How Robotics Will Affect the Experience of Health Care for Patients

Key Points

•

Patients will increasingly encounter robotic service delivery as they access health care.
•

This technology can fill in gaps of care and improve access to care such as with telemedicine.
•

Patients will encounter a responsiveness to their needs that is faster, more accurate, and more complete through enhanced data collection.
•

Home robotic systems will provide companionship and the delivery of services.
•

The “human touch” will not be lost as emotional support and responsiveness will be part of robotic behavior.

Health-care delivery is constantly evolving for the patient, resulting in increased demand from providers and technology developers. Health-care delivery and the technology used to perform these services have grown and been refined at an exponential pace. One only needs to casually peruse through technology journals or websites of various vendors to realize there is an electronic tool that can aid in various tasks, ranging from using utensils with artificial limbs, communicating over vast distances, or performing the most intricate of surgeries with minimal recovery time. Software improvements have enhanced our phones and tablets to deliver, collect, and monitor a wealth of medical information. The possibilities of utilizing robotics and technology for health care seem limitless at first glance. However, the practice of medicine is still rooted in a traditional, unidirectional transaction from a content expert (provider) to a recipient (patient). With more ubiquitous adoption of technology in daily life, patients have an ever-increasing stake in not only managing but delivering their own care. This begs the question: Should health care be delivered the same way it has been done for centuries?

Most of the robotic product development currently appears strongest in the physical rehabilitation industry, which encompasses clinical practice, product development, and the technology market. The industry as whole has a wide scope, for example, of improving mobility, function, and independence for patients with various physical and/or neurologic disabilities. Yet many products have a narrow focus on manual tasks or augmentation in gait. This is not to criticize the industry on prioritizing resources on fundamental tasks such as regaining hand function or transferring one’s weight from seated to standing. It does, however, illustrate a disconnect from what patients seek from their health-care providers and what we actually deliver. There is a need for a call to arms involving patients, their caretakers, medical teams, and robotics developers to better understand and deliver products that address what patients want the most.

To better understand how robotics could affect the experience of health care for patients, one should understand what patients and their caregivers truly want from health care. This is an incredibly complex and often misunderstood concept. Traditional Western medical training emphasizes predefined (prescribed) “models” of care, where a patient-provider relationship dyad (patient centric vs. paternalistic) trades information and creates a management plan. But what if robotics could step in and transform this for both parties? The dyad model could still exist but with replacement of a human clinician with a robotic medical provider, or a triad model with a patient–robotic assistant–clinician arrangement. The robotic assistant could serve to monitor a patient’s health, provide daily home health needs, and perhaps inform the clinician of health data elements or history that a patient could not provide in sufficient detail or in a competent way. The robotic assistant could also provide these updates in a more frequent fashion in between formal clinic visits. On the other hand, the clinician could instruct the robotic assistant to automatically intervene for certain health care needs such as wound care, medication administration, or more complex actions such as cardiopulmonary resuscitation when family or other health-care workers are not available.

Robotics goes beyond the past movie interpretations of a stiff, cold machine with slow-moving arms and a comical stammering digital voice that serves you a beverage. Transparent use of robotic technology today such as vacuums, food delivery, and various toys may indicate growing acceptance with consumers. Availability, affordability, and ease of use are only the most basic of requirements for technology to be integrated in everyday life, not just for health care. Much like a household appliance, the functions and more importantly the consumer (patient) experience and acceptance of robotics into one’s life would strongly influence a product’s market success in health-care delivery. Caution should be exercised, as market success may not necessarily translate to good clinical outcomes or high-quality care. In addition, cultural differences in the expectations and family involvement of a patient’s care should be considered. Most importantly, the large socioeconomic gap separating patients who not only get access to but are literate of health care and its capabilities is paramount. To assume that each patient gets the same kind of care even within one’s own facility would be as careless as assuming all types of water taste the same. In an ideal world, all health care is truly individualized, However, the so-called personalized care should be scrutinized for bias between the haves and have-nots.

