full list of authors and their declarations of interests...2019/10/01 · cancer incidence). when...

1

Web appendix 2:

Full list of authors and their declarations of interests

Chair:

Gordon Guyatt, MD, MSc (general internist, distinguished professor)

Department of Health Research Methods, Evidence, and Impact, McMaster University,

Hamilton, Canada

Methods co-chair:

Lise M. Helsingen, MD, PhD candidate

Clinical Effectiveness Research Group, Department of Transplantation Medicine, Oslo

University Hospital, Oslo, Norway

Clinical Effectiveness Research Group, Institute of Health and Society, University of Oslo,

Oslo, Norway

Frontier Science Foundation, Boston, Massachusetts, USA

Clinical experts:

Michael Bretthauer, MD, PhD (gastroenterologist, professor)




Oslo, Norway


Joseph C. Anderson, MD, PhD (gastroenterologist)

Veterans Affairs Medical Center, White River Junction, Vermont, USA

The Geisel School of Medicine at Dartmouth, Hanover, New Hampshire, USA

University of Connecticut Health Center, Farmington, USA

Reto Auer, MD, MAS (general practitioner)

Institute of Primary Health Care, University of Bern, Bern, Switzerland

Center for Primary Care and Public Health, University of Lausanne, Lausanne, Switzerland

Anja Fog Heen, MD, PhD candidate (general internist, methodologist)

Department of Medicine, Innlandet Hospital Trust-division, Gjøvik, Norway

Silje Bjerkelund Murphy. MSc, RN (registered nurse)

Diakonhjemmet Hospital, Oslo, Norway

Juliet Usher-Smith, MB BChir, PhD (general practitioner, researcher)

The Primary Care Unit, Department of Public Health and Primary Care, University of

Cambridge, Cambridge, UK

2

Majid Abdulrahman Almadi, MD, MSc (gastroenterologist, associate professor)

Division of Gastroenterology, Department of Medicine, King Khalid University Hospital,

King Saud University, Riyadh, Saudi Arabia.

Division of Gastroenterology, The McGill University Health Center, Montreal General

Hospital, McGill University, Montreal, Canada

Douglas A. Corley, MD, PhD (gastroenterologist)

Division of Research, Kaiser Permanente, Oakland, California, USA

Department of Gastroenterology, San Francisco Medical Center, California, USA

Patient partners:

Casey Quinlan

Society for Participatory Medicine, Boston, Massachusetts, USA

Mighty Casey Media, LLC, Richmond, Virginia, USA

Jonathan M. Fuchs (Fellow American College of Healthcare Executives)

Population Health and Health Policy Consultant, California, USA

Annette McKinnon

Patient Advisors Network, Founding Member, Canada

Methodology and research content experts:

Iris Lansdorp-Vogelaar, PhD (health economist, modeller, assistant professor)

Department of Public Health, Erasmus MC University Medical Center Rotterdam, Rotterdam,

the Netherlands

Henriette C. Jodal, MD, PhD candidate




Oslo, Norway


Mette Kalager, MD, PhD (surgeon, researcher, assistant professor)




Oslo, Norway


Amir Qaseem, MD (internist, methodologist)

American College of Physicians, Philadelphia, USA

Thomas Agoritsas, MD, PhD (general internist, methodologist, assistant professor)

3

Division General Internal Medicine & Division of Clinical Epidemiology, University

Hospitals of Geneva, Geneva, Switzerland


Hamilton, Canada

Lyubov Lytvyn, MSc, PhD candidate (patient partnership liaison)


Hamilton, Canada

Reed A.C. Siemieniuk, MD, PhD candidate (general internist, methodologist)


Hamilton, Canada

Per Olav Vandvik, MD, PhD (general internist, professor)

Institute of Health and Society, Faculty of Medicine, University of Oslo, Oslo, Norway

Department of Medicine, Innlandet Hospital Trust-division, Gjøvik, Norway

Disclosures

All panel members were pre-screened for conflicts of interest prior to the guideline process

that resulted in the BMJ Rapid Recommendations. The pre-screening was performed by the

RapidRecs Executive team from the non-profit organization MAGIC (www.magicproject.org)

with support and approval from at least two unconflicted BMJ editors. No financial conflicts

of interest were allowed (specifically, no financial ties to medical device companies with any

stake in colorectal cancer screening, or to public or private healthcare funders) and intellectual

and professional conflicts of interest were managed appropriately (see web appendix 7:

Methods for BMJ Rapid Recommendations). Panel members could not have a financial

conflict for the past three years and do not anticipate a conflict arising in the foreseeable

future, which we defined as at least one year.

Financial disclosures:

No panel members had any financial conflicts of interest to disclose related to this clinical

question.

Professional disclosures:

The majority of the panel members routinely see patients who are eligible for colorectal

cancer screening. Joseph C. Anderson, Majid Almadi and Douglas A. Corley regularly

performs screening colonoscopies. No other professional conflicts of interest to disclose.

Intellectual disclosures:

Henriette C. Jodal and Lise M. Helsingen were lead authors of the systematic review of

screening trials that formed part of the evidence base for this guideline, and Per Olav Vandvik

and Joseph C. Anderson were co-authors of the review. Iris Lansdorp-Vogelaar, Lise M.

Helsingen, Gordon Guyatt, Per Olav Vandvik and Michael Bretthauer were co-authors on the

paper describing the microsimulation modelling that formed the evidence base for this

guideline.

Published opinions: Michael Bretthauer and Mette Kalager have written opinion pieces being

critical of modelling approaches to guide recommendations (i.e. Annals of Internal Medicine

4

DOI: 10.7326/M16-1805). Reto Auer has published several papers promoting shared

decision-making in colorectal cancer screening decisions in primary care.

Reto Auer is the principal investigator of a trial funded by the Swiss National Science

Foundation promoting shared decision-making in colorectal cancer screening. He is the

principal investigator of analyses of claims data to study variation in screening rates in

Switzerland and is an unpaid member of the Swiss expert panel on colorectal screening,

coordinated by the Swiss Cancer League, where he is involved in making recommendations

for communication material to the public and healthcare professionals in Switzerland.

Iris Lansdorp-Vogelaar has received funding for monitoring, evaluation and cost-

effectiveness analysis of different colorectal cancer screening strategies, from the US National

Cancer Institute, European Commission, ZonMw, The Dutch Cancer Society, Netherlands

Institute of Public Health and NHMRC Australia. Her institution has received payments for

her lectures and educational events on screening for Erasmus University and Utrecht Medical

Center. She has contributed in writing a guidebook on colorectal cancer screening for the

International Agency for Research in Cancer (IARC) and has received travel expenses and per

diem from IARC. She has not directly contributed to other guidelines, but her work has served

as input to the deliberations of other guideline-issuing committees such as USPSTF and ACS.

Amir Qaseem is a board member of Guidelines International Network, Physician Consortium

for Performance Improvement, Medbiquitous and Dynamed. He is Vice President and leads

the clinical policy work at American College of Physicians including previously published the

American College of Physician's clinical guidance on colorectal cancer screening and is

working on a current update of that guidance

Majid Almadi is President of the Saudi Gastroenterology Association. He has participated in

the National Saudi Guidelines for colorectal cancer screening published in 2015. He has

received a fund from the Saudi Health Council for a stool based colorectal cancer screening

pilot study in addition to a cost-effective analysis on the initiation of colon cancer screening in

Saudi Arabia.

Douglas A. Corley is performing contractual research with Medial Research/EarlySign

validating a machine learning algorithm tool. The company is doing early research and has no

licensed products related to colorectal cancer screening. He is a member of the American

Gastroenterological Association, the American Society of Gastrointestinal Endoscopy and the

American College of Gastroenterology. These organizations have guidelines on colorectal

cancer, but he is not involved with the creation of these guidelines. He contributed to a

guidebook on colorectal cancer screening for the International Agency for Research in Cancer

(IARC) and received partial coverage of travel expenses for this project. He has received

grants from the United States National Cancer Institute for evaluating colorectal cancer

screening use and effectiveness.

