Print

Print
I think the contribution of the NEJM paper is not that it demonstrated that observations from EMR enabled a correct decision ( the best EMR-based data mining can do is to generate hypotheses) but to illustrate in some kind of a systematic way what the breadth of experience exist in a particular institution about managing a clinical problem.
 I see many patients in second opinion for problems/diseases (X) where evidence is sketchy at best. Sometimes, the answer can be to enroll the patients in clinical trials, but often that it is not possible. I then tell these patents and referring docs, "we really don't know how to manage X, but this is what we do". And, often residents/fellows do the projects based on EMR-data mining to provide my institution's  collective experience about management of these problems ( it, however, typically take more than 4 hours to do that!)
In fact, since Internet democratized accessed to scientific information, it is rather rare to have requests for the second opinions for diseases/problems where guidelines/randomized evidence exists. What EMR will enable is to obtain collective experience (familiarity) regarding the problems that are yet to be studied in scientific, rigorous, systemic fashion. (And, the numbers of these problems/analyses will be infinite, as argued in one of our recent papers).
Ben Djulbegovic 

Sent from my iPad

On Nov 4, 2011, at 6:36 AM, "Rakesh Biswas" <[log in to unmask]> wrote:

EMRs/Case Reports contain contextual information about an 'individual' that can have a psychological appeal for most clinicians as we often tend to think in terms of clinical pattern matching/recognition ( as in Amy's ...fingernails were bluish, he could barely stand, was dizzy etc or even ECG patterns for that matter?).

On the other hand current evidence based reviews contain information about a 'population' that is difficult to match with the patient at hand (due to the uniqueness of each individual case) and consequently at times may appear psychologically less appealing to clinicians?

Either way if there were a randomized controlled trial between a combined 'individual and population based information' (as in the evolving EMR model) and just 'population' based information ( as in the current model) to study their effect on clinical decision making in terms of patient related health care outcomes it is more likely that the combined approach would fare better?

:-)

regards,

rakesh

From: Dr. Amy Price <[log in to unmask]>
Date: Fri, Nov 4, 2011 at 9:10 AM
Subject: Re: Evidence-Based Medicine in the EMR Era
To: [log in to unmask]


Carlos,

 

I think Brian answered well. In the first instance the concept annoyed me due to an experience where I experienced physicians blindly accepting the printout on an ECG...as a minor finding of variations. Well the patients fingernails were bluish, he could barely stand, was dizzy etc so I sent him to the hospital, later at a surgeon friends home I went to his library and took out the cardiology book and matched the rhythms to the print out. When he asked what I was doing I told him and he said “Oh Amy, only the cardiologists bother with that, I am surprised you can read that”.  They were patterns with labels..a six year old could have gotten it. I apparently made a good call as the hospital took him and he had open heart surgery. This really kind of biased me against EMR as I figure if none of the chart helps existed they would have actually looked at the patient and all the evidence....you know a patient is more than a traffic light and may need help even if the diagnostic doesn’t show in bold or red...I think when there is too much automation perhaps people forget what they once knew...

 

Really, I only looked at the link because you sent it through and then I started thinking about data mining possibilities, I think the possibilities are impressive and I really appreciate your assistance today in opening a rather firmly closed mind!

 

Amy  

 

From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Brian Alper MD
Sent: 03 November 2011 02:06 PM


To: [log in to unmask]
Subject: Re: Evidence-Based Medicine in the EMR Era

 

We want the best available evidence at the time we are making a clinical decision.  Ideally we have the best available evidence in a clearly understandable form, including understanding the quality/certainty/limitations of that evidence.

 

When the best available evidence is observational data from EMR records it would be ideal to have that readily available (with an understanding of its limitations).

 

In our traditional approaches some clinicians and researchers will seek observational data like this and publish it in the form of cohort studies, case-control studies or case series.  This observational data may then be found by clinicians and may represent the best available evidence for a given concept.  It can also be found by researchers and may stimulate further research and understanding.

 

In the new model described (this “case report” of using observational data from an EMR for a local decision) the data is unpublished (except for what was shared in this special report) and is otherwise not available for others.  Likewise the clinicians in this example could not find the data that may be available in other EMR records at other institutions, and that data may be similar (corroborating) or different.  It also took the clinicians 4 hours to obtain and analyze the data which is not be feasible for many clinical questions during practice.

 

Perhaps we will see a future when observational data from EMRs can be obtained without compromising privacy or ownership concerns, compiled from multiple settings, represented in real time to fit specific information needs, and be interpretable with the understanding of its limitations as observational data, ideally presented in context with sharing of more reliable evidence when more reliable evidence is available.

