2.4. Statistical analysis

Baseline characteristics were presented using descriptive statistics. Baseline patient, tumour and treatment characteristics between respondents and non‐respondents were analysed. Continuous variables are depicted as means and standard deviations, and categorical variables as frequencies and percentages. Differences in characteristics and utilisation of care between short‐term (≤2y), mid‐term (3‐4y) and long‐term (≥5y) survivors were examined using chi‐square (categorical), t test (nominal) or ANOVA tests (continuous). Above follow‐up categories are commonly used categorisations in cancer research. Univariable logistic regression was used to identify possible associative factors. Hereafter, a multivariable logistic regression model was formed using the identified factors to analyse the association between healthcare use and overuse and age, sex, marital status, BMI, educational level, SES, comorbidity, stoma and (neo)adjuvant therapies (e.g. radiation and chemotherapy). Also, the relation between patient‐reported outcomes (anxiety, depression, fatigue) and healthcare utilisation was analysed using multivariable logistic regression analysis. Analyses were performed using Stata software (Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC). Two‐sided analysis with p < 0.05 considered significant was used. We adhered to the STROBE checklist for observational cohort studies (Elm et al., ²⁰⁰⁷).