Principal Findings
This rapid review revealed that chatbot roles in health care are diverse, ranging from patient support to administrative tasks, and they show great promise in improving health care accessibility, especially for groups considered marginalized. It also highlighted critical gaps in the literature, which are addressed in the following subsections.
Global Trends in Chatbot Research Indicate Its Predominance in Higher-Income Countries and Opportunities in Lower-Income Regions
With 35 countries represented by the studies in this review, the topic is clearly of global interest. However, more than a quarter of the included studies (46/161, 28.6%) originated from the United States, with the remainder conducted in high- or upper–middle-income countries across North America, Europe, and parts of Asia [250]. The concentration of chatbot research in high-income countries reflects underlying disparities with low- or lower–middle-income countries, particularly in parts of Africa, South America, and certain regions in Asia, in terms of technology access and health care investment. This gap highlights the need for more research focused on these regions, considering their unique digital infrastructure and resource challenges, to democratize health technology and address chronic conditions and health literacy [20,251-254].
Chatbots Have Varied Roles in the Enhancement of Health Care Delivery and User-Centric Services
Our review underscores the transformative roles of chatbots in health care, particularly in delivering remote health services and enhancing patient support, care management, and mental health support. Consistent with previous literature [254-257], our findings affirm chatbots’ potential to improve health care accessibility and patient management. The findings’ emphasis on education and skills building, particularly to enhance health literacy (which aligned with past literature [255,258]) and to support behavioral change (also highlighted by past research [255]), aligns with the growing need for patient empowerment in health care. The administrative efficiency of chatbots, noted in our review, resonates with previous findings [23,35,255,258] on the importance of resource optimization in health care settings.
Our findings indicate that chatbots also play a key role in facilitating clinical research, consistent with past work [259], a potential that needs further exploration, especially considering AI’s evolving role in health care [72,259-262].
The Diverse User Base of Chatbots Shows Their Potential to Support Equity and Bridge the Access Gap in Health Care Services
Our analysis indicates a broad and diverse user base for health care chatbots. From individuals focused on general well-being to those with specific health conditions, chatbots have been designed to cater to a wide array of needs. Notably, their use by demographic and family-centric groups and their accessibility to underserved populations underline the inclusive capacity of chatbots and their role in enhancing health care access and equity, especially for groups considered marginalized, in line with existing research [12,67,255,263-265].
In addition, our findings show the significant use of chatbots in mental health support for various age groups, reflecting the pressing need for accessible mental health services highlighted by others [4,8,12-17,29,30].
Furthermore, chatbots have emerged as tools for reducing stigma [12,265], linking users to health services [266-268], and protecting sensitive information [269]. Their empathetic and multilingual capabilities, as seen in our results [107,111,112,120,122,126-128,132] and past literature [270-276], are vital to reach diverse populations. They are particularly critical in light of the digital divide and the need for inclusive and accessible health care solutions [254,258,263,277,278].
The Use of AI in Chatbots Is a Promising but Still Evolving Field
The studies included in our review show a substantial number of AI-based chatbots, with fewer relying on non-AI platforms. AI in health care is recognized for its potential to improve health outcomes and the quality of life globally [260]. Given advances in machine learning and AI, expanding the scope of chatbots is expected to cause a mutation in their role in the health care system to assist clinicians and potentially take over some of their duties [72,261,262]. The synergy between big data and AI, coupled with the increasing availability of data in health care, suggests that AI-based chatbots could effectively use extensive health care data [259,279]. This aligns with 1 (0.6%) of the 161 included studies [94], which discusses the use of collected data as a key benefit of chatbots. However, ethical considerations such as data privacy and algorithmic biases must be addressed for responsible AI deployment, crucial for maintaining trust and fairness [73].
Studies included in this review indicate that using avatars in these chatbots to simulate social behaviors can enhance user engagement and trust. This form of chatbot technology is particularly appealing in patient interactions and medical education to establish trust and therapeutic alliances between health care professionals and patients and to improve the communication skills of medical students and health care professionals [118,123,130,131,280].
Balancing AI’s benefits to enhance data use and user interactions with its ethical concerns, including data privacy and algorithmic bias, is crucial for its implementation, shaping the future of patient care and medical education in an innovative and ethically sound way.
Despite the Potential Revolutionary Roles of Chatbots in Health Care, Critical Challenges and Limitations Exist
This review stresses that despite chatbots’ roles and benefits, their use comes with various challenges, including ethical, technical, medicolegal, and user experience concerns, as also discussed in past literature [3-5,23,25,30,72,74,95,281,282].
While the studies included in our review have highlighted chatbot use to address minor health concerns and provide off-hour information, there is a noticeable gap in evaluating their technical limitations, especially in complex health care scenarios, as underscored by past literature [3-5,23,25,30,72,74,95,281,282]. This raises concerns about patient safety and the accuracy of health management, emphasizing the need for comprehensive assessment and iterative improvement of chatbot technologies [22,25,68,72,95,254,283].