Juliet Usher-Smith has received grants for studies from Public Health England, The School

for Primary Care Research and Cancer Research UK. She has published a systematic review

of risk prediction models for colorectal cancer, and a paper validating a number of those

models in a UK population. The QCancer® risk calculator which is being suggested as part of

this Rapid Recommendation was included in her validation work. She has not developed any

of the risk models.

5

Mette Kalager and Michael Bretthauer are authors on the NORCCAP trial, one of the

trigger trials of this guideline with long-term follow-up of sigmoidoscopy screening. They are

investigators on the ongoing NordICC trial comparing no screening to colonoscopy screening.

No results have been published from this study yet.

Michael Bretthauer has taken part in writing a guidebook on colorectal cancer screening for

the International Agency for Research in Cancer (IARC) 2017/2018.

Joseph C. Anderson is a co-writer on the American College of Gastroenterologists 2008

colorectal cancer screening guidelines. He is a member of the American Gastroenterological

Association, American College of Gastroenterology, American Society of Gastrointestinal

Endoscopy and the US Multi Society Task Force for Colorectal Cancer Screening.

Thomas Agoritsas, Gordon Guyatt, Lyubov Lytvyn, Reed Siemieniuk, Amir Qaseem and Per

Olav Vandvik, are members of the GRADE Working Group – BMJ Rapid Recommendations

adheres to GRADE methods. No panel member had any other relevant intellectual conflict to

disclose.

6

Web appendix 3:

Values and preferences

Values and preferences literature

We searched for studies that explored individuals’ values and preferences regarding colorectal

cancer screening. The research question was: “How large a reduction in colorectal cancer

mortality or colorectal cancer incidence would individuals require to undergo screening?

We searched MEDLINE, EMBASE, and PsychINFO using a combination of medical subject

headings and key words related to colorectal cancer, combined with a validated search filter

for values and preferences studies up to May 1, 2018.1 We included studies with general

patient populations that addressed all of the four options considered in our guideline:

colonoscopy, sigmoidoscopy or faecal testing every or every two years (FIT or gFOBT), and

no screening. We included studies that provided quantitative estimates of the thresholds of

benefits and harms required for individuals to choose to undergo screening, including studies

on decision aids and discrete choice experiments. Only two studies addressed the population,

intervention, and outcomes of interest to inform our guideline.

One study from the Netherlands reported on preferences of 400 adults (mean age 60.7, SD

6.6; 52% male), of whom 23% had previously undergone screening.2 Investigators conducted

a discrete choice experiment design that looked at varying thresholds of mortality reduction,

assuming a baseline life-time colorectal cancer mortality risk of 3%. The study did not

explore potential harms associated with each screening test. When considering no screening

compared to faecal testing, sigmoidoscopy and colonoscopy, participants preferred no

screening to faecal testing, and either endoscopy option to no screening. Participants had

similar attitudes toward sigmoidoscopy and colonoscopy unless colonoscopy had a very high

relative risk reduction (translating to an absolute reduction from 3% to 0.5% depending on

cancer incidence). When considering annual faecal testing to five-yearly sigmoidoscopy, if

sigmoidoscopy had a greater relative risk reduction (absolute reduction from 3% to 1.8%)

than faecal testing (absolute reduction from 3% to 2.4%), then participants preferred

sigmoidoscopy. If both had the same relative risk reduction (absolute reduction from 3% to

1.8%), then faecal testing was preferred to sigmoidoscopy.

A study from the US reported on the preferences of 116 older adults (mean age 74, range 70

to 90; 46% male), almost all of whom have been previously screened (92%).3 All participants

were given the option to choose between four unlabeled tests, presenting benefits, burdens

and harms associated with colonoscopy, sigmoidoscopy, faecal testing every year and no

screening. None of the participants chose no screening. Faecal testing was most preferred

(40%), followed by colonoscopy (34%) and sigmoidoscopy (25%).

The limited data available clearly demonstrated substantial variability in values and

preferences across settings and individuals. In the Dutch study2, either endoscopy option was

preferred to faecal testing every year, and to no screening. In the American study3, faecal

testing was the most preferred screening option, and any screening option was preferred to no

screening. Both mean age and percentage of participants previously screened differed in the

two studies, which may explain some of the within-study variability.

7

What is the magnitude of benefit needed for most people to undergo screening?

The review of relevant literature for colorectal cancer screening values and preferences could

not answer this question, and we therefore had to rely on the panel’s experience: Panel

members chose thresholds of benefit above which they believed the majority of well-informed

individuals would choose screening and below which the majority would not.

To set this threshold, panelists completed surveys regarding their beliefs about choices

patients were likely to make given the magnitude of benefit from screening. GG and LMH

designed the surveys and LL conducted pre-testing with the patient partners before sending

the surveys to the entire panel. Neither the panelists, the chair, nor the methods co-chair had,

when designing and completing the surveys, reviewed the results of the microsimulation

modelling studies’ estimates of screening benefit.

The surveys addressed decisions on screening versus no screening, and decisions regarding

choosing one screening test over others. For each survey, panel members reviewed summaries

of the harms and burdens associated with screening, and provided their opinion among

different response alternatives. The panel members were instructed not to give their personal

opinion, but choose the response alternative they thought would be applicable to the majority

of well-informed people.

Below we summarize how these surveys were performed and what information was available

for the panelists at the time.

First survey

The practical issues associated with the screening options as well as preliminary data

regarding burdens and harms from screening were presented to the panel in a teleconference

and made available to the panelists by e-mail before the first survey

Preliminary data on screening harms and burden

Procedure-related mortality

• Colonoscopy: Fewer than 1 per 1000 procedures

• Sigmoidoscopy: Fewer than 1 per 1000 procedures

Gastrointestinal perforations

• Colonoscopy: Approximately 1 per 1000 procedures


Major gastrointestinal bleedings

• Colonoscopy: Approximately 3 per 1000 procedures

• With polypectomy:10 per 1000

• Without polypectomy: 1 per 1000


Number of screening or work-up colonoscopies in 15 years

• FIT every year: 300 per 1000 screened

• FIT every two years: 200 per 1000 screened

• Sigmoidoscopy once: 200 per 1000 screened

• Colonoscopy once: 1000 per 1000 screened

8

Number of surveillance colonoscopies in 15 years (NB, some people will have several and

some will have none)

• FIT every year: 200 per 1000 screened

• FIT every two years: 200 per 1000 screened

• Sigmoidoscopy once: 200 per 1000 screened

• Colonoscopy once: 300 per 1000 screened

First survey: Values and preferences – judgment exercises of hypothetical scenarios Replies: 21. Number of votes given after each response alternative

Introduction to the survey:

The purpose of the following exercises is to make you think through the questions and provide a common understanding of the magnitude

of effects. This survey will give us an idea of where people stand that would help us with the upcoming discussions. We will present your

answers before we start the discussion in our next panel meeting. We will only present votes, and not who voted for what. Please do not consider your answers here as final. You should all be open to change your minds as the discussion progresses.

We may make different recommendations for colorectal cancer screening for people with different risks of colorectal cancer and

colorectal cancer death. We might recommend against screening for people with a low absolute risk of being diagnosed with or dying

from colorectal cancer, and recommendation for screening in people at higher absolute risk. Therefore, we need to establish the threshold

where our recommendations shift.

Please do the following exercises with focus on the population – what would most people do if they were fully informed of the potential

burden related to screening.

It seems clear that values and preferences regarding screening for colorectal cancer differ markedly. Some people would choose to undergo screening even with a very small absolute benefit, and some will be reluctant to screen even when there is a large benefit. Do

you agree or disagree with this statement?

• Response: 1 disagree, 20 agree

Colorectal cancer mortality

1. A patient who is screened with colonoscopy, has a 1 in 1000 (0.1%) lower risk of dying from colorectal cancer at 15 years.

