Kennedy 2003.

Methods	Randomised, cluster, multisite (19 sites with randomisation by treatment centre) trial, 12 months
Participants	UK; 19 hospital outpatient clinics (700 participants); ulcerative colitis or Crohn's disease; 269 men, 366 women, mean age 46.34 years CONT, 44.37 years INT
Interventions	INT (n = 297): participants could telephone hospital if any problems/as needed. Participants given initial consultation and a detailed guidebook about their condition and what symptoms to be aware of and prepared a written self‐management plan. Safety net – none. CONT (n = 403): consultant‐led appointments as usual (schedule NR), availability of emergency appointments unclear.
Outcomes	Disease‐related quality of life: 12 months (IBDQ) Quality of life: 12 months (EQ‐5D, HADS, SF‐36, PEI) Satisfaction: 12 months (bespoke questionnaire) Interim events: 12 months (medical records and patient diaries) Symptoms: 12 months (patient diaries) Economic costs: 12 months (patient diaries and medical records, EQ‐5D)
Notes	Economic evaluation – health service resource use and cost effectiveness using EQ‐5D. Funded by NHS Research and Development Health Technology Assessment Programme. COI NR. Contacted authors for more information, further data provided for IBDQ and depression.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The 19 remaining hospital sites were randomly ordered from 1 to 19 (by a statistician not associated with the study using a prepared sequence of random numbers), and these numbers were randomly drawn out of a hat allocating each hospital either to the CONT group (10 sites) or INT group (9 sites).
Allocation concealment (selection bias)	Low risk	The 19 remaining hospital sites were randomly ordered from 1 to 19 (by a statistician not associated with the study using a prepared sequence of random numbers), and these numbers were randomly drawn out of a hat allocating each hospital either to the CONT group (10 sites) or INT group (9 sites).
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unclear if assessors blinded and outcomes mainly self‐reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for loss to follow‐up were listed and loss was similar between groups. ITT analysis conducted.
Selective reporting (reporting bias)	Low risk	All outcomes detailed in the methods section were reported.
Other bias	Low risk	No obvious signs of other bias.