Perspectives

Ask the Doctor: Today’s Technology Frustrations and Opportunities

Dr. Prateek Sharma
Topics
Share Article

Ask the Doctor: Health2047 advisor and practicing physician Dr. Christine Stock has embarked on this direct input gathering project. Her mission is to assure that physicians inform everything Health2047 does and that the physician’s perspective remains central in developing companies with scalable solutions aimed at our biggest healthcare challenges. She recently spoke with Dr. Prateek Sharma, who is a gastroenterologist, professor of medicine, and director for the gastroenterology fellowship program at the University of Kansas School of Medicine. They discussed unmet challenges and opportunities facing practicing physicians, and their conversation inspired in this blog post.

Today’s Technology Frustrations and Opportunities

The technology we use in the practice of medicine and the administration of healthcare has obviously evolved very quickly in recent years. As a gastroenterologist, I naturally view the attendant frustrations and opportunities through the lens of my specialty, but they undoubtedly apply to many disciplines.

EHR frustrations
For example, our current EHR systems are not really geared toward useable, safe, and high-quality data, especially when it comes to gastroenterology. There are still far too many routine needs that require manual data pulls, either by the physician or by physician extenders. It’s a pervasive frustration.

A good example of a useful, quality metric in gastroenterology is adenoma detection rate. Determining how many adenomas or precancerous polyps you are finding during your colonoscopy procedures gives you a measure of how good you are at performing a colonoscopy.

Calculating the rate requires combining information on what you detected during colonoscopy with subsequent details from pathology reports on the polyps that you removed. The EHR, although it may contain that information, provides no process to explore that metric. If I wanted to, say, find out how many adenomas or how many cancers were diagnosed during my procedures during a given time period — that information cannot be readily pulled from the EHR. Such capabilities might seem simple and obvious with all the other advances we’ve had in IT generally, but it ends up being quite a bit of a challenge in EHRs.

Something as rudimentary as checking for vaccination records (say, for our inflammatory bowel disease or cirrhosis patients) still requires laborious chart pulls and/or a designated healthcare worker do it — there’s not an easy, intuitive, or automated way of extracting the information from our systems. These are the types of frustrating challenges that we face in terms of EHR data quality, access, and function — essentially, there should be some intuitive way to query your own database.

When compared to consumer technology, the average EHR pales painfully. Beyond being able to easily perform useful queries, where are the handy features like reminders and automatic data flows on quarterly, semiannual, or annual bases (say, aligned with institutional certification or CMS requirements)?

It seems to me that those kinds of pulls could be easily implemented, as they are in social media platforms. Consider Facebook: When you initially log in, you enter your date of birth, and every year Facebook wishes you happy birthday. Is there any reason why similar but more constructive professional alerts couldn’t be built into automated EHR features?

With EHRs now, everything you do — even a brief phone call with a patient — is documented. So there’s a lot of physician effort devoted to feeding information to the EHR in documentation and note creation. But in return, the EHRs are very passive systems that don’t allow for some of these practical, active, or interactive ways of using all that information built into the software.

And when it comes to using data across different EHRs, the situation is even more grim. For referrals, for example, unless you’re using the same electronic record system, the flow of patient data is anachronistic — we’re talking paper and faxes! I still get quite a bit of my records from referred patients via snail mail. You’d think that file sharing among referring physicians through email or via Dropbox or similar technologies, in a HIPAA-compliant and confidential way, would be a standard capability in this day and age. But no.

Technological opportunities
While EHR and data exchange conditions in modern practice may leave a lot to be desired, not all of our technologies are stuck in the stone age.

A major improvement that I see emerging is the use of artificial intelligence in endoscopy. Of course, I’m emphasizing endoscopy, but I think AI will also have a major impact in many specialties and healthcare functions.

For example, we just submitted a protocol to our IRB on using machine-learning software applied during our colonoscopy procedures for enhanced automated detection of colon polyps. The concept is that there is an algorithm embedded in a software box attached to the processor and the endoscope. You still have your regular monitor while you’re doing the procedure, on which you can be looking at the screen and trying to find polyps with your own eyes as per usual. But then there’s a second screen right next to it, which is running the software and algorithm in parallel, collecting and processing information as the computer learns to identify the polyp circles from the process and assist in detection.

We’re just starting to see the additive value of machine learning in improving polyp detection. There’s data to show that cancers can be visually missed as well, especially when they’re early, subtle cancers. Machine learning can help.

You can see how this could be applied to all procedural specialties just as in gastroenterology. As the box is being continually taught to recognize precise areas of interest, very subtle abnormalities, T1 cancers even before they can metastasize — it can increasingly and more reliably help diagnose and assist during the procedure.

And to take it a step further, once we see something abnormal, we generally resect or remove it. Machine learning can elevate our effectiveness here as well. Sometimes, we don’t treat the affected area completely for multiple reasons, such as cautery artifact or an incomplete polypectomy, etc. I foresee machine-learning enhanced systems being able to tell us, “Hey, you missed this area here,” or “This cancer occupies a surface area of 4.8 square millimeters and you’ve removed only 4.6.”

Bridging the gap
Pragmatic technology designed to automatically assist — this is what practicing doctors long for. Whether it’s Siri or Google or Alexa, we need to take the types of consumer technology that we already use at home and start applying them to medicine.

While I’m seeing a patient or working in my office, I should be able to just communicate with the health system and say, “Hey Computer, pull up the patient’s last endoscopy report and read it to me; tell me if the patient had a polyp or not.” Or “Hey Computer, pull up the last patient’s last polyp picture on my desktop screen.” Or “Hey Computer, when did the patient last have vaccinations.”

This sort of real-time, interactive technology could really ease my workflow and it could also transform the patient experience. You could use it while consulting with your patient, looking at images together, actually using the information. The patient wouldn’t have to go to some “release of information” area or fetch their health records to be fully informed. Doctors and patients would be constructively served information for use by the system, instead of the other way around.

Moving into really futuristic innovations that are reasonably possible, there may come a day when capsules swallowed by patients would be able to trace a path from mouth to rectum and identify polyps, cancerous tissue, or even take samples and perform complete scans. Or maybe you’ll eventually just do a simple blood test that will return risk probabilities and prevention advice.

But in the here and now, there’s a lot more our current technological capabilities can be contributing to the standard practice of medicine and efficient function of our healthcare system. Physicians today face plenty of frustration with standard healthcare technologies, but that means there’s also plenty of opportunity for improvement.

— Dr. Prateek Sharma is a gastroenterologist, professor of medicine, and director of the gastroenterology fellowship program at the University of Kansas School of Medicine at Health2047.

Powering insights

Case Study: Health2047’s Model in Action

Podcast: So You Want to Transform Healthcare

HEAL Security is a cognitive cybersecurity intelligence platform custom-designed for the healthcare sector.

Founders

Charles Aunger

Milestones

Founded 2021

WEBSITE
Social

HOPPR is transforming healthcare diagnostics with AI-driven medical imaging technology.

Founders

Khan M. Siddiqui, MD
Oliver Chen, MD
Robert Bakos
Gerry Stegmaier

Milestones

Founded 2018

WEBSITE
Social