Lessons for healthcare providers from this newest patient survey
An enlightening survey conducted by the respected Pew Research Center in December 2022 reveals many perspectives held by Americans regarding the use of artificial intelligence (AI) in healthcare and medicine. More than 11,000 U.S. adults were surveyed between December 12 and 18, 2022.
Here we share some of the results for the benefit of physicians, nurses, surgeons, specialists, and other healthcare providers. As you contemplate adding AI-based health services, these findings might provide useful insights. If you have already incorporated AI-driven applications into aspects of your healthcare offering, perhaps these observations will help guide you in promoting them effectively to your patients.
Exciting Uses of AI in Healthcare
While there are numerous uses of artificial intelligence in healthcare, and almost unlimited potential for more, these three are arguably the most common applications in use today.
Surgical Robotics. AI-based robotics, which can guide doctors during minimally invasive surgical procedures, has been in active use for 40 years and has repeatedly proven its value. According to the National Institutes for Health, in the 1980s a guided system dubbed Robodoc was developed for use in hip replacement surgery. At the same time, a surgical robot called PUMA 560 was employed in a brain biopsy procedure, and a urologic robot was used in prostate surgery. Use of AI-driven robotics became popular among surgeons as they avoided the effects of hand tremors during delicate surgeries. Teleoperated robotic systems developed by the Defense Advanced Research Projects Agency (DARPA) to reduce battlefield casualties led to surgeon-controlled robotic telepresence surgeries that have become a standard of care.
Natural Language Processing. NLP got its start as a language translation app in World War II, and today is used in healthcare to comprehend, organize, and classify clinical documentation and data. NLP systems can analyze unstructured clinical notes about patients, for example, by understanding the language, terminology, and coding used by physicians. Because AI can analyze great volumes of data in real-time, it can improve both clinical and non-clinical healthcare decisions.
A prime example of AI in healthcare is Suki, an NLP-based voice assistant that learns from every interaction with physicians. Suki reduces documentation time by 50% to 70% or more; doctors simply dictate their notes and Suki creates the documentation quickly and accurately, learning as it goes.
Virtual Health Assistants. AI-programmed health assistants can be found on the websites of some physician and professional associations, as well as in several telehealth apps. Virtual health assistants answer questions from patients, guide them in managing their health, and reduce the frequency of face-to-face doctor visits. These functions add value for both patients and physicians. Virtual health assistants are chatbots programmed specifically for healthcare—think of a Doctor Siri or Nurse Alexa and you get the picture.
Telehealth a Recent Platform for Use of AI
Telehealth applications may be the most recent introduction of AI in healthcare. Many Americans grew accustomed to using telehealth or telemedicine options when the pandemic discouraged close personal encounters in 2020 and 2021. In 2021, more than one-third of adults (37%) reported having used telemedicine in the previous 12 months. No masks, sanitizing, or social distancing was required in order to obtain a professional medical opinion about a symptom or a prescription for a stubborn cold.
Telehealth enables patients to use their computers and smartphones to connect with physicians and nurses on the internet. Appointments can be conducted through video calls, online chats, and even email or phone call. This can save time, fuel, stress, and money. MDLIVE, Teladoc, WebMD, and Doctor on Demand are a few examples of telehealth apps.
Post-pandemic, telemedicine applications are useful for full-time office or plant workers, patients with young children at home, the disabled or elderly, and people living in remote areas. Telehealth apps can provide virtual doctor visits after-hours and regardless of patient location—as long as there’s an internet or telephone connection.
Some telehealth applications use artificial intelligence to allow doctors to make decisions that are more data-driven, with the goal of enhancing the patient experience and improving the health outcome. But the use of AI in telemedicine is still in its early stages.
As we have seen, the use of robotics in healthcare, enhanced by artificial intelligence, is hardly new. And general-purpose AI assistants like Alexa and Siri are familiar household names for many of us. However, other applications of AI in healthcare are more recent, and perhaps not as widely known or understood. This may explain some of the results of the Pew Research survey, which asked questions about the use of AI in the context of diagnosing disease or recommending treatments.
Comfort Level. The majority of adults (60%) queried in the Pew Research Center’s December 2022 survey “would not feel comfortable if their healthcare provider relied on AI for their medical care” in diagnosing disease or recommending treatments. Conversely, 39% said they would feel comfortable with this use of AI. However, three-quarters of Americans (75%) are concerned that healthcare may be moving too quickly to expand AI use without fully understanding the risks.
Health Outcomes. Asked how the use of AI in healthcare would affect health outcomes for patients, 38% thought it would lead to better health outcomes. But 60% indicated the outcomes either would not change (27%) or would actually be worse (33%).
