Coding and programming as a growing necessity for the doctors of the future?

[Matthew9Thirtyfive]

Both of you chill. An argument about what we come to SDN for is completely unnecessary.

Don’t make me change your avatar to something embarrassing.

---

Members don't see this ad.

 

[DD214]

" There's no value in physicians learning coding. Source: I used to be a software engineer (C/C++). "

False

I learned some VB on my own and developed my own program to write f/u notes while I am talking to patients

I can click on a few boxes on my program and have it generate a completely readable history with proper sentence structure including punctuation.

I then cut and paste right into the EMR and make any changes needed

Saves me a TON of time for f/u appts and probably increases my productivity by 20%

 

[IsleyOfTheNorth]

You could probably pay a programmer $200 to make the same program for you in a day.

I could write such a program myself but even I would probably outsource it as it's not worth my time.

 

[CocoMelon0531]

Waste of elective time

Financial literacy is much better use of time in order for a mass movement of physicians to unshackle themselves from their MBA overlords, be financial independent, be happier, and subsequently deliver better healthcare.

 

[zeloc]

What did you expect to get from posting here? No one in this thread has much background in anything to be able to give an informed opinion. Of course physicians are going to say, "What's the point of learning to code"? You might as well be asking car mechanics what they think of Biden's healthcare plan. Medical students and pre-med coders have never worked in medicine and so might not see the applications either.

I code and I believe it is a very important skill for everyone, including physicians. Physicians who can't code of course are not going to be able to think of any applications they could have used it for, because they don't know what coding is. But there is not time for it in medical school, at least not the way medical school is structured now, and I do not think it is really appropriate to have as a requirement. Ideally everyone would learn to code in high school or college, but it still shouldn't be a requirement.

The ppl saying, "leave the coding to the IT ppl." Have you never wondered why EMRs are so screwed up?

There are many physicians who code and have made the practice of medicine far easier than it would be, but the technology is all taken for granted by practicing physicians. There are tons of stuff that happens automatically behind the scenes. Also, medicine would be far more advanced with higher quality care taking less time for physicians if all the technological advances did not take 17 years to enter practice, which is due to regulatory/governmental/administrative bottlenecks.

Informatics, on the other hand, should be taught in all medical schools, and there are a decent number of schools that do have some teaching in it. Clinical informatics is the field dealing with the acquisition, storage, and use of clinical data for patient care. Having even a minimal background would make a huge difference when administrators are talking to physicians about any changes to EMR, billing, or documentation systems, making smarter choices, creating better clinical workflows, managing patient data more securely, saving physicians time, etc. I've seen a lot of horror stories of each of the above that has cost physicians, departments, and hospital systems a lot of time and money, as well as medical errors. Plus it would make for much better lobbying to the government. There are laws that affect physicians every day, but 99% of physicians are not aware of the law or the impact it is having on them because it's wrongly ascribed to some other factor, like greedy capitalists, highly complex sick patients, technological limitations, etc.

 

[deleted849476]

Software engineering major currently in med school. Maybe my background would help if I did really specific research. Command line helps for quick computations and massive file functions, parsing files, etc, sure, but unless I was doing like genomics or protein folding or something I don't see myself using it much. Not to say there aren't a ton of applications for it, but unless you're doing the nitty-gritty software dirty work and architecting systems etc., no one is going to need a med student so they can "hey, code that up for me really quick."

People toss around "learn to code" like it's learning to juggle. It's an engineering discipline. People who don't do bachelor's-->internship-->serious/difficult projects all while eating/breathing code for fun every night in college either A) never work on serious projects and work at IT jobs where they usually don't even write code, or B) spent their entire youth remaking video games and coding scientific calculators from scratch for fun at home. My graduating class was night-and-day. A few people going to google and Microsoft, some getting those "IT" jobs for $20-40k, majority just giving it up and doing something else entirely, basically including me going to med school. I mean you just have to WANT it. I saw freshmen in undergrad who were twice as serious about software than my peer med students are about anything. I think its because there's honestly a massive difference from being able to make something that is fun/useful/looks good/is novel/helps other people or other experts/demonstrates proficiency in CS and being able to get A's on CS exams and do your CS homework. A's just don't grant you the ability to make something good.

I think developing an app that does anything novel or significant and doing well in medical school would be phenomenal. I'm not talking about page-turner apps with information plastered in text boxes on the screen that people can whip up in an editor without evening writing a line of code.

