A series of opinions published on healthydebate.ca assert that the Canadian Task Force on Preventive Health Care (Task Force) is flawed, and should be disbanded or reformed to include experts on each topic addressed. The authors assert that Task Force recommendations are out of date and use out-of-date evidence. These advocates understandably want the issues of interest to them to be vigorously addressed.
Experts practising in any field see the worst cases of “their disease,” be it cancer, blindness, mental disorder, heart attacks or strokes, and want to prevent that – an understandable emotional reaction. They want to increase the sensitivity of testing and expand the ages or other criteria for screening to ensure no one is missed.
However, those without epidemiology background have difficulty understanding that increasing sensitivity comes at a cost, whether decreasing specificity (i.e., false positive tests), increased exposure to harms of screening, or diversion of resources from other endeavours. The writers of these opinion pieces appear unaware why the Task Force is formulated the way it is, how it uses evidence and reaches its conclusions.
Conflict of interest is a major issue in guideline development. The authors surprisingly assert that the only reason to exclude experts is to prevent financial conflict of interest, and that this is unimportant in Canada because screening activities do not directly benefit specialists. Over a long history, many guidelines from specialty societies have recommended more interventions that have resulted in referrals and work for those specialists, leading to overdiagnosis, overtreatment and increased health-care costs.
Conflict of interest also includes strong, preconceived opinions/research interests that can lead to biased assessment of evidence. Therefore, many groups have produced “guidelines for guidelines” to reduce conflicts of interest and ensure their scientific validity. The Canadian Medical Association Journal follows such policies.
In public decisions, one of the greatest difficulties is determining who is an “expert.” Many people who research a topic have strongly held beliefs derived from prior ideas, not necessarily from evidence. Their research and analysis are influenced by prior beliefs. The “experts” nominated by peers in the field may be biased.
How the Task Force works
This current iteration of the Task Force was established to advise family physicians and other primary care providers on delivering preventive health care. These health-care professionals therefore comprise the majority of members, and all members have expertise in methods for evaluating health-care interventions, especially screening. They are unpaid volunteers, with a salaried civil service secretariat; systematic reviews of evidence are performed by contracted organizations. The Task Force methods for analyzing data are available on the website and are based on current best methods to define the questions, search for evidence, collate the evidence, then use the evidence to reach a recommendation. The evidence collation relies mainly on randomized controlled trials (RCTs) since these are the best methods of evaluating the efficacy of interventions. However, other evidence is also used where appropriate. At each step, outside experts are invited to review and comment, but do not vote on the actual recommendations.
Task Force decisions rely on published evidence so will always lag behind the latest innovations, but that is usually good. It takes time to fully evaluate new ideas; many are rejected for not living up to early promise, causing extra harms, or increasing costs excessively. Others may need adaptation or new learning before widespread use.
The Task Force considers the benefits, harms and costs of interventions when drafting a recommendation for clinicians. When trying to find early cases of disease, whether through a blood test, an X-ray or a screening questionnaire, the test/instrument used must be highly sensitive, picking up many minor variations from normal. Many of these slight variations are unimportant and will revert to normal if left alone or may persist but never develop enough to cause problems. If found by screening, such variations may lead to a cascade of further activities to determine whether this is or could become a serious disease. This process may include invasive testing and can lead to unnecessary treatments – even operations – and anxiety. People may suspend or alter their normal lives while going through these processes. Thus, more is not necessarily better, and patient outcomes may be worse with more “care.” In addition, patients may be rejected for life insurance – for income assurance, for their mortgage, or for travel.
Task Force decisions rely on published evidence so will always lag behind the latest innovations, but that is usually good.
It is inappropriate to inflict such harms on people unless the potential benefit outweighs those harms. Therefore, the Task Force must try to assess and balance the harms of any intervention (often immediate and to many) against the benefits that often accrue to few people and in the distant future.
Furthermore, the Task Force recognizes that its recommendations largely must be undertaken by family physicians who are already overloaded by simply trying to deal with the sickness that patients bring through their doors.
How much benefit from screening?
