RapidMiner in Clinical Medicine: Market Needs
DocMusher
New Altair Community Member
Hi,
I have been thinking quite some time on the position of RapidMiner studio and Server in clinical medicine. Besides data-mining pur-sang as in genetics or genomics, the market needs for RapidMiner in clinical medicine are to be found in the fact that despite the accelerating pace of scientific discovery, the current clinical research enterprise does not sufficiently address pressing clinical questions. Given the constraints on clinical trials, for the majority of clinical questions the only relevant data available to aid in decision making are observational.
Is this a big issue? yes it is!! Even as physician, I am not in the ability to define which part of my clinical decision is based on randomized clinical trials, which part is based on experience and which part is "Fingerspitzengefühl" (instinctive feeling, gut feeling). This is not a reflection of personal lack of expertise, even in the well-studied field of cardiology, only 19 percent of published guidelines are based on randomized controlled trials (Tricoci P, Allen JM, Kramer JM, Califf RM, Smith SC Jr. Scientific evidence underlying the ACC/AHA clinical practice guidelines. JAMA. 2009;301(8):831–41) in an industrial sector, where health expenditure, total (% of GDP) in western civilization in between >10-17%!
This frustrates me: Imagine You are admitted to a critical care department, you are critically ill, your future depends on the best care available. I have to treat your critical condition without the use of all available knowledge and technology known to mankind. Recent technology allows the use of point of care tools such as the “patients like mine” button enabling the quantification and prioritization of unanswered clinical questions for EHR-enabled randomization supporting patient care decisions in the absence of published evidence. (Longhurst CA, Harrington RA, Shah NH. A “Green Button” For Using Aggregate Patient Data At The Point Of Care. Health Aff [Internet]. 2014;33(7):1229–35. Available from: http://content.healthaffairs.org/cgi/doi/10.1377/hlthaff.2014.0099)
That is where RapidMiner can make a difference based on the fact that databases (Patient Data Management Systems) become standard in all hospitals. Looking forward to see new webapps and dashboards being developed in the future in this sector.
To finish, a nice link I found (http://www.ohdsi.org/demos/)
Cheers Sven
I have been thinking quite some time on the position of RapidMiner studio and Server in clinical medicine. Besides data-mining pur-sang as in genetics or genomics, the market needs for RapidMiner in clinical medicine are to be found in the fact that despite the accelerating pace of scientific discovery, the current clinical research enterprise does not sufficiently address pressing clinical questions. Given the constraints on clinical trials, for the majority of clinical questions the only relevant data available to aid in decision making are observational.
Is this a big issue? yes it is!! Even as physician, I am not in the ability to define which part of my clinical decision is based on randomized clinical trials, which part is based on experience and which part is "Fingerspitzengefühl" (instinctive feeling, gut feeling). This is not a reflection of personal lack of expertise, even in the well-studied field of cardiology, only 19 percent of published guidelines are based on randomized controlled trials (Tricoci P, Allen JM, Kramer JM, Califf RM, Smith SC Jr. Scientific evidence underlying the ACC/AHA clinical practice guidelines. JAMA. 2009;301(8):831–41) in an industrial sector, where health expenditure, total (% of GDP) in western civilization in between >10-17%!
This frustrates me: Imagine You are admitted to a critical care department, you are critically ill, your future depends on the best care available. I have to treat your critical condition without the use of all available knowledge and technology known to mankind. Recent technology allows the use of point of care tools such as the “patients like mine” button enabling the quantification and prioritization of unanswered clinical questions for EHR-enabled randomization supporting patient care decisions in the absence of published evidence. (Longhurst CA, Harrington RA, Shah NH. A “Green Button” For Using Aggregate Patient Data At The Point Of Care. Health Aff [Internet]. 2014;33(7):1229–35. Available from: http://content.healthaffairs.org/cgi/doi/10.1377/hlthaff.2014.0099)
That is where RapidMiner can make a difference based on the fact that databases (Patient Data Management Systems) become standard in all hospitals. Looking forward to see new webapps and dashboards being developed in the future in this sector.
To finish, a nice link I found (http://www.ohdsi.org/demos/)
Cheers Sven
Tagged:
0
Answers
-
Hi Sven,
it is good to hear this form an insider. When it comes down to medical science it often feels like it is purely done from a statistical point of view.
Definitely something to think about. Maybe actual science and "treatment of patients" need to be separated more?
~Martin0 -
Hi there,
IBM's a big player in this area, checkout http://www-05.ibm.com/innovation/uk/watson/watson_in_healthcare.shtml , and they have teamed up with health insurers, so they aren't going to get smaller soon. The grammatical knowledge basis for Watson may well be misplaced, but it does have a nice natural language interface as a consequence. Sven, I remember your interest in images, check out http://www-03.ibm.com/press/us/en/pressrelease/47435.wss !
Perhaps not with Big Blue, but there's a convergent singularity emerging, OK, rather slowly.
H0 -
Hi All,
i read a nice article on watson yesterday: See https://t.co/WYIFicPfzS
Definitly worth a read.
~Martin0 -
Hi Martin,
Good article, one of the first to cast doubt; I think the last para gets it right..
I did some research which shows that Watson was really good at the buzzer in the Jeopardy challenges, less biased. Here's the address.the brand is being used to create a halo effect for a set of technologies that are not as revolutionary as claimed.
http://meme-machines.com/watson/
Good call on the article though, thanks.
H0 -
Sven,
There's going to be a PPT online at some point from the South West Kidney Institute who mixed RapidMiner & Tableau. If anyone was at RM Wisdom perhaps they can fill us in on the details.
http://www.itbusinessnet.com/article/RapidMiner-Extends-Cutting-Edge-Predictive-Analytics-for-Tableau-42568650 -
Hi,
This would be useful! I also think you are focused on a very important part of the analytics process: the result of a process should be represented, visualized etc in the way, the reader (from novice to expert, or from the public to the CIO, CEO) is able to get the message. Speaking for the medical sub-community in RM: Qlik and tableau, please contact us, there is much work ahead.
Sven0
Altair Employee