Robots and computers have fascinated most of us for decades. I fondly remember programming in Fortran using 88-character punch cards and being amazed at how rapidly a computer could solve equations that would take hours to solve manually. My fascination with the computer continued to be stoked by movie characters such as C3PO and R2D2 in the Star Wars series, by HAL in 2001 — A Space Odyssey, and by Captain Kirk’s conversations with the Enterprise’s computer. More recently, Hollywood treated the world to a biocomputer humanoid in Avatar.
Computers are not only tools to work on. They also interact with us, by talking us through a purchase at a grocery, for example. They have made our transactions easier by eliminating waiting for boarding passes at the airport and by taking our payment in a parking garage. They also entertain us in increasingly sophisticated video games, help us teach our children with online interactive education, and help us troubleshoot computer problems by using a voice-recognition-driven help desk.
Other computers reach to global positioning satellites (GPS), facilitating travel in unfamiliar areas. In most cases these applications are delivered by impersonal machines — albeit, machines with plenty of raw processing power.
Conversational virtual nurse agent conducts a bedside dialogue with a patient. “She” points to a medication and describes it; the patient clicks the”right” box, meaning, Right, I understand.
Medicine has lagged many industries in its use of computers, perhaps because of the very personal nature of medical care or the enormous variability of the human condition.
Raw processing power and software sophistication have reached the point that human emulation is now possible. I predict that computers will soon take over many of the repetitive data gathering and educational tasks that humans now perform in medical interactions.
In 2007, John W. Bachman, MD, wrote in Family Practice Management about his experience in having patients enter their own historical data in his electronic health record. It saved time and money and, he thinks, resulted in more accurate data.
A much more advanced model was described in a PhD thesis by Timothy W. Bickmore in 2003. Relational agents (animated conversational computer images with voice recognition capability) had conversations with patients. Bickmore’s relational agent, Laura, demonstrated both verbal and nonverbal communication capability, including head nods, eye gaze, hand gestures, posture shifts, eyebrow raises, and facial displays of emotion. This technology, in the hands of innovation officers at the major health plans, will change managed care. The implications are enormous; the applications, almost innumerable.
Intelligent virtual assistants
A small company, Next IT, headquartered in Spokane, Wash., is one of the pioneers in commercializing the use of human emulation for a variety of organizations, including Aetna, Continental Airlines, and the United States Army.
The company is focusing its efforts on a variety of health care sectors to provide a scalable, reliable service that complies with HIPAA. Next IT hopes that the use of its intelligent virtual assistants will revolutionize the health care industry, just as this technology has changed others. Its earliest efforts for Aetna, for example, involved transactional interactions such as benefit look-up. The company says that these agents are capable of being used in more complex interactions such as adherence and persistence programs and behavioral change.
Disease management efforts have been a mainstay of the managed care industry’s efforts to improve population health for two decades. Many large studies have proven that early intervention in blood pressure control, diabetes control, and lipid lowering can dramatically improve outcomes over long periods of time. But these efforts are very expensive for their modest return in the short term. This has led most commercial disease management vendors to focus most of their efforts on patients with advanced disease, as this population is more likely to have an avoidable expensive event soon.
We intuitively know that changing behavior and improving adherence to guidelines would make an enormous difference in people with early disease in a scalable, inexpensive way.
Keeping patients on expensive drugs costing thousands to tens of thousands of dollars per month has been a challenge. The tendency for many patients is to miss a dose or even cease a medication because of a variety of psychological, clinical, social, financial, or other barriers.
Specialty pharmacy companies also have the added task of ensuring that a patient will be home to accept an overnight delivery of a temperature-sensitive medication. An additional challenge is to find the optimal drug or drug combination for these patients, a task that demands constant measurement of the clinical status of the patient over time. Intelligent virtual agents could simplify many of these tasks and allow companies to offer a much more comprehensive program without paying for very expensive professionals.
Filling prescriptions for chronic medication is bread and butter for retail pharmacies. They are also positioned to determine whether a patient is obtaining refills as prescribed. But for a pharmacist to actually contact a patient either by phone or mail is very expensive. This technology offers a cost effective way to automate this process.
HEDIS has improved our health system through guideline-driven medical tasks, such as immunization schedules and cancer screening. HEDIS relies on data analysis and patient interaction. Much of the cost is in the actual human or postal interaction. These actions could be automated with the use of human emulation agents.
And finally, with the rapid growth and adoption of smart phone technology, your personal avatar could help you not only with medication adherence, but with weight loss, exercise, and smoking cessation.
For further reading
- Slack WV, Hicks GP, Reed CE, Van Cura LJ. A computer-based medical history system. N Engl J Med. 1966;274:194–198.
- Locke SE, Kowaloff HB, Hoff RG, Safran C, Popovsky MA, Cotton DJ, Finkelstein DM, Pate PL, Slack WV. Computer-based interview for screening blood donors for risk of HI transmission. JAMA. 1992;268:1301–1305.
- Bachman J. Improving care with an automated patient history. Fam Pract Manag. 2007;Jul–Aug 14(7):39–43.
- Bickmore TW. Relational agents: effecting change through human-computer relationships. Massachusetts Institute of Technology, Ph.D. thesis, February 2003. http://www.ccs.neu.edu/home/bickmore/bickmore-thesis.pdf