Otter.ai and other subscriptions services provide high quality "automatic transcripts" from audio files. For example, I recently fed a five-hour government CMS web meeting through this service.
With ordinary conversations, Auto Transcript is pretty spot-on. However, with detailed technical material, it's more likely to make mistakes.
Here's a workaround. Make the auto transcript of the public meeting (or youtube meeting) via a service like Otter.AI. It will do the best it can, but it will never add or subtract words from what it attempts to transcribe.
THEN, as step two, feed the Auto Transcript into CHAT GPT (or another AI) and ask it to "fix" the Auto Transcript. The result will be better, because the Chat GPT AI will feel able to make small changes for sense or common sense, and will probably improve sentence flow and paragraphing or headings.
Here's an example from a June 25, 2024, web-streamed CMS lab meeting on new test pricing policy. Note, I've manually inserted test names that were unavailable in the audio file.
###
PROMPT:
This is an Auto Transcript, which naturally has some errors or awkward grammar.
Please improve it, but make only minor changes for clarity and stay very close to the auto transcript.
[AUTO TRANSCRIPT otter.ai]
Speaker 6 40:23
Thank you. My name is Michelle Aldrich and I am the clinical
lab laboratory director for Europeans transplant genomics. I'm here today to
talk about the newly approved category one CPT code 815 6x for true graph
kidney. Next slide please. This test is a gene expression profile to be used
after kidney transplantation as a way to detect a subclinical rejection with
the goal of improving long term outcomes of kidney transplantation and reducing
the need to either go back on dialysis or have a second transplant the
methodology is listed there. And the other alternative for this test is it will
avoid the need for invasive and costly surveillance biopsies. Next slide. And
so this slide lists the resources necessary to perform the test both from a
reagent consumable equipment and a labor standpoint, as illustrated these
resources are substantial. In addition, of course, there was heavy r&d
involved in the development in many clinical studies that have been performed.
To demonstrate the clinical validity and clinical utility of this assay. It is
currently approved by modex under General LCD. Next slide please. So our
crosswalk recommendation is to crosswalk to CPT code 81595, which is a similar
test it is a gene expression profile in the cardiology transplant space. It
utilizes 20 genes 11, which are specific for the test nine which are
housekeeping and it does report a rejection risk. Next slide please. If we
wanted to compare the test a little closer, you can see that there are a couple
of differences. Initially, I'm from a sample type perspective, the heart
transplant test 81595 does utilize whole blood that has been undergone a
purification step to look at only PBMCs and then extracts RNA from those cells.
In contrast, 815 6x or our new kidney CPT code does extract RNA from whole
blood that has been collected in an RNA specific PAXgene to the other main
difference is the heart transplant looks only at 20 genes while we were looking
at 139 genes, both of them then go through a bio informatics algorithm with
similar types of reporting regarding the status of rejection, and of the
allograft. And that's all I have, are there any questions?
AI IMPROVED VERSION
Chat GPT:
#
AGENDA 26
TRANSPLANT GENOMICS
Thank you. My name is Michelle Aldrich, and I am the clinical laboratory
director for [Eurofins] Transplant Genomics. I'm here today to talk about the
newly approved Category One CPT code 8156X for TruGraf Kidney.
- · Transplantation medicine (allograft rejection, kidney), mRNA, gene
expression profiling by quantitative polymerase chain reaction (qPCR) of 139
genes, utilizing whole blood, algorithm reported as a binary categorization as
transplant excellence, which indicates immune quiescence, or not transplant
excellence, indicating subclinical rejection
Next slide, please.
This test is a gene expression profile used after kidney transplantation to
detect subclinical rejection, with the goal of improving long-term outcomes and
reducing the need to either return to dialysis or undergo a second transplant.
The methodology is listed here. An alternative benefit of this test is that it
will avoid the need for invasive and costly surveillance biopsies.
Next slide.
This slide lists the resources necessary to perform the test from a reagent,
consumable, equipment, and labor standpoint. As illustrated, these resources
are substantial. Additionally, there was significant R&D involved in the
development and many clinical studies performed to demonstrate the clinical
validity and utility of this assay. It is currently approved by MoDEX under
General LCD.
Next slide, please.
Our crosswalk recommendation is to crosswalk to CPT code 81595, which
is a similar test. It is a gene expression profile in the cardiology transplant
space, utilizing 20 genes—11 specific for the test and 9 housekeeping genes—and
it reports a rejection risk.
81595, $3240 (Allomap, CareDx) Cardiology (heart transplant), mrna, gene
expression profiling by real-time quantitative pcr of 20 genes (11 content and
9 housekeeping), utilizing subfraction of peripheral blood, algorithm reported
as a rejection risk score
Next slide, please.
If we compare the tests more closely, there are a few differences. From a
sample type perspective, the heart transplant test 81595 utilizes whole blood
that has undergone a purification step to isolate PBMCs and then extracts RNA
from those cells. In contrast, 8156X, our new kidney CPT code, extracts RNA
from whole blood collected in an RNA-specific PAXgene tube. Another key
difference is that the heart transplant test looks at 20 genes, while our test
looks at 139 genes. Both tests then go through a bioinformatics algorithm with
similar types of reporting regarding the status of rejection and the allograft.
That's all I have. Are there any questions?
