The relationship between a physician and a patient can at times be tenuous, complex, and central to the healing process. Particularly in primary care or general medicine, when the physician is expected to diagnose an incredible range of potential illness, communication between the physician and her patient needs to be empathetic. It’s this human-to-human connection that allows physicians to entertain all the possible causes of illness, leading to more complete diagnosis and treatment.
The complex nature of human communication — the ability to understand and develop context, to parse emotion with subtlety — means that machine automation at the point of care will remain a distant concept for some time.
This is not to say that automation has no part to play in the way physicians and patients communicate. To be sure, the U.S. healthcare system is overburdened with manual systems and processes that require too much oversight for too little reward. Automation is welcome, and sometimes warranted.
For example, one-off transfers of data from a spreadsheet to a more complex system, email or text reminders for upcoming appointments, and follow-up correspondences after appointments are all ideal scenarios for automation.
By their very nature, these administrative tasks are likely better off being automated. Letting a machine transfer data from one repository to another would take less time and reduce errors than having a person copy it. Similarly, creating a stock response for an email follow up that varies ever so slightly depending on the context of the patient’s visit can help create a sense of appreciation on the part of the patient, and may encourage better compliance with treatment plans.
Automating these tasks can reduce the demand on staff, and even positively affect outcomes. In at least one study, a heavier use of automation in patient records and notes was correlated with a 15 percent decrease in the likelihood of in-hospital deaths across all conditions studied.
A more famous example is the BD FocalPoint GS Imaging System, which can analyze 170 Pap tests for cancerous and precancerous cells in a single day. A technician can likely get through 80 to 90. Federal restrictions prohibit technicians from going through more than 100 in a single day. The machine doesn’t only work through the tests faster; it also spots abnormalities with greater accuracy. Providers working without the BD Focal reportedly spotted abnormalities at 79.2 percent — a percentage which rose to 85.8 percent with the aid of the robot.
While machines can’t do the difficult qualitative work with patients that humans can, they do excel at quantitative tasks, especially with high levels of repetition. That’s why the Mayo Clinic began harnessing the natural language processing power of IBM’s Watson machine in 2014. In order to more effectively fill clinical trials, physicians sought to use Watson to automate the cross referencing of patient data to clinical trial criteria. This process requires matching up a single patient to thousands of different criteria and drags the on-time completion rate of clinical trials into the single digits. Automating the process with Watson’s processors will help physicians fill more trials, gather more study data, and perhaps save more lives in less time.
On a more practical level, automation could supply the lifeline that will prevent independent practices and huge hospitals alike from drowning in the ocean of new ICD-10 codes scheduled to take effect in October. Computer assisted coding (CAC) could take a number of forms, and whether it’s overseen by a coding professional or generated at the point of care, these systems could be the difference between accurate, efficient coding and the meltdown of an entire profession.
CAC software offers a glimpse of what’s likely the feature of healthcare automation: a series of medical software systems — mostly diagnostic, though some surgical systems aren’t out of the realm of possibility — that process data at rapid rates to assist the technician or physician in delivering care to the patient. Truly automating the physician-patient interaction would mean replacing the physician with a machine, which, given the natural complexity of human communication, seems quite far off, if not wholly impossible in the foreseeable future.
About the Author: Zach Watson is the content manager at TechnologyAdvice. He covers marketing automation, healthcare IT, business intelligence, and other emerging technology. Connect with him on LinkedIn.
Edited by
Dominick Sorrentino