Adults talk about developing apps like kids do about playing in the NBA when they grow up. But I guess do your best to be the exception if you can.

And yes I'm kinda disillusioned with it all

 

[deleted849476]

What did you expect to get from posting here? No one in this thread has much background in anything to be able to give an informed opinion. Of course physicians are going to say, "What's the point of learning to code"? You might as well be asking car mechanics what they think of Biden's healthcare plan. Medical students and pre-med coders have never worked in medicine and so might not see the applications either.

I code and I believe it is a very important skill for everyone, including physicians. Physicians who can't code of course are not going to be able to think of any applications they could have used it for, because they don't know what coding is. But there is not time for it in medical school, at least not the way medical school is structured now, and I do not think it is really appropriate to have as a requirement. Ideally everyone would learn to code in high school or college, but it still shouldn't be a requirement.

The ppl saying, "leave the coding to the IT ppl." Have you never wondered why EMRs are so screwed up?

There are many physicians who code and have made the practice of medicine far easier than it would be, but the technology is all taken for granted by practicing physicians. There are tons of stuff that happens automatically behind the scenes. Also, medicine would be far more advanced with higher quality care taking less time for physicians if all the technological advances did not take 17 years to enter practice, which is due to regulatory/governmental/administrative bottlenecks.

Informatics, on the other hand, should be taught in all medical schools, and there are a decent number of schools that do have some teaching in it. Clinical informatics is the field dealing with the acquisition, storage, and use of clinical data for patient care. Having even a minimal background would make a huge difference when administrators are talking to physicians about any changes to EMR, billing, or documentation systems, making smarter choices, creating better clinical workflows, managing patient data more securely, saving physicians time, etc. I've seen a lot of horror stories of each of the above that has cost physicians, departments, and hospital systems a lot of time and money, as well as medical errors. Plus it would make for much better lobbying to the government. There are laws that affect physicians every day, but 99% of physicians are not aware of the law or the impact it is having on them because it's wrongly ascribed to some other factor, like greedy capitalists, highly complex sick patients, technological limitations, etc.

I essentially agree with all of this but at that point it sounds like this person is more of an administrator/liaison/researcher who is developing tools and making them better for clinicians than some doc who is stepping out of a patient room and also developing some lucrative app on his own independently of his main job.

I mostly agree with the pedagogy thing, but I also think that to get any transferable critical thinking you have to dedicate a lot of time to CS and problem solving and not just taking a crash course on learning to code. I just don't think there is anything special about learning to code that will give transferable skills to medicine (or anything else) any more than would solving puzzles or doing math or something.

 

[7331poas]

I code and I believe it is a very important skill for everyone, including physicians. Physicians who can't code of course are not going to be able to think of any applications they could have used it for, because they don't know what coding is. But there is not time for it in medical school, at least not the way medical school is structured now, and I do not think it is really appropriate to have as a requirement. Ideally everyone would learn to code in high school or college, but it still shouldn't be a requirement.

Name a single time you have ever coded and applied it to clinical practice.

 

[Lawpy]

What did you expect to get from posting here? No one in this thread has much background in anything to be able to give an informed opinion. Of course physicians are going to say, "What's the point of learning to code"? You might as well be asking car mechanics what they think of Biden's healthcare plan. Medical students and pre-med coders have never worked in medicine and so might not see the applications either.

I code and I believe it is a very important skill for everyone, including physicians. Physicians who can't code of course are not going to be able to think of any applications they could have used it for, because they don't know what coding is. But there is not time for it in medical school, at least not the way medical school is structured now, and I do not think it is really appropriate to have as a requirement. Ideally everyone would learn to code in high school or college, but it still shouldn't be a requirement.

The ppl saying, "leave the coding to the IT ppl." Have you never wondered why EMRs are so screwed up?

There are many physicians who code and have made the practice of medicine far easier than it would be, but the technology is all taken for granted by practicing physicians. There are tons of stuff that happens automatically behind the scenes. Also, medicine would be far more advanced with higher quality care taking less time for physicians if all the technological advances did not take 17 years to enter practice, which is due to regulatory/governmental/administrative bottlenecks.

