How Robotics Will Affect the Practice of Medicine for Practitioners

Key Points

•

Through the performance of repetitive and time-consuming tasks, robots can help prevent practitioner burnout.
•

Robots can add greater safety and efficiency such as in pharmacy systems.
•

Tremendous flexibility exists to automate functions such transport systems, hospital admissions, and infusion devices.
•

Devices can extend and enhance practitioner skills such as through gait trainers and robotic surgical systems.

Historically the practice of medicine has been consumed by repetitive, albeit important, tasks. In today’s world, the number of independent tasks a practitioner may need to accomplish per patient has increased dramatically. Especially with the emergence of electronic medical records, or electronic health records, and value-oriented healthcare management, the additive burden of small tasks and redundant charting has become an inordinate obstacle directly associated with increasing burnout across all fields in health care. It has been reported that primary care providers may spend as much as 2 hours on electronic medical record–related tasks for every hour of direct patient care. Robots are well suited for repetitive, technically defined tasks and are being increasingly adopted in the health-care field. The global medical robotic systems market is expected to surge to $10.7 billion by 2026. Although we are still far from the Star Trek/Star Wars vision of a robotic future, many advances have been made and robots are currently widely used throughout the globe in the multifaceted world of health care.

In this chapter we will look at the current applications of robotics, artificial intelligence (AI), and technologic advances improve the efficiency of various practitioners, attempt to reduce burnout, and most importantly improve patient outcomes. As the field of health care is incredibly broad, multidisciplinary, and traditionally grounded in physical and face-to-face interactions, we will also discuss the role these technologies play in the entire realm of health care delivery, including environmental services, dining and nutrition, and hospital logistics. Due to the breadth of the health-care field, and the myriad of robots available, it is beyond the scope of this chapter to discuss every currently available option. Instead, we will focus on one or two exemplars of robotic applications in each setting. Lastly, as the COVID-19 pandemic has changed the way people interact with health-care systems across the globe, we will even touch on applications in telemedicine, telehealth, remote health management, and new modalities of infection control.

“Keeping It Safe”

A pharmacy, both inpatient or outpatient varieties, is a prime example of where small mistakes can have big consequences. Pharmacists and their team (including pharmacy technicians, or techs) may be expected to serve many patients on a daily basis, attempting to both provide timely and useful service as well as mitigating or eliminating potentially dangerous situations. For reference, there are approximately 115,000 actively practicing full-time pharmacists in the United States. Pharmacists manage teams of technicians to help receive and process dozens to hundreds of prescriptions daily, whether brought in by hand by the customer or sent by electronic means, telephone calls from the health provider, or even mail order. The average volume of medications provided in the United States is 259 prescriptions per day based on a 2013 Pharmacy Survey.

Additionally the pharmacy may, and in many cases must, perform cross-checking of prescriptions against already prescribed medications, narcotics monitoring boards, and a myriad of potential side effects. Mistakes can be annoying at best and deadly at worst. More than 100,000 Americans die each year of adverse drug reactions. Adverse drug reactions have become one of the six leading causes of death in the United States. This is not unique to the United States: a 2018 report from the University of York estimates that as many as 237 million medication errors occur in England every year.

Teams must also maintain, catalog, and stock hundreds of varieties of medicine and supplies, some with incredibly similar shape and design. Similar medications can look alike or their names can sound alike, commonly referred to as LASA medications. There are currently more than 20,000 prescription medications approved by the Food and Drug Administration in the United States. These prescriptions change in quantity, packaging, generic versus brand name and can even be occasionally counterfeit. Pharmacies often change suppliers of a particular generic drug, or a pharmacy wholesaler may continually rotate supply companies, contributing to innocuous shifts in pill descriptions that may lead to further confusion. The US-based Drug Enforcement Administration reported in 2021 that there has been a sharp increase in fake prescription pills containing fentanyl and meth.