The findings in our review indicate the regulatory and ethical landscape for chatbots as another area of concern. This agrees with past studies highlighting the need for ethical use, data privacy, and transparent communication about chatbots’ capabilities and limitations [4,73,74,254,281,284,285]. The absence of specific laws and regulations addressing health care chatbot use introduces risks around liability and medicolegal issues [72,286,287]. These challenges are further complicated by ethical dilemmas, such as privacy and confidentiality in nonanonymous interactions [71,72,288,289] and safety concerns in medical emergencies due to limited chatbot expertise [72].
Technical issues identified by this review, including difficulty in language processing and a lack of empathic response, can lead to trust issues and increased clinical workload and align with past literature [3-5,68,72,73,280,290]. Overreliance on chatbots for self-diagnosis and health care decisions may lead to misjudgments, potentially exacerbating health issues [4,68,73]. In addition, the financial motives of private companies in the health sector raise ethical concerns about the primary purpose and application of health chatbots [73]. The requirement for sophisticated AI technology also implies increased demands on human resource expertise and storage services, potentially escalating costs [73,287].
Our results indicate that chatbots serve a wide range of populations from various groups in terms of age, gender, ethnicity, and socioeconomic and educational status due to their promising acceptability and usability [291]. However, the digital divide [292-294], algorithmic ethical concerns [295], and the potential misuse of chatbots in replacing established health services [296] present risks. These factors, along with social, economic, and political influences [297], could inadvertently widen health disparities, highlighting the importance of inclusive and equitable chatbot development and deployment.
The discussion on health care chatbots is fundamentally about their potential and promise, grounded in our exploration of current studies and developments. These digital tools could significantly enhance health care access, service quality, and efficiency. However, realizing their full potential hinges on addressing challenges such as ethical AI use, data privacy, and integration with health care systems.
Efforts moving forward should concentrate on incorporating AI responsibly and designing chatbots that cater to all user demographics, ensuring equitable health care access. Collaboration across technology, health care, and policy sectors is crucial to establish ethical guidelines and confirm chatbots’ efficacy and safety. Successfully navigating these challenges will enable chatbots to fulfill their promising role in health care, contributing to a more accessible and patient-focused system.
Limitations
This review, while insightful, is not without its limitations. Although rapid and systematic reviews are often considered comparable in their conclusions, each methodology has its own set of constraints [289,298]. Specifically, this rapid review was limited by a noncomprehensive search strategy that included only 2 databases. In addition, the inclusion criteria were restricted by date and language, which potentially led to the exclusion of some pertinent studies. Another limitation was the concentration of screening and analysis tasks on a single reviewer (ML), which might have introduced bias or overlooked nuances in the data. Moreover, a formal quality appraisal of the included studies was not conducted due to the descriptive nature of this review. Consequently, this limitation may affect the depth of understanding and the strength of the conclusions drawn.
One critical aspect of our methodology was the combination of empirical findings and opinion-based data from the discussions in the included studies. We did not distinguish between these 2 types of data but rather treated them as a unified source of information. This approach, while allowing for a comprehensive overview of chatbots in health care, might have led to a potential bias in favor of chatbot benefits because both empirical results and positive speculative insights were reported together. However, this potential bias is somewhat mitigated by our consistent reporting of the challenges associated with chatbots, as identified in the included studies. By presenting both the potential benefits and the challenges, we aimed to offer a balanced view, reducing the likelihood of a 1-sided interpretation favoring chatbot benefits.
In addition, this review might have overestimated the results due to the dependence on the discussion sections of each study, which may have overcounted the results and miscounted those that may have disagreed or contradicted the results of these included studies. However, this did not significantly impact the study’s aim to provide an exploratory and descriptive overview of health care chatbots, mapping out the landscape of their applications in health care. In such a context, a broad, inclusive approach that captures diverse opinions and trends is more important than precise quantification.
Moreover, one of the potential limitations of this review is the exclusion of generative AI and LLMs such as ChatGPT. However, among the studies we reviewed, a standout comparison involved a health care chatbot, specialized in medical terminology, and ChatGPT. This unique comparison serves to highlight the advanced capabilities of LLMs such as ChatGPT in enhancing the delivery and accuracy of remote health services [59,75]. Nonetheless, a significant challenge persists in guaranteeing the contextual relevance and appropriateness of chatbot responses, particularly in intricate medical scenarios [59,60]. In addition, the personalization of health care interactions and the precision of information provided by these AI-driven systems are critical areas necessitating extensive future research and rigorous evaluation of their outputs [59,60,299].
Finally, the results were presented solely as a narrative summary [77], which might limit the breadth of perspectives and interpretations that a more diverse methodological approach could have provided. Nevertheless, the inclusion of both benefits and challenges in our reporting suggests that the review may not be significantly biased toward a positive portrayal of chatbots, providing a more nuanced understanding of their role in health care.