[Systematic review of teen pregnancy prevention programs using websites and chatbotsProgramas de prevenção de...

To evaluate the effectiveness of artificial intelligence (AI)-based chatbots in preventing teenage pregnancy compared with traditional methods of sex education or no intervention.

A systematic review of original studies published between January 2010 and December 2024 was conducted based on the PRISMA guidelines and the population, intervention, comparator, outcome, time, and data (PICOT-D) question formulation strategy. Databases such as PubMed, the Cochrane Library, Embase, and Scopus were searched; DeCS and MeSH descriptors were used. Data were managed in Rayyan^® and in an Excel^® spreadsheet. The protocol was registered in PROSPERO (CRD42024580136).

We analyzed 14 studies involving 10,018 participants (71.1% were women). Chatbots demonstrated high usability (SUS score of 77.7, 82nd percentile). Of the total users, 83% interacted with the chatbot, 46% initiated hormonal contraceptives, and 56.8% rated the chatbot as easy to understand. A significant increase was seen in the use of contraceptives (adjusted odds ratio [aOR] = 1.60); p = 0.02) and condoms (aOR = 1.48; p <0.001). There were also improvements in knowledge of emergency contraception (57.3% vs. 47.5%; p <0.01) and access to contraceptive services (95.6% versus 91.8%; p = 0.01).

AI-based chatbots are effective in preventing teenage pregnancy and can be a key complement to traditional sex education methods.