This is brilliant. True empowerment indeed! I am glad the letter was not mailed verbatim. The generic language is a dead giveaway that the letter was either originally a template or AI-generated. Instances of “my client” and “they” should be changed to the person’s name, using formal titles like Mr. or Ms., and their correct pronoun. Also, I wonder who signed the letter, given that it implies the sender is an attorney at law. If I were the recipient, I would be looking for suffixes like Esq. or JD after the signature. But overall, this is an impressive example of how true empowerment is coming to all patients and how things are changing incredibly fast. Thanks for posting. I hope the issue is resolved to everyone’s satisfaction.
I was sort of looking for the punchline. I've lost track over how many letters / emails I've written exactly like that without AI. A good old Google search would give you templates as well. I would have waited for the outcome to see if it worked before doing the post.
Fair point. But I think the outcome isn't really the main focus here. As you imply, the quality of the letter doesn't necessarily determine the resolution. For me, this story is more about how AI is giving patients a fighting chance they didn't have before. Sure, templates have always been available, but now anyone can create something tailored to their needs with a bit of guidance. And that's definitely empowering.
I think Hugo expressed it well. Someone who doesn't have problems doesn't need tech tools to solve them. This was a case where the disabled patient was in distress and the advocate partner is short on time.
Grace is no dummy and I'm sure she wouldn't have used a tool like this if she already had the answer in hand...
Thanks for sharing your perspectives Erica. I've been crafting these letters personally and professionally for over 25 years. As a board certified patient advocate who works with immigrant populations, underserved communities, and with patients with limited English proficiency, these individuals need more support to draft letters of appeal and letters of medical necessity and usually can't quickly draft these letters nor can they search for the right templates. Templates are available but I find them to be of poor quality and outdated. From my perspective, searching for templates is not the best use of my time nor the most impactful use of my time to as quickly as possible help the patients I serve. LLMs have been a wonderful game changer as a tool for my advocacy work and as a tool supporting my role as the primary carepartner to 2 disabled adults. The Social Security Administration (SSA) also offers a library of forms to select from for various processes (but there's no guidance as to what form is needed for which type of denial): https://www.ssa.gov/forms/, the catch is if you select the wrong form, fill it out incorrectly, your appeal is denied. If your appeal is denied again, you do not get any information as to why it was denied, for example, if you filled out the wrong form or if something was filled out incorrectly. It can take months to get the 2nd denial notification. You are able to typically appeal one more time, depending on the state you are in and your personal circumstances, but that can take additional months. On average, 3 rounds of appeals can take 9-12 months. Most individuals can not afford to live without their disability benefits as well as without their Medicare coverage for that long. Financial toxicity and well as severe negative impacts on patient outcomes are direct dire consequences. Crafting a letter through prompting ChatGPT or other similar LLMs streamlines the process, drafts a letter structured to the circumstances of a patient's case, eliminates the risk of using the wrong template or form, and it can be translated to an individual's native language to bridge language barriers and limited English proficiency to help with patient education and shared decision making. The punchline is, all of this can be streamlined to under 5 minutes. This process used to take me approx 1.5 hours per case. For me as a professional patient advocate, this helps me not only scale the number of patients I can support but also reduces the patient administrative burden that falls on the individual and family, as well as increases the likelihood of the appeal being granted on the 1st attempt, rather than needing subsequent attempts. I can attest the process works quite well. I have been using this process for well over a year and half and highly recommend it. The punchline is having a customized, accurate, professionally structured letter of appeal in less an 5 minutes, printed, and ready to go in an envelope. By maintaining a library of optimized prompts, I can reuse prompts and get the letter done in less than 2 minutes in similar cases. I have yet to get an appeal denied using this process and probably perform about 20 appeals/month. As always with my posts and stories, I do follow up and report back on outcomes. My guess is it'll take 2-3 months to get a reply back from SSA and I will report back on what the outcome will be. Hope this helps.
Grace, that's an amazing exposition of the horrors of patient administrative burden. I know you had a very packed week, so thanks for taking the time to spell this all out for us.
Wow, Grace, thank you for the comprehensive and thoughtful response. The strategy of maintaining a library of prompts for reuse is excellent. In the near future, this may be the type of task that can be done by a virtual assistant or specialized AI agent.
So of COURSE I asked GPT-4o if it could recall all the prompts I've used with it, and it said:
I don’t have a way to display the first prompts used in all of your past sessions.
However, if you would like, I can help you build out a library of useful prompts based on your typical interests and needs. For example, we could create a set of prompts focused on AI, healthcare, patient empowerment, and your key areas of work such as statistical process improvement or health data standards.
Let me know what specific areas you want to include, and I can help draft a set of starter prompts!
_______________
NOTE/BEWARE: IT HALLUCINATED. I've never talked to it about statistical process improvement; in fact I've never *done* anything in that area.