Though health-care administrators use patient surveys to gain feedback following a hospital stay, Nepal et al. dove much deeper to engage patients in conversation during an acute hospitalization. This effort found that the patient experience is rooted in empowerment and independence, in understanding the medical problem at hand across the clinician-patient continuum, and in having a clear line of communication across stakeholders. Some of the issues raised in this study may not be an area easily filled by robotics, such as room design or the taste of hospital food; however, much of the patient experience involves engagement, compassion, as well as efficiency and overall sense of comprehensive care. This puts a large burden on clinical as well as ancillary health-care workers. Employees of health-care organizations thus need to deliver not only medical care, but also an experience much akin to what is provided in the hospitality industry.

An unfortunate reality of health-care delivery is that staffing has been, and continues to be, insufficient. This does not excuse a poor experience when in the clinic or hospital. A robotic or digital guide could to some extent supplement where the human touch is lacking. Routine cleaning could be automated much like appliances used in people’s homes. Medications could be routed quickly and efficiently within the medical center using high-capacity robotic carts that could deliver at least some of their cargo directly to patients (consumers) via self-navigating delivery vehicles and drones. Navigation within a large medical center could be assisted in an app-based or motorized companion that may guide or physically accompany or shuttle patients who have difficulty walking or decipher signs and directions to get to a clinic. Ma et al. studied how a real example of a companion robot could work in a large hospital or clinic setting. This intelligent guidance robot addressed various questions in a natural language environment. Though its initial intent was to serve almost as a tour guide, the majority of the inquiries involved requests for medical information. This is an intriguing concept and may be alarming to some. This clearly demonstrates that some populations have far more comfort engaging in a conversation to discuss and learn about their medical problems with a robot. It also demonstrates the thoroughness of the patient-clinician interaction may be incomplete and patients are left to fill those needs with a robot. There are various explanations, such as insufficient time, use of overly complex language by the clinician, or simply having many sometimes unrelated medical problems that may not have been asked by the patient or adequately answered by the clinician. This boils down to one thing, that the communication between clinician and patient was insufficient to comprehensively meet the patient’s needs.

Despite the innovation in physical robotics design, none of these devices would be of benefit or have as much adoption as we are seeing today without cutting-edge software. We live in an era where cell phone and wearable technology use is exponentially higher than just a few years ago. Though there is debate on whether this amount of screen time is appropriate, the reality is that digital devices are here to stay. Rather than fixating on the negatives of screen time, health-care workers and robotics developers could leverage our reliance on technology to bridge the patient experience. This adaptation would start even before a patient logs into a telehealth visit or leaves their home to come to the clinic or hospital and may continue after the visit is finished. Patient acceptance of robots or electronic interfaces with health-care providers is profiled by Chai et al. In that study, patients answered a survey regarding their general opinions on the utility of a robot for certain tasks such as taking vital signs or simple clinical tasks that a human nurse or assistant would routinely perform. This information may then be entered in the medical record and prepared for use by the provider prior to engaging with the patient. It was generally perceived useful, particularly in the advent of the COVID-19 pandemic, to use a robot to perform these tasks. Furthermore, human attention can be better dedicated to more critical needs, medical decision-making, and patient education and may promote social distancing and time-efficient evaluation (limiting exposure) between patient and provider. These perceptions were further assessed by deploying a quadruped robot with a tablet that would round on different patients and facilitate a screening evaluation by a human clinician. This allowed the delivery of care from human to human via a mobile communication platform. Overall acceptance and willingness of patients to use this system was high, ranging from mid-80% to 90% of survey respondents.

