Electronic health records are widely adopted in the hope of saving time and improving the quality of patient care. However, due to the fragmented interface and cumbersome data entry procedures, physicians often spend more time navigating these systems than interacting with patients.
Researchers at MIT and Beth Israel Dicones Medical Center are combined Machine learning When Human-computer interaction Create a better electronic health record (EHR).They have developed MedKnowts, a system that integrates the search process. Medical record Document patient information in a single interactive interface.
Driven by artificial intelligence, this “smart” EHR automatically displays customized patient-specific medical records when clinicians need them. MedKnowts also provides auto-complete for clinical terms and auto-fills the fields with the following information: Patient information To help doctors work more efficiently.
“In the origin of EHR, there was this tremendous enthusiasm that organizing all this information would help track billing records, report statistics to the government, and provide data for scientific research. Few people stopped asking deep questions. I think many clinicians feel that this EHR is burdened for the benefit of bureaucrats, scientists and accountants. We participated in this project asking how EHR could actually help clinicians, “said a professor of computer science at the Computer Science and Artificial Intelligence Laboratory (CSAIL). This is David Karger, the senior author of the paper.
This study was co-authored by CSAIL graduate students and lead authors Luke Murray, Divya Gopinath, and Monica Agrawal. Other authors include Stephen Horn, Emergency medical care David Sontag, Principal Doctor and Clinical Leader in Machine Learning at the Healthcare Delivery Science Center at the Beth Israel Deaconess Medical Center, and MIT Associate Professor of Electrical Engineering and Computer Science, and a member of the CSAIL and Institute for Medical Engineering and Science. It will be announced next month at the Association for Computing Machinery Symposium on User Interface Software and Technology.
To design an EHR that benefits doctors, researchers had to think like doctors.
They created a note-taking editor with a side panel that displays relevant information from the patient’s medical history. The historical information is displayed in the form of a card that focuses on a particular problem or concept.
For example, if MedKnowts identifies the clinical term “diabetes” in the text as the type of clinician, the system will automatically “diabetes card” containing medications, laboratory values, and excerpts from past records related to diabetes treatment. Is displayed.
Most EHRs store historical information on separate pages and list drugs or test values alphabetically or in chronological order, so clinicians need to search the data to find the information they need, Murray said. Mr. says. MedKnowts only displays information related to a particular concept written by the clinician.
“This is close to how doctors think about information. Often, doctors do this unknowingly. Doctors look at the medication page and focus only on medications that are relevant to their current condition. We help you do the following automatically, and hopefully move some things out of the doctor’s head to determine what’s wrong with the patient and plan treatment. You can spend more time thinking about the complex part of standing up, “says Murray.
Interactive text, called chips, acts as a link to the associated card. When the doctor enters a note, the autocomplete system recognizes clinical terms such as medication, test values, and condition and converts them into chips. Each tip is displayed as a word or phrase highlighted in a specific color, depending on the category (red for medical conditions, green for medicines, yellow for procedures, etc.).
With autocomplete, structured data about the patient’s condition, symptoms, and drug use is collected without the additional effort of a physician.
Sontag said he hopes that this advance will “change the paradigm of how to study the progression of the disease and create large health datasets to assess the actual effectiveness of treatment.” increase.
After a year of iterative design process, researchers deployed the software to the emergency department at the Beth Israel Deaconess Medical Center in Boston to test MedKnowts. They worked with an emergency physician and four hospital writers to enter notes in electronic health records.
Agrawar said that the introduction of software into the emergency department, where doctors perform surgery in high-stress environments, required a delicate balance.
“One of the biggest challenges we faced was trying to get people to change their current behavior. Doctors who used the same system and performed the same click dance many times were a kind of It forms muscle memory. I wonder if it’s worth it to make changes, and I’ve definitely found that some features are more utilized than others. “She says.
The COVID-19 pandemic also complicated the deployment. Researchers were visiting the emergency department to understand the workflow, but were forced to end those visits due to COVID-19 and were unable to stay in the hospital while the system was deployed. bottom.
Despite these first challenges, MedKnowts was popular with writers during the month of deployment. They gave the system an average rating of 83.75 (out of 100) for ease of use.
According to research results, Scribe found that the autocomplete feature was especially helpful in speeding up work. The color-coded chips also helped us quickly scan notes for relevant information.
These early results are promising, but researchers will consider the feedback and proceed cautiously to work on future iterations of MedKnowts.
“What we’re trying to do here is to smooth the path of doctors and accelerate them. There are risks. Some of the goals of bureaucracy are to slow things down and everything. Make sure i is scattered and all t intersect .. And if you have a computer that intersperses i and intersects t for the doctor, it is actually that the doctor It may be against the bureaucracy’s goal of having people think twice before making a decision. It protects doctors and patients from the consequences of making them more efficient, “says Karger.
Researchers are planning to improve the machine learning algorithms that drive MedKnowts so that the system can more effectively emphasize parts of the most relevant medical records, Agrawal said. increase.
We also want to consider the needs of different medical users.Researchers use MedKnowts Emergency department Keep in mind-a setting where a doctor usually sees a patient for the first time. A primary care physician who knows the patient better may have several different needs.
In the long run, researchers envision creating adaptive systems that clinicians can contribute to. For example, a doctor notices that a particular cardiovascular term is not in MedKnowts and adds that information to the card. This will update the system for all users.
The team is looking for commercialization as a means of further development.
“I want to build a tool that allows doctors to create their own tools. I don’t expect doctors to learn to be programmers, but with the right support, I want to build my own medical application. It has the potential to be radically customized to meet your needs, “says Karger.
Massachusetts Institute of Technology
This story is republished in courtesy of MIT News (web.mit.edu/newsoffice/), A popular site that covers news about MIT’s research, innovation and education.
Quote: AI-enhanced system reduces doctors spending time searching for clinical information, obtained from https://medicalxpress.com/news/2021-09-ai-enhanced-enables on September 23, 2021 You can spend more time treating the patient (September 23, 2021). -doctors-clinical-patients.html
This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.
The AI-enhanced system allows doctors to spend less time searching for clinical information and more time treating patients.
Source link The AI-enhanced system allows doctors to spend less time searching for clinical information and more time treating patients.