Fwd: Baysesian vs. Fisherian
Rakesh Biswas
From: Huw Llewelyn [hul2] <00006060f673088...@jiscmail.ac.uk>
Date: Thu, Oct 28, 2021, 1:47 AM
Subject: Re: Baysesian vs. Fisherian
To: <EVIDENCE-B...@jiscmail.ac.uk>
Dear Donald
I’m very flattered that you should ask me this! David Spiegelhalter (a national treasure in the U.K. for explaining the statistics of Covid pandemic on radio and TV) wrote in his recent book, The Art of Statistics: "I must now make an admission on behalf of the statistical community. The formal basis for learning from data is a bit of a mess."
The controversies about the various approaches have developed a life of their own and careers are being built on them. I wonder whether there is little wish for the matter to be settled!
Personally I think that the missing link is the principle that all the possible true outcomes of valid random selection after an infinite number of sample must be regarded as equally probable prior to the beginning of sampling (eg 1/6 before throwing a die). We can then deduce from this assumption that the probability of an average or a proportion after an infinite number of observations being less extreme than the null hypothesis conditional on the data observed so far is 1-P (one sided).
Bayesian argue that you should guess an outcome first (eg 9/12) and when the study result of 21/37 comes in, update the prior belief (eg (9+21)/(12+37) to give a posterior proportion of 30/49. I discuss this in detail in section 3.8: ‘Combining probability distributions using Bayes’ rule’ in my PLoS One paper: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0212302.
The Neyman Pearson approach seems to assume that there are only two possible outcomes (based on intervention and control) of unknown prior probabilities to consider and then it estimates likelihood ratios for these distributions.
The difference between these control and intervention distributions can be used to calculate the power of a test (eg 80%) to detect such a difference during a future study with a P value of up to 0.025 one sided. This is used in current NHST. It is then the P value that is quoted rather than a likelihood ratio.
Many criticise this as a ‘mishmash’ of two methods. However the P value allows us to estimate the probability of a difference greater than the null hypothesis (eg zero) after continuing the study until we have an infinite number of observations. We can also estimate the probability of getting a P value up to 0.025 by repeating the study with the same number of observations. It will be much lower than for an infinite number of observations (see the comments for my PLoS paper.
I must emphasise that the principle of equal prior probabilities only applies to random sampling. In clinical practice, the prior probabilities of clinical outcomes are rarely the same.
Best wishes
Huw
On 27 Oct 2021, at 10:45, dsta...@tidewater.net wrote:
[RHYBUDD! E-BOST ALLANOL / CAUTION! EXTERNAL E-MAIL]
Dear Huw,
With your training and experience are you convinced of Fisherian over Neyman-Pearson still to be the preferred methodology?
Frequentist statistics has so many tests available, but can you name one overarching principle, one integrative principle, one synthesizing methodology, perhaps with the exception, of huge data sets, that Fisherian stats possesses?
Appeals to clinical acumen seem so much more convincing using inverse probability.
Regards,
Donald
I read and appreciated your editorial ( now known as blog!)
Get Outlook for iOS
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK> on behalf of Huw Llewelyn [hul2] <00006060f673088...@JISCMAIL.AC.UK>
Sent: Thursday, October 21, 2021 01:15
To: EVIDENCE-B...@JISCMAIL.AC.UK
Subject: Re: What does EBP look like
Hi Rajesh
Thank you. You, David and everyone might be interested in this Royal College of Physicians blog that I was invited to write in 2018: https://www.rcplondon.ac.uk/news/ai-and-healthcare-gathering-and-interpreting-data-future.
Huw
On 21 Oct 2021, at 04:04, Rakesh Biswas <rakesh...@gmail.com> wrote:
[RHYBUDD! E-BOST ALLANOL / CAUTION! EXTERNAL E-MAIL]
Excellent insights Huw. I'm a biased fan of your work.
