the sigmoid colon); total colonic aganglionosis (TCA); and total intestinal aganglionosis (TIA) [34]. Separate structured fields will capture: syndromic status (yes/no; if yes, syndrome name including trisomy 21, VACTERL syndrome association, RET proto-oncogene variants, and other named syndromes); and associated anomalies, documented using a structured checklist covering vertebral, cardiac, tracheo-oesophageal, renal, and limb anomalies (VACTERL) [35], with additional free-text fields for genitourinary, spinal, and other anomalies outside this classification.

Patient-reported outcomes

During clinic visits or between encounters (online or paper), families are asked to complete validated surveys addressing multiple domains of well-being (Table 3). Surveys are available in English and French. Instruments include the Hirschsprung's and Anorectal Malformation Quality of Life (HAQL) questionnaire (disease-specific symptoms and quality of life, with age-specific and stoma-specific versions) [18]; the Baylor Continence Scale (BCS) [16], evaluating the social and functional impact of continence issues; the PedsQL General Well-being Scale [36]; the PedsQL Family Impact Module [36]; the Parent Experience of Child Illness (PECI) [37]; and, for those in bowel management programmes, the Patient-Reported Experience and Outcome Measures in Bowel Management Programmes (PREOM-BMP) [38]. Both French and English validated versions are available and will be used according to family language preference. Proxy-report by the parent or caregiver will be used for children aged 6–11 years; self-report will be used from age 12 onward, consistent with the validated age thresholds for each version.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Summary of patient-reported outcome surveys.

https://doi.org/10.1371/journal.pone.0356813.t003

For the purposes of this registry, a bowel management programme is defined as a structured, clinician-supervised regimen of enemas, laxatives, or dietary modification,  prescribed and monitored by the COCOE team,  for the management of fecal incontinence or constipation that has not responded to first-line measures. Eligibility for the PREOM-BMP is determined at each clinic visit by the responsible clinician and recorded in REDCap. Based on current COCOE caseload, approximately 40–50% of enrolled participants are expected to be in a formal bowel management programme at any given assessment visit.

Survey administration and burden mitigation

The six instruments have a combined estimated completion time of 54–95 minutes depending on age and programme status. To reduce burden, instruments are administered in a staggered fashion across visits: the HAQL, BCS and BFS are prioritised at each visit; the PedsQL modules and PECI are administered at alternating annual visits; and the PREOM-BMP is administered only to eligible families. Where possible, surveys are distributed electronically via REDCap prior to the clinic visit to allow completion at home, reducing in-clinic time. Paper versions are available for families without internet access, and a trained research assistant is available to assist with completion where literacy or language is a barrier. Surveys are available in English and French. In the event of partial completion, available items will be scored according to the instrument's published scoring guidelines; instruments with fewer than the minimum required items completed will be recorded as missing and addressed per the missing data plan in the Statistical Analysis section. Based on experience from comparable paediatric surgical registries, we anticipate an instrument completion rate of approximately 70–80% per visit; completion rates will be reported at each annual analysis.

Missing data

Missing data will be characterised in terms of amount and pattern at each planned analysis. For multivariable analyses, data missing at random (MAR) will be handled using multiple imputation by chained equations (MICE) with 20 imputed datasets pooled using Rubin's rules [42]. Complete-case analyses will serve as the primary approach and MICE results as sensitivity analyses. Where data missing not at random (MNAR) cannot be excluded (particularly for SDOH variables where “prefer not to answer” may be non-random), best-case and worst-case sensitivity analyses will be conducted [42].

Missing patient-reported outcome data and incomplete follow-up are anticipated to be the principal analytic challenges over the observation window to age 18, and are addressed by a pre-specified hierarchy of methods. At the item level, partially completed questionnaires will be prorated according to the developer’s published scoring rule, and scale scores will be set to missing where fewer than the minimum required items are available. At the visit level, longitudinal analyses of PRO scores will use linear mixed-effects models and generalised estimating equations, which retain participants with incomplete visit series and yield valid inference under a missing-at-random assumption [41]. For multivariable models requiring complete covariate vectors, multiple imputation by chained equations will be used, with the imputation model including the analysis model variables, the outcome, and auxiliary variables predictive of non-response (age band, distance from COCOE, deprivation quintile, language of administration, and number of previously missed appointments) [42].

As loss to follow-up and PRO non-response may both be informative, inverse probability of censoring weights estimated from these same predictors will be applied as a sensitivity analysis to the survival and longitudinal models [26,27]. Robustness to departures from the missing-at-random assumption will be assessed using pattern-mixture models in which imputed PRO scores are delta-adjusted by 0.2, 0.5, and 0.8 standard deviations in the unfavourable direction. The amount and pattern of missingness, by instrument, visit, and disease cohort, will be tabulated in every registry publication, and no outcome analysis will be reported without the accompanying completeness figures. A pre-specified threshold of more than 40% missingness for a given instrument at a given assessment point will preclude model-based inference for that instrument at that point, which will instead be reported descriptively.

This protocol paper has additionally been prepared with reference to the SPIRIT 2025 statement (Standard Protocol Items: Recommendations for Interventional Trials) [43], adapted for observational registry protocols. A completed SPIRIT checklist is provided as S1 File.

Ethics approval

This study is conducted in accordance with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS2) [23]. Ethics approval has been granted by the McGill University Health Centre Research Ethics Board (MUHC REB No. 2026–10997), which is responsible for monitoring the study at all participating institutions in the health and social services network in Québec. Authorisation to access patient charts is obtained from the Director of Professional Services (DPS). Access occurs only after consent has been obtained from the patient and family.

Dissemination policy

Findings will be disseminated regardless of direction or magnitude of results.

Confidentiality and data governance

All data collected from a participant are considered personal information and are kept confidential within the limits of the law. Upon enrollment, each participant is assigned a unique registry identifier known only to the principal investigator, co-investigators, and research assistants. Participant names do not appear on any data collection forms. Data within the registry are de-identified. The only protected health information collected includes dates (birth, discharge, death); for publication, dates are converted to relative units (e.g., days from birth), birth year is provided, and no centre identifiers are released.

The data registry is situated within and under the auspices of the MUHC Department of Paediatric Surgery. Electronic data are stored on password-protected computers in administrative offices. Paper records and signed consent forms are stored in a locked filing cabinet in a locked room. The REDCap server is located within the IT department of the Research Institute of the MUHC. Access to the registry is restricted to authorised personnel (the principal investigator (PI), co-principal investigators, and students/trainees granted explicit access). The master linking log is stored on a password-protected encrypted external USB drive kept in a locked cabinet in the principal investigator's office.

Participants may withdraw from the registry at any time by verbal or written request. Upon withdrawal, clinical data are destroyed and personal information deleted from paper and computer records. A confirmation letter is provided. Research results already obtained from analysis of registry data will not be destroyed and may be used for other research projects (subject to ethical approval).

The guardian of the data is the principal investigator, Dr. Hussein Wissanji. If Dr. Wissanji leaves the Montreal Children's Hospital, guardianship is transferred to the new principal investigator. The data will be stored for as long as the registry has scientific interest and the Director can ensure its management.

Potential benefits and harms

There are no direct benefits to patients or caregivers from participating in this registry; however, participants may derive satisfaction from contributing to a national dataset that will improve understanding of colorectal disease management and outcomes. There are no known harms. The potential for privacy breach is minimised by strict confidentiality and de-identification protocols. Participants may have access to their research file to verify and correct their data. Third parties such as insurers or employers may not access registry data unless authorised by a court of law or by the person who contributed the data. Participants will not incur additional costs and will not receive compensation, though a certificate of participation is available upon request.