There have been many strategies aimed to prevent pharmacy errors and adverse drug events as well as increase overall efficiency. Robotics is playing a significant role in accomplishing these goals. One approach is the use of AI-powered pharmacy dispensing platforms. The Fred Dispense Plus platform, recently deployed in Australia, is a computer-based tool that “uses AI and big data to support clinical and business decision making. As it predicts medication directions, the AI tool allows pharmacists to quickly and safely enter directions for dispensing labels through shortcuts instead of manual typing. Another smart tool on the same software platform can perform pricing comparison.” This platform also integrates with other third-party software to help provide comprehensive medication management.

Automated pharmacy dispensing and robotic filling devices are becoming more ubiquitous in developed countries, and as the capabilities and storage potential of these robots grow so will their role in health care. A relatively small, nonmobile robot such as the Kirby Lester KL60 Dispensing Robot (from Capsa Healthcare) ( Fig. 2.1 ) fully automates 60 high-moving oral solids; that is 35% to 40% (or more) of a pharmacy’s total daily prescriptions in a device that is less than 4′ deep. With the KL60 automatically handling more than one-third of orders, pharmacy staff are freed up to concentrate on important priorities like immunizations and customer service. Larger robotic prescription robots are also available that can handle up to 108 high-moving tablets or capsules stored in locked cassettes, which can constitute automation of 50% or more of average daily prescriptions.

Even larger systems, such as the PillPick by Swisslog Healthcare, can automate unit-dose packaging, storage, and dispensing, while also providing additional safety measures like a vision-based sorting and verification system ( Fig. 2.2A and B ). Their solution is termed PickView and can help reduce or eliminate dispensing of look-alike or soundalike medications, as well as enabling hospitals to provide documentation that the right drug was in the right package.

Smart infusion pumps have become more common in health-care systems and are a popular method of delivering medication during inpatient hospital or outpatient clinic visits. These pumps may be combined with bar code–based scanning of medications and patients to develop “closed-loop” systems. Closed-loop systems help automatically regulate a process, in this case drug delivery, without human intervention. As human error is a significant factor in 70% to 80% of industrial adverse events, the more routine human interference can be automated, the greater the opportunity for improved efficiency and decreased mistakes.

There are multiple groups working on bringing AI-enabled infusion pumps to market. One such effort began in an attempt to coordinate complex and real-time changes in patients’ blood pressures during labor and delivery and notably during Cesarean sections. “We’re doing all of these things simultaneously while taking care of an anxious, awake patient and preparing for major abdominal surgery. In a couple of minutes, the math becomes too complex for any human to do, and we may under- or over treat the patient” as described by Dr. Kovacheva with Brigham and Women’s Hospital.

The current health-care and business landscapes dictate the priority of dispensing medications accurately, precisely, and on an ever increasingly tight schedule. A 2016 look at time utilization of pharmacists at a typical US pharmacy found that dispensing still consumes more of a pharmacist’s time than any other activity—an average 39%, versus 22% for consultation, 15% for medical record interaction, and 19% for drug therapy management.

Interacting with insurance companies can also take an immense amount of time to check approvals, arrange preauthorizations, assess copayments, and collect payments. Additionally, patient interaction and education by a pharmacy employee can be up to 2 hours per day. Reducing a pharmacy worker’s labor burden in any way possible is critically to ensuring comprehensive and robust health care.

“Keeping It Effortless”

Health care providers in hospital-based or outpatient clinic–based settings encounter significant barriers to consistent and efficient patient care. For many patients the first step toward care is either establishing care with a provider or undergoing admittance to a hospital system. Admission to a hospital is no simple or small feat, whether by way of the emergency department or direct admission from another health-care provider or facility. An analysis led by Thibodaux Regional Medical Center, located near New Orleans, Louisiana, looked at the hospital’s inpatient admission process. Although every medical center has its own unique population, needs, and challenges, many of the issues the analysis found can be generalized to many hospitals in the United States. According to their report: “While Thibodaux had invested millions in its EHR [electronic health record] and other information technologies designed to improve patient care, the hospital’s inpatient admission process was still entirely manual, involving documentation on paper and faxing. This led to errors such as lost faxes, and to cumbersome processes such as multiple phone calls to track down information while patients waited. The manual process also led to inefficient handoffs, redundant data collection, as well as delays in getting necessary information, doctor’s orders, and bed assignments so care could be initiated.”