Many advocates of screening make the logical leap that if a disease has been discovered, it will lead to cure and a “normal” life. For most programs, the reality is that only a proportion of disease found is helped by early detection: much disease will proceed regardless, while even those helped may be left with scarring and after-effects and require long-term follow up. On the other hand, without screening, when disease is detected after symptoms appear, a proportion would still be cured. The absolute numbers suggest the benefits of screening are usually marginal.
When the preventable proportion of disease is higher, the benefits of screening can be greater than the harms. If the target population is expanded to lower risk groups, – for example younger people with low incidence of disease – the probability of benefit will be lower, but for most screening tests the harms remain similar, meaning the harms outweigh the benefits. Thus, choices of who to screen should not be based on finding every case, but on when this balance is likely beneficial.
The role of experts
During the process of preparing its guidelines, the Task Force requests experts in the field to participate in topic specific working groups as well as in discussions about evidence and reviewing draft guidelines. Most times, experts provide useful input and critiques that are incorporated into the guideline. But opinions unsubstantiated by evidence are disregarded.
An example of the opposition of experts came after the Task Force 2013 recommendations on cervical cancer screening (which was led by one of the authors of this article). In the 2000s, in Canada, most provinces recommended that women be screened annually, often including uncomfortable manual pelvic examinations, from about age 18, as did the United States. Many other countries initiated screening at later ages – 25 or even 30 years – and screened much less often. Yet, they did not have higher cancer rates. We found that in Canada, even before screening, invasive cancer before the age of 30 was rare, and especially before the age of 25. Based on that evidence, our working group recommended strongly against routine screening for women below age 20, weakly against for those under 25, weakly in favour for those 25 to 29 years, and strongly in favour for those over 30 years, with intervals of three years.
The response was vehement opposition from many gynecologists and advocacy groups. They opposed both raising the age for starting screening and longer intervals, as well as the cessation of “routine” pelvic examination. Provincial guidelines, which are mostly influenced by “experts” in the disease (but not experts in primary care, epidemiology or evidence), generally ignored the Task Force and continued recommending early and frequent screening. Change occurred slowly, with Alberta and British Columbia leading the way. Only now, more than 10 years later, have most provinces made those changes. Meanwhile, many young women had unnecessary examinations, and many older women have had them too frequently. As a result, many have been unnecessarily harmed by having multiple intrusive tests and invasive procedures for positive test results that led to further action. Some may have experienced pregnancy loss after surgery of the cervix to remove minor abnormalities found on Pap testing.
One of the experts asserted that the Task Force should have changed to HPV tests in 2013. Unfortunately, at that time, the evidence was limited, and the balance of benefits and harms was not yet clear. The cost was high, and the infrastructure was not yet available in Canada. Fortunately, change is occurring. HPV samples that can be taken by the women themselves are now replacing Pap tests in many provinces with research supporting longer intervals between tests. More than 10 years later, evidence to address these gaps is much improved and the Task Force is currently updating this guideline.
The Task Force reassesses its guidelines as new evidence emerges but is only funded to review and publish up to three guidelines per year. Thus, priorities must be set.
The future challenge
One of the benefits of any systematic review of evidence is to demonstrate the gaps in knowledge and what research is needed to address these gaps. The challenge to those who do not like Task Force recommendations is this: Can you provide or research the new evidence required to show how a different policy is better? Selectively picking and choosing evidence is inappropriate to provide sound and effective policy. If “experts” spent as much time and effort in improving the quality of their services and testing additional screening options, there could be real advances in the health of the public.
Canadians developed the concept of evidence-based medicine, and the first iteration of the Task Force was a pioneer in furthering these methods. Canadians continue to be leaders in developing more refined ways to assess evidence from more than just randomized controlled trials, and the current Task Force uses these methods.
The Task Force is focused on assisting the front line, where primary care clinicians must prioritize among the multitude of issues patients bring to the encounter. It would be a tragedy and disservice to all Canadians if lobbies for special interest groups were successful in disrupting this work.
Meantime, rather than pressing to over-screen those who already receive too much medicine, health-care professionals should be working to ensure recommended guidelines are applied for people who currently receive little or no primary care.
We are puzzled by this response. It appears that we disagree on what evidence-based medicine should be. We use peer-reviewed evidence (imperfect as that may be) to reach conclusions, whereas Dr Appavoo says this is too rigid.