How would patients view such benefits?

• Everyone would choose screening: 0

• Most would choose screening: 1

• A majority would choose screening: 3

• A majority would decline screening: 10

• Most would decline screening: 5

• All would decline screening: 2

Given this scenario, are you in favor or against recommending colonoscopy screening?

• Response: 16 against, 5 in favor

2. A patient who is screened with colonoscopy, has a 10 in 1000 (1%) lower risk of dying from colorectal cancer at 15 years.










3. A patient who is screened with colonoscopy, has a 20 in 1000 (2%) lower risk of dying from colorectal cancer at 15 years. How would patients view such benefits?









9

4. A patient who is screened with colonoscopy, has a 30 in 1000 (3%) lower risk of dying from colorectal cancer at 15 years.










Thanks for providing your answers when the screening mode was colonoscopy. Would anything change if the screening mode were

different? Would patients demand a larger or smaller benefit from screening?

5. Would patients be more or less reluctant to screen if the screening mode were sigmoidoscopy? Please state your opinion.

• Much more reluctant: 4

• A little more reluctant: 1

• Equally reluctant (or enthusiastic): 9

• A little less reluctant: 6

• Much less reluctant: 1

6. Would patients be more or less reluctant to screen if the screening mode were faecal testing? Please state your opinion

• Much more reluctant: 0

• A little more reluctant: 3

• Equally reluctant (or enthusiastic): 2

• A little less reluctant: 5

• Much less reluctant: 11

Conclusions

Based on the results of the first survey and the following panel discussion, the panel

concluded they believe most people would want a benefit of at least a 10 in 1000 (1%)

reduction in colorectal cancer mortality OR colorectal cancer incidence to screen for

colorectal cancer. The panel thought this would be true for any of the four screening options

(FIT every, FIT every two years, sigmoidoscopy or colonoscopy). The panel discussed

whether they should consider various combinations of reduction in colorectal cancer

incidence and mortality. However, considering issues of complexity, the panel decided to

proceeded to address these two benefits separately.

Second survey

Practical issues, burdens and harms of screening was presented to the panel in a separate

document by e-mail, as follows:

Burdens and harms of screening

Here we give a summary of the burdens and harms of colorectal cancer screening that you

should bear in mind when you answer the survey.

Faecal immunochemical testing (FIT) must be repeated every year or every two years.

The patient conducts the test at home by taking a stool sample that is mailed or delivered

for analysis. No preparations are needed. If a test is positive the patient will be referred for

colonoscopy.

10

Sigmoidoscopy is performed once, at a hospital or clinic, and examines the lower part of

the large bowel. Typically, the patient will receive a bowel enema on the procedure day.

Sedation is uncommon, and no recovery time is necessary. If a polyp is detected, the patient

will be referred for colonoscopy.

Colonoscopy is performed once, at a hospital or clinic, and examines the full length of the

large bowel. The patient must clean the bowel with oral laxatives either starting the day

before the procedure, or in the early morning of the procedure. Sedation practices vary,

from no sedation to deep sedation. Depending on sedation regime, there is need for some

recovery time after the procedure in which activities are limited and one cannot drive. Deep

sedation may slightly increase the risk of complications from colonoscopy.

Both colonoscopy and sigmoidoscopy may cause moderate to severe discomfort or pain

during and after the procedure, depending on sedation regime, endoscopy technique and

endoscopist. However, even when performed without sedation the majority will experience

only mild or no discomfort or pain.

For all the screening tests, the result of the (work-up) colonoscopy trigger the need for

colonoscopy surveillance at regular intervals, typically every 3-5 years, depending on polyp

findings. A positive primary screening test may increase psychological distress, but this is

probably of short duration (

11

1 Perforations, gastrointestinal bleeding or transfusions requiring hospitalization or an emergency department

visit within 30 days after screening/work-up or surveillance colonoscopy 2Paralytic ileus, nausea, vomiting and dehydration or abdominal pain requiring emergency department visit or

hospitalization requiring hospitalization or an emergency department visit within 30 days after

screening/work-up or surveillance colonoscopy 3 Myocardial infarction or angina, arrhythmias, congestive heart failure, cardiac or respiratory arrest, syncope,

hypotension, or shock requiring hospitalization or an emergency department visit within 30 days after

screening-, work-up or surveillance colonoscopy

No

screening

FIT every

two years

FIT every

year

Sigmoidoscopy Colonoscopy

Procedure related

mortality

15 years

0 per 1000 No apparent

difference

No apparent

difference

No apparent

difference

No apparent

difference

Gastrointestinal

perforation or bleeding

1

15 years

0 per 1000 1 more 2 more 2 more 2 more

Other gastrointestinal

adverse events2

15 years


Cardiovascular events3

15 years


Number of screening

tests

15 years


Number of participants

with one or more

colonoscopies

15 years


Number of participants

with two or more

colonoscopies

15 years


Second survey: Colorectal cancer screening - choosing between tests

Replies: 18 (19 replied to the question for FIT every two years). Number of votes given after each response option

Introduction to the survey

When completing the following survey bear in mind the burdens and harms of the four screening options that has been provided by e-mail

in a separate document. Try to think of what most people would do if they were fully informed of the potential burdens and harms related

to the four screening methods. Please consider that the response options translate into the following percentages:

-"All or almost all" means over 90% would choose this option

-"Most" means 75 to 90% would choose this option

12

-"Majority" means 51 to 74% would choose this option

We will present the results of this survey in our next panel meeting. We will only present votes, and not who voted for what.

Colonoscopy versus sigmoidoscopy

For the following questions bear in mind that we have previously decided that typical patients consider the burdens and harms of

sigmoidoscopy and colonoscopy more or less equivalent.

1. Evidence suggests a reduction in colorectal cancer mortality of 1 per 1,000 (0.1%) over 15 years with colonoscopy versus

sigmoidoscopy, what would patients choose?

• All or almost all would choose colonoscopy: 1

• Most would choose colonoscopy: 1

• The majority would choose colonoscopy: 6

• The majority would choose sigmoidoscopy: 5

• Most would choose sigmoidoscopy: 5

• All or almost all would choose sigmoidoscopy: 0









3. Evidence suggests a reduction in colorectal cancer mortality of 10 per 1,000 (1%) over 15 years with colonoscopy versus








Colonoscopy versus FIT


yearly FIT, what would patients choose?




• The majority would choose yearly FIT: 6

• Most would choose yearly FIT: 9

• All or almost all would choose yearly FIT: 1

















7. Evidence suggests a reduction in colorectal cancer mortality of 1 per 1,000 (0.1%) over 15 years with colonoscopy versus FIT

every two years, what would patients choose?




• The majority would choose FIT every two years: 5

13

• Most would choose FIT every two years: 9

• All or almost all would choose yearly FIT every two years: 2

8. Evidence suggests a reduction in colorectal cancer mortality of 5 per 1,000 (0.5%) over 15 years with colonoscopy versus FIT






• Most would choose yearly FIT every two years: 5

• All or almost all would choose FIT every two years: 0


FIT every two years, what would patients choose?







Sigmoidoscopy versus FIT

10. Evidence suggests a reduction in colorectal cancer mortality of 1 per 1,000 (0.1%) over 15 years with sigmoidoscopy versus yearly FIT, what would patients choose?







11. Evidence suggests a reduction in colorectal cancer mortality of 5 per 1,000 (0.5%) over 15 years with sigmoidoscopy versus








12. Evidence suggests a reduction in colorectal cancer mortality of 10 per 1,000 (1.0%) over 15 years with sigmoidoscopy versus yearly FIT, what would patients choose?

























14







FIT every year versus FIT every two years

16. Evidence suggests a reduction in colon cancer mortality of 1 per 1,000 (0.1%) over 15 years with FIT every year versus FIT every two years, what would patients choose?

