As you probably know, machine learning refers to the process through which a system can use a programmed process to "learn" and become more accurate over time. It's different from the way most of us think of a robot or a computer program because in this case, the programming changes based on the results it gets and new inputs available.
Experts say that the extensive use of machine learning in healthcare is inevitable. But how might that affect healthcare and what will the consequences of it be?
Machine learning is not a new field -- in fact, it was described by Arthur Samuel in 1959 as, “the field of study that gives computers the ability to learn without being explicitly programmed.” But exponential improvements in computing, plus innovations in AI and technology in general, have meant that the field has started burgeoning over the last few years.
Machine learning is used in applications from cyber security to healthcare to, well, the movies that Netflix suggests you watch after a long week. In medicine, we're seeing applications in everything from cancer prediction and prognosis, to diagnosis in medical imaging, to treatment suggestions.
For instance, when you search for, say, peach pie recipe on Google, you'll see search results based on what Google thinks matches those words. That's a regular computer algorithm. Machine learning comes into play, however, in that Google also "watches" (actually a computer collecting data) which search results you click on, and how much time you spend on the pages you visit, to re-rank and re-sort the search results it gives the next person searching for peach pie recipe.
There is a lot that is scary and even threatening about this new world of technology. It's disturbing to think that the best doctors are sometimes worse diagnosticians than computers. And it's true that, as this JAMA paper identifies, "comparative studies on the effectiveness of machine learning–based decision support systems (ML-DSS) in medicine are lacking, especially regarding the effects on health outcomes."
The JAMA paper looks at several potential unintended consequences of the use of machine learning in medicine, for instance:
- Reducing the skills of physicians
For instance, a study of 30 internal medicine residents showed that the residents exhibited a decrease in diagnostic accuracy (from 57% to 48%) when electrocardiograms were annotated with inaccurate computer-aided diagnoses."
- Overreliance on text.
The JAMA paper reports that machine learning "could lead to reduced interest in and decreased ability to perform holistic evaluations of patients, with loss of valuable and irreducible aspects of the human experience such as psychological, relational, social, and organizational issues."
Ultimately, machine learning is like any other technology in medicine -- the scalpel, anesthesia, or telehealth. These tools are powerful, and come with almost inalterable impacts on how we practice medicine. They're also in our control, we're supposed to wield responsibly and manage the negative consequences.
In the field of telemedicine, we at ClickCare are determined advocates for healthcare providers to be thoughtful about what technology they use for telehealth, and how they use it.
We advocate for healthcare providers to do 2 key things when it comes to technology in telehealth and telemedicine:
- Be open to new technology even as you're discerning about what to use.
Of course, technology comes with unintended consequences and challenges. But it can also enrich and enhance the practice of medicine. We advocate for providers to be proactive in selecting technology that really works for them -- neither a "bury your head in the sand" approach, nor a "take whatever comes along" approach. You know what's best for you and your patients.
- Be open to redefining some aspects of your role as a medical provider.
As technology shifts, we have the opportunity and responsibility of shifting our roles as providers, too. Machine learning may create the ability for providers to focus less on the mechanics of medicine and more on the human art of it. Telemedicine based medical collaboration allows providers to step out of their silos and treat patients as a true team. We believe that the most successful, happiest providers are those that are willing to find new ways of working as our tools change.
We believe the same principles apply to our adoption and use of machine learning in medicine. And we look forward to being on that journey with you.
Want to get all the information before making a telehealth or telemedicine decision? Get our free summary of hybrid store-and-forward telemedicine so you can be informed:
Photo by Samuel Zeller on Unsplash