Reflection - Finale: A Piece of Our Minds

From Xiruo:

Telemedicine, in this very special time ever, becomes a topic that every caregiver is considering to be the future, at least in part. This is a big topic that basically includes every aspects of clinical informatics: EHR, information system, health information exchange, management, deployment, analysis, and so on. It is probably also a good time to reflect on our current healthcare systems, like whether they’re working (well clearly not very well) in this pandemic, whether some new norms should be introduced for telemedicine, and what effects it will bring to traditional healthcares. For all the topics we’ve covered for this primer, hopefully it could help address some of the problems in this whole giant complex system.

From Yue:

Telehealthcare refers to the use of mobile communication technology to provide medical services and information to the general publication. Compared with traditional medical care, telemedicine optimizes the allocation and utilization of medical resources, and cooperates among patients, doctors, and hospital administrators to save time and money. China already has over 1,000 telehealth companies, according to data firm Tianyancha, including some run by tech giants JD.com, Baidu, Tencent and Alibaba. The main telemedicine applications in China includes appointment making, health consultation, prescription service, chronic disease management, and electronic health record. However, there are still some limitations of applying telemedicine in China right now.

Due to the uneven medical resources and remote geographical location, telemedicine is the most in-demand technology for patients in rural area. However, the main obstacle coming with this in China rural area is the limited access to the laptop, cell phone, or internet, which is accompanied with the low economic status. Therefore, instead of solving the accessing of telemedicine, it is crucial to reduce poverty and increase the technology accessing first. While the coronavirus has stretched medical services around the world to breaking point, the virus has also fostered a boom in telemedicine. The industry is predicted to be worth almost $30 billion this year in China alone. It has the potential to transform Chinese healthcare by reducing strain on urban hospitals and providing a stop-gap solution for rural dwellers.

Although telemedicine has been expected to reach out to people as many as possible, it is still a tough issue to educate people that telemedicine is an effective and safer alternative. Before the pandemic, JD Health took 10,000 online consultations per day. But as hospitals and clinics became swamped with suspected coronavirus patients, that has rocketed to 150,000, with JD Health’s pharmacy delivering medicines directly to patients’ homes. The Good Doctor subsidiary of Ping An Insurance, which claimed in September to have 300 million registered users. All are seeing a boom in consultations due to lockdown measures. Thanks to the coronavirus, people have developed the habit of getting diagnosis and treatment online, which significantly reduces the pressure on traditional hospitals.

But problems persist, especially as a severely aging population increases the demand for treating chronic conditions such as arthritis, cancer, and heart disease. China has only 1.8 doctors for every 1,000 people, compared with 2.4 in the U.S. and 2.8 in the U.K. Compounding matters, China’s doctors are unevenly weighted towards specialties to the detriment of primary care. Besides realizing the low availability of doctors in China, understanding the willingness to use telemedicine service in the older population is necessary to expand that service. According to the Telehealth index: 2019 Consumer Survey, the elderly population more aged than 65 years-old would like to integrate telehealth service into health plan apps and favorite the idea to have a video visit for a prescription renewal. More studies are needed to figure out the needs of Chinese older adults population in telemedicine.

Still, significant challenges remain. The biggest challenge of telehealth comes from diagnosis, which often requires specialist equipment. It is an emerging industry and may take a long time to solve those problems.

From Jake:

"We are drowning in information while starving for wisdom. The world henceforth will be run by synthesizers, people able to put together the right information at the right time, think critically about it, and make important choices wisely."

E. O. Wilson

In my brief career in informatics I’ve come to come to a few realizations that I’ll talk about in this piece. In my previous Op-ed I wrote how the clinician will drive the field of informatics and the integration of natural language processing in informatics. In this op-ed, I will not only share my updated view on that, but also focus on the informatician and the role of Natural Language Processing for the future of medicine.

The first time I wrote an op-ed was after taking a course with healthcare workers such as doctors and nurses. The course was titled Introduction to Health Informatics and Systems Thinking and it introduced me to the field of health informatics. This course also completely changed my perspective on the field of informatics. I initially thought that large retrospective cohort studies would be done on the EHR by statisticians. This is not the case. It became clear to me that there was a big obstacle in the way of the integration of statistics into healthcare. The obstacle was the trust in statistics by physicians. This isn’t by any means the physicians fault. They’ve been sold pipe dreams of AI since the inception of computers and these promises have consistently failed them. I argued before, and I’d like to continue to argue, that the new wave of AI in medicine will be centered around doctors. I firmly believe that the informatician, biostatistician, or computer engineer is simply the point-guard, distributing information to the most talented on the front lines.

While the clinician may be the integral factor in the implementation of statistical analysis for tasks such as clinical decision support, informatics is so much more than that. After taking all the core courses in the Biomedical and Health Informatics PhD program at the University of Washington, I have first hand witnessed the brilliance of my fellow informaticians. While I initially argued in my first op-ed that all informaticians would have to kowtow to the physician, I no longer believe this is the ideal solution. In the quote above it states, “the world henceforth will be run by synthesizers”, and I believe that is the key function of the informatician. The informatician still needs to put information in the right hands, but to wait for the entirety of clinicians to understand machine learning and natural language processing on their own is unrealistic. The growing role of the informatician will be to analyze, synthesize, and then teach their findings to multiple generations of clinicians who may not have been trained in mathematics and computations. Informatics is a team sport, but it may not be like basketball like I initially anticipated. While I still see the informatician as a facilitator, I now see them as the quarterback, driving the innovation and working with the dirty mechanics of computers and machine learning to create elegant tools and algorithms that can be implemented. I’ve come to this conclusion because I’ve been surrounded by talented informaticians for two years and as the amount of data and technology rises, so will the skill levels of the next generation of informaticians.

The aspect of medical informatics that I’m most excited to see and focused on in this primer is the future of Natural Language Processing. Informatics is not just natural language processing and there were several topics that I learned about in the course that didn’t generalize to those performing NLP such as security and managing change. However, the vast majority of health-related data is text. It’s clear to me after writing this primer that NLP will be at the forefront of innovations around the Electronic Health Records, Clinical Decision Support, Analytics, and Telemedicine. In these cases, the data are complex and so are the methodologies. Natural Language Processing applications is a key example of how the informatician is going to have to apply extensive knowledge while also distributing the findings of this in an explainable way to multiple physicians. While there are several aspects of an informatician's job, the natural language processing engineer will be a key role that will be driven by the informaticist in the field of medicine.

All in all, I want to convey that the increased use of technology in medicine has created a demand for interdisciplinary teams that possess a NLP engineer. While I initially attempted to tell you this approach would be driven by physicians, I now realize it’s unrealistic to expect physicians to master an orthogonal field. Instead, the informatician will have to drive innovation in NLP and garner trust from physicians who may not be predisposed to trusting statistics and machine learning. After going through the Biomedical and Health Informatics PhD program with multiple talented cohorts, I believe there are plenty of budding informaticians poised to create new tools and insights from language that will garner trust and broach the medical field. The informaticians will be the synthesizers that EO Wilson was referencing, putting together the right information at the right times so physicians and nurses can make informed decisions from data.