Establishing a task force promptly, including musicians and members of the local community, can critically contribute to the success of an intervention. Failure to do so will likely limit the delivery of the preventive messages, even if these are based on scientific evidence (e.g., hand washing). The inclusion of scientists representing a range of disciplines is beneficial (20, 69, 70). However, in practice, cultural divergences and mismatched expectations may create barriers. Misunderstandings often arise over financial matters in that economic precarity on the part of musicians may compromise their involvement, while scientists might struggle to obtain funding matching musicians' expectations. The motivation to deliver a musically-sound message to help society might outweigh many obstacles, because many musicians are conscious of their important role in contributing to overall societal wellbeing [e.g., rappers and griots in Africa (71)] and, to their local community (Box 1). In fact, several initiatives related to disease prevention emanate from the musicians themselves (e.g., MC Fioti's COVID-19 “vaccine anthem” in Brazil). Collaborations with scientists during these initiatives can refine evidence-based messages and test the effectiveness of campaigns.

For many infectious diseases, preventive measures to reduce transmission risks are well-known, including hand washing, social distancing, using personal protective equipment, and sanitation (6, 67). In such cases, selecting behaviors to be promoted in songs is relatively straightforward. However, when EIDs emerge, preventive measures are not necessarily well-known. Thus, the task force will have to balance the need for early message broadcasting with that of gathering sufficient scientific evidence to convey the appropriate message. As seen during the COVID-19 pandemic, an accurate understanding of the most relevant preventive behaviors and interventions (e.g., use of masks, hand washing, and vaccination) requires time and messages might need changing iteratively depending on the epidemiological situation. In addition, it might not be possible to convey multiple messages in a single song, to avoid both confusion and lack of message retention (66). Thus, we recommend selecting preventive messages during song composition and adapting campaigns to changing epidemiological patterns. This might require the development of multiple songs over time or tailored to populations with different levels of EID risk. Finally, there is a risk of conveying a message that is misinterpreted by the population leading to unintended consequences, such as negative attitudes about infected individuals or suspect wildlife reservoirs. These risks should be considered during an open dialogue between social scientists, the local community, and public health authorities, especially given the limited time for extensive piloting in crisis situations. Overall, balancing possible risks and benefits might limit the amount and type of information delivered in each song.

Effective mass media campaigns typically require repeat broadcasting of preventive messages through the radio or television. For example, previous mass health media campaigns have suggested using a “message saturation” strategy for radio broadcasting to penetrate the audience, with 6–12 messages per day over several months (72, 73). The type of funding necessary to attain such coverage is often unavailable at the start of an outbreak, particularly when an EID affects a low-income country (e.g., Ebola outbreaks). Early engagement of national and international funders and the media industry is therefore essential to obtain the funds required to achieve the desired broadcasting frequency and population exposure. Wide broadcasting of health awareness messages becomes even more critical when EIDs spread over large geographic areas and have the potential to become a pandemic. Additional challenges arise from adapting songs to further cultural backgrounds, languages, and populations (e.g., from rural to urban areas). Furthermore, society often responds to both musical messages and to performers. Thus, engaging appropriate messengers or influencers with sufficient notoriety to engage a large population and promote desired behaviors is key (68, 74), as demonstrated by the “Natural Opinion leader model” used in HIV prevention (25). However, this further entails securing adequate funding and motivating prominent musicians to promote health messages that might not generate high coverage and interest. Open dialogue with musicians through online round tables, seminars, and direct communication is an essential contributor to such initiatives.

A key issue of music-based interventions is how to monitor and test their effectiveness in improving preventative practices and reducing disease burden at the population level. This is particularly challenging when dealing with the often-unpredictable nature of EID because data prior to an intervention are often unavailable. Moreover, the need to act rapidly precludes the design and deployment of scientifically sound methods to monitor awareness programs as they are implemented. The need for a rapid response can also make the use of “control groups” (that do not receive the intervention) unethical. Disentangling the specific effect of musical campaigns on reducing population-level disease burden is challenging, since music is often used in addition to other preventive measures (e.g., distribution of soap or vaccination campaigns). Indeed, the role of music campaigns is typically to improve the uptake or adherence to these other measures, and therefore will very much depend on their availability, accessibility, and appropriateness. Knowledge alone will not necessarily translate into improved practices for many reasons that need to be better understood.

Despite these challenges, several promising initiatives are attempting to test the effectiveness of art-based campaigns in reducing disease burden. Methods include the use of quantitative disease incidence data (63) and of approaches to obtain qualitative data through surveys/interviews (55) that could be adapted to the specific challenges of EIDs. Work led by the Development Media International (https://www.developmentmedia.net/) has focused on evaluating mass media campaigns delivered by radio, television, and mobile videos using cluster randomized controlled trials (RCTs). In this design, a randomized group of subjects (cluster) is used for the evaluation, which enables large-scale assessments. For example, Sarrassat et al. (63) used data from health facilities and from cross-sectional household surveys conducted before, during, and after radio campaigns to evaluate whether they could reduce under-five child mortality in Burkina Faso. Similarly, Lemieux et al. (25) used questionnaires to evaluate a more localized music-based HIV prevention intervention among urban adolescents of three high schools in the USA.