It seems that Dr Appavoo thinks the cervical screening committee of the Task Force that published the January 2013 guidelines (the writing was done in the summer and fall of 2012), should have included experts who could have influenced the conclusions based on their predictions of research in progress. Trials underway in 2012 were published in 2015 to 2018. They were testing scientific questions whose answer was not yet known. Trials often find an answer different from what was expected. Even if the main effect is as expected, some aspects of it may differ, such as which age groups have the best response, or the duration of effect. It is therefore completely inappropriate to include experts who will “anticipate forthcoming ground-breaking evidence”. That is antithetical to the empirical scientific endeavour. Why do trials if experts can predict the outcomes in advance?
Dr Appavoo reiterates her view that experts do not have conflicts of interest that would influence their opinions or clinical practice. This also contradicts repeated empirical findings that guidelines written by experts tend to favour actions that lead to greater income . This may in part be due to cognitive biases that favour test sensitivity without as much attention to potential harms. Once completed, the trials of HPV tests showed their value in detection, but recall rates due to positive results were higher than for cytology testing. This would increase referrals to gynecologists to perform more colposcopy, biopsies, and downstream procedures. Thus, changing to HPV tests would benefit gynecologists and harm some women, unless counterbalanced by later start age and a longer interval between screening tests. Gynecologists opposed these latter changes though the trials, international evidence, and evaluation demonstrated their safety.
Dr Appavoo suggests that the Task Force recommendations caused delay in provinces changing to HPV testing. We hope the Task Force recommendations are read by decision-makers in provincial and territory screening guidelines groups, but they are also influenced by “experts” in gynecology and laboratory practice. The main recommendations of the Task Force were for later starting age and longer intervals of testing: which still have not been fully adopted in all Canadian jurisdictions. We understand that the reasons for delay in adopting HPV tests mainly came from laboratory concerns; that the costs of HPV testing systems and reagents is high, some provinces needed to transition from using glass slides to liquid cytology systems, and that once the decision to change was made it would be irreversible, due to loss of expertise and cytotechnology workforce. Provinces moved to change their test only when it was clear that the higher costs of HPV technology and positive test rates could be counterbalanced by performing tests less often.
The argument about PSA screening is too complex to take up in this response. We agree that the US trial of PSA screening was flawed, but so was the European trial, in differing ways. Many bodies assessing this decision have concluded that such screening provides marginal benefit, and even urology groups have become less emphatic than previously about recommending all men be screened.
Dr Appavoo refers to studies of young women’s mammography screening, and the paper by Wilkinson et al. This is a creditable collation of evidence between 2010 and 2018 from across the country. It shows a small association between provinces with a policy of earlier screening and changed spectrum of breast cancer cases. It concludes that perhaps early screening policies increase the proportion of breast cancers diagnosed at stage 1 and reduce those at stage 4. However, it does not show whether the outcomes are changed for these women: that is, whether there are fewer breast cancer deaths or women live longer. The findings could be due at least partly to lead-time bias. Nor does it take into account changes that occurred at varying times in different provinces in imaging technology, in better training for technical staff and radiologists, and changes in pathology techniques. Some of the findings may also be due to changes in the population screened, for example increased participation of people from different racial backgrounds. This study shows how critical it is to undertake high quality comparison studies to address such questions, so in future we will have better evidence to make decisions.
We all hope to reduce premature mortality from breast cancer, or indeed any disease, and it is unworthy for Dr Appavoo to accuse task force members of being anti-screening. She references a press report of the chair expressing doubt about the modelling approach used by the US Task Force in 2023, not about screening itself. If you examine all topics evaluated, when there is high quality evidence and the effect is substantial, the Task Force makes strong recommendations.
She also accuses the Task Force of “shifting the goalposts” in assessing evidence about women’s preferences and values, since they reached a conclusion that differs from what she prefers. We recognise that in a pluralistic society, there is a role for special interest groups to put their case forward. But such opinions alone are not enough to order interventions that have limited or no value (e.g., thyroid testing in the absence of risk factors or symptoms). Screening is a complex process, where intuitive, “common sense” ideas, even of “experts”, are often wrong. As Dr Appavoo calls for, interventions with a balance between potential benefits and harms make it desirable for patients to understand the evidence so they can make their own informed decisions. She was the lead author for the guidelines of the Canadian Association of Radiologists that do not mention that balance, nor of informing women about evidence.