Now if only the lifetime clinical case notes with categorical and numerical particular patient evidence generated by doctors like you globally were available through easy online search engines, in a deidentified case report format, just imagine what a great help it would be for any similar patient's caregiver to figure out various similar patient trajectories and outcomes and be able to predict with better accuracy close to precision?
Even after decades of this seminal paper on Case based reasoning (more here : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC544898/), there appears to have been little headway in this direction?
best,
rb
On Wed, Oct 20, 2021, 5:22 AM Huw Llewelyn [hul2] <00006060f673088...@jiscmail.ac.uk> wrote:
Hi again David, Ben, Donald, Michael and everyone
My personal understanding of what I have been doing in clinical practice over the years is that I was trying to support and supplement the patient’s homeostatic (eg social interactions, cardiac output or blood glucose control) and reparative feedback processes (eg overcoming infection or neoplastic processes)
The data inputs were both categorical and numerical, the diagnostic and decision making process created the supplementary feedback loop resulting in a decision and action that involved a choice of intervention with degrees of urgency or drug dose selection etc. This action resulted in an outcome that was further evidence (categorical or numerical) that often changed or modified the diagnosis, resulting in another decision and so on.
RCT evidence usually provides information for the first cycle in ‘fresh’ patients but not so often for second line and subsequent attempts after previous treatments have failed. Also a failure to take into account heterogeneous treatment effect (especially the effect of disease severity or age) undermines the process. Kent et al make this point (see https://www.acpjournals.org/doi/full/10.7326/M18-3667)
It has been my practice to specify the particular evidence (categorical and numerical) for that particular patient that I had used for each diagnosis and decision for others to understand and if necessary to agree or change in the light of new evidence. I provided a continuously updated summary as I went along. This is what I understood and taught (eg in the Oxford Handbook of Clinical Diagnosis) to be clinical evidence based medicine (c-EBM) . I would really like to see c-EBM combined with traditional research evidence based medicine (r-EBM 1) to become EBM 2.0 (or EIHC 2.0 as you suggest in your blog David).
Best
Huw
On 19 Oct 2021, at 20:49, Donald E. Stanley <STA...@mmc.org> wrote:
[RHYBUDD! E-BOST ALLANOL / CAUTION! EXTERNAL E-MAIL]
Some, perhaps, confusion on categorical and gradational?
Jury decision is categorical.
Evidence is gradational.
Get Outlook for iOS<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Faka.ms%2Fo0ukef&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=CpDooLCbQQ1sM3okkqGrr643hCzAao4gQVOeKcgzdMU%3D&reserved=0>
________________________________
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK> on behalf of David Nunan <david...@PHC.OX.AC.UK>
Sent: Tuesday, October 19, 2021 11:50:34 AM
To: EVIDENCE-B...@JISCMAIL.AC.UK <EVIDENCE-B...@JISCMAIL.AC.UK>
Subject: Re: What does EBP look like
CAUTION - EXTERNAL EMAIL
Dear Colleagues,
Thank you all for your replies to my original post.
I’ve written my blog on EBM 2.0, which was richly informed by the discussions here. I welcome your comments.
As part of world EBHC day tomorrow (20th), I will be leading a Twitter chat (between 10-11 and 20-21 BST) about the points raised in my blog if anyone is interested/has the time: https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Ftwitter.com%2Fdnunan79%2Fstatus%2F1450487541258276868%3Fs%3D20&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=29sX8MKRgwQi4gTScfVgToNnlWT2GpsqKA04GCkdbcM%3D&reserved=0<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Ftwitter.com%2Fdnunan79%2Fstatus%2F1450487541258276868%3Fs%3D20__%3B!!Npd4GBrkbw!nbRzqTL8-rol_gKxwEXqL93wzSv66qBI0J9YVJ9niO3pHvnmLi-Nw4sSfQ0%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=STCzNqD3CFEX8IbazrA1mf7pjSJt4%2BIeze%2BqirqNHo0%3D&reserved=0>
David
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK> on behalf of Donald E. Stanley <STA...@MMC.ORG>
Date: Tuesday, 19 October 2021 at 16:07
To: EVIDENCE-B...@JISCMAIL.AC.UK <EVIDENCE-B...@JISCMAIL.AC.UK>
Subject: Re: What does Disease and treatment entail?
