This trial is a multi-centre randomized controlled study where patients will either be randomized to follow-up at the hospital (control group) or by their GP (intervention group) (Figure 1).
Patients are eligible if they undergo surgery for colon cancer with histological grade Dukes's Stage A, B or C. Patients receiving postsurgical adjuvant chemotherapy (Dukes C) are also eligible. We define a colon cancer as a tumour located 15 cm above the anal verge by coloscopy or a tumour at or above the level of the sacral promontory as seen at the time of surgery. Annually, approximately 200 patients with colon cancer undergo surgery in the catchment area of the North Norwegian Hospital Trust (a population of 470,000).
There are three hospital trusts and one university hospital (UNN HF) (in total 11 hospital units) located within the North-Norwegian Health Region. Three hospitals (located in the cities of Harstad, Mo, and Bodø) have an annual volume of more than 20 colon cancer procedures; the University Hospital operates on about 110 patients with suspected CRC cancer annually and approximately 60 of them have a colon cancer disclosed. The University Hospital and these three local hospitals consented to participate. All patients who undergo surgery for colon cancer are followed up at the surgical outpatient clinic at each hospital.
In North Norway, 550 GPs refer patients with suspected colon cancer to the local hospitals. They are situated in all communities in North Norway. All GPs were informed by mail about the purpose of the study and invited to participate. None refused to participate.
Patients are randomized to follow-up either by their GP (intervention) or at the surgical outpatient clinic. All patients will receive their first postoperative check-up at the hospital which performed the surgery (baseline, 1 month). A clinical examination will then be performed and information about the histology and the results of the surgery will be provided. If the patient gives informed consent, randomization is performed. Patients randomized to GP follow-up are referred to their GP. This referral contains information about the surgery and any complications, Dukes's staging, guidelines for follow-up and behavioural strategy in the case of a SCE.
Guidelines for follow-up
NGICG revised its guidelines in 2007  (Table 1), and the GPs were provided with a guideline form based upon this recommendation. Most relapses of colon cancer are detected within the first three years of follow-up. On the basis of this knowledge, the regular check-ups are performed at three-month intervals for the first two years and then every six months.
All patients with elevated CEA prior to surgery are requested to undergo this test at every postoperative clinical examination. Chest x-ray and ultrasound are performed on a regular basis. Colonoscopy is performed twice during the follow-up period. The follow-up guideline is similar in both arms (Table 1).
When recurrence is suspected
All GPs are given written instructions on what to do if recurrence is suspected. A set of serious clinical events (SCEs) is defined as: Colonoscopy verifies recurrence of CRC. Elevated levels of CEA are shown by repeated measurements. Blood in stool is detected by the Hemofec test. Unexplained abdominal pain. Unexplained weight loss of 5 kg during the last three months. Cancer-suspect lesions detected by rectal examination. Lymphandenopathy. Metastasis-suspect lesions shown by chest x-ray or ultrasound of liver. If an SCE occurs, the patient must be referred promptly to hospital.
Information to the patient
All patients included in the study are given a folder explaining the purpose of the study. Follow-up guidelines are provided with detailed information concerning the clinical examinations, CEA measurements, x-ray/ultrasound procedures and colonoscopy.
The EQ-5D and EORTC-QLQ C30 questionnaires for the two-year follow-up period (at 1, 3, 6, 9, 12, 15, 18, 21, 24 month follow-up) are given to the patient at baseline examination.
Economic evaluation questionnaire: A questionnaire has been developed to enable calculation of costs related to patient examinations. This questionnaire is incorporated in the QOL questionnaire, and patients are requested to fill out the questionnaire at 1, 3 6, 9, 12, 15, 18, 21 and 24 months postoperatively. The cost elements will include patient- and family-related costs due to outpatient visits, GP visits, laboratory tests, radiographs/ultrasound, examinations due to suspected relapse, treatment of relapse, travelling, production losses, co-payments and other patient expenses.
Patients in the control group consist of patients randomized to regular follow-up at the hospital's surgical outpatient clinic. All these patients are followed up according to the NGICG guidelines. This follow-up is performed by consultants or internship doctors in digestive surgery.
We aim to compare quality of life and costs of follow-up among patients followed up by their local GP or at the surgical outpatient clinic. We hypothesize that patients followed up by their GP will experience similar or higher scores on quality of life measures and lower costs during follow-up.
We aim to test whether the incidence of SCEs will be similar for patients followed up by their GP and patients in the control group. We hypothesize that patients followed up by their GP will not have an increase in time to detection or in the frequency of SCEs.
Primary outcome (measured at 1, 3, 6,9,12,15,18,21 and 24 months)
Quality of life
EQ-5D is a standardized generic instrument for use as a measure of health outcome. Applicable to a wide range of health conditions and treatments, it provides a simple descriptive profile and a single index value for health status. EQ-5D measures five dimensions of health-related quality of life (HRQOL): mobility, self-care, usual activities, pain/discomfort and anxiety/depression. EQ Visual Analogue Scale: The EQ VAS records the respondent's self-rated health status on a vertically graduated (0–100) visual analogue scale . Quality of life assessment based upon the European Organization for Research and Treatment of Cancer Quality of Life questionnaire (EORTC QLQ C-30); EORTC QLQ C-30 incorporates nine multi-item scales: five functional scales (physical, role, cognitive, emotional and social); three symptom scales (fatigue, pain, nausea/vomiting); and a global health status/QOL scale. Six single-item scales are also included (dyspnoea, insomnia, appetite loss, constipation, diarrhoea and financial difficulties) .