An unpopular but real issue is also the current primary care crisis. Given the strains on health care in Canada, limited resources must be allocated appropriately to focus effort where most can be gained, and minimal harm created. We need to recognise that many people across Canada have difficulty obtaining care for manifest sickness. For these persons screening is a lower priority. Given the considerable effort of screening for small (but for some conditions, we agree real) benefit, consideration of this complexity is needed to guide front line clinicians in their interactions with patients. The role of the Canadian Task Force on Preventive Health Care is to assist them in this process.
James A Dickinson, Harminder Singh, Roland Grad.
Thank you for your article, I strongly agree with it. As a female family physician, I often counsel women on screening and routinely use the Canadian Task Force guidelines to discuss benefits and the harms of screening. I find their balanced, evidence-based, primary care approach is ideal in the setting of unclear evidence/lack of evidence and competing interests that is inherent in developing population screening guidelines. I very much hope the Canadian Task Force continues in its current form with primary care leading the development of its guidelines and continuing to emphasize an objective review of benefits vs. harms and opportunity cost of screening.
Thank you for the insightful article.
While acknowledging the importance of methodologists’
collaboration with content experts in shaping healthcare
guidelines, I respectfully argue against the exclusion of the latter.
Speaking to the authors’ example, the Canadian Task Force
cervical cancer guideline exemplifies a paradigm of rigid
methodology, which, I believe, hinders progress in Canada's
healthcare landscape. As mentioned by the authors, in 2013 a
leading gynecologist expressed objections to the guidelines upon
their publication. Keep in mind that financial conflict of interest is
minimal for gynecologists recommending home HPV tests.
At a juncture when HPV screening was gaining traction among
experts, the Task Force recommended against it, revealing a lack
of nuance and expertise in their approach. This deficiency was
evident when subsequent studies like NTCC (2015) and FOCAL
(2018) confirmed the benefits of HPV screening, contradicting the
Task Force's stance only 2 years after publication of the guideline.
Despite assurances of revising guidelines when new information
becomes available
(https://www.stcatharinesstandard.ca/ts/life/health_wellness/2021/
11/25/canadas-breast-cancer-screening-policy-based-off-flawed-
study-researchers.html), the 2013 recommendations persist in
2024, very likely contributing to Canada's lag in implementing
HPV screening programs compared to countries like Australia, the
Netherlands, and the UK. Had content experts been consulted
over a decade ago, Canada might have emerged as a leader in
this domain.
The failure to anticipate forthcoming groundbreaking evidence
underscores the importance of involving content experts at all
stages of guideline planning and development. Specialists’
insights could have potentially averted a substantial delay in HPV
screening rollout, thereby preventing avoidable cases of cervical
cancer among Canadian women.
The notion that content experts, due to their immersion in the
subject matter, are incapable of objectivity lacks empirical
substantiation. On the contrary, experts’ nuanced understanding
of diseases and research is invaluable in discerning the context
and validity of studies. For instance, the inclusion of the non-
blinded outlier Canadian National Breast Screening Studies
(CNBSS) in RCT analyses of mammographic screening led to
uncertainty resulting from mixed evidence related to its
methodological flaws
(https://academic.oup.com/jbi/article/4/2/135/6555326?login=false
). These flaws were known for decades, even explored by a
prominent statistician in the 1990s
(https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/1097-
0142%2819950215%2975%3A4%3C997%3A%3AAID-
CNCR2820750415%3E3.0.CO%3B2-M). Expert input could have
highlighted these shortcomings and prevented the perpetuation of
doubt regarding the efficacy of screening.
Similarly, the pooling of poorly performed studies—such as the
PLCO study with 85% contamination of the screening arm—with
the more robust ERSPC has engendered uncertainty regarding
the benefits of PSA screening. Had experts been involved, such
methodological pitfalls might have been averted. In the US the
screening recommendations changed back in favour of screening
with PSA after observation of increasing metastatic disease in
non-screened men. Canada has not followed suit.
