Photo Credit: Outflow Designs
A systematic review aimed to evaluate the impact of chatbots on contraceptive knowledge, attitudes, and behaviors, as well as to identify best practices for their development. Chatbots, defined as computer programs simulating human conversation, are increasingly used in health care to offer responsive and private counseling. They hold the potential to improve access to contraceptive information and services, facilitating better contraceptive uptake and communication. The review assessed existing evidence to determine whether chatbots effectively support contraceptive outcomes and gathered recommendations for their development, including considerations of cost-effectiveness.
The researchers conducted a thorough search of peer-reviewed and gray literature from 2010 to 2022, focusing on sources that evaluated chatbots offering contraceptive information or services. Studies were included if they discussed aspects like contraceptive uptake, knowledge, or negotiation skills. Data were extracted using a specific framework, and a narrative synthesis was used due to insufficient quantitative data. Out of the identified literature, 15 sources—including eight original research papers and seven gray literature reports—were reviewed, representing 16 unique chatbots.
The results highlighted mixed evidence regarding the efficacy of chatbots. A large randomized controlled trial (RCT) indicated no effect on the intention to use contraception. In contrast, a small cohort study showed increased contraceptive uptake among adolescent girls, though its lack of a control group limited the strength of these findings. A development report suggested that chatbots had little impact on improving access to contraceptive services, based on a suboptimal measure—redeemed referral vouchers. Additionally, there was weak evidence suggesting that chatbots might enhance contraceptive knowledge through user interactions, though these findings lacked control groups and long-term follow-up data.
User engagement with chatbots varied significantly, with some chatbots reaching large audiences while others had limited reach. Feedback indicated that users appreciated the convenience, anonymity, and privacy of chatbots. However, negative perceptions were also common, with some users finding the chatbots to be technically unreliable, impersonal, or lacking empathy. Despite these mixed responses, the review found that best practice recommendations for developing chatbots in the field of contraception were consistent with broader guidelines on chatbot use in health care.
The conclusions of the review emphasized the limited and often conflicting evidence regarding the effectiveness of chatbots in improving contraceptive outcomes. While chatbots offer certain advantages like privacy and rapid information delivery, there is a need for more rigorous research to determine their true impact. Specifically, future studies should explore the scalability of chatbots, their effectiveness compared to other digital interventions, and the factors that influence user engagement. Additionally, research should focus on understanding the varying degrees of interactivity between different types of chatbots and their respective impacts on users.
As new technologies, like generative AI chatbots, emerge, the potential for more personalized and effective interactions grows, but so does the need for thoughtful evaluation to ensure these innovations are both effective and safe.