Of course. You are on the mark, Ben.
Metaphor: suppose we display three bullseye targets,one superimposed on another. The targets consist of 1. Patient, 2. Disease, 3 treatment.
We shoot our arrows and hit the bullseye in the exact center of the bullseye time after time. Impossible!That is the misnomer called precision medicine.
The best we can aim for is the bullseye, not the same hole as earlier, but clustered close to the center of the bullseye.
It is an symptomatic curve that we cannot straighten.
Donald
Get Outlook for iOS<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Faka.ms%2Fo0ukef&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=CpDooLCbQQ1sM3okkqGrr643hCzAao4gQVOeKcgzdMU%3D&reserved=0<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Faka.ms%2Fo0ukef__%3B!!Npd4GBrkbw!nbRzqTL8-rol_gKxwEXqL93wzSv66qBI0J9YVJ9niO3pHvnmLi-N9yHukLY%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=P64IDaIfTwDjWWTnkMrTIUt5yfe1TTCKqO2tZ6%2FZs6w%3D&reserved=0>>
________________________________
From: Benjamin Djulbegovic MD <bdjulb...@coh.org>
Sent: Tuesday, October 19, 2021 10:54:30 AM
To: Donald E. Stanley <STA...@MMC.ORG>; EVIDENCE-B...@JISCMAIL.AC.UK <EVIDENCE-B...@JISCMAIL.AC.UK>
Subject: Re: What does Disease and treatment entail?
CAUTION - EXTERNAL EMAIL
Donald,
Putting Gettier’s problem aside, two quick comments:
1. a decision to administer treatment is always categorical (you decide Rx or not, while evidence about its effects indeed presents on the continuum )
2. Your AML example is right on the mark and requires more discussion. It reflect increasing demand to practice “precision” (individualized) medicine to the point that we continue to slice categories into subcategories and inevitably to single individuals. One one hand, we are all unique individuals with distinct experiences, settings and genome. On other hand, can you really ever draw any accurate conclusion from the observations in one individual? To infer causality, we need some sort of (sub)group, aggregate observations (which can include, repeated aggregate observations in single individuals) - and this remain a paradox of medicine: we need the group observations to manage individual patients.
Ben
Get Outlook for iOS<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Faka.ms%2Fo0ukef__%3B!!Npd4GBrkbw!gEdomCjJqUQpcFoEoMY-aMKg4791onXZilXQ2YMdcZXqONIZnYi3qTLZO5o%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=PWkXGXezqn7Z%2FrhW4OPmHG52tJcPAMVRo9L6eenOXGI%3D&reserved=0<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2Faka.ms%2Fo0ukef__%3B!!Npd4GBrkbw!gEdomCjJqUQpcFoEoMY-aMKg4791onXZilXQ2YMdcZXqONIZnYi3qTLZO5o%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=MxdqDraIj34fpu4zIZDE7rACL8jUCiOF63hUbLxe0Zs%3D&reserved=0>>
________________________________
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK> on behalf of Donald E. Stanley <STA...@MMC.ORG>
Sent: Monday, October 18, 2021 11:17 AM
To: EVIDENCE-B...@JISCMAIL.AC.UK
Subject: Re: What does Disease and treatment entail?
Dear Huw,
Yes numerical.
You seem to know much outside medicine so I refer you to the similar problem in epistemology ie. Gettier problem.
