EHR & Artificial Intelligence Can Reduce Medical Errors - InformationWeek
IoT
IoT
Healthcare // Electronic Health Records
Commentary
11/11/2014
03:58 PM
Mansur Hasib
Mansur Hasib
Commentary
Connect Directly
Twitter
RSS
100%
0%

EHR & Artificial Intelligence Can Reduce Medical Errors

Electronic health records save lives by collecting patient data in one place. Artificial intelligence takes it a step further by calling on the expertise of multiple doctors.

While I had heard that almost 400,000 Americans die each year because of medical mistakes, in a recent article Forbes contributor Dan Munro underscored that volume when he asked readers to imagine the largest commercial aircraft -- an Airbus A-380 -- crashing every day for a year: The number of passengers who would perish aboard those imaginary crashes compares to the number of patients really dying annually in our hospitals due to blunders.

People who want nothing to change usually dispute the number of deaths. For the sake of argument, let us assume the actual number could be represented, then, by one crash every four days. Even then, surely it is worthwhile trying to figure out how to prevent these errors.

Certainly, procedural failures or pure accident causes some errors but incomplete or incorrect information about the patient is at the heart of a large percentage of these mistakes.

As Munro points out, a major problem is that the current healthcare industry is incentivized by revenue and profits -- not safety and quality. Therefore, as newly re-elected Florida Governor (and former healthcare CEO) Rick Scott said at a recent meeting to discuss cutting costs in healthcare, the industry has been unwilling to voluntarily reduce profits. Since safety and quality using current methods would be expensive and slash profits, perhaps electronic health records (EHRs) and health information technology (HIT) could  accomplish the  goals of all stakeholders.

EHRs can maintain patients' complete medical histories, along with all known allergies and medications. The record should travel with patients, no matter where they go for treatment. Doctors do not have to rely on the patient's fallible memory at every encounter. The record speaks for patients, even if patients are incapacitated for any reason.

[How is technology fighting ebola? Read IBM Brings Open Data Tech To Ebola Fight.]

We must recognize that doctors often face points of no return -- and patients get no second chances. Choosing the right medicine or treatment is frequently a game of probabilities. Choose the right medicine and the patient will live. Choose the wrong one and the patient will die. This is why even the most qualified doctors often seek second or third opinions before embarking on a risky treatment plan. Doctors have told me countless stories about their ability to save patients because a complete EHR was available. In these cases multiple doctors were able to view the same information at the same time, often while residing thousands of miles apart. They collaboratively agreed on the best option -- and saved the patient's life.

EHRs also facilitate artificial intelligence. A patient's medical history often is full of reams of data; manually winnowing through that information is a daunting task. Today, teams of top doctors help develop artificial intelligence systems that can quickly determine if a proposed medicine, food, or medical procedure will likely cause the patient greater harm than good. This will reduce a large number of medical mistakes.

There is no cause for concern. Decisions suggested by artificial intelligence systems developed by top-notch doctors likely are more accurate than decisions made solely by humans. Watch Vinod Khosla discuss this fascinating issue. All doctors are not created equal. As Khosla pointed out, studies show that if you give the same data on a patient to a random group of 10 doctors and ask them if surgery is recommended, half will choose surgery while the other half will choose not to perform surgery.

If artificial intelligence systems are built using the medical minds of the doctors that choose the right answers, these technological solutions sift through an incredible amount of data and provide more medically reliable recommendations. Of course, a human doctor still makes the ultimate decision. However, the doctor has the benefit of a large amount of data analysis and is much more likely to make a decision based on complete information, not incomplete data.

Perhaps EHRs plus AI will save many more lives and dramatically reduce medical errors without increasing costs too much.

The owners of electronic health records aren't necessarily the patients. How much control should they have? Get the new Who Owns Patient Data? issue of InformationWeek Healthcare today.