Informatics, on the other hand, should be taught in all medical schools, and there are a decent number of schools that do have some teaching in it. Clinical informatics is the field dealing with the acquisition, storage, and use of clinical data for patient care. Having even a minimal background would make a huge difference when administrators are talking to physicians about any changes to EMR, billing, or documentation systems, making smarter choices, creating better clinical workflows, managing patient data more securely, saving physicians time, etc. I've seen a lot of horror stories of each of the above that has cost physicians, departments, and hospital systems a lot of time and money, as well as medical errors. Plus it would make for much better lobbying to the government. There are laws that affect physicians every day, but 99% of physicians are not aware of the law or the impact it is having on them because it's wrongly ascribed to some other factor, like greedy capitalists, highly complex sick patients, technological limitations, etc.

I get the applications of coding but i'm going to 100% oppose extremely strongly any attempt to make coding required for students, trainees and doctors.

 

[wholeheartedly]

What did you expect to get from posting here? No one in this thread has much background in anything to be able to give an informed opinion. Of course physicians are going to say, "What's the point of learning to code"? You might as well be asking car mechanics what they think of Biden's healthcare plan. Medical students and pre-med coders have never worked in medicine and so might not see the applications either.

I code and I believe it is a very important skill for everyone, including physicians. Physicians who can't code of course are not going to be able to think of any applications they could have used it for, because they don't know what coding is. But there is not time for it in medical school, at least not the way medical school is structured now, and I do not think it is really appropriate to have as a requirement. Ideally everyone would learn to code in high school or college, but it still shouldn't be a requirement.

The ppl saying, "leave the coding to the IT ppl." Have you never wondered why EMRs are so screwed up?

There are many physicians who code and have made the practice of medicine far easier than it would be, but the technology is all taken for granted by practicing physicians. There are tons of stuff that happens automatically behind the scenes. Also, medicine would be far more advanced with higher quality care taking less time for physicians if all the technological advances did not take 17 years to enter practice, which is due to regulatory/governmental/administrative bottlenecks.

Informatics, on the other hand, should be taught in all medical schools, and there are a decent number of schools that do have some teaching in it. Clinical informatics is the field dealing with the acquisition, storage, and use of clinical data for patient care. Having even a minimal background would make a huge difference when administrators are talking to physicians about any changes to EMR, billing, or documentation systems, making smarter choices, creating better clinical workflows, managing patient data more securely, saving physicians time, etc. I've seen a lot of horror stories of each of the above that has cost physicians, departments, and hospital systems a lot of time and money, as well as medical errors. Plus it would make for much better lobbying to the government. There are laws that affect physicians every day, but 99% of physicians are not aware of the law or the impact it is having on them because it's wrongly ascribed to some other factor, like greedy capitalists, highly complex sick patients, technological limitations, etc.

EMRs are horrendous because they weren’t designed to facilitate better patient care. They were designed to maximize billing and to pull data to meet regulatory and quality improvement crap. Not blaming the IT folks for that.

I agree that informatics could be harnessed in amazing ways, but currently there’s major issues with the accuracy of the information documented so that’s going to screw up your results when pulling massive amounts of data to work with. Epidemiologists deal with that problem frequently.

 

[Redpancreas]

Full disclosure I did not read your full post and responses to give you the most tailored answer and am replying to the premise your title implies.

Medicine itself is a core competency. It's a language, then there's a reasoning process, followed by instincts, then people/business skills to develop +/- procedural skills, etc. Coding is another core competency. Sure there's going to be some Leonardo Da Vinci's out there that crossed over from CS to Medicine or just want to learn coding and do it. I personally feel though that it's a waste of time and in the world we live in, it would be better to just get someone else on board instead of trying to juggle two core competencies. In order to move medicine forward or start your own initiative in the future it may be good to be versed on 1) what coding is 2) its potential 3) what kind of language/software you need to tackle XYZ problem you're having, but I don't think there's much value in even physician-innovators to know all the technical details . Some purists may disagree. In my experience though, those people haven't medical school yet and lack perspective on what medical training is.

 

[Redpancreas]

Software engineering major currently in med school. Maybe my background would help if I did really specific research. Command line helps for quick computations and massive file functions, parsing files, etc, sure, but unless I was doing like genomics or protein folding or something I don't see myself using it much. Not to say there aren't a ton of applications for it, but unless you're doing the nitty-gritty software dirty work and architecting systems etc., no one is going to need a med student so they can "hey, code that up for me really quick."

