New research sheds light on the attitudes of healthcare professionals towards AI-led interventions and identifies potential barriers to the implementation of AI in the field of sexual and reproductive health services. While there was recognition of the potential benefits of automation and AI, particularly in administrative tasks, concerns were raised about the limitations of AI in handling complex cases and understanding patients’ emotional states. The study has been published in the International Journal of STD & AIDS.
“The surge in sexually transmitted infections to unprecedented levels and the underfunding faced by the UK National Health Service underscores the pressing need for reliable, effective, and safe sexual health education,” said study author Tom Nadarzynski, an associate professor of digital health at the University of Westminster.
“I perceive the advent of AI as a game-changing opportunity to delegate parts of health promotion to chatbots, especially for marginalized communities. These groups often experience apprehension when discussing sexual health with healthcare professionals due to fears of stigma, embarrassment, and discrimination.”
To conduct the study, the researchers distributed an online survey to 150 sexual and reproductive healthcare professionals working in England, mainly through online advertisements and professional networks. The sample included a diversity of roles, including consultants, specialist doctors, nurses, health advisors, psychologists, support workers, healthcare assistants, commissioners, service managers, and health promotion practitioners.
The survey included demographic questions, assessment of attitudes towards digital sexual health services during the COVID-19 pandemic, exploration of the perceived usefulness of automation in sexual and reproductive health services, and assessment of attitudes towards AI and chatbots. Additionally, 24 participants from the survey completed qualitative interviews to gain more in-depth insights.
The findings of the study revealed mixed attitudes towards the automation of sexual and reproductive health services. While some automation services were seen as useful, such as appointment booking, patient triage, and HIV medication adherence, others were perceived as less useful, including services requiring psychological and emotional support.
Similarly, regarding AI-enabled chatbots, participants had mixed attitudes. While some believed chatbots could provide more personalized treatment and improve access to sexual and reproductive health services, others were skeptical about the use of AI in medicine and the potential negative effects of chatbots on their work.
Participants also expressed uncertainty about how patients would disclose relevant information to chatbots and whether chatbots could prevent unnecessary visits or reduce travel time to healthcare providers.
“Our study illuminates the potential of AI chatbots in facilitating access to dependable sexual health information, such as STI screening, and coordinating appointments with healthcare professionals,” Nadarzynski told PsyPost. “Although this is still an emerging technology, efforts are ongoing to ensure the safety and acceptability of chatbots across diverse communities. Healthcare professionals see the potential of AI in healthcare delivery, but many express reservations about endorsing these tools without substantial evidence of their efficacy.”
The qualitative analysis resulted in three themes. Firstly, healthcare professionals acknowledged the need for innovation and digitalization in sexual and reproductive health services, but many had limited understanding of AI technology.
Secondly, participants saw potential benefits of AI and chatbots in automating repetitive administrative tasks and providing basic generic advice, but they were cautious about their role in complex cases requiring clinical input. Lastly, healthcare professionals emphasized the importance of maintaining a human connection in healthcare and ensuring that AI-led interventions do not create additional barriers for patients.
Less than half the sample (40%) reported personal experience with a chatbot, and only 5% reported experience with a dedicated sexual health chatbot. About a third (34%) reported not being able to understand how sexual chatbots work.
“It was surprising to find that many healthcare professionals lack an understanding of how AI chatbots function,” Nadarzynski said. “Until we develop mechanisms for chatbots to elucidate their processes, healthcare providers may remain skeptical of this technology.”
“Their concerns stem from their commitment to patient safety and their duty to provide the highest quality of healthcare. Poorly designed or implemented chatbots might risk compromising the quality of sexual and reproductive health services, which emphasizes the importance of collective involvement in their development from the onset.”
The findings suggest that AI-led interventions can be valuable in sexual and reproductive health services but should be designed and integrated in a way that complements the work of healthcare professionals and maintains a patient-centered approach.
“An intriguing avenue for future exploration is the influence of sophisticated language models like ChatGPT on the accessibility of accurate sexual health information,” Nadarzynski added. “Current statistics show that only 9% of people using Google for self-diagnosis arrive at correct conclusions. This implies a failure rate of over 90%. Newer chatbots like Bing or Bard may offer improved accuracy in providing health information by engaging in more comprehensive questioning. It will be captivating to observe the progression in this field over the upcoming year.”
The study, ““But can chatbots understand sex?” Attitudes towards artificial intelligence chatbots amongst sexual and reproductive health professionals: An exploratory mixed-methods study“, was authored by Tom Nadarzynski, Alexandria Lunt, Nicky Knights, Jake Bayley, and Carrie Llewellyn.