I work as a coder for a major midwest hospital. Just got out of a large staff meeting where management was encouraging us to work harder, faster and as accurately as possible as the management folks were looking at "other options" such as AI, off-shoring and/or some combination of those methods. Personally, not as worried as that five year suggestion off obselence implies. First: If you code on a regular basis you learn how much you have to read between the lines. Most Doctors have writing skills that would make a fourth grader blush. Misspellings, context errors, incomplete sentences etc. etc. etc. ad nauseum. The idea of suggesting that providers will be taught to adapt to make their notes more "AI friendly" is a heavy lift. I work with brilliant physicians who have been doing things the same way for twenty-five to thirty years and have no interest in adapting to make a computer program's life easier. Second: Medical diagnosis coding is currently based on a code set known as ICD-10. ICD-10 had been used around the world for nearly a decade before the folks who make these decisions moved from ICD-9 to ICD-10. Change in hospitals and physician offices moves at a glacial pace. Third: I've played with chat-gpt a bit, some coding questions that beginers tend to get wrong. For example, essential hypertension vs hypertension with a patient that has chronic kidney disease. It got the first right, the second wrong. It's going to take awhile to get it up to speed. Right now, AI is more the threat of the "bogeyman" to get currently employed coders to work harder for less. Dragon software (voice to text) has been around for more than a decade and it still makes basic spelling and context erros. AI is coming, just not as fast as folks think.
I found your comment insightful, especially with your first-hand experience in medical coding and the challenges AI faces in this space. I'm a student at MIT and currently developing an AI-driven medical coding solution and would love to hear more about your perspective. It would be great to discuss how coders like you see the future of AI in healthcare and learn from your experience. If you're open to it, I'd really appreciate the opportunity to connect for a quick conversation. Let me know if you'd be interested! Thanks, Bilal
Medical coding is the only industry I can think of where the structured data -- the procedure codes, the P2P and MUE edits, the modifiers, all the mappings -- exist on a computer somewhere and it's just the introduction of humans that keeps it inaccessible.
@@shubhamdorlikar8025 Yes ... in two ways, two stages. Firstly, by automating the process of coding, training specialised LLMs to translate the text in EMRs into codes. This is probably happening now. Then in 5 years or less people will realise they can use AI directly with medical records, so the justification for coding will disappear. You can place a long bet on this happening.
I feel like health care professionals are going to be trained to document procedures in ways ai can extract diagnosis and procedures and then turn them into code.
I work as a coder for a major midwest hospital. Just got out of a large staff meeting where management was encouraging us to work harder, faster and as accurately as possible as the management folks were looking at "other options" such as AI, off-shoring and/or some combination of those methods. Personally, not as worried as that five year suggestion off obselence implies.
First: If you code on a regular basis you learn how much you have to read between the lines. Most Doctors have writing skills that would make a fourth grader blush. Misspellings, context errors, incomplete sentences etc. etc. etc. ad nauseum. The idea of suggesting that providers will be taught to adapt to make their notes more "AI friendly" is a heavy lift. I work with brilliant physicians who have been doing things the same way for twenty-five to thirty years and have no interest in adapting to make a computer program's life easier.
Second: Medical diagnosis coding is currently based on a code set known as ICD-10. ICD-10 had been used around the world for nearly a decade before the folks who make these decisions moved from ICD-9 to ICD-10. Change in hospitals and physician offices moves at a glacial pace.
Third: I've played with chat-gpt a bit, some coding questions that beginers tend to get wrong. For example, essential hypertension vs hypertension with a patient that has chronic kidney disease. It got the first right, the second wrong. It's going to take awhile to get it up to speed.
Right now, AI is more the threat of the "bogeyman" to get currently employed coders to work harder for less. Dragon software (voice to text) has been around for more than a decade and it still makes basic spelling and context erros. AI is coming, just not as fast as folks think.
I found your comment insightful, especially with your first-hand experience in medical coding and the challenges AI faces in this space. I'm a student at MIT and currently developing an AI-driven medical coding solution and would love to hear more about your perspective. It would be great to discuss how coders like you see the future of AI in healthcare and learn from your experience.
If you're open to it, I'd really appreciate the opportunity to connect for a quick conversation. Let me know if you'd be interested!
Thanks,
Bilal
Medical coding is the only industry I can think of where the structured data -- the procedure codes, the P2P and MUE edits, the modifiers, all the mappings -- exist on a computer somewhere and it's just the introduction of humans that keeps it inaccessible.
Your're asking the wrong question. In 5 years clinical coding will be irrelevant . I really suggest you get on board with GPT next week.
It's really affect the jobs of medical coding?
@@shubhamdorlikar8025 Yes ... in two ways, two stages. Firstly, by automating the process of coding, training specialised LLMs to translate the text in EMRs into codes. This is probably happening now. Then in 5 years or less people will realise they can use AI directly with medical records, so the justification for coding will disappear. You can place a long bet on this happening.
I feel like health care professionals are going to be trained to document procedures in ways ai can extract diagnosis and procedures and then turn them into code.
Only auditors will be needed
Thank you for your replying guys