People toss around "learn to code" like it's learning to juggle. It's an engineering discipline. People who don't do bachelor's-->internship-->serious/difficult projects all while eating/breathing code for fun every night in college either A) never work on serious projects and work at IT jobs where they usually don't even write code, or B) spent their entire youth remaking video games and coding scientific calculators from scratch for fun at home. My graduating class was night-and-day. A few people going to google and Microsoft, some getting those "IT" jobs for $20-40k, majority just giving it up and doing something else entirely, basically including me going to med school. I mean you just have to WANT it. I saw freshmen in undergrad who were twice as serious about software than my peer med students are about anything. I think its because there's honestly a massive difference from being able to make something that is fun/useful/looks good/is novel/helps other people or other experts/demonstrates proficiency in CS and being able to get A's on CS exams and do your CS homework. A's just don't grant you the ability to make something good.

I think developing an app that does anything novel or significant and doing well in medical school would be phenomenal. I'm not talking about page-turner apps with information plastered in text boxes on the screen that people can whip up in an editor without evening writing a line of code.

Adults talk about developing apps like kids do about playing in the NBA when they grow up. But I guess do your best to be the exception if you can.

And yes I'm kinda disillusioned with it all

This is exactly it OP. Coding like medicine is a core competency.

 

[radsisrad]

The markets appear to believe that Greenwich Life Sciences has cured cancer. So maybe docs do need to learn how to code.

 

[zeloc]

Software engineering major currently in med school. Maybe my background would help if I did really specific research. Command line helps for quick computations and massive file functions, parsing files, etc, sure, but unless I was doing like genomics or protein folding or something I don't see myself using it much. Not to say there aren't a ton of applications for it, but unless you're doing the nitty-gritty software dirty work and architecting systems etc., no one is going to need a med student so they can "hey, code that up for me really quick."

Believe it or not, as a resident I had a medical student on my service who knew how to code. After being on the team for a few days, rounding, etc, he realized how he could dramatically increase the speed of pre-rounding. With my permission he spent a few days on it instead of working on clinical responsibilities and coded up something that we overlayed on the EMR that literally saved hours of time in pre-rounding (multiply the time saved per pt by the number of pts seen per day). It was one of the most incredible things I'd seen. We ended up sharing it with a bunch of other residents and it not only saved oodles of time but made patient care more safe by preventing things from being missed. Of course it's not like any of the residents arrived later, so we were using the time saved on patient care.

I essentially agree with all of this but at that point it sounds like this person is more of an administrator/liaison/researcher who is developing tools and making them better for clinicians than some doc who is stepping out of a patient room and also developing some lucrative app on his own independently of his main job.

Not sure what you mean by this or who it is aimed at, but I am a practicing physician and I code.

Name a single time you have ever coded and applied it to clinical practice.

Is this a joke? To name a single example, I wrote code to place an entire admission order set in literally a few seconds. How long does it take you to write admission orders for a patient using your EMR? Maybe time it and see how long it takes?
This is exactly what I meant when I was saying that ppl who don't know how to code can't see its applications.

I get the applications of coding but i'm going to 100% oppose extremely strongly any attempt to make coding required for students, trainees and doctors.

Agree.

EMRs are horrendous because they weren’t designed to facilitate better patient care. They were designed to maximize billing and to pull data to meet regulatory and quality improvement crap. Not blaming the IT folks for that.

Misleading statement. That's like saying that CAD is caused by smoking. Yes, but there are many other causes like aging, diet, stress, genetics, and lack of physical exercise. While it's true that EMRs were initially developed for billing purposes and clinical functionality was added later, another significant reason for bad EMRs is that many of them were coded by engineers and others that had no clinical background, leading to very many specific deficiencies based on ideas in engineering that are different to ideas in medicine. While over time various things have been corrected, it's still obvious to see today how EMR workflow in general is designed in such a way that the physician has to adapt to the EMR and the EMR way of doing things. Physicians are clinical experts and have specific workflows and EMRs should be designed to accommodate to physician workflow, not the other way around.

 

[zeloc]

I agree that informatics could be harnessed in amazing ways, but currently there’s major issues with the accuracy of the information documented so that’s going to screw up your results when pulling massive amounts of data to work with. Epidemiologists deal with that problem frequently.

This is a misunderstanding of informatics, and exactly why it's important that it be taught more widely.