• All or almost all would choose FIT every year: 0

• Most would choose FIT every year: 3

• The majority would choose FIT every year: 2




17. Evidence suggests a reduction in colon cancer mortality of 5 per 1,000 (0.5%) over 15 years with FIT every year versus FIT








18. Evidence suggests a reduction in colon cancer mortality of 10 per 1,000 (1.0%) over 15 years with FIT every year versus FIT








Conclusions

Based on the results of the second survey and the following panel discussion, the panel

decided on guidance for choosing between tests to be applied when evaluating the estimates

of colorectal cancer mortality with the different screening options:

• If the difference in colorectal cancer mortality reduction is 1 per 1000 (0,1%) with

colonoscopy vs. sigmoidoscopy the panel will not recommend one option over the

other; however, with a colorectal cancer mortality reduction of 5 per 1000 (0,5%) or

more with colonoscopy vs. sigmoidoscopy, the panel will recommend colonoscopy.

• If the difference in colorectal cancer mortality reduction is 10 per 1000 (1%) or more

with colonoscopy vs. FIT every year or every two years, the panel will recommend

colonoscopy.

• If the difference in colorectal cancer mortality reduction is 10 per 1000 (1%) or more

with sigmoidoscopy versus FIT every year or FIT every two years, the panel will

recommend sigmoidoscopy

• The thresholds where it flips from choosing sigmoidoscopy or colonoscopy over FIT

may be lower depending on potential incidence benefits with the different tests.

• If the difference in colorectal cancer mortality reduction is 5 per 1000 (0,5%) with FIT

every year vs. FIT every second year, the panel will recommend FIT every year.

15

Third survey

After the guideline panel was presented with estimates of benefits of screening in terms of

colorectal cancer mortality and incidence reductions, it became clear that there was no longer

consensus for requiring a colorectal cancer mortality and/or incidence reduction of 10 per

1000 (1%) to recommend screening with FIT. The majority of the guideline panel felt that the

burdens and harms of FIT screening were considerably smaller than the burdens and harms

from sigmoidoscopy or colonoscopy, and the panel therefore decided to re-visit the threshold

decision from the first survey for FIT.

Third survey: Re-visiting FIT threshold

Replies: 16. Number of votes given after each response option

Introduction to the survey

Panel members vary in their views regarding the threshold for recommending for and against FIT. To try and move toward consensus we

need your help in completing the following exercise. This is fairly long: Please bear with us. We will be happiest if you complete all the

questions, but if you feel that you have expressed your views clearly it is fine if you stop before finishing all the questions.

As you complete the survey, bear in mind that we have a threshold of 10 per 1000 (1%) reduction for both colorectal mortality and/or

incidence for recommending for sigmoidoscopy and colonoscopy, and we have decided to stay with this threshold. Please also review the

burdens and harms of screening and try to think of what most people would do if they were fully informed of these potential burdens and

harms.

Remember that the response options translate into the following percentages:

-"All or almost all" means over 90% would choose this option

-"Most" means 75 to 90% would choose this option

-"Majority" means 51 to 74% would choose this option

We will present the results of this survey in our next panel meeting. We will only present votes, and not who voted for what.

Below is a table of burdens and harms for each of the four screening options. The estimates are derived from a microsimulation model

(MISCAN-Colon) and all represent low certainty evidence. Estimates apply to a population between 50 to 79 years, with a 4% risk for

colorectal cancer within the next 15 years (which corresponds to approximately the 50th centile for men in UK). The simulation assumes

100% compliance to the chosen screening method and follow-up regime, and a timeframe of 15 years.

No

screening

FIT every two

years

FIT every

year

Sigmoidoscopy Colonoscopy

Procedure related mortality

15 years

0 per 1000 No apparent

difference

No apparent

difference

No apparent

difference

No apparent

difference

Gastrointestinal perforation or

bleeding 1

15 years


16

FIT every year versus no screening

1. A patient who is screened with FIT every year for 15 years, has a 1 in 1000 (0.1%) lower risk of dying from or being

diagnosed with colorectal cancer at 15 years. Given the harms and burdens of screening, how would patients view such

benefits?

• All or almost all would choose screening: 0


• The majority would choose screening: 4

• The majority would decline screening: 5


• All or almost all would decline screening: 4



benefits?









benefits?




Other gastrointestinal adverse

events2

15 years


Cardiovascular events3

15 years


Number of screening tests

15 years


Number of participants with one

or more colonoscopies

15 years


Number of participants with two

or more colonoscopies

15 years


17






benefits?









benefits?







6. A patient who is screened with FIT every year for 15 years, has an 8 in 1000 (0.8%) lower risk of dying from or being


benefits?









benefits?









benefits?



18







benefits?









benefits?







Conclusions

Based on the results of the third survey and the following panel discussion, the panel decided

they believe most people would want a benefit of at least a 5 in 1000 (0.5%) reduction in

either colorectal cancer mortality or colorectal cancer incidence to screen with FIT every year

or FIT every two years.

References

1. Selva A, Solà Yuan, Pardo-Hernandez H, et al. Development and use of a content search strategy for

retrieving studies on patients’ views and preferences. Health and Quality of Life Outcomes 2017;15:126

2. Hol L, de Bekker-Grob EW, van Dam L, et al. Preferences for colorectal cancer screening strategies: a

discrete choice experiment. Br J Cancer 2010;102(6):972-80. doi: 10.1038/sj.bjc.6605566 [published

Online First: 2010/03/04]

3. Kistler CE, Hess TM, Howard K, et al. Older adults' preferences for colorectal cancer-screening test

attributes and test choice. Patient preference and adherence 2015;9:1005-16. doi: 10.2147/ppa.S82203

[published Online First: 2015/07/24]

19

Web appendix 4:

Screening benefits by sex

Sigmoidoscopy trials - the randomised screening trials suggested difference in colorectal

cancer incidence and mortality by sex.1 There was high certainty evidence for a reduction in

colorectal cancer mortality for both women and men, but the effect was smaller in women

(RR 0.85, 95 % confidence interval 0.71 to 0.96) compared to men (RR 0.67, 95 %

confidence interval 0.61 to 0.75). There was also a smaller reduction in colorectal cancer

incidence in women (RR 0.86, 95 % confidence interval 0.81 to 0.92) compared to men (RR

0.75, 95 % confidence interval 0.71 to 0.79). The review team assessed the credibility of the

observed subgroup differences to be moderate, supporting a greater relative effect of

sigmoidoscopy in men than in women for colorectal cancer incidence and mortality.1

gFOBT - there was no available data from the trials to do sex-stratified analyses of the

benefits of gFOBT.

Microsimulation - the microsimulation model could not replicate the sex-specific differences

in colorectal cancer incidence and mortality reduction as observed in the meta-analysis of

sigmoidoscopy trials.2

Implications for clinical practice - examples with translation to absolute numbers:

Let us consider a man and a woman with a 2% risk of colorectal cancer over 15 years.

According to the trial results summarized above, sigmoidoscopy screening for the male

individual would result in a 25% relative risk reduction for colorectal cancer incidence,

resulting in an absolute risk reduction of 0.5%, thus reducing his colorectal cancer risk from

2% to 1.5%. The woman can expect a 14% relative reduction, resulting in an absolute risk

reduction of 0.28%, thus reducing her colorectal cancer risk from 2% to 1.72%.

Let us now consider a man and a woman with a 4% absolute risk of colorectal cancer over 15

years. According to the trial results summarized above, sigmoidoscopy screening for the male

individual would result in a 25% relative risk reduction for colorectal cancer incidence,

resulting in an absolute risk reduction of 1%, thus reducing his colorectal cancer risk from 4%

to 3%. The woman can expect a 14% relative reduction, resulting in an absolute risk reduction

of 0.56%, thus reducing her colorectal cancer risk from 4% to 3.44%.