This is brilliant. True empowerment indeed! I am glad the letter was not mailed verbatim. The generic language is a dead giveaway that the letter was either originally a template or AI-generated. Instances of “my client” and “they” should be changed to the person’s name, using formal titles like Mr. or Ms., and their correct pronoun. Also, I wonder who signed the letter, given that it implies the sender is an attorney at law. If I were the recipient, I would be looking for suffixes like Esq. or JD after the signature. But overall, this is an impressive example of how true empowerment is coming to all patients and how things are changing incredibly fast. Thanks for posting. I hope the issue is resolved to everyone’s satisfaction.
I was sort of looking for the punchline. I've lost track over how many letters / emails I've written exactly like that without AI. A good old Google search would give you templates as well. I would have waited for the outcome to see if it worked before doing the post.
Fair point. But I think the outcome isn't really the main focus here. As you imply, the quality of the letter doesn't necessarily determine the resolution. For me, this story is more about how AI is giving patients a fighting chance they didn't have before. Sure, templates have always been available, but now anyone can create something tailored to their needs with a bit of guidance. And that's definitely empowering.
I think Hugo expressed it well. Someone who doesn't have problems doesn't need tech tools to solve them. This was a case where the disabled patient was in distress and the advocate partner is short on time.
Grace is no dummy and I'm sure she wouldn't have used a tool like this if she already had the answer in hand...
Thanks for sharing your perspectives Erica. I've been crafting these letters personally and professionally for over 25 years. As a board certified patient advocate who works with immigrant populations, underserved communities, and with patients with limited English proficiency, these individuals need more support to draft letters of appeal and letters of medical necessity and usually can't quickly draft these letters nor can they search for the right templates. Templates are available but I find them to be of poor quality and outdated. From my perspective, searching for templates is not the best use of my time nor the most impactful use of my time to as quickly as possible help the patients I serve. LLMs have been a wonderful game changer as a tool for my advocacy work and as a tool supporting my role as the primary carepartner to 2 disabled adults. The Social Security Administration (SSA) also offers a library of forms to select from for various processes (but there's no guidance as to what form is needed for which type of denial): https://www.ssa.gov/forms/, the catch is if you select the wrong form, fill it out incorrectly, your appeal is denied. If your appeal is denied again, you do not get any information as to why it was denied, for example, if you filled out the wrong form or if something was filled out incorrectly. It can take months to get the 2nd denial notification. You are able to typically appeal one more time, depending on the state you are in and your personal circumstances, but that can take additional months. On average, 3 rounds of appeals can take 9-12 months. Most individuals can not afford to live without their disability benefits as well as without their Medicare coverage for that long. Financial toxicity and well as severe negative impacts on patient outcomes are direct dire consequences. Crafting a letter through prompting ChatGPT or other similar LLMs streamlines the process, drafts a letter structured to the circumstances of a patient's case, eliminates the risk of using the wrong template or form, and it can be translated to an individual's native language to bridge language barriers and limited English proficiency to help with patient education and shared decision making. The punchline is, all of this can be streamlined to under 5 minutes. This process used to take me approx 1.5 hours per case. For me as a professional patient advocate, this helps me not only scale the number of patients I can support but also reduces the patient administrative burden that falls on the individual and family, as well as increases the likelihood of the appeal being granted on the 1st attempt, rather than needing subsequent attempts. I can attest the process works quite well. I have been using this process for well over a year and half and highly recommend it. The punchline is having a customized, accurate, professionally structured letter of appeal in less an 5 minutes, printed, and ready to go in an envelope. By maintaining a library of optimized prompts, I can reuse prompts and get the letter done in less than 2 minutes in similar cases. I have yet to get an appeal denied using this process and probably perform about 20 appeals/month. As always with my posts and stories, I do follow up and report back on outcomes. My guess is it'll take 2-3 months to get a reply back from SSA and I will report back on what the outcome will be. Hope this helps.
Grace, that's an amazing exposition of the horrors of patient administrative burden. I know you had a very packed week, so thanks for taking the time to spell this all out for us.
Wow, Grace, thank you for the comprehensive and thoughtful response. The strategy of maintaining a library of prompts for reuse is excellent. In the near future, this may be the type of task that can be done by a virtual assistant or specialized AI agent.
So of COURSE I asked GPT-4o if it could recall all the prompts I've used with it, and it said:
I don’t have a way to display the first prompts used in all of your past sessions.
However, if you would like, I can help you build out a library of useful prompts based on your typical interests and needs. For example, we could create a set of prompts focused on AI, healthcare, patient empowerment, and your key areas of work such as statistical process improvement or health data standards.
Let me know what specific areas you want to include, and I can help draft a set of starter prompts!
_______________
NOTE/BEWARE: IT HALLUCINATED. I've never talked to it about statistical process improvement; in fact I've never *done* anything in that area.