Improving Decisions and Outcomes in Parastomal Hernia Surgery

IDOPS is a prospective observational cohort and clinical registry designed to establish a longitudinal evidence base for patient selection and treatment of parastomal hernia. The registry captures the clinical pathway from initial specialist assessment through treatment decisions, reassessments, elective surgery where applicable, and emergency presentations. Treatment is determined through routine clinical care.

Registry structure and procedures

Data are collected in a dedicated REDCap database. The initial assessment is recorded once, while repeatable instruments capture subsequent treatment decisions, outpatient reassessments, operations, postoperative complications, acute admissions, and acute interventions. This structure allows multiple events to be recorded for each participant while preserving their chronological relationship.

Treatment decisions include referral for multidisciplinary hernia discussion and the reasons for non-operative management. These distinguish between an unfavourable balance of symptom burden and surgical risk, lack of suitability for surgery, and a patient's decision to decline an offered operation. Changes in management are recorded prospectively.

Patient-reported data are collected during outpatient assessments and through automated questionnaire invitations delivered via secure digital mail. Annual invitations for patients who have not undergone surgery are scheduled relative to enrolment. Postoperative invitations are scheduled relative to the first operation recorded in the registry. Invitation logic incorporates questionnaire completion status and whether follow-up remains active.

Data dictionary and validation

The REDCap data dictionary defines variable names, field labels, response options, coding, branching logic, required fields, and applicable validation rules. Calculated fields are used for derived variables, including body mass index and questionnaire scores. Hernia characteristics are recorded using the European Hernia Society classification, with primary and recurrent hernias distinguished.

Data-entry controls include predefined response categories, date validation, required fields, and numerical range checks where appropriate. For example, height and weight fields have predefined permitted ranges. Conditional fields are displayed according to preceding responses to support consistent recording of treatment decisions and clinical events.

Data quality assurance and source verification

The study team will review registry data for completeness, duplicate records, implausible values, inconsistent responses, and discrepancies in the chronological sequence of clinical events. Identified discrepancies will be checked against the relevant source and corrected where possible.

Medical records, imaging reports, and operative reports will be used to verify clinical data when entries are incomplete, inconsistent, or uncertain. Patient-reported questionnaire responses will be retained as reported by the participant. Corrections to clinical data and changes to the database structure will be documented using REDCap's logging functionality.

Registry operations and change management

The study team will maintain documented procedures for recruitment and consent recording, data entry, follow-up administration, discrepancy resolution, data extraction, analysis, and reporting. Postoperative complications and acute events will be captured from routine clinical documentation. Database changes will be reviewed for their effects on existing records, branching logic, calculated fields, and automated invitations before implementation. Analyses will use dated data extracts and documented variable definitions.

Sample size considerations

The anticipated cohort size is approximately 250 participants recruited over five years, based on expected clinical volume. The study is exploratory and is not powered to test a single treatment effect. The number of participants with available follow-up and the number of observed events will determine the precision of estimates and the complexity of adjusted analyses. Development of prediction models will depend on whether sufficient data and events are available; subsequent validation would require an appropriate validation dataset.

Missing data

Missing data will be assessed by variable and follow-up time point. Where possible, missing clinical information will be retrieved from source records. Unavailable information will remain missing rather than being interpreted as the absence of an event. Unresolved inconsistent or implausible values will be treated as missing for the relevant analysis.

Questionnaire response rates will be reported at each scheduled assessment. Scoring will follow the applicable questionnaire instructions. Analyses will distinguish non-response from assessments that are no longer applicable because of death, withdrawal, or stoma reversal. The choice between complete-case analysis and methods such as multiple imputation will depend on the extent and likely mechanism of missingness. Material assumptions will be examined through sensitivity analyses where feasible.

Statistical analysis

Analyses will describe patient characteristics, treatment pathways, operative practice, and longitudinal clinical and patient-reported data. Estimates will be presented with measures of uncertainty. Time-to-event analyses will use a clearly defined time origin appropriate to the question, such as enrolment or surgery. Death will be considered as a competing event where relevant.

Changes in patient-reported scores will be analysed using methods that account for repeated measurements within participants. Multivariable analyses will explore associations between patient characteristics, treatment decisions, operative techniques, and subsequent outcomes, with model complexity limited by the available sample and event counts.

Patients may transition from non-operative management to surgery during follow-up. Analyses involving treatment status will account for the timing of this transition where relevant. Comparisons between management strategies or operative techniques will be interpreted in light of confounding by indication and will not, by themselves, establish causal treatment effects.