Although it is not clear whether this mode of video clinical communication and evaluation would have garnered such acceptance prior to the COVID-19 pandemic, it appears now patients are far more willing to receive their medical care through a digital interface. From a bird’s-eye view, the use of a motorized robot to transport an audiovisual communication device around an emergency room or clinic would solve some of the problems of a strained health-care infrastructure and may allow direct and efficient care delivery. Depending on your point of view, the existence of a motorized, robotic communication device may exemplify that on-demand patient care delivery equipment is in high demand. Patients can already contact and participate in a video evaluation with providers through their cell phones. Thus, it would not be unreasonable for video evaluations to replace or supplement parts of the patient evaluation and reallocate direct human interaction to when physical examination or contact is absolutely needed. These technologies can potentially allow patients to seek and obtain emergent and/or specialist care within a safe space and over long distances. It is unclear whether the robot transporting the telecommunications device or just the option to telecommunicate itself was perceived positively by the patients in the study by Chai et al. A motorized patient interview platform deployed within a clinic or from one home to another would have a place in health care when telecommunication is not readily available or is too costly. A few dedicated platforms could move from site to site to fulfill the needs of patient evaluation and management. Such a technology may even fill health-care gaps in rural or low-income areas of the world, where a dedicated clinic may not exist.

The acceptance of human interactions through medical telecommunications devices and the actual autonomous use, assessment, and intervention by a medical robot could radically transform health-care delivery. The current simplicity, portability, and accessibility of video telecommunication with a health-care provider today may indicate that future evaluations with a medical robot, perhaps with humanlike appearance could be done the same way. But is it enough that the robot looks human? Cresswell et al. interviewed select volunteers of various ethnic and professional backgrounds on their general views of robotics in health care. The respondents were mostly highly educated individuals from Europe, the United States, and Australia who either had some understanding or experience discussing and thinking about how robots play a significant role. Though this population of respondents may not represent the world’s view, one thing is clear—there is a general lack of experience and a fear of the unknown when it comes to an integrated human-robot interaction. In this study, references were made to a reportedly negative perception of robots, based on their appearance, mistrust of its reliability, potential threat to replacing or significant change to established human tasks, ethical concerns revolving supervision, and liability toward the robot and/or its operator.

The demand and use case for telemedicine was witnessed in full swing during the COVID-19 pandemic; however, its existence and benefit have been established long before the pandemic. One key benefit is the delivery of specialized care across long distances and into the safety and comfort of patients’ homes. The need for lengthy and potentially expensive transportation costs to the patient would also be reduced. Patients could potentially attend clinic visits with minimal interruption to their workday or other responsibilities. The social distancing requirements put forth by COVID-19 were also mitigated safely by telemedicine. Depending on the type of care needed, telemedicine has some distinct advantages and disadvantages. Clinicians need to rapidly pivot their style of patient examination, relying more on the interview and elaborately described symptoms and physical details when direct contact is not possible. Visual inspection of various parts of the body rather than tactile examination has become the mainstay. On the other hand, video evaluations may not lend well to visits needing detailed examination of tissue texture, shape, or consistency; joint manipulation; or close inspection of the body such as the skin or genital regions.

Though the current delivery of video telemedicine may not seem pertinent to the robotics industry, the overall simplification of the patient visit, its repetitive nature, and delivery of care on demand and in the patients’ desired location may lend well to robotic integration. Clinicians have a typical limited set of interview questions to determine what course of action would need to be taken. If the questions are broken down into their components—what, where, when, exacerbating or alleviating factors, and so on—an artificial intelligence programmed to analyze language could potentially determine a most likely diagnosis and initiate a testing or treatment recommendation. With the current state of mobile phone and Internet-based apps possessing seemingly unlimited data resources, now more than ever, there is a possibility that simple medical visits could be fulfilled by a robot. With the appropriate set of detailed and adaptive questions; attached tools such as a thermometer, blood pressure cuff, pulse oximeter; advanced microphones to auscultate the body; and cameras and software to detect superficial lesions or abnormal movements, a physical examination could be performed on demand for patients without ever making contact with the human clinician.

