Using Knowtex at the University of Rochester Medical Center

** If you would like to view/download the complete case study, please click here **

The Problem

Clinician burnout has reached critical levels, with recent studies estimating over 40% of clinicians experience burnout.¹ A major contributor is the mounting administrative workload clinicians face, especially clinical documentation; for every 1 hour clinicians spend face to face with patients, they spend nearly 2 hours on EHR and desk work, fueling frustration and burnout.²

Many healthcare organizations have tried alleviating documentation burdens by hiring medical scribes or utilizing transcription and dictation.While scribing has enabled some physicians to focus more on patient care during visits, the costs of scribes can be prohibitive and their presence in the exam room can interrupt the patient-physician interaction. Additionally, scribes have very high turnover, with the average tenure of a scribe being under 2 years.³ This results in high operational costs and significant training challenges for the scribe, especially in creating personalized notes tailored to each physician's style. Other options like transcription and dictation require the physician workflow and do not capture the patient interaction and verbatim details.

The Solution

Knowtex empowers physicians with voice AI workflows to capture clinical encounters naturally through voice, automatically generating real-time medical notes and codes from the doctor-patient encounter. Through its proprietary and advanced machine learning models, Knowtex is able to tailor the note by medical specialty and can customize to the style of individual physicians to accurately represent their clinical voice.

To ensure accurate reimbursement, Knowtex also interprets the visit interaction to intelligently generate diagnostic and billing codes.

Knowtex is accessible via any device including mobile app, tablet, or computer, requiring no dedicated hardware, and is designed to seamlessly integrate with team-based care and any clinical workflow.

About URMC

To tackle administrative workload challenges, the University of Rochester Medical Center (URMC) partnered with Knowtex in February 2023 to pilot an innovative documentation solution powered by AI.

The University of Rochester Medical Center (URMC) is a leading academic medical center located in Rochester, NY. With over 26,000 employees, URMC is the region’s largest employer. Within URMC, the orthopedics department manages a high patient volume across inpatient and ambulatory settings.However, URMC orthopedic surgeons were spending excessive time on EHR documentation, which hampered productivity and increased frustration.

Methods

The goals were two-fold:

| Reduce the documentation burden for orthopedic surgeons to improve job satisfaction and productivity

| Enable greater focus on patient care and improve the patient experience by eliminating typing or dictating into the EHR during exams.

By using Knowtex’s voice-driven documentation capabilities, URMC aimed to optimize clinical workflows for orthopedic care. Knowtex was deployed to a cohort of orthopedic providers across different sub-specialties to measure usage of the platform. Provider interviews were conducted to collect high-quality qualitative feedback.URMC’s Signal data from Epic provided detailed provider-level data on In Basket,Orders, Notes & Letters, and ClinicalReview use in Epic.

Results

The implementation of Knowtex at URMC yielded promising results regarding the platform’s significant impact on physician time savings, patient focus, and overall documentation quality.

1. Yielded significant time savings for physicians.

    a. Pre-survey, providers reported spending at least 15 minutes per patient to write their notes. Post-survey, generation of the note was completed in under 1 minute on average upon completing the patient encounter.

2. Documentation quality exceeded provider expectations, and improved as providers used it.

    a. Knowtex’s proprietary machine learning is trained on significant medical specialty data and uses provider feedback to create increasingly accurateand provider-specific notes over time. Note accuracy improved from 75%to 90% based on physician edits.

3. Improved patient focus and promoted the patient-provider interaction.

    a. Most providers reported not taking notes during the visit while using Knowtex and almost all patients welcomed its use.

4. 100% utilization by all providers in the pilot program, who recommended expansion to colleagues, resulting in 70% adoption rate in two years.

URMC's Signal Data

URMC's signal data from Epic⁵ was reviewed and analyzed to measure pre/post Knowtex use. This report reviews provider performance before and after the use of Knowtex to gain an understanding of provider efficiency through a Quality Improvement (QI) lens.

-29% decrease in pajama time.

-1300 in character count in notes on average, reducing note bloat

-10% decrease in time in notes per note

AI Medical Coding

Knowtex automatically suggests applicable ICD-10, CPT, and E/M codes from thenatural clinician-patient conversation to accurately capture diagnosis, procedures, and visit complexity for quick review and approval by the provider. During the URMC pilot, Knowtex's coding functionality analyzed the context of the AI documented patient encounters and identified significant potential value for the orthopedic department.

+92,800 per month in capture of potential revenue leakage.⁶

1. Knowtex generates smart phrases based on procedures completed during the visit that are optimized for medical coding.

    a. If the provider reviews and interprets X-rays during the visit in the presence of the patient, Knowtex will generate the phrase, “x-rays from today's visit were personally, reviewed, interpreted, and showed.”

2. Worker’s compensation notes are thoroughly captured, complete with date of injury and working status.

3. The patient’s pertinent medical history is automatically included within the note and all addressed issues are coded appropriately.

4. A high level of detail about the visit is captured.

    a. Medication counseling, surgical discussion including risks and recovery timelines.