These two examples show that the risk of disease has a much larger influence on the absolute

effects of screening than the presumable relative differences of sigmoidoscopy screening

effectiveness between men and women. Therefore, the panel concluded that these differences

are not large enough to impact on the recommendations and thus did not make separate

recommendations for men and women.

References

1. Jodal HC ea. Colorectal cancer screening with faecal testing, sigmoidoscopy or colonoscopy: a

systematic review and network meta-analysis. In submission

2. Buskermolen M ea. Colorectal cancer screening with faecal immunochemical testing, sigmoidoscopy or colonoscopy: a microsimulation modelling study. In submission

20

Web Appendix 5:

Harms and burden of screening

Literature review

The panel requested a systematic survey of reviews of screening-related burdens and harms

that included observational studies designed to evaluate harms after screening, because data

from randomised trials were limited.

We searched for systematic reviews of observational studies reporting on harms and burdens

of colorectal cancer screening in asymptomatic individuals undergoing screening for

colorectal cancer. We searched Medline, EMBASE and CINAHL using a combination of

medical subject headings and key words related to harms and burdens of colorectal cancer

screening and limited to “best balance reviews”, up to April 12, 2018. The full search

strategies are included at the end of this appendix.

We included reviews of studies of asymptomatic populations that were screened with one of

the options considered by this guideline (faecal testing (gFOBT or FIT) every year or every

two years, sigmoidoscopy or colonoscopy). The outcomes we considered included screening-

related mortality, gastrointestinal perforation and bleeding and other serious adverse events.

In addition, we looked for evidence of psychological impact of a positive screening test, pain

and discomfort related to sigmoidoscopy or colonoscopy, gastrointestinal complications

related to sedation during colonoscopy, as well as time requirements for the different

screening options. The panel asked the review team to explore potential subgroup effects of

screening related harms for age

Reviewers (LMH, HCJ) screened titles and abstracts in duplicate and assessed eligibility from

the full text for all possibly eligible articles. By using Amstar2 criteria1, five reviews with at

least moderate overall confidence were chosen to inform this guideline.2-6 The results of this

review are summarized here per outcome of interest.

21

Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow

diagram of studies included in the survey of systematic reviews of screening-related harms

and burdens

Procedure-related mortality: One review gave a pooled estimate of procedure-related

mortality after colonoscopy defined as death within three months after procedure from

cardiorespiratory events, gastrointestinal perforation or bleeding.2 The review included 18

observational studies examining complications after colonoscopy for mixed indications in

949,209 individuals and found an overall colonoscopy-related mortality rate of 0.03 per 1000

colonoscopies (95% confidence interval 0.01 to 0.06).

Colorectal perforation and bleeding: One review provided pooled estimates of

gastrointestinal perforation and bleeding in the setting of screening with faecal testing

(gFOBT and FIT) and sigmoidoscopy.3 The perforation rate was 0.8 per 1000 colonoscopies

22

(95% confidence interval from 0.2 to 3.2) performed after a positive faecal test, and 1.4 per

1000 screened with sigmoidoscopy (95% confidence interval 0.9 to 2.6). The bleeding rate

was 1.9 per 1000 colonoscopies (95% confidence interval 0.5 to 6.4) performed after a

positive faecal test, and 3-4 per 1000 screened with sigmoidoscopy (95% confidence interval

0.5 to 6.3)

Three systematic reviews had pooled estimates of gastrointestinal perforation and bleeding

after colonoscopy screening/surveillance.2-4 The perforation rates were similar for all reviews:

0.3 per 1000 colonoscopies (95% confidence interval 0.2 to 0.5)2 ; 0.4 per 1000 colonoscopies

(95% confidence interval 0.2 to 0.5)3; and 0 to 2.2 per 1000 colonoscopies.4 Likewise, the

bleeding rates were similar for all reviews: 2.4 per 1000 colonoscopies (95% confidence

interval from 0.9 to 4.6)2; 0.8 per 1000 colonoscopies (95% confidence interval from 0.5 to

1.4)3; and 0 to 1.9 per 1000 colonoscopies4.

Other gastrointestinal adverse events: We found no systematic review with pooled

estimates of other gastrointestinal adverse events. One large American registry study based on

data from SEER-Medicare (n=53 220 , age 66 to 95), was included in two of the systematic

reviews2,3 and provided estimates of the risk of other gastrointestinal adverse events defined

as paralytic ileus, nausea and vomiting, dehydration and abdominal pain causing an

emergency room visit or hospitalization within 30 days after colonoscopy.7 The study found

that only colonoscopy with polypectomy was associated with an increased risk of

complications at 30 days. Estimates per 1000 individuals as follows:

• No colonoscopy: 5.7 (95% confidence interval 5.0 to 6.3)

• Colonoscopy without polypectomy: 6.5 (95% confidence interval 4.2 to 8.9)

• Colonoscopy with polypectomy: 13 (95% confidence interval 11.7 to 14.4)

Cardiovascular adverse events: None of the systematic reviews provided pooled estimates

of cardiovascular events after colorectal cancer screening, and only two of the included

primary studies compared harms other than perforation and bleeding with a control group.7,8

These two studies did not find a statistically significantly higher risk of serious harms due to

screening colonoscopy (including myocardial infarction, cerebrovascular accident, other

cardiovascular events, and mortality). The study based on data from Medicare7 found that

colonoscopy with polypectomy was associated with an increased risk of cardiovascular

events. Per 1000 individuals in the no-colonoscopy group, the 30-day risk of cardiovascular

events was 15.9 (95% confidence interval 14.8 to 16.9) versus 23.3 (95% confidence interval

21.6 to 25.1) among those with a colonoscopy with polypectomy.

Anxiety related to a positive screening test: Two systematic reviews with large overlap in

included studies, examined psychological distress related to colorectal cancer screening with

faecal testing or sigmoidoscopy.5,6 None of the reviews were able to pool data due to a wide

variation in outcomes measured. Five out of seven prospective studies included in the

reviews, reported adverse effects on psychological well-being after a positive test result. The

largest effects were observed before the screening test, in anticipation of and shortly after

being informed about a positive test result. This effect declined after work-up colonoscopy

and disappeared after one month to one year.

23

Harms by age: One systematic review reported on harms of screening for different age

groups3. Of the included studies, 18 provided analyses of harms of colonoscopy by age

groups, but the authors were not able to pool data for any of the outcomes. One American

registry-based study with 55,000 patients indicated that ages 75 to 84 may be associated with

higher odds of serious bleeding, perforation and cardiovascular adverse events than ages 66 to

74.9

The study based on Medicare data provided the only estimates of complications for different

age groups relative to a matched non-colonoscopy population.7 This study found a small

increase in gastrointestinal adverse events with increasing age. For gastrointestinal

perforations and bleeding it differed from 5.0 per 1,000 (95% confidence interval 3,8 - 6,2)

for ages 66 to 69, to 7,2 per 1000 (95% confidence interval 5,9 - 8,6) for ages 75 to 79.7 The

risk of cardiovascular events also increased with age but differed very little with the observed

risks in the matched non-colonoscopy group.

No systematic reviews reported on pain and discomfort related to sigmoidoscopy or

colonoscopy, gastrointestinal complications related to sedation during colonoscopy or time

requirements for the different screening options.

Estimates of adverse events in a 15-year period

The panel decided that the time frame under consideration for this guideline was 15 years.

None of the identified reviews, nor any of the randomised screening trials informing this

guideline provided estimates of screening complications after long-term follow-up.10

Therefore, to provide estimates for a 15-year period, including complications due to

surveillance colonoscopies, we were dependent on modelling.