Before one starts to fear that robots will replace the medical provider workforce, consider this—many medical visits could be evaluated in a primary care or urgent care setting with quick and limited interview and examination. The use of a health screening or triage robot could potentially deliver simple, preventive primary care functions regardless of whether a patient has access to a human clinician or even a physical clinic. By offsetting simple medical visits to robots, this could potentially allow human clinicians to focus their time and effort on more complex cases. The notion of “simple visit” is obviously up for interpretation and subject to debate. On one hand, a complex case could genuinely involve the presence of too many or too difficult a problem for the robots’ programming to solve. It could alternatively involve care for a patient who simply cannot operate or interact with a robot appropriately to benefit from its care algorithms. Finally, the level or detail or truthfulness of information provided by patients would influence the nature of the visit. Note that this is not simply a limitation in the robots’ functions, but also of human willingness, acceptance, and access to a robot.

If the simplest of primary care visits involve a series of health screening questions, vital signs, and point-of-care blood tests (biometrics), it is plausible that visits may be fully automated. Medication and vaccine administration could be performed on demand. Prescription generation and monitoring could be automated as well. Supplies could be delivered by mail in bulk, reordered and replenished automatically, and medication adherence could be monitored with accessory devices at home. Adherence monitoring would require additional ingenuity and forethought. We currently employ simple methods to approximate treatment adherence based on the interview, use of a pill box, physical examination findings, or blood work. Through the clinician-patient relationship, we could then counsel a patient based on the history or test data whether there is evidence to confirm or doubt a patient’s adherence to a treatment plan. How would a medical robot reconcile this? The accuracy of interpreting a patient’s responses could lead a robot to correctly or incorrectly determine that a patient has been compliant with health recommendations and medications. Take, for example, the act of taking an anticoagulant medication for atrial fibrillation. Optimal dosing of the anticoagulant is paramount to ensure the patient is sufficiently protected from a devastating stroke but also from catastrophic bleeding. If a medical robot is tasked with monitoring the administration of anticoagulants, how would it verify this? Monitoring the pill count may not be enough. Would a camera monitoring the ingestion of the medication be needed? Results of point-of-care testing to confirm a drug is therapeutic may determine whether the robot releases the next dose. What would the robot do if the patient did not take his or her medication? Would the robot be allowed to enforce medication compliance against the will of a patient? The ethics of how a medical robot could monitor and enforce patient compliance needs to be well thought out.

The clinician-patient relationship is delicate and based on trust. The patient must not only be provided with but must also follow a clinician’s recommendations. With a detailed history, certain physical examination findings, or laboratory tests, the physician could monitor patient adherence to medication guidelines. This information could be used to counsel patients on their performance. In certain clinical scenarios, accurate information from the patient is paramount for safety. For example, if anticoagulant pill count or video monitoring of pill ingestion happens to be circumvented, without verification via blood test a robot could potentially administer an inappropriate dose and risk injury to the patient. To what end will the robot be allowed to intervene, adjust the medication dose, or counsel (encourage) the patient to adhere to the treatment plan? Would this be the decision point where the robot requires contact by a human clinician? Perhaps this could be done without human verification? Would the robot reprimand the patient for not strictly adhering to the treatment plan? What if the therapeutic relationship (transaction) were to not be conducive? A highly trained robot could thus play a pivotal role in on-demand clinical decision-making.