A cost-effectiveness analysis will be performed alongside the randomized controlled trial. A Markov model will be employed. If the trial reveals no difference in health outcomes or quality of life of the patients, a cost-minimization analysis will be carried out. If we reveal a difference in quality of life between the two groups, we will implement a cost-utility analysis. The economic evaluation will have a health care and societal perspective. Resources used in both treatment arms will be registered prospectively based on reports by the patients and on the hospital patient records. Costs related to consultations at the primary and secondary care level, both at baseline and up to 24 months after randomization, including direct medical costs, indirect costs (production losses), travel costs and patient-/family-related costs will be registered. Quality of life measurements will be collected both at baseline and at several time points during the intervention. A 3% discount rate will be used to discount future costs and benefits. The cost elements will include costs related to outpatient visits, GP visits, laboratory tests, radiographs/ultrasound, examinations due to suspected relapse, treatment of relapse, travelling/transportation, production losses, co-payments and other patient/family expenses. The net costs and outcomes of the two treatment options will be compared and presented as incremental cost-effectiveness ratios.
Secondary outcome measures
Serious clinical events (SCEs)
Time to detection of recurrences and serious clinical events (SCEs) defined as: Colonoscopy-verified recurrence of cancer disease. Increase in CEA measurements shown by repeated measurements. Blood in stool detected by the Hemofec test. Unexplained abdominal pain. Unexplained weight loss of 5 kg during the last three months. Cancer-suspect lesions detected in rectal examination. Lymphandenopathy. Metastatic suspect lesions shown by chest x-ray or ultrasound of liver.
Sample size calculations are according to Campbell's guidelines 
Alpha and beta were set at 0.05 and 0.2 respectively and tests were two-sided. We hypothesize that patients followed up by their GP will have a minor to moderate increase in the Global QoL scores (EORTOC-QLQ C30). This moderate effect size equals an increase of 10 units on the Global QOL . Patients with colon cancer (local disease) show a standard deviation of approximately 20 (global QOL) . Allowing for a dropout rate of 20%, to detect this difference with confidence we required a total sample size of 170 patients. Sample size calculations based upon EQ-5D showed similar results .
Randomization will be done at the 1-month check-up (baseline) after the patients have given informed consent. The randomization service is web-based and managed by the Norwegian University of Science and Technology. Patients are stratified according to the Dukes's staging and whether they have a stoma. Patients allocated to general practice follow-up may be referred back to surgical clinics at any point in the study; similarly, patients in the surgeon-led follow-up group are free to consult their GP at any time during the study.
Data are being collected for patients in the intervention and the control group in identical ways. QoL and "cost questionnaires" are sent every three months (after each follow-up appointment) by the patients to the trial centre up to 24 months postoperatively. These questionnaires are optically readable, data being consecutively collected in a trial database.
SCEs are registered via hospital chart review at 24 months postoperatively. This follow-up period will allow identification of most SCEs, since tumour recurrence typically occurs within two years of surgery .
The study has been approved by the Norwegian Data Inspectorate. All data will be handled with strict confidentiality, and study reports or presentations will maintain the anonymity of patients, surgeons, GPs and hospitals. Data collection will be complete by the end of 2010.
We will use the intention-to-treat principle when analysing data. Furthermore, we will employ Fayers guidelines to handle missing data . Treatment arms will be compared with respect to potential covariates using continuous and categorical univariable analyses. These will include variables relating to patients (age, sex, comorbidities, cost), treatment (surgical resection, surgical complications, adjuvant therapy), tumour (Dukes's Stage), hospital surgical outpatient clinic (cost, patient satisfaction, QoL), GP practice (cost, patient satisfaction, QoL). The cost elements will include costs related to outpatient visits, GP visits, laboratory tests, radiographs/ultrasound, and examinations due to suspected relapse, treatment of relapse, travelling/transportation, production losses, co-payments and other patient expenses. The net costs and outcomes of the two treatment options will be compared and presented as cost-effectiveness ratios. If the trial reveals no difference in health outcomes or quality of life for the patients, a cost-minimization analysis will be carried out. If we find a difference in quality of life or overall survival rates between the two groups, we will use a cost-utility approach. The economic evaluation will have a societal and health care perspective.
The method of analysis, including adjusting for covariates will comply with the CONSORT statement . The results will be expressed as odds ratios for binary outcomes, hazard ratios for time- to-event outcomes or mean differences for continuous outcomes with corresponding standard errors, 95% confidence intervals, and associated p-values. P-values will be reported to three decimal places with p-values less than 0.001 reported as p < 0.001. For all tests we will use alpha = 0.05 level of significance.