The MISCAN model includes predictions of gastrointestinal perforation and bleedings

(perforation, gastrointestinal bleeding or transfusion), other gastrointestinal events (paralytic

ileus, nausea, vomiting and dehydration or abdominal pain) and cardiovascular adverse

events (myocardial infarction or angina, arrhythmias, congestive heart failure, cardiac or

respiratory arrest, syncope, hypotension, or shock) requiring hospitalization within 30 days

after colonoscopy. The risk of these complications is calculated through additional analyses of

data from the American registry study utilizing data from SEER-Medicare.7,11 This registry

study included data from 53 220 colonoscopies performed between 2001 to 2005 on

individuals aged 66-95. The literature review gave support to keep the pre-defined modelling

assumptions for these outcomes, as the estimates from the registry cohort were in line with the

best current evidence.

References

1. Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that

include randomised or non-randomised studies of healthcare interventions, or both. BMJ

2017;358:j4008. doi: 10.1136/bmj.j4008Reumkens 2016

24

2. Reumkens A, Rondagh EJ, Bakker CM, et al. Post-Colonoscopy Complications: A Systematic Review,

Time Trends, and Meta-Analysis of Population-Based Studies. Am J Gastroenterol 2016;111(8):1092-

101. doi: 10.1038/ajg.2016.234 [published Online First: 2016/06/15]

3. Lin JS, Piper MA, Perdue LA, et al. U.S. Preventive Services Task Force Evidence Syntheses, formerly

Systematic Evidence Reviews. Screening for Colorectal Cancer: A Systematic Review for the US

Preventive Services Task Force. Rockville (MD): Agency for Healthcare Research and Quality (US)

2016.

4. Tinmouth J, Vella ET, Baxter NN, et al. Colorectal Cancer Screening in Average Risk Populations:

Evidence Summary. Canadian journal of gastroenterology & hepatology 2016;2016:2878149. doi:

10.1155/2016/2878149 [published Online First: 2016/09/07]

5. van der Velde JL, Blanker MH, Stegmann ME, et al. A systematic review of the psychological impact

of false-positive colorectal cancer screening: What is the role of the general practitioner? Eur J Cancer

Care (Engl) 2017;26(3) doi: 10.1111/ecc.12709 [published Online First: 2017/05/12]

6. Vermeer NC, Snijders HS, Holman FA, et al. Colorectal cancer screening: Systematic review of screen-

related morbidity and mortality. Cancer Treat Rev 2017;54:87-98. doi: 10.1016/j.ctrv.2017.02.002


7. Warren JL, Klabunde CN, Mariotto AB, et al. Adverse events after outpatient colonoscopy in the

Medicare population. Ann Intern Med 2009;150(12):849-57, w152. [published Online First:

2009/06/17]

8. Stock C, Ihle P, Sieg A, et al. Adverse events requiring hospitalization within 30 days after outpatient

screening and nonscreening colonoscopies. Gastrointest Endosc 2013;77(3):419-29. doi:

10.1016/j.gie.2012.10.028 [published Online First: 2013/02/16]

9. Zafar HM, Harhay MO, Yang J, et al. Adverse events Following Computed Tomographic

Colonography compared to Optical Colonoscopy in the Elderly. Preventive medicine reports 2014;1:3-

8. doi: 10.1016/j.pmedr.2014.08.001 [published Online First: 2014/12/23]

10. Jodal HC ea. Colorectal cancer screening with faecal testing, sigmoidoscopy or colonoscopy: a

systematic review and network meta-analysis. In submission

11. van Hees F, Zauber AG, Klabunde CN, et al. The appropriateness of more intensive colonoscopy

screening than recommended in Medicare beneficiaries: a modeling study. JAMA internal medicine

2014;174(10):1568-76. doi: 10.1001/jamainternmed.2014.3889 [published Online First: 2014/08/19]

Full search strategies for literature review

Database: Ovid MEDLINE, 1946 to April 12 2018

1 exp colorectal neoplasms/ or colonic neoplasms/ or sigmoid neoplasms/ or colorectal neoplasms,

hereditary nonpolyposis/ or exp rectal neoplasms/ or intestinal polyps/ or colonic polyps/

2 ((colorectal or rectal or colon* or rectum* or coli or anus or anal or sigmoid* or bowel* or

intestin*) and (tumour* or tumor* or neoplasm* or cancer* or carcinoma* or adenocarcinom* or

adenom* or polyp* or lesion*)).tw,kf.

3 Precancerous Conditions/ or (precancer* or pre cancer*).tw,kf.

4 (colorectal or rectal or colon* or rectum* or coli or anus or anal or sigmoid* or bowel* or

intestin*).tw,kf.

5 3 and 4

6 1 or 2 or 5

7 Mass Screening/

8 exp Population Surveillance/

9 "Early Detection of Cancer"/

10 (screen* or (population* adj2 surveillance) or (early adj3 detect*) or (early adj3 prevent*)).tw,kf.

11 or/7-10

12 6 and 11

13 Occult Blood/

25

14 (gFOBT or FOBT or FOB or FIT or IFOBT or haemoccult or hemoccult or sensa or

heamoccultsensa or hemocare or hema screen or hemascreen or hemacheck or hema check or

hemawipe or hema wipe or hemofec or hemofecia or fecatest or fecatwin or coloscreen or seracult or

colocare or flexsure or hemmoquant or immocare or hemochaser or bayer detect or hemeselect or

immudia or monohaem or insure or hemodia or immocare or magstream or guaiac or occult blood or

(stool adj3 occult) or (gaiac* adj2 smear*)).tw,kf.

15 ((faecal or fecal or feces or faeces) adj3 test*).tw,kf.

16 ((immunochemical adj3 test*) or (stool adj3 test*)).tw,kf.

17 colonoscopy/ or sigmoidoscopy/ or (colonoscop* or sigmoidoscop*).tw,kf.

18 or/13-17

19 12 and 18

20 Sigmoidoscopy/mo [Mortality]

21 Colonoscopy/mo [Mortality]

22 mortality/ or "cause of death"/ or fatal outcome/ or survival rate/ or mortality.ti,kf.

23 (crc mortality or colorectal cancer mortality).tw,kf.

24 (complications or adverse effects).fs.

25 (complicat* or harm* or adverse effect* or adverse event* or side effect*).tw,kf.

26 Intestinal Perforation/ or perforat*.tw,kf.

27 ((postpolypectom* or post polypectom*) adj4 (bleeding or blood loss or hemorrhage*)).tw,kf.

28 (remov* adj2 polyp* adj4 (bleeding or hemorrhage* or blood loss)).tw,kf.

29 ((bleeding or blood loss or hemorrhage*) adj4 (colon* or sigmoid*)).tw,kf.

30 Postoperative Hemorrhage/ or ((postoperativ* or post operativ*) adj4 (bleeding or blood loss or

hemorrhage*)).tw,kf.

31 exp Myocardial Infarction/ or (myocardial infarction* or heart infarction* or heart attack* or

ischemi* or cardiovascular).tw,kf.

32 exp Pulmonary Embolism/ or embol*.tw,kf.

33 diverticulitis/ or diverticulitis, colonic/ or diverticulitis.tw,kf.

34 Pancreatitis/ or pancreatitis.tw,kf.

35 Abdominal Pain/ or pain*.tw,kf.

36 exp "Hypnotics and Sedatives"/ or (sedation or sedated or sedatives).tw,kf.

37 "Quality of Life"/

38 (patient reported outcome* or PROMS).tw,kf.

39 Stress, Psychological/

40 (stress* or anxiety or psycho* or anxious* or fear or quality of life or distress* or discomfort* or

burden*).tw,kf.

41 psychology.fs.

42 anxiety/ or fear/

43 return to work/ or work/ or ((abscen* or return* or back) adj3 (work or job)).tw,kf.

44 "Recovery of Function"/ or recover*.tw,kf.