Processes to help automate admissions are being implemented in hospitals around the world. These automations tend to focus on improvements in data acquisition (for example, from surrounding third-party health provider records), patient reporting of medical histories, present illness, and reviews of systems, as well as integration of health-care reporting measures. Although fascinating and incredibly potent catalysts within the health-care field, most are software-driven or data science–driven processes and beyond the scope of this chapter. However, there are also many attempts to integrate robotics in health care in a customer service–based perspective.

In a 2016 study published in the Journal of Medical Practice Management , customer service in health-care settings was by far the most frequent complaint by patients. While having a (knowledgeable and friendly) human as a point of contact with in a health-care setting is currently indispensable, several companies are developing potential solutions to the “always available” customer service need. First created in 2014, Pepper is an interactive robot designed to identify, process, and empathize with human emotions. With the inclusion of multiple cameras, 19 different languages, and 20 degrees of freedom, Pepper is able to meaningfully interact with people in many settings. The robot is being trialed in several environments but with its emotion-processing capacity and ability to speak and direct people in multiple languages, a hospital or health-care setting is a great fit. Several hospitals in Belgium and Canada are using Pepper robots to meet, greet, and direct incoming visitors. As previously mentioned, creating a consistent and helpful welcome to a hospital is paramount to establishing an overall positive experience.

Among the many changes instilled by the COVID-19 global pandemic, and secondary to the critical need to minimize human exposure to pathogens, institutions initiated new methods of collecting vitals and screening hospital employees and visitors. One hospital system, Brigham and Women’s Hospital in Boston, Massachusetts, collaborated with researchers to place contactless vital sign monitors on autonomous quadruped robots. These technologic chimeras, referred to as Dr. Spot by the development teams, have proven viable, and there are plans to for them to deploy the robots in the emergency department and potentially on inpatient floors as well. The machines that collect vital signs without direct touch are not novel, but integrating them onto a robotics platform that allows safer and more effective deployment of resources exemplifies several of the advantages of robotics in the health-care system.

Health care is much more than direct interactions with a patient; on a daily basis a typical health-care provider may be expected to check a plethora of laboratory values, obtain and evaluate imaging, access and navigate medical records (both physical and electronic in nature), communicate various test results to patients, document physical examination findings, write prescriptions, and cross-check against other current prescriptions including from outside providers, among other responsibilities. These tasks are in addition to the face-to-face time spent with the patient, whether in person or virtually, and can add up to immense amounts of time. One study cited by the American Medical Association reports that up to a third of physicians spend 20 or more hours a week performing tasks outside of direct patient contact.

Companies are both currently deploying and developing robots in an effort to offload provider burdens, especially those involving repetitive motions or labor-intense but cognitively light tasks. One common theme is transportation of materials within a large hospital. For example: Relay, built by Savioke, provides trackable and secure chain of custody for biospecimens such as collected blood samples, especially for ones that cannot be transported via pneumatic tube systems existing in many hospitals. Deployed examples have reported an average of 400 deliveries per month with quicker and more cost-effective results than human full-time equivalents. Notably, this example comes from before the COVID-19 SARS-CoV-2 pandemic with the concurrent increased need for aseptic techniques and a global shortage in health-care workers. Savioke estimates that one Relay machine can save close to $500,000 in costs over 3 years compared with a full-time labor equivalent.

Patients presenting for outpatient clinic appointments with physical, occupational, or other therapists often find themselves interacting with robots. Advanced treadmill platforms with weight support and robotic functions are growing in popularity, with some version present at many rehabilitation centers. Sensors and motors help provide a safe and variable gait pattern to help reproduce and reinforce normal ambulation. A popular treadmill in the United States is the AutoAmbulator by Encompass Health ( Fig. 2.3 ). It can provide robotic therapy assistance for patients of various conditions at levels that would normally take three or more human therapists. Additionally, the precise and consistent feedback provided by the treadmill on patient contribution to the task has proven incredible helpful to health-care providers in prescribing exercise or monitoring results of interventions. Potential patient conditions include spinal cord injuries and disorders, brain injury, multiple sclerosis, stroke, Parkinson disease, and other orthopedic or neurologic conditions resulting in mobility changes.