Thanks for citing my discussion of these guidelines, which in ID practice show extraordinary variability from clinic to clinic -- and even within clinics among the same providers. Prior to supporting such broad screening, I'd suggest focusing on the higher risk populations -- say older, smoking history, HIV -- and referring directly for high-resolution anoscopy, which is much more accurate than the PAP.
I agree you can’t (and won’t) have RCT evidence for every screening test in every conceivable at risk population.
And in the absence of such evidence, the effectiveness and endpoint reduction benefits of any such strategies will remain unknown.
The answer is to have (far far) fewer of these “recommendations” that are non-evidence based.
It’s a sign of intelligence to recognize and admit when you don’t know something. Then you have a discussion with the patient regarding those unknowns and uncertainties. That’s seems far preferable (and more honest) than to pretend to know, and to then recommend stuff in which some of these guidelines writers have vested interests.
Your questions bubble up in a stew of uncertainty. The frequent absence of compelling evidence for efficacy and risk of treatment, imperfect screening tests with false positives and false negatives, Bayesian issues making comprehension of the applicability of screening horribly opaque, widely variable lethality and morbidity of diseases that are screened for, the essential subjectiveness of determining the value of finding a particular problem before it manifests, the variable and uncertain value from any perspective of early discovery depending on the disease, the use of relative statistics to present results. Stand back just a little and it’s a horrid mess.
The practicing physician advising patients and the advisory organizations making recommendations can be most helpful by providing specific findings, in Bayesian context, using absolute not relative statistical comparisons (including derivations like number-needed-to-screen), of the best available studies, along with an honest assessment of the shortcomings of the data. As long as support for the screening test (or prevention therapy) is anything short of widely compelling nothing less is ethical.
From any patient’s point of view, undergoing testing (or taking treatment for prevention) in order to reduce his risk to the small degree that many screening tests do may not be worth it, and that should be his call after being thoroughly informed.
I guess it's a good thing that Sensible Medicine and the internet didn't exist in Dr. Papanicolau's day - the Pap test clearly wouldn't have made it out of the gate . All those pelvics and smears every year on every woman most of them normal , but a boat load of false positives - and we didn't even know what caused cervical cancer at that time. The nice thing about science is that screening can be refined over time - as it has with cervical cancer. And lots of women alive because of a screening program that in retrospect would have been described as flawed and overreaching .And the whole idea behind the screening was that this was a terrible disease and deserved our attention . But if you never screen, then there is no refinement , just a shoulder shrug " well nothing we can do" . Don't get me wrong , I am not a fan of a lot of " routine " testing - but I think we should avoid adopting a nihilistic philosophy about it as well -
Sorry, wrong question, Ken. No clinician, especially ones in primary care, should be practising or collaborating with ‘population medicine’. Calling a test a ‘screening’ test is a deliberately subtle attempt to divert attention from, and ideally bypass, the ethical/legal requirement for the individual to give their informed consent to a test/treatment, having decided to do so on the basis of unbiased evidence/information about the prognostic pathway supplied by their doctor. Any recommendation is necessarily preference-sensitive and based on somebody else’s preferences - usually either the clinician’s or those of some guideline panel of content experts - and therefore necessarily undermines this process. Whether or not the person decides to have a test will be determined by their informed preferences, especially time preference and degree of risk aversion - and often relational concerns. On the basis of the same evidence lots may choose to do (e.g.) a PSA test, lots may not. The evidence is never enough, though essential.
Sensible take on screening. I think the NNS should always be part of the decision process. Patients should be aware of how many people need to be screened to save one life. Where quality of life matters the NNS to prevent one cancer specific death may also be important.
Thanks for citing my discussion of these guidelines, which in ID practice show extraordinary variability from clinic to clinic -- and even within clinics among the same providers. Prior to supporting such broad screening, I'd suggest focusing on the higher risk populations -- say older, smoking history, HIV -- and referring directly for high-resolution anoscopy, which is much more accurate than the PAP.
I agree you can’t (and won’t) have RCT evidence for every screening test in every conceivable at risk population.
And in the absence of such evidence, the effectiveness and endpoint reduction benefits of any such strategies will remain unknown.
The answer is to have (far far) fewer of these “recommendations” that are non-evidence based.
It’s a sign of intelligence to recognize and admit when you don’t know something. Then you have a discussion with the patient regarding those unknowns and uncertainties. That’s seems far preferable (and more honest) than to pretend to know, and to then recommend stuff in which some of these guidelines writers have vested interests.
Thanks for a thoughtful analysis.
Your questions bubble up in a stew of uncertainty. The frequent absence of compelling evidence for efficacy and risk of treatment, imperfect screening tests with false positives and false negatives, Bayesian issues making comprehension of the applicability of screening horribly opaque, widely variable lethality and morbidity of diseases that are screened for, the essential subjectiveness of determining the value of finding a particular problem before it manifests, the variable and uncertain value from any perspective of early discovery depending on the disease, the use of relative statistics to present results. Stand back just a little and it’s a horrid mess.
The practicing physician advising patients and the advisory organizations making recommendations can be most helpful by providing specific findings, in Bayesian context, using absolute not relative statistical comparisons (including derivations like number-needed-to-screen), of the best available studies, along with an honest assessment of the shortcomings of the data. As long as support for the screening test (or prevention therapy) is anything short of widely compelling nothing less is ethical.
From any patient’s point of view, undergoing testing (or taking treatment for prevention) in order to reduce his risk to the small degree that many screening tests do may not be worth it, and that should be his call after being thoroughly informed.
I am so excited about the website collection of screening articles. Thank you!
I guess it's a good thing that Sensible Medicine and the internet didn't exist in Dr. Papanicolau's day - the Pap test clearly wouldn't have made it out of the gate . All those pelvics and smears every year on every woman most of them normal , but a boat load of false positives - and we didn't even know what caused cervical cancer at that time. The nice thing about science is that screening can be refined over time - as it has with cervical cancer. And lots of women alive because of a screening program that in retrospect would have been described as flawed and overreaching .And the whole idea behind the screening was that this was a terrible disease and deserved our attention . But if you never screen, then there is no refinement , just a shoulder shrug " well nothing we can do" . Don't get me wrong , I am not a fan of a lot of " routine " testing - but I think we should avoid adopting a nihilistic philosophy about it as well -
Sorry, wrong question, Ken. No clinician, especially ones in primary care, should be practising or collaborating with ‘population medicine’. Calling a test a ‘screening’ test is a deliberately subtle attempt to divert attention from, and ideally bypass, the ethical/legal requirement for the individual to give their informed consent to a test/treatment, having decided to do so on the basis of unbiased evidence/information about the prognostic pathway supplied by their doctor. Any recommendation is necessarily preference-sensitive and based on somebody else’s preferences - usually either the clinician’s or those of some guideline panel of content experts - and therefore necessarily undermines this process. Whether or not the person decides to have a test will be determined by their informed preferences, especially time preference and degree of risk aversion - and often relational concerns. On the basis of the same evidence lots may choose to do (e.g.) a PSA test, lots may not. The evidence is never enough, though essential.
Sensible take on screening. I think the NNS should always be part of the decision process. Patients should be aware of how many people need to be screened to save one life. Where quality of life matters the NNS to prevent one cancer specific death may also be important.