Informatics is not only used in large datasets like what is used for populations or genomics. It can be used to improve patient care in your ER or inpatient or outpatient setting.
Informatics tools can be used to identify inaccurate information in clinical records.
Informatics can be used to improve the accuracy of the information documented.

It is a practical set of tools and techniques that can be applied to specific clinical purposes. Keep in mind that clinical informaticists are doctors. So they do not have illusions about what happens in medicine, nor idealized views of the accuracy of the data.

 

[7331poas]

Is this a joke? To name a single example, I wrote code to place an entire admission order set in literally a few seconds. How long does it take you to write admission orders for a patient using your EMR? Maybe time it and see how long it takes?
This is exactly what I meant when I was saying that ppl who don't know how to code can't see its applications.

Post the source code in here please. Would like to see it.

 

[zeloc]

Post the source code in here please. Would like to see it.

Another joke? For one, I would not post something that could result in liability if someone else tried to use it and then tried to blame a patient outcome on my code. As you asked and so are aware, this is code I wrote for clinical care.
Besides, I would not post something that you or anyone else could take and commercialize.

 

[Alyssum]

Probably not for 100% clinicians, but some background in programming is very useful (if not required) for anyone doing genomics type of research. You can't always count on having bioinformaticians willing to assist you (what do you have to offer to them?). But probably not relevant for most medical students.

 

[TelemarketingEnigma]

Probably not for 100% clinicians, but some background in programming is very useful (if not required) for anyone doing genomics type of research. You can't always count on having bioinformaticians willing to assist you (what do you have to offer to them?). But probably not relevant for most medical students.

There's always going to be ways to apply coding to various forms of research, many of which are of interest to some med students/doctors, but I don't think that makes it necessary for the average doctor or med student like you said - most people doing high level genomics research do not have MDs (or if they do, it's often MD/PhD)

personally I think it's useful to have at least the base level of familiarity where you may not be able to do complex stuff, but at least understand some concepts enough that you could google your way through simple things or have a basic level conversation with an actual CS trained person. But I think it's more important to learn how to work in concert with people with a much higher level of training than it is to force all med students to learn to code

 

[zeloc]

Post the source code in here please. Would like to see it.

If you are interested in seeing something, there are message boards where physicians write/post code. Because EMRs are different, you will tend to find such information in EMR-specific forums. Eg, there are ones for eclinical works, epic, etc.

Another possibility is that some EMRs/software have programming languages built into it, which it can recognize. If you start exploring any advanced functionality that may be available to you in your hospital's software, you might find references to a particular language, and then you can explore that language. If it is a custom language for that software, there will be sometimes be a manual or instructions on how to use it.

Also keep in mind that the code you come across may not be understandable. If you are doing something like making commonly-used phrases to call up with a few keys, then you will recognize those aspects of the code because you'll see the doctor phrases that are supposed to be called up. But if you are writing code to execute orders, then the code will only contain what it needs to to activate an order, and you won't necessarily see anything recognizable.

 

[7331poas]

If you are interested in seeing something, there are message boards where physicians write/post code. Because EMRs are different, you will tend to find such information in EMR-specific forums. Eg, there are ones for eclinical works, epic, etc.

Another possibility is that some EMRs/software have programming languages built into it, which it can recognize. If you start exploring any advanced functionality that may be available to you in your hospital's software, you might find references to a particular language, and then you can explore that language. If it is a custom language for that software, there will be sometimes be a manual or instructions on how to use it.

Also keep in mind that the code you come across may not be understandable. If you are doing something like making commonly-used phrases to call up with a few keys, then you will recognize those aspects of the code because you'll see the doctor phrases that are supposed to be called up. But if you are writing code to execute orders, then the code will only contain what it needs to to activate an order, and you won't necessarily see anything recognizable.

Not really responsive to my statement. Lets see the source code that you have written. Most of the stuff you just referenced in your post is not "programming", its basically just creating macros which does not require any computer science education.

 

[aldol16]

There's a reason why so there's such a growing field in digital health. There are people for this who aren't doctors. Doctors spend enough time in training - we can serve as advisors and consultants for the people who actually do the coding. Just like how a doctor doesn't have to go down with the patient to MRI, inject gad, and run the scan because there are people who can specialize in that.

 

[zeloc]

Not really responsive to my statement. Lets see the source code that you have written. Most of the stuff you just referenced in your post is not "programming", its basically just creating macros which does not require any computer science education.