Interventions may become automated, ranging from simple checking of vital signs to more skilled tasks such as interpreting laboratory or imaging data, mental health counseling and crisis intervention, and emergent procedures such as endotracheal intubation or minor surgery. This question can be made further complex if the industry allows robots to perform these tasks independently with artificial intelligence. Perhaps future clinical robots can synthesize and act on care plans generated on their own rather than being supervised by attending clinicians. Would robots have a role to play in the actual early diagnosis and management of a patient or would it simply augment the decisions made by a human clinician? The process to verify whether a medical robot can independently solve medical problems is discussed by Ashrafian et al. The age-old Turing test is discussed, which challenges whether a human can distinguish between human (control) or computer-generated responses. Though no actual decision-making comparison is performed, they propose a relatively simple and powerful modification. Rather than having a blinded examiner determine whether two interpretations of a medical problem was either computer or human generated, a third response that is human generated is randomly and blindly mixed with the computer-generated response. As a result, the blinded examiner could be tasked to distinguish between responses from a human and a computer (replicating the classic Turing test) or two different humans. This delivers a more stringent examination generating true-or-false, positive-or-negative detection of a computer response, as the response combination would not be consistently and predictably originating from either a computer or a single human. The authors challenge the industry by proposing this methodology be used to distinguish responses during a hypothetical medical interview of a patient to a clinician, where the clinician responses could be computer generated.

An oversimplified but current example can be seen in various “chat bots” that offer customer service on websites. Here, customers are engaged via text messaging where a keyword-based program interprets customers’ queries to navigate questions regarding a product or service. Take this example and develop it with robust natural verbal conversations with a lifelike display of a human clinician. If designed to exhibit emotion and body language, coupled with the ability to interpret spoken language to perform a medical interview, this could simulate a video telemedicine visit that has become so omnipresent today. Though there are some limitations to examinations over video, history taking remains key, and the automated medical interview may potentially transform the way patients receive and manage their own medical care. What is unclear is whether it is enough for an intelligent robot to simply display a photorealistic image of a clinician, with background office decor to replicate the experience of a human-to-human video visit. If it were enough to simulate human appearance and behavior on video, with enough medical proficiency built in, a true virtual medical visit could be conducted without ever needing human-to-human contact.

One needs to periodically look at his or her own profession or daily routine and ask, “Is there a way I can do this better with a robot?” This is not to say that human actions will all be replaced or improved with robots. The dichotomy of human- and robot-specific functions has become blurred over time, where actions we take for granted could either not be performed at all or not be performed as proficiently without the assistance of robots. Performance of repetitive health-care tasks by robots are just the tip of iceberg. The real benefit is the endurance and reproducible precision of performing said repetitive task. At the time of this writing, robots do not complain of excessive work hours or poor working conditions. Robots do not fatigue and decline in task performance as readily as humans unless there is equipment failure. The expectation of robot endurance is far higher with less empathy compared with human performance. There is no influence on robotic performance as a result of criticism or praise. Many may view technology and robots as tireless workhorses that are replaceable when it no longer serves its purpose. To treat a human in such a fashion could be viewed as discrimination.

Robotic assistance in health care is now readily seen in the operating room. Surgical robots have allowed the delivery of extremely precise, minimally invasive operations where a surgeon may not even need to make prolonged physical contact with the patient. Very small incisions are all that may remain to indicate a patient had radical surgery that previously may have resulted in large unsightly scarring. In my own profession of epilepsy neurology, our practice has advanced to collaborate with neurologic surgeons to plan very complex implantations of electrodes in the brain to target precisely, on the order of millimeters, the areas of seizure onset. Though these and other surgeries may be performed manually by a well-trained surgeon, the use of robots to assist in a reproducible and tireless manner has markedly improved not only the accuracy but the aggressiveness, confidence, and complexity of these procedures. However, how can we ensure medical centers or surgeons are competent with robot-assisted surgery? Many may be unaware that these options even exist, so how can the medical community expect patients are given all of these options? Say a patient needs surgery to remove a tumor or other potentially life-threatening lesion located in a delicate area where the margin for error is very small. Say this procedure’s risk can be lowered or recovery time can be decreased in half with a robot-assisted surgery. Will patients be influenced to seek this care at the most advanced centers, potentially traveling great distances and at cost to obtain this procedure? We live in a time where online reviews, elaborate infographics, and incredibly high, sometimes unrealistic, expectations are put upon clinicians to deliver the most complex, comprehensive care. Will patients instead drive the advancement for surgeons’ use of robotics, or will they trust tried-and-true manual techniques that may have been performed for decades?