Too difficult, in a nutshell to explain subjunctive conditionals: those beliefs arrived at by our knowledge of the disease as currently defined: the category. But if e.g. AML is understood to be one condition, most therapies using categorical recommendation are marginally effective. However if we understood AML on the genomic level, a shift in belief, we would choose alternatives for treatment. Our diagnosis is conditional on understanding the causal nature of disease and this changes the category. The definition of treatment is not categorical but better thought of as a plan of using therapies following results from RCTs. None result in categorical recommendations, only suggestions based on the subjunctive conditionals of the disease categories. If you believe p-type AML use t- type therapy for p-typeAML.
The paradox of belief is the same in medicine.
Medical diagnosis is categorical, but treatment is gradational. And we confuse the two.
Get Outlook for iOS<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Faka.ms%2Fo0ukef__%3B!!Fou38LsQmgU!6nRRzDOCCv7Iruf05cQQjq7AukM2YpqXaiFsg5sI7DWFxZLqST2LSj-wmL73COYjzQ%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=Vk88Orkzch4qMHSkIfl2FlDZgMiPtrVKLzk%2FFcj947I%3D&reserved=0 >
________________________________
From: Huw Llewelyn [hul2] <hu...@aber.ac.uk>
Sent: Monday, October 18, 2021 10:31:28 AM
To: Donald E. Stanley <STA...@mmc.org>
Cc: EVIDENCE-B...@jiscmail.ac.uk <EVIDENCE-B...@jiscmail.ac.uk>
Subject: Re: What does EBP look like?
CAUTION - EXTERNAL EMAIL
A particular patient’s evidence is of course a mixture of categorical (eg symptoms, signs, genetic markers as in AML) and continuous variables (eg albumin excretion rate, a result that I often use to illustrate concepts because I have extensive data).
A diagnostic criterion eg for ‘Albuminuria’ is usually categorical - based on being above 2 standard deviations of the log albumin excretion rate. However the particular patient’s result will be numerical, it’s degree of severity indicating the risk of nephropathy and absolute risk reduction on treatment. Having said that symptoms vary in severity, duration and speed of onset, which makes them numerical evidence (I’m assuming that by ‘gradational’ you mean ‘numerical’).
Regards
Huw
On 18 Oct 2021, at 13:30, Donald E. Stanley <STA...@mmc.org> wrote:
[RHYBUDD! E-BOST ALLANOL / CAUTION! EXTERNAL E-MAIL]
Last post ought to have offered an example of categorical vs. gradational Evidence.
A decade ago AML was categorized as one disease; now it is categorized by heterogeneous clonal rearrangements hence the evidence for Rex. Is graded according to the category ofAML.
That is why evidence got treatment changes as categories, defined by genome analysis, change.
________________________________
From: Donald E. Stanley <STA...@mmc.org>
Sent: Thursday, October 7, 2021 1:56:51 PM
To: EVIDENCE-B...@JISCMAIL.AC.UK <EVIDENCE-B...@JISCMAIL.AC.UK>; Benjamin Djulbegovic MD <bdjulb...@COH.ORG>
Subject: Re: What does EBP look like?
Sorry Ben, but that is Aristotle.
Donald
Dr. Donald E. Stanley
Associates in Pathology
500 West Neck Road
Nobleboro, ME.04555
________________________________
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK> on behalf of Benjamin Djulbegovic MD <0000643a2dcbdb3...@JISCMAIL.AC.UK>
Sent: Thursday, October 7, 2021 1:50:03 PM
Subject: Re: What does EBP look like?
CAUTION - EXTERNAL EMAIL
Sorry, James
But, I am afraid that you are “ overly simplifying things”, which is not to say that we should not try to make things as simple as possible (but not simpler than it is required, to quote Einstein😊
Best
ben
From: McCormack, James <james.m...@ubc.ca>
Sent: Thursday, October 7, 2021 10:33 AM
To: Benjamin Djulbegovic MD <bdjulb...@coh.org>
Subject: Re: What does EBP look like?