Dr. Mansur Hasib is the only cybersecurity professional in the world with 12 years' experience as CIO; a Doctor of Science (DSc) in Cybersecurity; CISSP (cybersecurity); PMP (project management), and CPHIMS (healthcare) certifications, who has written two books on the ... View Full Bio
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
Ariella
50%
50%
Ariella,
User Rank: Author
12/1/2014 | 9:36:00 AM
Re: First Step
@Alison Yes! I noticed the same thing when my son was in the hospital. I spent one night there and go maybe 3 hours of sleep because of the constant noise that the nurses seem to easily tune out. I do have an idea for a solution for that -- not to get them to respond if they choose to ignore the alert but getting it to be quieter for patients -- that I've included in a blog I wrote on adapting technology for a better patient experience. 
Alison_Diana
50%
50%
Alison_Diana,
User Rank: Author
12/1/2014 | 9:33:04 AM
Re: First Step
Alert fatigue is a HUGE deal. When we were at the hospital with a family member, we were overwhelmed by the buzzers and lights going off -- and that was just for ONE patient, in ONE room, and it did not include the EHR (although I admit, I peeked at the computer screen!). I cannot imagine how clinicians deal with the endless cacophony of lights, buzzers, bells, vibrations, and other devices trying to capture their attention -- most of which are valid and important. 
Ariella
50%
50%
Ariella,
User Rank: Author
11/19/2014 | 12:08:01 PM
Re: First Step
"
@Ariella: That could be solved by clever software designing. Software diagnosis and drug prescriptions can be made fairly accurate if there are numerous design steps involved. What you are talking about are human/software errors. While the human error cannot be minimized (because someone or the other would always click the wrong drug) it can be made limited using clever software  design that makes the doctor review his decisions."

It's not quite that simple. What you refer to does exist, but it leads to alert fatigue. If  at least half the time, one gets such an alert, one learns to ignore the alerts. That is a real and documented effect. That's why the system has to be much more sophisticated. It also has to look beyond the obvious errors to spot prescriptions that would have been given just due to similarity in name, say fertility drugs to a man, and such, which has occurred.
SachinEE
50%
50%
SachinEE,
User Rank: Ninja
11/19/2014 | 11:13:48 AM
Re: First Step
"On the number of deaths: As reported in "Medical Errors: a Report by the Staff of U.S. Senator Barbara Boxer"  there 210,000 to 440,000 deaths in America each year that result from "medical errors and other preventable harm at hospitals."  EHR does not really eliminate the danger. I interviewed a doctor behind the company that designed a better prescription alert system, for such errors are a big factor in those preventable deaths.  Dr. Gidi Stein explained, doctors can feel "drowned by the numbers and figures on each and every patient," and so may end up overlooking some small item that is, in fact, critical -- a blood test result, a particular condition, or a drug. The pull-down menu in EMR systems can be problematic, too. A doctor might inadvertently click on the wrong drug in an alphabetical list or select the patient in Room 9 instead of Room 10."

@Ariella: That could be solved by clever software designing. Software diagnosis and drug prescriptions can be made fairly accurate if there are numerous design steps involved. What you are talking about are human/software errors. While the human error cannot be minimized (because someone or the other would always click the wrong drug) it can be made limited using clever software  design that makes the doctor review his decisions.
SachinEE
50%
50%
SachinEE,
User Rank: Ninja
11/19/2014 | 11:10:42 AM
Re: Documentation side
"The world we live in today has people confusing the real world with virtual values which was what the article did and I don't mean that in a bad way at all but rather it only saw one side and not how the software works in the "real" world with people using it so again models are important and don't confused virtual world values with the real world as they need to work together and people need to know the difference an today with steroid marketing and flat out lies, it's a challenge."

@MedicalQuack: The reason why people want documentation with software supervision is because they want the best of both worlds where margin for error would be low. However in software diagnosis alone, the ER MD wouldn't have to have documentation because that is drastically accurate, and people fail to recognize this.
MedicalQuack
50%
50%
MedicalQuack,
User Rank: Moderator
11/12/2014 | 12:54:22 PM
Documentation side
Now we come to the part to where we talk about how decisin making software is actually used in the field and what other areas of a hospital or doctor's practices touches this...ok so let's talk models now as healthcare likes to avoid a lot of model talk but we should do more of it.

Hospitalist I know is responsible for the hand off from the ER room to admit patient for his group or make another decision.  You do have to remember it's a different group of doctors in the ER than the group the hospitalist is in, keep that in check here.  This is real and a hospitalist I know.  Patient is in the ER and has issues and they use decision making software.  I think it's a good thing but again look at the whole model which consumers don't even think about at all but the folks that use the software sure do in addition to caring for the patient.