Lam et al. writes that with “digital surgery” comes massive amounts of data, not only regarding the surgeon’s performance but also for patient information as well. They emphasize that scrutiny of technique, data security, and the ability for surgeons to adapt to new techniques are core pillars to develop and protect. Yet what entity should champion this? Is this push for innovation in the best interest of the patient and driven by patient demand? If robot-assisted surgeries have better outcomes and shorter recovery times, then shorter hospital stays and increased profitability may result. Yet at what other costs? With the drive to shorten hospital length of stay from a payment, quality, and complication perspective, clinicians and administrators also need to strongly consider the risks and overall health-care experience for the patient undergoing technically complicated surgeries. The cost of robotic surgical assistants is currently incredibly high. Though the prospect of minimal invasiveness and shorter recovery time may be attractive to the patient, the robotic operation itself, given its complexity, will likely cost much more in terms of upfront equipment, training, and maintenance compared with a traditional manual operation.

In the home setting, robots may be categorized into (1) service type or assistants for daily activity or (2) companions that are designed with the intent to engage humans socially and emotionally when a human or animal equivalent may not be capable or available. At its face value, this is an admirable goal: to allow patients to live independently with minimal hardship in the comfort of their own home. Chiu et al. profiled a Taiwanese population’s opinions following video observation of robots of various physical forms, human, animal, or nonbiological. They showed that acceptance of companion robots into one’s life was high regardless of one’s previous experience or lack thereof. Superficially, patients appeared to gravitate to certain physical forms that may be visually appealing, such as a friendly face or a cute animal. This would be important for product design to understand what drives the market to a certain type of robot. But should service or companion robots exist purely for the patient’s pleasure, or should they have a specific purpose such as performing physical tasks that a disabled or socially isolated human cannot reliably perform? Ideally both purposes would be satisfied.

In a 2017 interview of Maja Matarić, PhD, socially assistive robots and the intended gaps in health care they are intended to fill are profiled. One important point she makes is that a robot’s discrete functions or features are less important than how it motivates and encourages patients to facilitate their own recovery. This is a more holistic approach to robotics development than a simple appliance that has unique attributes to improve sales figures. It bridges the human touch, that warm feeling someone gets when given the right words of encouragement to take control of one’s health and to actually get better. Traditionally this may be achieved by separate visits with a provider, a therapist, or a social worker. With the right programming, this therapeutic relationship could be combined and achievable at the home regardless of the presence or absence of advanced technical features. We should all be realistic about the physical, mental, and emotional effort involved as well as the limitations to motivate a patient to be adherent to a treatment regimen. Sometimes the right human or right type of motivation may not be available. Dr. Matarić realizes very well that there is a disconnect when care is simply delivered via a large video display. Having a robot small enough for a patient to hold, with different textures or colored “skin,” possibly temperature changes, all add to the experience of interacting with a living being. Consider how emotional support animals have different appearances, hair color, breathing patterns, mannerisms in response to a patient’s state of well-being. Robots that detect these physiologic changes in humans and adapt their responses and tangible characteristics accordingly could provide very therapeutic effects. She recognizes that to pioneer robots to help people, the robots in essence may need to become more “human.”

This also illustrates a complex social problem impacting patients and the role of robots at home. Human interactions, not just between clinician and patient but on a larger societal scale, could be supplemented, if not replaced, by robots. Depending on your point of view, this may or may not be desirable. One needs to exercise caution and determine whether living with a robot is not only appropriate but truly desirable by the patient. The obvious altruistic explanation would be that the patient has a genuine physical or psychologic benefit from a robot or companion. Alternatively, this may be the result of insufficient human effort or financial resources to provide this type of care direct at the bedside or in the home environment. This is not to say that human-to-human care is not possible—some patients may prefer the interaction with a support animal or robot for a number of reasons. However, health-care workers, administrators, robot developers, and most importantly patients’ family members should consider whether the use of a robot is truly of direct benefit to the patient and not preferential to other stakeholders seeking respite from the caretaker role.