Hi Ben - not exactly sure what you mean by the “issue is not an easy one”
For most non-life threatening conditions shouldn’t the approach just be that systematic reviews provide the best estimates (with acknowledged limitations) of benefits and harms and present them in a way that clinicians can use to work with patients to make decisions. Obviously some patients just want to be told what to do - and that is fine as long as the clinician knows the best available evidence.
Or maybe I’m overly simplifying things.
James
On Oct 7, 2021, at 10:11 AM, Benjamin Djulbegovic MD <bdjulb...@coh.org<mailto:bdjulb...@coh.org%3E%3E wrote:
[CAUTION: Non-UBC Email]
Thanks, James
The problem, of course, is that the issue is not easy one. There are hundreds of decision theoretical approaches (and the one that you and I discussed re discussion vs treatment threshold is only one tiny subsets of potential approaches). But, we need to tackle this issue more forcibly (and formally) than it has been done so far…This includes teaching that perfect decision is theoretically not possible…
Best
ben
From: McCormack, James <james.m...@ubc.ca<mailto:james.m...@ubc.ca%3E%3E
Sent: Thursday, October 7, 2021 9:16 AM
To: Benjamin Djulbegovic MD <bdjulb...@coh.org<mailto:bdjulb...@coh.org%3E%3E
Cc: EVIDENCE-B...@jiscmail.ac.uk<mailto:EVIDENCE-B...@jiscmail.ac.uk%3E
Subject: Re: What does EBP look like?
Hi Ben: Totally agree that it is the thought process (dare I say the agenda) around decision making that is what needs to be practiced and learned.
1) I think in general the process around doing solid systematic reviews and meta-analyses is really quite good and fairly reproducible - the work everybody has done around this over the last 25 or so years should be applauded.
2) The problem in my honest opinion is when guidelines that use or develop these systematic reviews are told (by GRADE and others) that recommendations need to be dichotomized into something like a “strong” or a “weak” (sometimes called “conditional”) category. Right away, this is problematic due to the subjective nature of the terms “strong” and “weak” - all you have to do is look at the words used to define strong and weak to see the problem amplified. Words like “most" and “many” are used as definitions and these indefinite numeral adjectives simply add to the confusion.
3) Personally I think recommendations described like this get in the way of true shared-decision making - a practitioner may feel the need to convince a patient to use a “strong” recommendation etc
4) Why not simply say this is the best available evidence, here is a ballpark idea of the magnitude of the benefit and harm, here are some clinician and patient decision aids to help with the discussion and what ever decision is made is the correct one. Discussion thresholds NOT treatment thresholds.
5) I realize there are alway exceptions to this approach, however as a general approach if we taught and guided clinicians to do this we would truly be doing EBP 2.0
Thoughts?
James
On Oct 7, 2021, at 8:42 AM, Benjamin Djulbegovic MD <0000643a2dcbdb3...@JISCMAIL.AC.UK<mailto:0000643a2dcbdb3...@JISCMAIL.AC.UK%3E%3E wrote:
[CAUTION: Non-UBC Email]
After getting several private e-mails related to my post below, it appears that I did not explain the main point I was trying to make. So, let me clarify.
People care about decisions ( “ I just need to know what should I do”), and not so much about evidence (and other ingredients that may affect decisions). (Quality, certainty) of evidence is relevant in so far it affects decisions. Both first principle of EBM (“not all evidence is created equal”) and Cochrane collaboration refer to the importance of evidence for decision-making. In fact, the Cochrane logo reads as ”Trusted evidence. Informed decisions. Better health.” But, the Cochrane (and EBM for most part) is all about getting evidence right, with no (or, at best, rudimentary) attempt to teach people how to make better decisions. So, teaching how to make better decisions is what is missing in the current EBM curricula. (This would also include answering my questions in the postscripts of my original message below).
Hope I clarified the point I was trying to make
Thanks
ben
From: Benjamin Djulbegovic MD <bdjulb...@coh.org<mailto:bdjulb...@coh.org%3E%3E
Sent: Thursday, October 7, 2021 5:50 AM
To: Dr Amy Price <heali...@MSN.COM<mailto:heali...@MSN.COM%3E%3E; EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E
Subject: Re: What does EBP look like?
