Overview of approach and methods
The DETECT study will utilise both qualitative and quantitative methods to investigate screening colorectal cancer in individuals with CKD. A diagnostic test accuracy study, a cost-effectiveness/cost-utility analysis and qualitative in-depth interviews will be conducted to examine the efficacy, the efficiency and patient perspectives and preferences of colorectal cancer screening in CKD.
Study design 1: Diagnostic test accuracy study
A diagnostic test accuracy study will be conducted in the CKD population. The design, the conduct and the reporting of this study is in accordance with the STARD initiatives(33).
Patient recruitment and eligibility
Patients ages 35-70, with CKD (stages III-V), CKD-dialysis and CKD-transplant will be recruited from the three main area health services in N.S.W. Although the recommended age for colorectal cancer screening in the general population commences at 50(11), but given the relative increased risk of cancer in the younger transplant population(6) and the greater risk of dying from cardiovascular related causes in the older CKD individuals(34), expansion of the study to include the younger pre-dialysis, dialysis and transplant population is clinically appropriate, and will capture the population that may benefit the most from early cancer detection. Informed consent (verbal and written) will be obtained from all participants.
Exclusion criteria
Patients who have a first-degree relative with colorectal cancer or a personal history of colorectal cancer and inflammatory bowel disease, a recent FOBT test (less than one year), who have had a colonoscopy performed within two years, who are medically unfit for a colonoscopy, who are pregnant and who have active gastrointestinal bleeding will be excluded from the study.
Baseline data collection
The following information will be collected at baseline. They include: age, gender, self-reported race, co-morbidities, medication use, self and any family history of cancer.
Quality of life (QoL) data will also be collected from all enrolled participants at baselines, after the initial IFOBT screen and after the diagnostic colonoscopy at day 14. We will be using two generic QoL assessment tools: the SF-36 and the Euro-Qol (EQ5D) to capture the QoL associated with screening and the diagnostic procedures.
Screening procedure
Eligible participants will be invited to perform the screening tests using the iFOBT kit(35). Two consecutive faecal samples will be required for a single test kit. Test positivity is defined as 100 ng/ml of haemoglobin in either one of the two stool samples
Diagnostic procedures
All participants with positive iFOBT screens will be invited to undergo subsequent diagnostic colonoscopies. If polyps are found, polypectomies will be conducted at the time of the diagnostic procedure. If advanced mucosal neoplasia is found, it will be planned for endoscopic mucosal resection (EMR) at a separate procedure. Advanced neoplasia is defined as adenoma with at least one of the following features: 1 cm or more in size, tubulovillous or villous components, or high-grade dysplasia, or any advanced neoplasm 1 cm or more in diameter. EMR is used increasingly frequently for minimally invasive curative resection of benign and early-stage malignant lesions (T1a) throughout the gastrointestinal tract. EMR has the advantage of managing large, sessile polyps in the outpatient setting, which is potentially cost-saving and may improve clinical outcomes in this high risk cohort(36).
All cancers identified will be staged and the participants will be referred to the colorectal surgical and oncology team, depending upon the stage of initial diagnoses. All cancer diagnoses will be reported to the Central Cancer Registry of New South Wales (located within the Cancer Institute of NSW) for all patients with CKD, and the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) for those on renal replacement therapy (CKD-dialysis and CKD-transplant). The ANZDATA registry is comprehensive database that prospectively collects information on all patients on renal replacement therapy in Australia and New Zealand since 1963. The clinical data includes records of all new cancers except for squamous and basal cell carcinomas. Notification of malignant cancers is a statutory requirement for all health-related institutions in New South Wales. The Central Cancer Registry of New South Wales contains all cancer records and the identifying information for patients diagnosed and treated with cancer within the state of New South Wales since 1972.
Reference standard
Clinical follow-up will be the reference standard for all participants. All participants, with or without screen positive results, will be followed clinically two years after their initial screen. To ensure adequate follow-up and accurate calculation of the screening test performance characteristics of cancer, we will compare our records with that of the Central Cancer Registry (CCR) of NSW through data linkage at 2, 5 and 7 years after the initial screens with the attempt to capture all cancer diagnoses.
Outcomes
The outcomes of the study will include the following:
1. Prevalence of colorectal cancer and advanced colorectal neoplasia in patients with CKD
2. Screen positivity rate: defined as the proportion of participants with positive screens in the total screened study population
3. Test sensitivity: defined as the number of colorectal cancers and/or advanced neoplasms detected through screening divided by the total number of colorectal cancers and/or advanced neoplasms detected through screening and the total number of cancers and/or advanced neoplasms occurring within the delay in a given period (the follow-up time) after a negative screen.
4. Test specificity: defined as the number of participants with no colorectal cancers and/or advanced neoplasms within the follow-up period divided by the number of participants with no colorectal cancers and/or advanced neoplasms after a negative screen and the number of participants without colorectal cancers and/or advanced neoplasms after a positive screen within the follow-up period.
5. Participation rate of screening among individuals with CKD.
6. Potential harms of screening and the diagnostic colonoscopies, such as bleeding, bowel perforation and the inherent risks of peritonitis, particularly among peritoneal dialysis patients.
