6 Ways Machine Learning-Enabled Virtual Assistants Can Serve Doctors and Patients

With the advancement of healthcare software development, many futuristic applications have become possible. One such application is virtual medical assistants alleviating the pressure of medical staff. Healthcare professionals are starting to experience burnout due to the increasing number of patients and administrative tasks. This is largely due to the aging population and healthcare conditions such as a shortage of healthcare professionals and widespread chronic diseases.
 
Technology-related factors such as availability of smartphones and internet access have also contributed to this virtual medical assistance trend. According to ResearchAndMarkets, the virtual medical assistant market is expected to grow to $1.73 billion at a CAGR of 34.6% from 2019 to 2024. 
 
Preventing Doctors from Becoming Data-Entering Clerks
Doctors always take notes while examining patients. They also need to access patients’ health records and process large amounts of data. These tasks steal doctors’ time away from patients and drown them in administrative work. Doctors tend to spend two hours in front of the computer screen for every hour spent with patients. As a result, doctors are slowly becoming data-entering clerks.
 
This is where virtual assistants come into play.
 
1. Transcription
Machine learning-powered virtual assistants can be trained to perform transcriptions of conversations between doctors and patients. Such a program works without the need for special dictation or prompting words. This assistant delivers a fully transcribed and formatted clinical note, which is then reviewed and signed by the doctor. Afterwards, it is integrated into the patient’s electronic health records.
 
For example, Robin Healthcare has developed a medical transcription assistant, Robin, that uses machine learning to document useful clinical information based on a real-time dialogue between the doctor and the patient. To use this transcription assistant, doctors need to place the speaker device in their practice room. As they interact with patients, the device records the sounds while the built-in algorithm performs the transcription.
 
2. Medical Record Keeping
Medical record keeping is another time-consuming, tedious task to be automated with virtual assistants.
 
Nuance, a software company based in Massachusetts, has developed a medical assistant that relies on machine learning and voice recognition to automate clinical record documentation. This assistant is EHR-integrated and helps doctors navigate patient records through a voice command interface.
 
Suki is another virtual medical assistant powered by machine learning. Suki can be customized on demand to suit every doctor’s needs. When Suki was piloted in 2018, it was integrated into three EHR systems. Preliminary results showed that it reduced the amount of time doctors spent on paperwork by 60%. This led to more job satisfaction and more accurate records at once.
 
3. Assistance in Radiology
While coming up with a diagnosis, radiologists are often overwhelmed with the amount of data they receive from different medical data sources. There is a lot of processing and analysis involved to produce the required radiology report.
 
Nuance has developed PowerScribe One virtual medical assistant using a machine learning and human language processing platform. This assistant converts unstructured text registered from a doctor’s voice into structured data. PowerScribe One helps radiologists at different stages of their workflows, enabling data integration and lookup.
 
“We are committed to transforming the practice of diagnostic imaging with products, such as PowerScribe One, that help radiologists amplify their expertise and demonstrate their value through their primary work product—the diagnostic report,” said Karen Holzberger, vice president and general manager of Nuance’s Healthcare Diagnostics.
 
4. Virtual Medical Assistants for Patient Wellbeing 
Patients who will benefit the most from this type of virtual assistants are the elderly, patients with chronic diseases, and those in rural areas where access to care is limited. Another important aspect is that patients can feel less embarrassed and more honest when sharing private details with a virtual assistant rather than with a human out of the fear of being judged.
 
5. Virtual Nurse Helps with Recovery and Discharge
According to Harvard Business Review, virtual nursing assistants will save $20 billion annually on healthcare costs by 2026, by reducing the time nurses need to spend on their patients.
 
Financial gain is not the only benefit of using virtual nurses. The nursing staff is often overloaded with responsibilities, which harms patients’ wellbeing. When a nurse’s workload exceeds the optimal level, the chances of patient safety incidents increase by around 30%. By automating repetitive tasks, virtual nurses bring relief to human staff.
 
A virtual nurse can monitor post-surgery patients to make sure that their condition does not aggravate. It can collect vital signs, answer patients’ questions, and schedule an appointment with the doctor. Virtual nurses are available 24/7 and are programmed to show empathy. As they are based on machine learning algorithms, they can recognize when the recovery process goes wrong and alert doctors immediately. 
 
The discharge nurse is a popular type of virtual nurse. Many hospitals have to deal with unnecessary readmission cases which can be avoided if the discharge process is improved.
 
Elizabeth, a virtual discharge nurse, guides patients through the discharge process using voice and animation. Patients can participate in the conversation by selecting options from the menu. While a human nurse would spend around eight minutes on the discharge process, the virtual assistant can dedicate from 30 minutes to an hour. At the end of the conversation, Elizabeth prints the list of action points for the human discharge nurse to elaborate upon with the patient in a quick conversation. Elizabeth was developed at Northeastern university and trialed at Boston Medical Center.
 
6. Healthcare Companions  
Healthcare companions offer medical assistance to patients from the comfort of their home. After discharge, a healthcare companion calls in to check whether patients are following their prescriptions and stick to the recommended lifestyle.
 
Additionally, patients can share their symptoms with a health companion, and it will suggest a diagnosis to be verified. This is faster and more convenient than booking an appointment with the doctor at the initial stage. It also gives more accurate results compared to the internet search. One example of such a companion is Ada. Ada is used in 130 countries and has a reliable database of symptoms and diagnoses.
 
Virtual Assistant in Your Practice: Benefits and Factors to Consider
Virtual medical assistants offer a new opportunity for healthcare institutions to reduce the costs of human capital by automating trivial and repetitive tasks. These assistants help medical personnel by taking over non-care related tasks, freeing doctors for what they are trained and qualified to do. Simultaneously, virtual assistants help patients by providing a more convenient medical care, projecting unconditional empathy and patience.
 
Virtual assistants can help your practice tremendously. However, there are factors you need to consider while pursuing this avenue:
 

• Make sure everyone is comfortable: practitioners might feel being directed by their virtual assistants, and patients might worry about their health being monitored by machines.
• Enforce security measures: hackers might target your electronic health record system, and country-specific data privacy regulations must not be violated.
• Educate your patients: not every patient is tech-savvy, and some will need more guidance than others.
• Make sure technology and infrastructure are reliable: if a patient loses the internet connection, they cannot access their medical assistants.

 
Virtual assistants were invented to bring relief to medical personnel and to fulfill the need of those who cannot conveniently access healthcare. As these assistants are gaining ground in the healthcare sector, some medical professionals fear being replaced by virtual assistants in the future. However, imagining that a chatbot would completely replace humans is far-fetched even for inventors of those chatbots.
 
According to Adam Odessky, co-founder and CEO of Sensely, “the whole value of human care is a special sort of relationship that a machine can’t replace. We try to respond in a human way, but not overly human. We know we’re a bot. But we communicate in a way that’s reassuring to patients. When you feel sick, you want to be reassured.”

---

Nadejda Alkhaldi is an innovation analyst at Itransition. She obtained a Ph.D. in Business Economics from Belgium and worked in three different countries. Currently, Alkhaldi specializes in machine learning applications within different industries.