Aiming to avoid the delayed community awareness and devastating consequences of the previous outbreak in Guinea and neighboring countries (6), we created the “Ebola Gondwana task force” the day a new outbreak was officially announced [14^th of February (75)]. The name “Gondwana” is derived from the supercontinental block that included both Africa and Latin America during the Proterozoic era, reflecting the inclusion of both Latin American and African musicians—from Chile, Brazil, Liberia, and Guinea—and scientists—from Chile and Cameroon. The mandate of the task force was to work with local communities to create awareness about the outbreak as quickly as possible through musical content that would translate science-based prevention guidelines into more accessible messages developed by community members and local musicians.

Due to the COVID-19 pandemic, all communication was done virtually. Our task force immediately contacted several stakeholders responsible for handling the outbreak, including the Regional Office for Africa of the World Health Organization (WHO), the Liberia Ministry of Health, and the United Nations Children's Emergency Fund (UNICEF) in Guinea to make them aware of our initiative and keep them posted on the development of our musical material. Since a member of our group (P.T.D) was based in the city of Zor Zor in Liberia, 65 km east from the first case declared in Nzerekore-Guinea, we hired one local assistant in Liberia and Guinea who contacted local musicians and members of the community to co-develop preventive messages, discuss airtime with local radios, and continuously monitor the epidemiological situation.

We first created preventive messages based on the WHO's guidelines for Ebola prevention as well as lessons from previous outbreaks. This promoted hand washing with soap, avoiding contact with corpses during funerals and visiting a doctor if people develop symptoms indicative of Ebola disease (76). The message content was discussed with an anthropologist working in the area, members of the local community and other musicians working in Africa (F.D.S). In Liberia, our local assistant engaged with two community members (female and male) recording 3–4 min Ebola preventive messages in Loma language (spoken in Liberia and Guinea). Two Latin American musicians (Skyfonik from Chile and Choro de Favela from Brazil) of the task force then added African-style background music to the lyrics. In Guinea, our local assistant in Nzerekore recorded three women giving 1-min prevention messages in three languages (Kpelle, Mania, and Loma) to which we added a musical beat.

A member of our team (J.A.B) composed the lyrics and chords of the “Ebola is back” song in French. This was subsequently arranged, and studio recorded in Nzerekore by the locally known group “Combattant compétent au mic” comprising three musicians (Competent, La Kaka, and Charly). This group was supported by a collective of musicians that contributed to the song in five local languages. The song included key Ebola and COVID-19 prevention messages such as “Ebola is back but we can kick it out, let's wash our hands,” “Ebola is back so we shouldn't drag our feet, let's not touch the dead” or “Let's wear masks, night and day,” and “Ebola and your friend COVID, this is our home, the vaccine will kick you out.”

In Liberia, we paid the community-radio station in Zor Zor (Radio Life) to broadcast these messages for a 2-week period twice a day (morning and afternoon) from February 18th. The radio also broadcast the message at other times of the day, hosted live discussions with the community on Ebola and continued broadcasting the message after the initial 2 weeks. In Guinea, we paid one of Nzerekore's largest radio stations, “Radio Espace Forêt,” to broadcast the 1-min preventive messages once a day for 2 weeks from February 25^th. We also shared the message with UNICEF in Nzerekore. The “Ebola is back” song was broadcast on two main radio stations in Nzerekore particularly suited to reach rural and isolated communities: “Radio Djoma FM” (broadcast free of charge from March 17^th and including a live interview with the musicians) and “Emergence FM” (2-week broadcasting twice a day from April 17^th). Finally, the song was made freely available online (https://soundcloud.com/unabenradio/ebola-volvio-ebola-est-de-retour) so that it could be shared with NGOs and researchers. The campaign was also posted in our research group's social media (@themonkey_lab). No further action was undertaken until Guinea declared that the outbreak was over on the 19th of June 2021 (76). The project cost 1,200 dollars, raised in Chile through contributions by family and friends. Funds were transferred directly to local assistants and musicians to cover related field activities, studio recording, and musician time. Half of the money was needed to pay for radio airtime. All Latin American musicians and scientists contributed free of charge. The enthusiasm and daily communication between researchers and musicians largely contributed to the rapid and successful broadcasting. However, no funds were available to broadcast the messages for a longer period or at higher daily frequency.

Given the urgency of the response and our limited funding to hire local assistants for long periods of time, we were unable to monitor the effectiveness of this campaign in terms of adoption of individual and social preventive behaviors. Similar limitations likely apply to other outbreak situations, which could help explaining the limited information available on the effectiveness of these music interventions.