45 or/20-44

46 19 and 45

47 limit 46 to (danish or english or norwegian or swedish)

48 limit 47 to (yr="2008 -Current" and (danish or english or norwegian or swedish) and "reviews

(best balance of sensitivity and specificity)")

49 47 not 48

Database: Embase, 1974 to April 12 2018

1 intestine polyp/ or exp colon polyp/ or colorectal polyp/ or colon polyposis/ or duodenum polyp/

or rectum polyp/

2 colon tumor/ or colon adenoma/ or exp colon cancer/ or exp colon polyp/ or exp colorectal tumor/

26

3 exp rectum tumor/ or anus tumor/ or anus cancer/

4 ((colorectal or rectal or colon* or rectum* or coli or anus or anal or sigmoid* or bowel* or

intestin*) and (tumour* or tumor* or neoplasm* or cancer* or carcinoma* or adenocarcinom* or

adenom* or polyp* or lesion*)).tw,kw.

5 precancer/ or (precancer* or pre cancer*).tw,kw.

6 (colorectal or rectal or colon* or rectum* or coli or anus or anal or sigmoid* or bowel* or

intestin*).tw,kw.

7 5 and 6

8 or/1-4,7

9 mass screening/

10 health survey/

11 early cancer diagnosis/

12 (screen* or (population* adj2 surveillance) or (early adj3 detect*) or (early adj3

prevent*)).tw,kw.

13 or/9-12

14 8 and 13

15 occult blood/

16 (gFOBT or FOBT or FOB or FIT or IFOBT or haemoccult or hemoccult or sensa or





(stool adj3 occult) or (gaiac* adj2 smear*)).ti,kw.

17 ((faecal or fecal or feces or faeces) adj3 test*).tw,kw.

18 ((immunochemical adj3 test*) or (stool adj3 test*)).tw,kw.

19 colonoscopy/ or sigmoidoscopy/ or (colonoscop* or sigmoidoscop*).tw,kw.

20 or/15-19

21 14 and 20

22 mortality/ or all cause mortality/ or mortality rate/ or surgical mortality/ or mortality.ti,kw.

23 (si or co).fx.

24 (complicat* or harm* or adverse effect* or adverse event* or side effect*).tw,kw.

25 exp intestine perforation/ or perforat*.tw,kw.

26 ((postpolypectom* or post polypectom*) adj4 (bleeding or blood loss or hemorrhage*)).tw,kw.

27 ((bleeding or blood loss or hemorrhage*) adj3 (colon* or sigmoid*)).tw,kw.

28 postoperative hemorrhage/ or postoperative complication/ or ((postoperativ* or post operativ*)

adj4 (bleeding or blood loss or hemorrhage*)).tw,kw.

29 (remov* adj2 polyp* adj4 (bleeding or hemorrhage* or blood loss)).tw,kw.

30 exp heart infarction/ or ischemic heart disease/ or (myocardial infarction* or heart infarction* or

heart attack* or ischemi* or cardiovascular).tw,kw.

31 exp embolism/ or embol*.tw,kw.

32 diverticulitis/ or diverticulitis.tw,kw.

33 exp pancreatitis/ or pancreatitis.tw,kw.

34 exp abdominal pain/ or pain*.tw,kw.

35 exp hypnotic sedative agent/ or (sedation or sedated or sedatives).tw,kw.

36 exp "quality of life"/

37 emotional stress/

38 (stress* or anxiety or psycho* or anxious* or fear or quality of life or distress* or discomfort* or

burden*).tw,kw.

39 anxiety/ or fear/

40 return to work/ or work/ or ((abscen* or return* or back) adj3 (work or job)).tw,kw.

41 convalescence/ or recover*.tw,kw.

27

42 (patient reported outcome* or PROMS).tw,kw.

43 or/22-42

44 21 and 43

45 limit 44 to (danish or english or norwegian or swedish)

46 limit 45 to ("reviews (best balance of sensitivity and specificity)" and (danish or english or

norwegian or swedish) and yr="2008 -Current")

47 45 not 46

48 limit 47 to conference abstract

49 47 not 48

Database: CINAHL 1981 to April 12 2018

S1 (MH "Colorectal Neoplasms+")

S2 TX ( (colorectal or rectal or colon* or rectum* or coli or anus or anal or sigmoid* or bowel* or

intestin*) ) AND TX ( (tumour* or tumor* or neoplasm* or cancer* or carcinoma* or adenocarcinom*

or adenom* or polyp* or lesion*) )

S3 S1 OR S2

S4 (MH "Precancerous Conditions")

S5 TI ( precancer* or pre cancer* ) OR AB ( precancer* or pre cancer* )

S6 S4 AND S5

S7 S3 OR S6

S8 (MH "Disease Surveillance")

S9 (MH "Cancer Screening")

S10 (MH "Early Detection of Cancer)

S11 TI ( (screen* or (population* N2 surveillance) or (early N3 detect*) or (early N3 prevent*) ) OR

AB ( (screen* or (population* N2 surveillance) or (early N3 detect*) or (early N3 prevent*) )

S12 S8 OR S9 OR S10 OR S11

S13 S7 AND S12

S14 (MH "Occult Blood")

S15 TI ( gFOBT or FOBT or FOB or FIT or IFOBT or haemoccult or hemoccult or sensa or





(stool N3 occult) or (gaiac* N2 smear*) ) OR AB ( gFOBT or FOBT or FOB or FIT or IFOBT or

haemoccult or hemoccult or sensa or heamoccultsensa or hemocare or hema screen or hemascreen or

hemacheck or hema check or hemawipe or hema wipe or hemofec or hemofecia or fecatest or fecatwin

or coloscreen or seracult or colocare or flexsure or hemmoquant or immocare or hemochaser or bayer

detect or hemeselect or immudia or monohaem or insure or hemodia or immocare or magstream or

guaiac or occult blood or (stool N3 occult) or (gaiac* N2 smear*) )

S16 TI ( (faecal or fecal or feces or faeces) N3 test* ) OR AB ( (faecal or fecal or feces or faeces) N3

test* )

S17 TI ( (immunochemical N3 test*) or (stool N3 test*) ) OR AB ( (immunochemical N3 test*) or

(stool N3 test*) )

S18 (MH "Colonoscopy+") OR (MH "Sigmoidoscopy")

S19 TI ( colonoscop* or sigmoidoscop* ) OR AB ( colonoscop* or sigmoidoscop* )

S20 S14 OR S15 OR S16 OR S17 OR S18 OR S19

S21 S13 AND S20

S22 (MH "Mortality+")

S23 (MH "Cause of Death")

S24 TI mortality

28

S25 TI ( complicat* or harm* or adverse effect* or adverse event* or side effect* ) OR AB (

complicat* or harm* or adverse effect* or adverse event* or side effect* )

S26 (MH "Colonoscopy+/AE

S27 (MH "Sigmoidoscopy/AE")

S28 TI perforat* OR AB perforat* OR (MH "Intestinal Perforation")

S29 TI ( (postpolypectom* or post polypectom*) N4 (bleeding or blood loss or hemorrhage*) ) OR

AB ( (postpolypectom* or post polypectom*) N4 (bleeding or blood loss or hemorrhage*) )

S30 TI ( remov* N2 polyp* N4 (bleeding or hemorrhage* or blood loss) ) OR AB ( remov* N2

polyp* N4 (bleeding or hemorrhage* or blood loss) )

S31 TI ( (bleeding or blood loss or hemorrhage*) N4 (colon* or sigmoid*) ) OR AB ( (bleeding or

blood loss or hemorrhage*) N4 (colon* or sigmoid*) )

S32 TI ( (postoperativ* or post operativ*) N4 (bleeding or blood loss or hemorrhage*) ) OR AB (

(postoperativ* or post operativ*) N4 (bleeding or blood loss or hemorrhage*) )

S33 (MH "Postoperative Hemorrhage")

S34 TI ( myocardial infarction* or heart infarction* or heart attack* or ischemi* or cardiovascular )

OR AB ( myocardial infarction* or heart infarction* or heart attack* or ischemi* or cardiovascular )

S35 (MH "Myocardial Ischemia+")

S36 (MH "Pulmonary Embolism")