5. Personal Review of Data Section.

    a. This section is auto-generated to include lab testing (HbA1c, CBC, Lipid Panel, etc.) as well as outside imaging to support generation of billing codes.

Physician Testimonials

Dr. Digiovanni, MD - Foot & Ankle Orthopedist

What is your favorite thing about Knowtex?

I am energized by it. I am over 60 and it can keep me going longer as a provider, as you can be tired from typing in your 60s. There is lots of potential to keep physicians engaged longer in their careers. I think for my junior partners and my daughters in medical school, if they can use them in practice probably less time in notes over their career.

How has your experience using Knowtex changed during the pilot?

It’s been great, getting smarter. The adjustments being made are helping, follow-ups, notes are segmented more so easier to read, interpretation of x-rays and describing it is much more how they would say it instead of broken bones...I’ve noticed that it’s really good at the history and exam impression. Overall its been great, I am loving the experience over time.

How have your patient interactions been affected by Knowtex?

I feel as if it is it returned the joy of connecting with patients in the room and not being distracted. I think the patient felt that and I also felt that.

Hours Spent on Documentation (weekly)

Before Knowtex: 10 hours    Using Knowtex: 2 hours

Dr. McDowell, MD - Orthopedic Oncologist

How quickly are you closing notes after a patient encounter?

Since I started using Knowtex, I've been able to close my clinical notes much more efficiently. For a new patient visit, it used to take me about 20-25 minutes to manually dictate a note. But with Knowtex, it cuts my documentation time down by half, if not more than that. On days whenI'm using Knowtex consistently for all my patient visits, I'm now able to close all my notes by the end of clinic. The time savings have been really significant and make such a difference in my productivity and workflow.

What do patients think about Knowtex?

With patients, there's been no issues, most of them say oh that's awesome, glad we're using it.

Hours Spent on Documentation (weekly)

Before Knowtex: 9 hours   Using Knowtex: 1 hour

Dr. Gruttadauria, DPM - Podiatrist

Is it easy to get started with Knowtex?

In my experience it’s not complicated to use. You just hit record and start going.

How have you integrated Knowtex into your workflow?

I dictate the exam a lot. I bring Knowtex into the room for HPI, and will do [the] exam while patient is talking to me.  Then when the I leaves the room, I will quickly dictate the physical exam, and any imaging findings so it basically will get the HPI, assessment, and plan. I often see more than 40 patients a day and Knowtex has saved me an immense amount of time.

What do patients think about Knowtex?

I have not had any patients say no or complain about using Knowtex. I've had a few who didn't fully understand what I was asking them at first when I explained Knowtex. But most patients are pretty fascinated by it and think it's neat technology. I've even had some patients who were fascinated enough that I told them to go home and read the note Knowtex generated from our visit, and send me feedback on it. I let them know that if the note doesn't reflect what we talked about, they should let me know. So far I have not gotten any negative feedback from patients on the Knowtex documentation.

Hours Spent on Documentation (weekly)

Before Knowtex: 10 hours    Using Knowtex: 2 hours

Conclusion

Knowtex implementation at URMC demonstrates evidence that the ambient AI documentation and coding solution can benefit physicians, patients, and healthcare organizations.

Knowtex captures clinician-patient encounter audio and generates real-time, accurate documentation with codes. The automation of this administrative work results in significant time savings and productivity gains to potentially see more patients or perform more procedures. With Knowtex, URMC physicians describe being fully present during visits and have 8 additional hours of personal time each week. URMC’s Signal data from Epic shows a 29% decrease in pajama time, saving clinician time after hours for reduced burnout and improved wellness.

With automated coding functionality, Knowtex is also able to optimized note phrasing and suggest the most specific and accurate medical codes to capture diagnoses, procedures, and visit complexity. Analysis showed that Knowtex identified potential for significant revenue gains for URMC orthopedics.

A common issue with healthcare technology is difficulty for health systems to adopt new platforms, however, Knowtex has demonstrated high provider satisfaction and strong provider retention for 2 years. The clinical quality and accuracy of the solution also exceeds user expectations and is able to customize further by medical specialty and individual clinician style.

By leveraging cutting-edge AI and machine learning, Knowtex delivers streamlined clinical documentation workflows, improved physician and patient experiences, increased provider capacity and revenues, and enhanced documentation quality, to provide significant value that can be scaled across health systems.

References

1. https://www.medscape.com/slideshow/2020-lifestyle-burnout-6012460#1
2. https://pubmed.ncbi.nlm.nih.gov/27595430b/
3. https://www.zippia.com/medical-scribe-jobs/demographics/#employment-statistics
4. Accuracy of medical notes was determined using the rolling average similarity score. Higher similarity scoresindicates that the doctor’s final submitted note had very few edits to the Knowtex-generated medical note.
5. Signal Data User Guide, available at https://signal.epic.com/Documentation/MetricReference#IntroDiv. Epic is a registered trademark of Epic Systems Corporation
6. Revenue leakage per month identified by Knowtex: 16,000 total appointments per month * 10% appointments with undercoding * approx. $58 revenue leakage per appointment = $92,800 revenue leakage per month
7. Orthopedic Surgeons Salary in Rochester, NY | Medscape Salary Explorer