Many of you are probably aware of the Lancet piece where Gordon Guyatt and I reviewed the progress in EBM during last quarter of century (https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Feur02.safelinks.protection.outlook.com%2F%3Furl%3Dhttps*3A*2F*2Fpubmed.ncbi.nlm.nih.gov*2F28215660*2F%26data%3D04*7C01*7C*7C1cfc03f1724449bb228e08d992330e41*7Cd47b090e3f5a4ca084d09f89d269f175*7C0*7C1*7C637701570253349393*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000%26sdata%3DVw0QMUgywtkaGR1pn*2BxLt2GnGNYO75KQQ4G6zucs5BI*3D%26reserved%3D0__%3BJSUlJSUlJSUlJSUlJSUlJSUl!!Npd4GBrkbw!iarC9bbJSx2GWPsthO5B50n8zqYES6f1bkyVG6i-dZFqeYd6H7R-lEylCNU%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=TqIFyhxy4Y3C0K0%2BbhTgAgYW1BiSm7ALOzW3qEWIJZE%3D&reserved=0 <https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Feur02.safelinks.protection.outlook.com%2F%3Furl%3Dhttps*3A*2F*2Furldefense.com*2Fv3*2F__https*3A*2F*2Fpubmed.ncbi.nlm.nih.gov*2F28215660*2F__*3B!!Npd4GBrkbw!lay9jOd_pHW5mNUFn3BLNcic4ptPm2PcBoXHzVfossVoBY_qoroHY5MupsY*24%26data%3D04*7C01*7C*7C1cfc03f1724449bb228e08d992330e41*7Cd47b090e3f5a4ca084d09f89d269f175*7C0*7C1*7C637701570253349393*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000%26sdata%3DbliHxbW2BB6ZLezgEBYD8NsnJbm8CFDU4o6b*2Frnv7Qk*3D%26reserved%3D0__%3BJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJQ!!Npd4GBrkbw!iarC9bbJSx2GWPsthO5B50n8zqYES6f1bkyVG6i-dZFqeYd6H7R-Qu_dItQ%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799780169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=8yxqMBjnTvaXFWR9svty8uFvY7gX474ck0JBC8lcxnc%3D&reserved=0 > )
We have identified some key paths forward, most important of which is the need for developments of the methods and techniques for better integration of EBM with (theories) of decision-making. Currently, the process of decision-making (within EBM) largely remains a black-box operation. Another issue is our poor understanding of the value of the content expertise (in relation to methodological expertise). When should we listen to the content vs methodological experts and we should ignore their advices , remains an unresolved issue (and perhaps, will eternally remain so).
EBM has made huge advances in methods of research synthesis, but not so much at the level integration of evidence in decision-making.
Ben
Ps two practical questions that may identify some additional areas for research:
A) assessment of relevance is usually important to formulate a question of interest and application of evidence. However, how important is content expertise for literature search and critical appraisal ?
B) how often physicians refuse to give treatment because it is supported by very low quality evidence ? (this is common at policy levels, but it seems to be extremely rare in practice of medicine; why is it so?)
________________________________
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E%3E on behalf of Dr Amy Price <heali...@MSN.COM<mailto:heali...@MSN.COM%3E%3E
Sent: Thursday, October 7, 2021 8:16 AM
To: EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E
Subject: Re: What does EBP look like?
[Attention: This email came from an external source. Do not open attachments or click on links from unknown senders or unexpected emails.]
----------------------------------------------------------------------
For me, the value of EBP was learning to think critically from the lens of evidence and methods and to put these principles into practice in a less than ideal world. It is what we make it, instead of searching for a role model, let us become mentors and role models, not so people can copy us, but so they can think and apply the knowledge in communities of practice and change their worlds. That is the whole concept of practice. Could relevant inferences be extended to clinicians, researchers and policy makers, otherwise we risk silos with gaps in inclusion, equity and practice.