Care of patients with cognitive impairment can be very demanding, and caregivers deserve respite from these duties, which many times can be an all-day affair. This is a very popular area of research focusing on neuropsychiatric disorders such as autism spectrum disorder, attention deficit hyperactivity disorder, and various forms of cognitive impairment and dementia. Having a robot engage and occupy a patient would allow the caretaker to satisfy other needs of the household, perhaps go to work or engage in his or her own social needs while the loved one is kept occupied. There would be a delicate balance between augmenting the daily activities/interactions of a patient with impaired cognition and simply offloading the effort involved to monitor this patient. This challenge is not unlike parenting in general. Young children have individual likes/dislikes and needs. Not all of these may be met to their satisfaction by the parent or caretaker. The complexity of these interactions may not need to be very sophisticated. Like toys that are designed to entertain and occupy a child, robots may provide simple activities to patients with neurocognitive impairment. The caretaker as well as the clinician would need to decide whether this is appropriate, as it may not be socially acceptable for an adult to behave and interact with objects like a child. There is negative social stigma toward patients with disabilities in the general public, particularly cognitive impairment. The root of this is most commonly insufficient knowledge, experience, and understanding by the lay public on the challenges patients and their caretakers face. The crude example of a “large baby” who can only communicate in simple grunts or vocalizations and only interacts with playthings does not do this justice. Older children or adults with neuropsychiatric dysfunction may exhibit disinhibited behavior, require very constant or repetitive stimulation, and may act out or behave negatively when their needs are not met. This is typically due to insufficient understanding of a disabled patient’s needs and assumptions made by caretakers and clinicians alike—such as hunger, pain, television, and so on.

Robots can also provide more reliable and trackable data on a patient’s behavioral patterns. This can range from tacking daily habits such as feeding, toileting, and medication administration but can also be very individualized such as patterns of interest toward certain activities, other people, or interactions with the environment. It may be possible that robots and artificial intelligence can continuously monitor these and detect and predict trends in a patient’s behavior better than a human caretaker. These human-robot interactions are not without merit and go beyond simple entertainment or consumption of time. Some case reports have shown improvement in independence, initiation of social interaction, and improved (more detailed) expression of wants and needs in patients with autism. Other studies reported improvement in following or reproducing movements and following musical cues. Objective findings of reduced cortical thinning involving the anterior cingulate have been reported in the elderly. Though some of these achievements may seem simple, given the severity of cognitive dysfunction within this population, any improvement in social interaction would be deemed significant given the time and effort necessary to appreciate even the smallest of gains.

Perhaps some patients have personalities or mannerisms that are not conducive to companionship with another human. This would then question whether a patient’s actions resulting in incompatibility with a human are appropriate to subject to an animal or robot companion. Where do robots stand in this spectrum of patient-companion relationship? Say the barrier to a patient-human caregiver relationship was due to negative, possibly abusive, mannerisms by the patient. Would it be appropriate to then subject an animal or a robot to this kind of behavior? One can assume that “robots don’t have feelings, therefore patients can treat them however they please,” yet this may miss the point of what a companion robot was initially designed for. We assume that companions help and enhance a human’s behavior such that future interactions—be they with other humans, animals, or objects—regardless of whether real or virtual, occur in a societally accepted manner. However, if a patient with abusive or negatively interpreted mannerisms is allowed to continue this behavior onto a robot that may not act in defense, what rehabilitative benefit is provided to the patient? In this hypothetical situation, negatively interpreted behavior may be perpetuated, potentially at the detriment of the patient’s contribution to or reintegration with society.