S37 TI embol* OR AB embol*

S38 (MH "Diverticulitis") OR TI diverticulitis OR AB diverticulitis

S39 (MH "Pancreatitis") OR TI pancreatitis OR AB pancreatitis

S40 ( (MH "Abdominal Pain+") OR (MH "Colic+") ) OR TI pain* OR AB pain*

S41 ( (MH "Hypnotics and Sedatives+") ) OR TI ( sedation or sedated or sedatives ) OR AB (

sedation or sedated or sedatives )

S42 (MH "Quality of Life+")

S43 TI ( patient reported outcome* or PROMS ) OR AB ( patient reported outcome* or PROMS )

S44 (MH "Stress, Psychological+")

S45 TI ( stress* or anxiety or psycho* or anxious* or fear or quality of life or distress* or discomfort*

or burden* ) OR AB ( stress* or anxiety or psycho* or anxious* or fear or quality of life or distress*

or discomfort* or burden* )

S46 (MH "Anxiety+") OR (MH "Fear+")

S47 (MH "Job Re-Entry")

S48 (MH "Work")

S49 TI ( (abscen* or return* or back) N3 (work or job) ) OR AB ( (abscen* or return* or back) N3

(work or job) )

S50 TI recover* OR AB recover* OR (MH "Recovery")

S51 S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31 OR S32 OR

S33 OR S34 OR S35 OR S36 OR S37 OR S38 OR S39 OR S40 OR S41 OR S42 OR S43 OR S44 OR

S45 OR S46 OR S47 OR S48 OR S49 OR S50 S52 S21 AND S51

S53 S21 AND S51 Limiters - Published Date: 20080101-20180531; Clinical Queries: Review - Best

Balance; Language: Danish, English, Norwegian, Swedish

S54 S52 NOT S53

29

Web appendix 6:

Criteria for choice of risk calculator

The guideline panel identified a need for a tool to estimate 15-year colorectal cancer risk in

individuals faced with a decision to undergo screening. Such a tool would allow tailoring

recommendations for screening according to individual risk factors and the proposed

threshold of 3% 15-year risk. Here we outline the rationale for why the panel suggests the

QCancer® (15 yr, colorectal) risk calculator to estimate the risk of colorectal cancer in

individual patients.

A set of criteria to guide the choice of colorectal cancer risk prediction model were defined

before evaluating existing evidence for such models. The criteria were based on GRADE

guidance1 and the specific needs for this clinical practice guideline, and defined as follows:

• The risk prediction model must have an online, user friendly calculator

• The model should have undergone at least one validation in a population separate from

the derivation set

• The discriminatory ability should be reported as the area under the receiver operating

curve (AUC), and preferably > 0.65 in all data sets used to derive and validate the

model.

• A visual representation must suggest a consistent good fit between predicted and

observed outcomes in patients at different risk categories

• The purpose of the model must be prediction of incident colorectal cancer in an

average risk population

• Risk factors included in the model must be confined to those available in routine

health care or patient questionnaires (excludes those that requires genetic information

as this is not available in routine health care)

• Preference for a model not including biochemical data as this makes it more

complicated to use for the target group of the guideline

• The risk calculator could contain information on previous screening/polyps as long as

“no” is a possible answer (because the guideline is intended for people with no history

of screening).

We identified a publication by Juliet Usher-Smith and colleagues as the most up to date and

reliable systematic review on prediction models for colorectal cancer2. In an external

validation of different risk calculators, Usher-Smith and colleagues found that for men, the

best performing models, in addition to the QCancer10, were models by Tao et

a.l (2014), Driver et al. (2007) and Ma et al. (2010), all having AUCs over 0.67.3 For women,

QCancer10, Tao et al. (2014), Guesmi et al. (2010) and Wells et al. (2014) models were the

best performing with AUCs between 0.63 and 0.66.3 Of all the models, the QCancer104, had

the best fit with the above defined criteria.

30

References

1. Alba AC, Agoritsas T, Walsh M, Hanna S, Iorio A, Devereaux PJ, McGinn T, Guyatt G. Discrimination

and Calibration of Clinical Prediction Models: Users' Guides to the Medical Literature. JAMA. 2017

Oct 10;318(14):1377-1384. doi: 10.1001/jama.2017.12126.

2. Usher-Smith JA, Walter FM, Emery JD, et al. Risk Prediction Models for Colorectal Cancer: A

Systematic Review. Cancer Prev Res (Phila) 2016;9(1):13-26. doi: 10.1158/1940-6207.Capr-15-0274


3. Usher-Smith JA, Harshfield A, Saunders CL, Sharp SJ, Emery J, Walter FM, Muir K, Griffin SJ. (2018)

External validation of risk prediction models for incident colorectal cancer using UK Biobank. British

Journal of Cancer doi: 10.1038/bjc.2017.463

4. Hippsley-Cox J, Coupland C. Development and validation of risk prediction algorithms to estimate

future risk of common cancers in men and women: prospective cohort study. BMJ Open 2015;5:

e007825. doi:10.1136/bmjopen-2015-007825

31

Web appendix 7:

Methodology for development of BMJ Rapid Recommendations

About BMJ Rapid Recommendations

Translating research to clinical practice is challenging. Trustworthy clinical practice

recommendations are one useful knowledge translation strategy. Organisations creating

systematic reviews and guidelines often struggle to deliver timely and trustworthy

recommendations in response to potentially practice-changing evidence. BMJ Rapid

Recommendations aims to create trustworthy clinical practice recommendations based on the

highest quality evidence in record time. The project is supported by an international network

of systematic review and guideline methodologists, people with lived experience of the

diseases or conditions, clinical specialists, and front-line clinicians. This overview is one of a

package that includes recommendations and one or more systematic reviews published by the

BMJ group and in MAGICapp (http://www.magicapp.org). The goal is to translate evidence

into recommendations for clinical practice in a timely and transparent way, minimizing bias

and centred around the experience of patients. BMJ Rapid Recommendations will consider

both new and old evidence that might alter established clinical practice.

Process overview

1. On a daily basis, we monitor the literature for practice-changing evidence:

a. Formal monitoring through McMaster Premium LiteratUre Service (PLUS)

b. Informal monitoring the literature by BMJ Rapid Recommendations expert

groups, including clinician specialists and patients

32

2. The RapidRecs executive team and editors at The BMJ choose which clinical questions

to pursue among the identified potentially-practice changing evidence, based on relevance to a

wide audience, widespread interest, and likelihood to change practice.

3. We incorporate the evidence into the existing body of evidence and broader context of

clinical practice via:

a. a rapid and high-quality systematic review and meta-analysis on the benefits and

harms with a focus on the outcomes that matter to patients

b. parallel rapid recommendations that meet the standards for trustworthy guidelines1

by an international panel of people with relevant lived experience, front-line

clinicians, clinical content experts, and methodologists.

c. The systematic review and the recommendation panel will apply standards for

trustworthy guidelines.1,2 They use the GRADE approach, which has developed a

transparent process to rate the quality (or certainty) of evidence and grade the

strength of recommendations.3,4

d. Further research may be conducted including:

i. A systematic review of observational studies to identify baseline risk

estimates that most closely represent the population at the heart of the

clinical question, a key component when calculating the estimates of

absolute effects of the intervention

ii. A systematic review on the preferences and values of patients on the topic.

4. Disseminate the rapid recommendations through

a. publication of the research in BMJ journals

b. short summary of recommendations for clinicians published in The BMJ

c. press release and/or marketing to media outlets and relevant parties such as patient

groups

d. Links to BMJ Group’s Best Practice point of care resource

e. MAGICapp which provides recommendations and all underlying content

in digitally structured multilayered formats for clinicians and others who wish to

re-examine or consider national or local adaptation of the recommendations.

Who is involved?

Researchers, systematic review and guideline authors, clinicians, and patients often work in

silos. Academic journals may publish work from any one or combinations of these groups of

people and findings may also be published in the media. But it is rare that these g

full list of authors and their declarations of interests...2019/10/01 · cancer incidence). when...

Documents