7. Direct healthcare costs, including individually- collected screening, diagnostic, treatment and overhead costs.
Statistical analyses and sample size calculations
Sensitvity and specificity of iFOBT screening for advanced colorectal neoplasia and cancer will be estimated for (i) CKD (stages 3-5) patients, (ii) dialysis patients, and (iii) transplant patients. For each estimate, the required sample size will be determined by the combined expected prevalence of advanced neoplasia and cancer, the expected sensitivity and specificity, and the required precision of the estimated maximum 90% confidence interval width. For each of the three patient groups, the sensitivity is expected to be 75% and the maximum required 90% confidence interval width is ± 10%. Therefore, 51 cases of advanced neoplasms and cancer will be required. The total sample size and the precision of the estimates of specificity (which is expected to be 90%) for each patient group will be determined by the expected prevalence for that group. In the CKD stages (3-5) group (with a one-year combined estimated prevalence of disease equals to 3.1%), a total of 1637 patients would yield 51 cases and 1586 non-cases. The maximum 90% confidence interval width for specificity is ± 1.3%. Among those on dialysis (with a one-year combined prevalence of disease equals to 3.94%), a total of 1288 patients would yield 51 cases and 1237 non-cases, giving a maximum 90% confidence interval width for specificity of ± 1.4%. Assuming a one-year combined prevalence of disease equals to 4.2% in the transplant population, a total of 1208 patients will again yield a total of 51 cases and 1157 non-cases, giving a maximum 90% confidence interval width for specificity of ± 1.5%.
Across all 3 groups, a total of 4133 participants are required over a 5-year screening period. Assuming a participation rate of 68%, a population size of 6077 CKD patients is required to achieve the target sample size of 4133 for any meaningful analyses.
These sample sizes will provide 80% power to detect a difference of between 3% and 4% in specificity between groups of patients. The small number of expected cases does not allow comparisons in sensitivity to be made between groups.
Study design 2: Cost-effectiveness and cost-utility analyses
Using individually-collected clinical estimates such as the prevalence of disease, the test sensitivity and specificity of iFOBT, the screening participation rate, the probability of cancer and adenoma diagnoses, direct costs estimates and utility weights from the diagnostic test accuracy study and from published clinical estimates of randomised controlled trials of screening, trial and modelled -based cost-effectiveness and cost-utility analyses will be conducted to estimates the efficiency of colorectal cancer screening in CKD.
Probabilistic decision analytical models will be developed to estimate the incremental costs and health benefits of screening compared with no screening in the CKD population. Using time-dependent transition probabilities, the models will provide the analytical framework to simulate the natural history of colorectal neoplasms, the screening and the diagnostic process and the outcomes of colorectal neoplasms in individuals with CKD.
Outcomes
The outcomes of the study will include the following
1. Healthcare costs of screening and no screening
2. Health outcomes (measured in survival and quality adjusted survival) of screening and no screening
3. Incremental cost-effectiveness ratios as cost per life year saved and cost per quality-adjusted life year gained
Study design 3: Qualitative in-depth interviews
This is a qualitative study to investigate the perspectives, experiences, attitudes beliefs and preferences regarding bowel cancer screening in individuals with CKD. Semi-structured face-to-face interviews will be conducted initially with 60 patients (30 participants, 30 non-participants will be sampled from all participating centres). Participants will be purposively selected from the participants and the non-participants of the diagnostic test study (Study 1) to ensure a range of age, ethnicity, co-morbidities and a balance of gender. Data collection will cease when theoretical saturation is reached in the concurrent analysis, which is when little or no new concepts emerge in subsequent interviews. The participants will be asked for their perspectives on: a) knowledge about bowel cancer risk and screening, b) reasons for participating/not participating in the diagnostic screening trial, c) experiences of participating in the screening diagnostic study, d) perceived benefits and harms in participating in screening. All interviews will be audio recorded and transcribed verbatim. The computer software 'HyperRESEARCH 3.0' will be used to assist with storage, coding and searching of qualitative data. Coding and analysis will accord to thematic analysis. The analysis will be performed for all participants collectively; then a sub-analysis will be conducted for screening participants and non-participants to identify and compare differences between both groups.
Outcomes
The outcomes of the study will include the following:
1. Identification of barriers and facilitators to participation in bowel cancer screening
2. Understanding of patients decision-making underpinning participation in cancer screening
3. Information to inform policy- and decision-makers about a patient focussed screening program that takes into consideration their needs, priorities and preferences.
Ethical considerations
The DETECT study protocol has been approved by the Sydney West Area Health Service Ethics Review Committee (HREC10/WMEAD/13 and SSA/10/WMEAD/54). The University of Sydney, Human Research Ethics committee has also been notified of the approval. The screening test itself is minimally invasive and will not impose any significant harms to the patients. The diagnostic colonoscopies will be performed by experienced gastroenterologists from all participating centres, to prevent any significant complications such as catastrophic bleeding and bowel perforations. The investigator team will monitor all potential electrolytes derrangement resulting from the bowel preparation by collecting and monitoring pre and post-operative serum biochemistry. Information sheets, clearly and succinctly outlining the screening procedure and the associated harms will be given to all eligible patients to read. Patients are in no way obliged to participate and if they do, they can withdraw at any time from the study. They will be informed that their decision to withdraw will not result in any consequences relating to their care provided. Summary of the monthly progress report, outlining any potential complications/harms will be given to all participants and their treating physicians.