Doring et al. provide a very detailed review of the current literature of sex dolls and sex robots and their impact on human sexuality. Where this article truly exemplifies itself is how it challenges the reader and the medical/science community as a whole to consider how advances in robotics can impact not just how we interact with one another, but how human-robot interaction can mold the way humans seek relationships. On one hand, the unidirectional relationship of user and robot (product) can be seen as a predictable transaction of commands resulting in set responses. The sexual health and desires of the user are either satisfied or unsatisfied by the constraints of what the robot is designed to perform. The human-robot relationship could be taken a step further if the supportive or companion functions of the robot surpass those that a human partner or caretaker is capable or willing to provide. This again partly returns to the notion that robots in this day and age are primarily compliant objects. This can have significant benefits as well as consequences. Whereas the tireless nature of robots would be of benefit for repetitive tasks, if the relationship is more involved, say with sexual health, the user or patient in this case could become accustomed to not only the functions (service) but the compliance (willingness) of the robot. Doring et al. discuss existing literature raising concerns that this could perpetuate dysfunctional sexual behavior. Say a patient uses a sex doll and explores his or her desires fully and without judgment. This may involve intercourse as well as nurturing functions such as supportive conversations, which is assumed to be therapeutic. Should the patient instead exhibit verbally or physically abusive tendencies onto an otherwise compliant robot, this would potentially perpetuate deviant behavior, which may possibly be inflicted on future human partners. Robot developers, clinicians, patients, and caretakers alike would need to acknowledge and agree that robots with advanced companion functions may not only supplement relationships but can facilitate pathologic behavior as well.

In conclusion, the use of a medical robot would not just revolve around patient monitoring, trust or mistrust, but most importantly independent self-care. Medical robot use has the potential to accelerate the delivery of care simply by increasing access. The analysis and guidance a medical robot provides could facilitate earlier intervention to common medical comorbidities such as obesity, hypertension, and diabetes and its sequelae on a daily basis rather than from one clinic visit to another. Robotics developers and clinicians would need to determine how much of simpler, primary care should be automated. Medical errors and accuracy would need to be scrutinized. Should there be future determination that the robot misdiagnosed or incorrectly administered a medication, what entity would be responsible? Would the robot undergo human supervision of some or all recommendations? If an error was made, would the robot developer or its human supervisor (if one exists) reconcile this? How much would user (patient) error be a factor? We would need to imagine that given the technology gap that already exists in the general population, there will be at least an initial limitation to the comfort and engagement by patients in their own care delivered by a robot. Socioeconomic differences are likely to impact the availability and acceptance of medical robots. The research and development of an artificial intelligence that is capable of accurately diagnosing basic conditions or is compact enough to readily deploy in the homes of patients will likely come at great cost. It is thus assumed that early adoption of this technology would be limited to patients who not only are accepting and competent with technology but who have the purchasing power to obtain these devices and services.

Services, both medical and technical, will likely require upkeep and may necessitate (or monetize) a subscription akin to Internet, phone, or television services. This is bound to impact how health insurance companies in the United States or government health services regulate the reimbursement of medical robotics. If a robot with the promise of personalized care can provide lower-cost preventive care in an automated fashion, it would be advantageous for developers to leverage this not only for profit but also for a reasonable cost to patients. The potential for lower cost per patient care via a medical robot may influence insurance companies to reward patients with the use of a robot over a human clinician. Clinicians who are early to adopt these technologies for supervision could potentially monitor the health of a greater number of patients remotely. Remote monitoring with a theoretically higher volume may result in higher productivity and reimbursement. If used effectively, this may fill the primary care gap. If not implemented appropriately, these systems may instead add to the workload burden and burnout of clinicians practicing today. The prospect of medical robotics is boundless and no longer just a dream or a science fiction fantasy; however, strong collaboration among patients, clinicians, and developers is more important than ever to ensure its implementation is successful and beneficial.