Srabani Roy Chowdhury (MAKAUT, WB)
The introduction of electronic health records has risen the hope of saving time and improving the quality of patient care. But due to fragmented interfaces and hectic procedures involved in data entry, doctors and other medical staff require more time for navigating these systems than interacting with patients.
Machine Learning in EHR:
Researchers are working on the development of combining machine learning and human-computer interaction. This will create a better electronic health record (EHR). MedKnowts is one such developed system that helps in the unification of looking up medical records and documenting patient information processes into a single, interactive interface. This “smart” EHR is driven by artificial intelligence. It provides automatic displays of medical records that are customized and patient-specific.
Salient features of MedKnowts:
To help the doctors to work more efficiently, MedKnowts also provides autocomplete for clinical terms and auto-populates fields with patient information. To help the doctors work efficiently, the researchers had to think like doctors to design an EHR. There is a note-taking editor which has a side panel. It displays relevant information related to the patient’s medical history. That historical information appears in the form of cards. Each of these cards is focused on a particular type of disease or problem. For example, MedKnowts can identify the clinical term “diabetes” in the text from a clinician or doctor’s typing. On identification, the system automatically displays a “diabetes card”. It contains medications, lab values, and past records that are relevant to diabetes treatment.
However, most of the EHRs store historical information on separate pages and there is an alphabetical or chronological list of medications or lab values. This helps the clinician in searching through the data to find the information they need. MedKnowts is capable of displaying only the information relevant to the particular concept the clinician is writing about. Chips are pieces of interactive texts. They serve as links to related cards. When a doctor or clinician is typing a note, the autocomplete system recognizes clinical terms and transforms them into chips.
The clinical terms can be medications, lab values, or conditions. Each of those chips is displayed as a word or phrase. These phrases are highlighted in a specific color relating to a specific category (for example, a medical condition is highlighted in red, a medication is highlighted in green, etc.) Autocomplete helps in collecting structured data on the patient’s conditions, symptoms, and medication usage without any additional effort from the physician.
Applications of MedKnowts:
After a year-long design process, MedKnowts was tested by the researchers. The software was deployed in the emergency department of Beth Israel Deaconess Medical Center in Boston. They worked with a physician and four hospital staff who made data entries into the EHR. Due to the high-stress environment in the emergency, the working pattern had to be delicate enough. However, this setup got complicated due to the COVID-19 pandemic. The regular visits of the researchers were forced to an end because of the pandemic which affected their workflow inspection. Despite these challenges, MedKnowts became a very popular EHR. According to the researchers, there’s always room for improvement for the machine learning algorithms that run MedKnowts to highlight relevant medical inputs more effectively. As MedKnowts was particularly designed for an emergency department where doctors see patients for the first time, researchers want to modify it for different medical uses.
Conclusion
In the long run, researchers plan to create a better adaptive system where doctors and clinicians can also contribute. For example, if a doctor realizes a certain cardiology term is missing from MedKnowts or feels the need to add any relevant term, he can add that information to a card. This would update the system for all medical users.
Also read: Tetrachromacy: A World of 100 million colors
Reference:
- Toward a smarter electronic health record. (n.d.). MIT News | Massachusetts Institute of Technology. Retrieved September 25, 2021, from https://news.mit.edu/2021/medknowts-electronic-health-record-0923
- MedKnowts: Unified Documentation and Information Retrieval for Electronic Health Records: https://arxiv.org/abs/2109.11451
- The Corrosion Prediction from the Corrosion Product Performance
- Nitrogen Resilience in Waterlogged Soybean plants
- Cell Senescence in Type II Diabetes: Therapeutic Potential
- Transgene-Free Canker-Resistant Citrus sinensis with Cas12/RNP
- AI Literacy in Early Childhood Education: Challenges and Opportunities
One thought on “MedKnowts, an AI enhanced system – A boon to the doctors”