ER room examines patient and uses software to ensure they didn't miss any possible diagnosis and remember they have contracts with the hospital and malpractice insurance to pay.  Ok so ER MD to cover everything picks and loads up every single diagnosis it could possibly be, again they want to be covered.  This ends up being a quite a bit list of things that could be wrong, and it's exaggerated too as nobody wants to miss that diagnosis so even if it is related, the ER MD checks it "on" the list for the hospitalist to review with the hand off.

Hospitalist gets the medical record information in the hand off and again this is a different group of MDs from the ER and says "what a mess" "now I have to go through and remove tons of these as it clogs up the chart and most of its is not relative as the ER has done a CYA, so before hospitalist begins the job, clean up the chart and document the removal of all the non needed remote diagnosis codes added by the ER doctors.


This is the way it works in the "real" world of medicine with documentation.  There needs to be documentation as such for legal cases so while the software is a good thing for look up, you can see how other elements enter the picture and they do all the time.  It's always good to look at both sides and see how your model works as like this above, you're in for some surprises and this why models are important. 

The world we live in today has people confusing the real world with virtual values which was what the article did and I don't mean that in a bad way at all but rather it only saw one side and not how the software works in the "real" world with people using it so again models are important and don't confused virtual world values with the real world as they need to work together and people need to know the difference an today with steroid marketing and flat out lies, it's a challenge.

http://ducknetweb.blogspot.fr/2014/03/virtual-worlds-real-world-we-have.html

Larry Ellison made a very good point about artificial intelligence a while back and said to be careful as you do get fooled, if you don't think about what's really real and what's virtual. 

http://ducknetweb.blogspot.fr/2014/02/larry-ellison-ceo-oracle-hcm.html

 
Ariella
50%
50%
Ariella,
User Rank: Author
11/12/2014 | 12:27:58 PM
Re: First Step
 On the number of deaths: As reported in "Medical Errors: a Report by the Staff of U.S. Senator Barbara Boxer"  there 210,000 to 440,000 deaths in America each year that result from "medical errors and other preventable harm at hospitals."  EHR does not really eliminate the danger. I interviewed a doctor behind the company that designed a better prescription alert system, for such errors are a big factor in those preventable deaths.  Dr. Gidi Stein explained, doctors can feel "drowned by the numbers and figures on each and every patient," and so may end up overlooking some small item that is, in fact, critical -- a blood test result, a particular condition, or a drug. The pull-down menu in EMR systems can be problematic, too. A doctor might inadvertently click on the wrong drug in an alphabetical list or select the patient in Room 9 instead of Room 10.
Ariella
50%
50%
Ariella,
User Rank: Author
11/12/2014 | 12:26:29 PM
Re: First Step
@Gary_EL yes, that's true. I was rather shocked when a doctor told us to bring him the CD of the X-Ray image from a lab. He explained that the systems used don't readily share information. So much for the efficiency of EHR with everything accessible in one point. 
Gary_EL
50%
50%
Gary_EL,
User Rank: Ninja
11/12/2014 | 12:05:39 PM
First Step
Before EHR's can help make decisions collaboratively, they need to be able to communicate with each other. EHR's, as they exist now, are "Islands of Automation", to borrow a term from yesteryear. One hospital's system can rarely communicate with that of another. There needs to be a requirement that each new system support some kind of portal, or API, that makes communication with other EHR system possible.
PedroGonzales
50%
50%
PedroGonzales,
User Rank: Ninja
11/11/2014 | 6:45:15 PM
Re: errors
@ Thomas.  You are absolutely right.  As the saying goes, garbage in, garbage out.  I would think that health care institutions would place greater emphasis on methods to make sure data entry into the EHR is accurate.  I think EHRs would become more complex if AI is integrated into them. 
Page 1 / 2   >   >>
How Enterprises Are Attacking the IT Security Enterprise
How Enterprises Are Attacking the IT Security Enterprise
To learn more about what organizations are doing to tackle attacks and threats we surveyed a group of 300 IT and infosec professionals to find out what their biggest IT security challenges are and what they're doing to defend against today's threats. Download the report to see what they're saying.
Register for InformationWeek Newsletters
White Papers
Current Issue
Digital Transformation Myths & Truths
Transformation is on every IT organization's to-do list, but effectively transforming IT means a major shift in technology as well as business models and culture. In this IT Trend Report, we examine some of the misconceptions of digital transformation and look at steps you can take to succeed technically and culturally.
Video
Slideshows
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Flash Poll