AI in Health Diagnostics : 

We'll talk about more that I think is really critical and how doctors are offering patients the ability to  to have access to clinical trials and to think about the complexity of care delivery choices in the therapeutic Journey  and then as the therapeutic Journey changes as it often does in C Cancer Care how we think about that with clinical decision support how we get the patients the right information that they need for them in their caregivers so from the care delivery piece I think that there are so many places and and where I see it today in terms of just outside of the Care delivery pieces in drug Discovery so you know whether it's designing the right proteins uh for  understanding the folding of proteins and designing drugs  to really think about how you begin clinical trials or even clinical trial optimization because clinical trials are very expensive and you can probably optimize clinical trials so they're powered effectively to give you the result that you need to understand if a therapy is going to be effective by changing certain parameters and AI is used in all of those spaces today  but to your question more directly how it impacts physicians in sort of our work burden because it has become administratively burdensome to a working doctor.

Risk of AI in Healthcare : 

I also am a working doctor so I'm a breast medical oncologist in Austin Texas at Texas Oncology.  we use AI in the clinic today so we have I use an AI scribe uso I have  ambient listening as I sit there and talk to patients that allows to a note to be created that about 30 seconds after I finished a visit a note is scribed to reflect the things that we we've talked about in our visit so that could be anything from a patient understanding their stage their their therapeutic plan uh risks and benefits uh of Alternatives of various different treatments and so by having an AI scribe not only does it I call all these pivots the home for dinner campaign or how do we sort of get doctors home for dinner and not you know dictating for two hours after after a clinic  but not only does it allow me to be more efficient and effic clinic but by that note being available to my patient the very evening that I see them on their patient portal they're able to discuss their  their clinical issues with their caregivers and really get a greater understanding of their disease you know it may be unique to oncology but  but this asymmetry of information exists a lot between doctors and patients and getting patients realtime information about their disease for them and their caregivers allows them to be better informed about their Healthcare journey and that really facilitates better adoption of treatments  and better compliance with treatments that becomes really important so I use an AI scrib and clinic and across Texas Oncology we've had Champions that have really championed this we have 500 providers that are using an AI scribe service to generate notes that's been really meaningful and not only in that care delivery efficiency piece but also in the documentation piece so when I see a patient I tend to document an icd10 code that correlates to their diagnosis right so if it's a left upper outer quadrant breast cancer that is what I put  but in my note there's all sorts of other information characterizing their diabetes and their osteoporosis and their obesity and  you know their hypothyroidism and I don't take the time to code those things which is a limitation in how I'm assessed in treating the patient but who has time to really document all of those other things but my scribe service does generate all of those other icd10 codes that gives better color even for payers and others that touch my patients in understanding really the complexity of their disease and disease management that's incredible you know I think about that as you're speaking and the number of times.

AI in Healthcare Monitoring 

 I'll you know forget to say you know encounter for therapeutic monitoring right of of the long-term management of something or anti-coagulation like these things matter because they can actually have a uh an influence on what is uh compensated a return not even to the physician but even for the patient like oh was this necessary was this unnecessary and just the thought to be able to say as I'm saying out loud you're here to follow up your hospitalization for you know forbid covid you know pneumonitis or or pneumonitis from therapy like that automatically comes you know toxicity grade three because they were admitted for pneumonitis like and and and as I'm saying that out loud I I that's that's pretty much here it sounds like you're using that already which is amazing one other part I would say even you painted this beautiful picture of the diagnosing the cancer how well and accurately and hopefully how early we diagnose it to then our treatment decisions to making sure we don't miss anything to also considering variables that could be uh relevant  in the response or toxicities that sometimes it's just hard to keep up with as trucks come out even after all of that including the patient room and the discussions and then getting the consent is then this whole you know laborous process of of just getting it approved and insurance companies and you know you're a patient and listening to this trust me when I say it is a awfully annoying thing for us as well for those Physicians that like are deeply invested in in the clinical outcome it hurts us when things are being delayed and it's always at the mercy of sometimes somebody the person in your office the person that may be out is it coded correctly is it with drawn correctly they have a big lift because they don't have a medical degree all of that is already happening as well so hopefully be able to get scans quicker and approvals for surgeries quicker and all of these things so it is a true you know I'm told as they say in sales endtoend kind of implementation that helps the entire process you know from one point to the other it's a great point that you bring up Sanjay and I think most doctors actually aren't aware of it but  managing the administrative burden of prior authorization delays appropriate care and  is uh financially a big lift for most Healthcare organizations and so being able to contribute to means to do that more efficiently and effectively really matters so  like when I'm starting a patient like  when I was in clinic on Monday  I I saw a new stage three triple negative breast cancer patient and I started her on the keynote 522 regimen that ply aab with taxol carboplatin adamy and cytoxan you know standard of care but what I did for her was multicomponent right because I had to Stage her cancer and talk to her about that educate her  right for her chemotherapy and do all of the appropriate workup so one of the things I did is I tested her thyroid function right because I and and M that's nothing to do with her cancer I know I'm going to put her on immunotherapy.

 I know that a potential toxicity of immunotherapy is I could cause her to have hypothyroidism and what routinely happens is when doctors right for a thyroid function evaluation in but in the laboratory it's a reimbursement Hazard because you haven't coded for an appropriate diagnosis to make that appropriate but because I use an ambient AI uh listening service I've explained to the patient uh you know why I'm ordering these different tests and it's there so when my team wants to understand what codes are there to to better justify the laboratory tests that I've ordered they have it available and so it helps manage the prior authorization of those evaluations much better and then on the back end the revenue cycle of getting  a a billing appropriately and making sure you get paid appropriately for the work that you're doing it manages that too and you know.

Ethical and Regulatory Challenges

 I think many doctors again don't appreciate this but  in a lot of our value-based care models were compensated appropriately based on the complexity of patients we don't it's uh we will get paid less if our patients look like 93y old marathon runners with lung cancer  you know and if the if the payer does not have insight into their multiple comorbid illnesses and so our failure to document that  hurts Us in terms of that reimbursement we have to manage all those things right but when we don't document it and who has time you know most of us have a huge administrative burden already um that if you're not documenting appropriately reflective of patients coorbit illness your reimbursement is less so  you probably are familiar with cmi's uh enhancing oncology model which is cmi's alternative payment model pilot in value based care for for Cancer Care  so if you have one less hierarchical condition category which is like less comorbid illness on average it's about $7,000 less  of compensation that a practice receives for the care of that patient and so there are a lot of aligned incentives to try to make sure that your medical documentation is really accurate not just sufficient to treat the cancer but is an accurate reflection of the picture of the patient as it possibly can be so I think AI helps us there because most of us are spending a lot of time jumping through the administrative hurdles trying to get our patients what they need  and something as easy as a as ambient AI listening as a scribe service that can provide additional documentation for things like icd1 codes can augment your practice