I already explained why I wasn't posting my code above.
What EMR are you using? Let's see you write macros to do anything that I mentioned. Please post the code here. (Hint - it's not possible. )

There's a reason why so there's such a growing field in digital health. There are people for this who aren't doctors. Doctors spend enough time in training - we can serve as advisors and consultants for the people who actually do the coding. Just like how a doctor doesn't have to go down with the patient to MRI, inject gad, and run the scan because there are people who can specialize in that.

Doctors aren't qualified to act as advisors and consultants for the ppl who do the coding. You need to understand database design, data, cost of acquiring data, a number of research considerations which most physicians have no background in, statistics, computer programming, leadership, quality improvement, usability (which physicians know very little about), etc, etc. That's why there are informaticists.

 

[7331poas]

I already explained why I wasn't posting my code above.

then ill remain skeptical

 

[zeloc]

Yep.

And if you end up doing lots of research and want to learn coding you can.

I am a 100% clinical outpatient FP. Coding classes would not benefit me in any way that I can see.

I've met and worked with a large number of researchers and very few of them knew how to code. Statistics and coding are quite distinct.

 

[TelemarketingEnigma]

I've met and worked with a large number of researchers and very few of them knew how to code. Statistics and coding are quite distinct.

This really depends on the type of research. My personal familiarity on the coding front is with giant genomics projects that used a ton of coding, but this was usually specific teams of phd type folks, who might collaborate with MDs on the clinical aspects of the projects.

Doctors aren't qualified to act as advisors and consultants for the ppl who do the coding.

MDs are useful as consultants for coding projects in the sense that they can advise on what kind of things to do with the code, and what's actually useful in a clinical setting, but they should not be in charge of how to do the coding. Collaborative teams with multiple skills are more effective than one physician trying to do it all.

 

[aldol16]

Doctors aren't qualified to act as advisors and consultants for the ppl who do the coding. You need to understand database design, data, cost of acquiring data, a number of research considerations which most physicians have no background in, statistics, computer programming, leadership, quality improvement, usability (which physicians know very little about), etc, etc. That's why there are informaticists.

You're misunderstanding me. MDs aren't there to advise or consult on the technical details on coding. There are software engineers who are the experts on that. But software and algorithms need clinical input on what's important and what's not. A software programmer can probably create an algorithm that tells you when PACs are occurring with arbitrary certainty. Is that particularly useful? The software programmer probably won't know that. You need someone with an MD - or at least with telemetry experience - to tell them what they should be looking at.

To use a real-world example, I'm sure you've heard of a well-known academic center that's been on the news recently for its SARS-CoV-2 vaccine distribution policies. It subsequently tried to blame an "algorithm" for the distribution. But you can't blame an algorithm. An algorithm does exactly what the programmer wants it to do. I'm sure if they actually had more clinical advice from residents, the algorithm would have worked much better and saved them from national embarassment.

 

[vancoremed]

I personally don’t see why people would think that learning a program like R wouldn’t be helpful for physicians. We live in an increasingly data driven world. There are so many good diagnostic/prognostic/treatment models out there in the literature that is just not being used. They aren’t used because most physicians don’t understand basic regression or machine learning models and certainly wouldn’t have a clue how to implement them. There are some calculators out there that makes things easy but they are all over the place. If physicians could learn a programming language like R, researchers could add their models to a package that would be easily able to be used at the bedside. This would allow all models to be in one place easily accessible. Just type in the patients data and you get a reccomendation that combined with the physicians own judgement will improve patient care.
AI won’t replace physicians but physicians who use AI will replace physicians who don’t.

 

[aldol16]

That's not how it works. You're never going to have physicians at the bedside or in the workroom typing patient data into a regression model to get some output number that they use to individualize care. That's not the way AI is going. What's happening is that all of this information is being incorporated into the EMR itself. The EMR already has all that information in a usable format so all the programming and implementation can happen on the back end. The physician can get a readout with just a few clicks of the mouse. That's where AI-driven healthcare is going.

Also, physicians understanding the basic concepts of regression modeling and machine learning might be an okay ask but in terms of really understanding it? Not necessary. That's why you have software engineers at Epic.