Dr Stanley life and death choices are decided through our value judgements, when they are combined with the best available current evidence on methods and practice we have progress. When people make the decision to act on their value judgement and use evidence and practice to change medicine we have innovation and best available implementation. A foundation of values can open the door for trust and change, positive and negative.
Best
Amy
On 10/7/21, 7:35 AM, "Evidence based health (EBH) on behalf of Donald E. Stanley" <EVIDENCE-B...@JISCMAIL.AC.UK on behalf of STA...@MMC.ORG<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%20on%20behalf%20of%20ST...@MMC.ORG>> wrote:
These suggestions are good and relevant, but they are value judgments!
Donald E. Stanley FCAP
500 west Neck Road
Nobleboro, ME
Sta...@mmc.org<mailto:Sta...@mmc.org>
dsta...@tidewater.net<mailto:dsta...@tidewater.net%3E
Get Outlook for iOS<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Feur02.safelinks.protection.outlook.com%2F%3Furl%3Dhttps*3A*2F*2Furldefense.com*2Fv3*2F__https*3A*2F*2Faka.ms*2Fo0ukef__*3B!!Fou38LsQmgU!6Q_B9x0JyqtmLm4tk_IxPJIWhW7xccKWiYiyXlLeoN5z9rw3DMLUI2beGJlZuLzynQ*24%26data%3D04*7C01*7C*7C1cfc03f1724449bb228e08d992330e41*7Cd47b090e3f5a4ca084d09f89d269f175*7C0*7C1*7C637701570253349393*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000%26sdata%3D*2FWH4wiCFpe30DGUseLpwemjsmQrK5YA*2FK5o*2F830kWRg*3D%26reserved%3D0__%3BJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSU!!Npd4GBrkbw!iarC9bbJSx2GWPsthO5B50n8zqYES6f1bkyVG6i-dZFqeYd6H7R-UgfF3Lg%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799790124%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=WaBGZIYnjuv8mL5ZNKsOKYIB4GaSdyXkJgG9GQmMBhk%3D&reserved=0 <https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Feur02.safelinks.protection.outlook.com%2F%3Furl%3Dhttps*3A*2F*2Furldefense.com*2Fv3*2F__https*3A*2Faka.ms*2Fo0ukef__*3B!!Fou38LsQmgU!6Q_B9x0JyqtmLm4tk_IxPJIWhW7xccKWiYiyXlLeoN5z9rw3DMLUI2beGJlZuLzynQ*24*2520%26data%3D04*7C01*7C*7C1cfc03f1724449bb228e08d992330e41*7Cd47b090e3f5a4ca084d09f89d269f175*7C0*7C1*7C637701570253349393*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000%26sdata%3D38AteH17yJhnAQ7*2BJrNCJshtGAn2D*2Bwr3e3Njeojglg*3D%26reserved%3D0__%3BJSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUlJQ!!Npd4GBrkbw!iarC9bbJSx2GWPsthO5B50n8zqYES6f1bkyVG6i-dZFqeYd6H7R-YuBKF6w%24&data=04%7C01%7C%7C8228cfa6611041a5db4c08d993399465%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637702697799790124%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=01haAE0OHqLfwDkFM6aJWrX3%2BF9cCL%2F4jSlBaiqIwoE%3D&reserved=0 > >
________________________________
From: Evidence based health (EBH) <EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E%3E on behalf of Michael Power <hmicha...@GMAIL.COM<mailto:hmicha...@GMAIL.COM%3E%3E
Sent: Thursday, October 7, 2021 7:25:17 AM
To: EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E <EVIDENCE-B...@JISCMAIL.AC.UK<mailto:EVIDENCE-B...@JISCMAIL.AC.UK%3E%3E
Subject: Re: What does EBP look like?
CAUTION - EXTERNAL EMAIL