Mistakes. On the plus side, 40% felt that the use of AI in healthcare would reduce the number of errors made by healthcare providers. Slightly more than a quarter (27%) held the opinion that using AI would add to the number of mistakes.
Patient-Physician Relationships. The majority of respondents raised concerns about what the use of AI in healthcare could do to the personal relationships between patients and healthcare providers. These connections are often as important to patients as the actual care they receive. More than half (57%) believe those relationships would be negatively affected. (Only 13% thought they would be improved.) There is opportunity here for physicians to address this concern as they move to introduce AI-based apps in their practices.
Information Security. In light of the volume, frequency, and visibility of data breaches in the healthcare industry, it is surprising that more Americans aren’t concerned about the security of their information. About a third (37%) think using AI in health and medicine would make patient records and information less secure. Only 22% think using AI would improve security.
In addition to the specific response categories above, comments from survey respondents also reflected general views about the use of AI in healthcare.
- Some believe that AI apps are only as good as their sources, since they are “programmed by humans, and humans are biased.”
- Others observe that “AI isn’t going to have preconceived ideas about a patient’s symptoms” and “AI is not biased toward a person’s characteristics” in ways that humans may be.
- Still other comments suggest that a “real person/doctor would be primarily treating people, even if we adopted AI” and “human judgment is needed in medicine.”
These responses reflect an openness to the use of AI in healthcare and that it may be highly beneficial in certain cases. Three specific scenarios are addressed below.
Use of AI in Skin Cancer Screening
Among the specific examples of the use of AI to diagnose disease, its use in helping to screen for skin cancer is very promising. AI used for skin cancer detection can scan images of an individual’s skin and flag areas that may warrant testing or biopsy. A physician or physician’s assistant would be required to double-check the area, decide whether a biopsy or other procedure is warranted, and perform that procedure. However, the use of AI could be a beneficial first step in the process.
Nearly two-thirds of U.S. adults (65%) say they would definitely or probably want AI to be used for their own skin cancer screening. And a little more than half (55%) believe that AI would make skin cancer diagnoses more accurate. Only 13% believe it would lead to less accurate diagnoses, while 30% think it wouldn’t make much difference in the accuracy.
Americans who are aware of this use of AI in skin cancer detection view it as an advance for medical care. More than half (52%) call it a major advance, while 27% call it a minor advance.
Use of Robotics in Surgery
The Pew Research survey confirms that public familiarity with the use of AI-based surgical robots is higher than for many other health and medical applications—possibly because they’ve been in use and in the news since the 1980s. Almost 60% of adults indicated that they have heard about the use of surgical robots.
However, this doesn’t mean we are embracing the idea of surgical robots for use in our own individual surgeries. In fact, only 40% indicate they would be comfortable with AI-based robotics being used in their own surgery. The majority (59%) said no thanks.
Use of AI Chatbots to Support Mental Health
The use of AI-based chatbots in support of mental health was another use of AI specifically addressed in the survey. Although this application is in its early stages, chatbots such as Woebot, Wysa, and Youper use cognitive behavioral therapy techniques to help users manage their mental health.
However, survey responses to the idea of using an AI chatbot for mental health support are extremely negative. Almost eight out of every ten U.S. adults (79%) say they would not want to use an AI chatbot if they were seeking mental health support. Only 20% say they would use this kind of tool.
Responses to additional questions about using AI chatbots for mental health support were similar. For example, 46% of U.S. adults believe AI chatbots should only be used by people who are also seeing a live mental health therapist. Another 28% do not think AI chatbots should be available for this use under any circumstances. Less than one-quarter of respondents (23%) feel this type of chatbot should be available to individuals whether or not they are seeing a therapist.
Surgical robots, natural language processing, and virtual health assistants are popular uses of artificial intelligence in healthcare today. Surgical robots have been in use for 40 years and continue to prove their value in minor and major surgical procedures. Surprisingly, however, a recent survey conducted by the Pew Research Center revealed a form of “not in my backyard” syndrome when nearly 60% of adult Americans said they would not be comfortable with the use of AI-guided robots in their own individual surgeries. Still, AI is perceived as a useful option in specific scenarios, such as skin cancer screenings. In this case, 65% of adults would be comfortable with the use of AI in their own skin cancer screening.
The use of AI holds great promise and presents numerous opportunities for both patients and physicians. For these reasons, we expect to see advances in the use of AI in many health and medical scenarios. And while we’re on the subject of AI in healthcare, we must mention that AI has also proven useful in enhancing cybersecurity in hospitals and other healthcare organizations. Ask your CISO about it—or give us a call.