The disconnect between the Task Force's recommendations and
subsequent patient outcomes underscores a sense of authority
without commensurate responsibility. Ultimately, it is the content
experts who bear the burden of caring for patients affected by
guideline shortcomings, as evidenced by documented trends in
cancer stages among younger women since the 2011
recommendations (https://www.mdpi.com/1718-7729/29/8/444).
By prioritizing antiquated RCTs over expert advice and assessing
guideline quality on methodological rigor rather than patient
outcomes, Canada's healthcare guidelines have regrettably fallen
behind the curve instead of leading the way.
All humans have bias, even methodologists. Acknowledging
biases, such as anti-screening bias, is important. It is betrayed in
many ways, such as a Task Force cochair having made up their
mind before evidence is reviewed
(https://www.thestar.com/life/health-wellness/some-doctors-
patients-want-canada-to-follow-u-s-proposal-for-earlier-
mammograms/article_9bfba9c8-2ec7-5c05-ab9b-
e906b39bc580.html) and in the shifting of goalposts in the
interpretation of women’s values surveys
(https://canadiantaskforce.ca/wp-
content/uploads/2018/11/Womens-Values-and-Preferences-on-
Breast-Cancer-Screening_FINAL.pdf). There is no such thing as a
lack of bias. Even AI can be biased, based on input
(https://www.forbes.com/sites/ariannajohnson/2023/05/25/racism-
and-ai-heres-how-its-been-criticized-for-amplifying-
bias/?sh=78d7f834269d). Balance is required to make cogent
decisions.
Guidelines transcend mere methodological exercises; they have
tangible implications for human lives and should prioritize
considerations of both morbidity and mortality. Thus, they
necessitate a nuanced approach that not only values evidence
but also empathizes with patient preferences and outcomes. The
archaic notion that physician objectivity necessitates detachment
from empathic emotions is both paternalistic and anachronistic. In
contemporary healthcare, empathy and expertise must harmonize
to ensure optimal patient care and guideline development.
Ultimately and intuitively, content experts should lead guideline
development in their fields and methodologists should play a
critical supportive role.
Thank you for this important question.
The Task Force mandate is to improve public health and health care provided to Canadians. It involves members of the public in its processes. https://canadiantaskforce.ca/methods/patient-preferences-protocol/
Regularly the Task Force takes suggestions from members of the public on what topics should be assessed, and what outcomes should be selected to guide the search for evidence on each topic. A patient advisory group provides input. You can join that if you are interested. https://canadiantaskforce.ca/get-involved/tf-pan/
The collection and assessment of evidence is difficult. At all stages, the balance between desirable and undesirable effects is sought; taking into account the variability or uncertainty in values and preferences of citizens.
In Task Force recommendations, the grades of decision are designed to focus attention on when patient preferences affect the choices made. The grades are: strong or conditional, for or against an activity, producing a four point scale – strong for, conditional for, conditional against and strong against.
A strong recommendation is made when the Task Force believes most people in this situation would want the recommended course of action and only a small number would not.
A conditional recommendation means that most people in this situation would want the recommended course of action, but many would not. Different choices are acceptable for each person, and clinicians should support patients and discuss their values and preferences to reach a n informed decision. Decision aids may support people in reaching these decisions.
Since there is such an important role for variations in patient preferences the Task Force aims to incorporate those preferences when the final guideline recommendations and knowledge translation (KT) tools are developed. Then patient KT tools are usability tested and members of the public provide feedback on them.
All materials are available on the Task Force website, for community physicians and members of the public. https://canadiantaskforce.ca/tools-resources/
The Task Force differs from most lobby groups for specific diseases, and even many provincial public health organizations, in that it provides information designed to enable members of the public to make their own decisions, with numerical estimates of the benefits and harms of screening activities by major risk groups: usually by age. They demonstrate the uncertainty of many decisions, especially at the margin.
It is easier to simply make a rule: e.g. every woman over age 40, 45, or 50 must have a mammogram. But nothing magic occurs on your birthday (sadly) not everyone is “average”, and people make personal judgements about what risks they prefer to take. Those who wish to learn more can read the rationale and background information on the Task Force website, in order to make the decision that is right for them.
Where would patients, their families , or the public fit?