 

Brian S. Alper, MD, MSPH

Editor-in-Chief, DynaMed (www.ebscohost.com/dynamed)

 

From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Dr. Carlos Cuello
Sent: Thursday, November 03, 2011 1:21 PM
To: [log in to unmask]
Subject: Re: Evidence-Based Medicine in the EMR Era

 

Hi all

 

A nice discussion erupted here with some colleagues in Mx

 

At a first glance I said "what a load of bias" but on second thought: It was all they got. It was either EMR or Expert Advise... 

 

They both have the possibility (risk) of bias and in the end they had to make the call.

 

On a GRADE profile this still would be very quality of evidence (we have very little confidence of the estimate based on these results), and then you would have to weigh the desirable vs the undesirable effects (risks, cost, etc.) to make a weak recommendation.

 

 

What do you think?

 

 

 

On Thu, Nov 3, 2011 at 10:54, Rakesh Biswas <[log in to unmask]> wrote:

The authors have summed it up interestingly...

"Did we make the correct decision for our patient? ...we may never really know.

“...in the light of experience as guided by intelligence...”

In the practice of medicine, one can't do better than that."

:-)

 

On Thu, Nov 3, 2011 at 10:04 PM, [log in to unmask] <[log in to unmask]> wrote:

Dear Carlos and Ernesto,

I am also intrigued with this field not only for traditional medical care but also for use in the realm of mental health. this over time could provide patterns that could lead to defining research instead of researchers beating the same dead horse without resurrection tactics from multiple angles it would also cross reference objectively behaviours and medication effects other than psychotropic agents although of course could include these

 

Amy

Amy Price 

Empower 2 Go 

Building Brain Potential

Sent from my iPad


On 3 Nov 2011, at 12:24, Ernesto Barrera <[log in to unmask]> wrote:

Hi Carlos,

We have millions and millions of data already stored in medical records. It seems reasonable to ask questions and find answers that will help us in the absence of evidence... and with evidences.

I think software could be used to establish causal relationships in an automated manner, and generate warning signs.

And generate causal associations and predictive models that, in future could replace, for example, risk tables constructed with another populations.

This has a clear practical application in pharmacovigilance, particularly given unusual adverse effects in phase III trials. So, medical records become converted into real data fields of observational studies

As I have previously commented , from my point of view, data mining can be the seventh model S of the pyramid of evidence (predictive Statistics).

Thank you,

Ernesto Barrera

Family Physician
Madrid (Spain)




El 03/11/2011 15:40, Dr. Carlos Cuello escribió:

Quite interesting on how electronic medical records can become a part of observational evidence and apply when there is no other evidence at hand.

 

Good read

 

http://www.nejm.org/doi/full/10.1056/NEJMp1108726

 

--
Carlos A. Cuello-García, MD
Centre for Evidence-Based Practice & Knowledge Translation

Tec de Monterrey School of Medicine

Cochrane Collaboration Iberoamerican branch

CITES piso 3. Morones Prieto 3000 pte. Col. Doctores 64710 

Monterrey, NL. Mexico. 

☎ +52.81.8888.2223 & 2154. Fax: +52.81.8888.2052 Skype: dr.carlos.cuello
www.cmbe.net ⚫ Twitter ⚫ Linkedin 

 

The content of this data transmission must not be considered an offer, proposal, understanding or agreement unless it is confirmed in a document signed by a legal representative of ITESM. The content of this data transmission is confidential and is intended to be delivered only to the addressees. Therefore, it shall not be distributed and/or disclosed through any means without the authorization of the original sender. If you are not the addressee, you are forbidden from using it, either totally or partially, for any purpose

 

 



 

--
Carlos A. Cuello-García, MD
Centre for Evidence-Based Practice & Knowledge Translation

Tec de Monterrey School of Medicine

Cochrane Collaboration Iberoamerican branch

CITES piso 3. Morones Prieto 3000 pte. Col. Doctores 64710 

Monterrey, NL. Mexico. 

☎ +52.81.8888.2223 & 2154. Fax: +52.81.8888.2052 Skype: dr.carlos.cuello
www.cmbe.net ⚫ Twitter ⚫ Linkedin 

 

The content of this data transmission must not be considered an offer, proposal, understanding or agreement unless it is confirmed in a document signed by a legal representative of ITESM. The content of this data transmission is confidential and is intended to be delivered only to the addressees. Therefore, it shall not be distributed and/or disclosed through any means without the authorization of the original sender. If you are not the addressee, you are forbidden from using it, either totally or partially, for any purpose