 

[Lawpy]

I personally don’t see why people would think that learning a program like R wouldn’t be helpful for physicians. We live in an increasingly data driven world. There are so many good diagnostic/prognostic/treatment models out there in the literature that is just not being used. They aren’t used because most physicians don’t understand basic regression or machine learning models and certainly wouldn’t have a clue how to implement them. There are some calculators out there that makes things easy but they are all over the place. If physicians could learn a programming language like R, researchers could add their models to a package that would be easily able to be used at the bedside. This would allow all models to be in one place easily accessible. Just type in the patients data and you get a reccomendation that combined with the physicians own judgement will improve patient care.
AI won’t replace physicians but physicians who use AI will replace physicians who don’t.

Look i get the benefits of AI but can we not make it another requirement in an already bloated curriculum?

 

[sliceofbread136]

I’d rather have mechanical engineering knowledge and be able to truly understand the physics of a collapsible tube in a pressurized space. Would be much more useful IMO

 

[Yhwach]

I’d rather have mechanical engineering knowledge and be able to truly understand the physics of a collapsible tube in a pressurized space. Would be much more useful IMO

I feel like CS/SWE allows you to produce products with better economies of scale though. Like if I wanted to be a physician entrepreneur id want to go into healthtech rather than devices.

 

[RangerBob]

We could all come up with a list of things we think docs should learn in addition to the basic curriculum.

Spanish is clearly first. That will get the clinical-based docs farther in work and life than any coding language.

After that we could argue about the next most important, whether it’s world religions/cultures, more basic science (such as the physics example above), more complementary/alternative medicine, statistics (we all “learn” it but most of us have a very limited understanding), AI, coding, etc.

At a certain point we run out of brain space to learn/retain the most important thing, which is medicine.

 

[zeloc]

You're misunderstanding me. MDs aren't there to advise or consult on the technical details on coding. There are software engineers who are the experts on that. But software and algorithms need clinical input on what's important and what's not. A software programmer can probably create an algorithm that tells you when PACs are occurring with arbitrary certainty. Is that particularly useful? The software programmer probably won't know that. You need someone with an MD - or at least with telemetry experience - to tell them what they should be looking at.

To use a real-world example, I'm sure you've heard of a well-known academic center that's been on the news recently for its SARS-CoV-2 vaccine distribution policies. It subsequently tried to blame an "algorithm" for the distribution. But you can't blame an algorithm. An algorithm does exactly what the programmer wants it to do. I'm sure if they actually had more clinical advice from residents, the algorithm would have worked much better and saved them from national embarassment.

I'm not misunderstanding you. Every physician thinks that they think like every other physician. But it is not the case. If a physician is advising a software engineer, they are not advising for their individual use, but for a large number of physicians/hospital system. I gave you an abbreviated list of what an MD needs to know to advise, and it was a very long list that did not include technical details of coding. I've seen a lot of projects go wrong because physicians thought they knew what they were doing in their advising. Physicians are masters of clinical care, not usability, workflow, leadership (you need to convince the other physicians in the organization to go along with whatever changes are made, and it is not easy, and it cannot just be mandated unless you want disgruntled/angry MDs), or the plethora of other things I mentioned.

 

[zeloc]

That's not how it works. You're never going to have physicians at the bedside or in the workroom typing patient data into a regression model to get some output number that they use to individualize care. That's not the way AI is going. What's happening is that all of this information is being incorporated into the EMR itself. The EMR already has all that information in a usable format so all the programming and implementation can happen on the back end. The physician can get a readout with just a few clicks of the mouse. That's where AI-driven healthcare is going.

Also, physicians understanding the basic concepts of regression modeling and machine learning might be an okay ask but in terms of really understanding it? Not necessary. That's why you have software engineers at Epic.

You're right, but not addressing what the previous poster said, the specific words were "I personally don’t see why people would think that learning a program like R wouldn’t be helpful for physicians." Of course it would be helpful for physicians. For one, you would have much faster translation from evidenced-based clinical research to actual practice.

 

[KnuxNole]

I wouldn't have the time for that ****. I would want the experts to spend their time on that while I do more important things.

 

[zeloc]

Can any of you compsci/engineering experts in this thread point me towards compsci/biostats learning material I can get through in the next couple of months that may help for clinical research / medical device innovation / multidisciplinary innovation?

Question is far too broad. 100 different kinds of materials/courses could be recommended. You need to further define what your interest is. Clinical research is extremely broad by itself, and totally different to medical device innovation which is different to multidisciplinary innovation. I would make a new topic.