Thais Catalani Morata

A ciência enfrenta um contexto sistemático de baixa reprodutibilidade. Esse cenário tem o potencial de impulsionar a adoção de práticas de ciência aberta. Contudo, a adesão a essas práticas na pesquisa audiológica brasileira permanece desconhecida. Este estudo objetivou avaliar a prevalência e a tendência temporal na adoção de indicadores de reprodutibilidade em artigos de audiologia publicados em periódicos brasileiros. Trata-se de um estudo de metaciência, transversal e documental, que analisou seis indicadores de reprodutibilidade em 250 artigos empíricos publicados em três periódicos brasileiros de audiologia entre 2021 e 2025. Os resultados confirmaram parcialmente a hipótese de baixa adoção estrutural, com frequências variando entre 0% e 13,6%. Embora tenha ocorrido um aumento nas declarações e no compartilhamento efetivo de dados no período, dimensões fundamentais para a transparência da pesquisa continuam negligenciadas. O pré-registro de protocolos (0,8%), a adesão a recomendações metodológicas (1,2%), a disponibilização de código aberto (0%) e os estudos de replicação (0%) apresentaram-se praticamente inexistentes no período analisado. Foi verificada tendência de aumento nos indicadores de declaração e compartilhamento de dados. O desenvolvimento de iniciativas de reprodução de resultados e replicação metodológica de estudos são recomendados como caminhos adotados por outras áreas das ciências da saúde. A promoção de diretrizes editoriais alinhadas aos princípios de ciência aberta e a inclusão do letramento em ciência aberta configuram elementos relevantes para a qualificação das práticas de produção e comunicação científica.

: Accurate health information is ineffective if patients cannot understand it. Large Language Model (LLM) health research values veridical precision; however, linguistic accessibility remains an under-examined component of output quality and usability. This study investigated two sources of variability in readability classification: differences across LLM systems and across readability metrics. The analysis tested 1,120 data points from seven systems in English and Portuguese, comparing baseline responses with a Wikipedia-grounded condition. Content was assessed using five standard readability metrics that measure distinct aspects of text complexity. Systems were statistically homogeneous at baseline but became significantly heterogeneous under Wikipedia grounding, indicating variability in the combination of Retrieval-Augmented Generation (differential readability effects of the same source-grounding instruction across systems). Significant metric variability was observed in all conditions, showing that readability metrics are not interchangeable. Although retrieval grounding is commonly used to improve accuracy, our findings show a trade-off: verified-source grounding can yield inconsistent readability. Therefore, evaluation protocols should use transparent, vendor-agnostic criteria, with metric-specific and language-aware thresholds, and be applied whenever models or grounding configurations change to support accessible cross-language health communication.

Extensive training has been shown to improve the effectiveness of hearing protection devices (HPDs) in reducing noise exposure for workers. This study aimed to (1) assess the evidence for short- and long-term effectiveness of HPD fit testing paired with extensive instructions and (2) identify the type and frequency of refresher instruction associated with maintaining its effectiveness. A literature search was completed on February 5, 2025. The inclusion criteria focused on uncontrolled before-and-after (UBA) studies that reported personal attenuation ratings (PARs) provided by HPDs in the field. A total of 23 UBA studies involving 5,575 workers were included, meta-analyses were performed when multiple comparable studies were located. For earplugs, retesting immediately after extensive training increased the mean PAR by 12.88 dB (13 studies) compared with the pre-intervention PARs without instructions. The improvement decreased to 8.84 dB (2 studies) within 6 months and further reduced to 4.46 dB (2 studies) at 12 months. For earmuffs, only one study reported increase in PARs of 5, 2, and 3 dB for immediate, short-term, and long-term assessment, respectively. Introducing extensive refresher training within a 6-month interval increased the mean PAR from earplugs by 1.53 dB (2 studies) compared with the post-intervention PAR retested immediately after extensive training, with a further increase of 3.45 dB (1 study) following annual extensive instruction. Simple refresher instructions within a 6-month interval maintained the PAR values (1 study). However, without any refresher training, the PAR decreased by 2.64 dB (2 studies) within 6 months and 3.09 dB (2 studies) over 6 months or longer. Refresher training may increase the probability that workers receive efficient protection from their HPDs. However, more high-quality studies could help inform the optimal type and frequency of refresher training.

The concept of an encyclopedia is useful for young people, the readers of tomorrow, and Klexikon is such a wiki encyclopedia (in German). Minors are a special audience group. Therefore, special considerations are necessary when we address minors and especially those younger than 14 years. The Wikimedia Foundation has recently acknowledged this with its Child Rights Impact Assessment report. We, therefore, introduce the topic and present a concept for an encyclopedia for children. The concept makes it possible to create a useful website within two or three years with just a handful of authors but also invites rethinking editing processes and expectations about an encyclopedia. We will especially have a look at opportunities in the Spanish speaking world. Creating such an encyclopedia is a challenge, but rewarding and above all realistic, also in your language.

The presentation covered the Colombian version of Reading Wikipedia in the Classroom (RWC). Our adaptation of RWC to the Colombian context, in particular to work in two small towns, aimed to have cultural exchange to incorporate the knowledge of the region and to recognize the existence of a collective subject, since the participants are not isolated in the world. The collective of Colombia Wikimedians should assume the co-responsibility of implementation, promoting the creation of value networks of mutual commitment. This adaptation of the program kept three modules from its initial version: access to information, evaluation of information, and creation of information. The deployment in Colombia, called “Leamos y Escribamos Wikipedia” (Reading and Writing Wikipedia), has led to conceptual, methodological, and implementation innovations based on asking what the real needs of people, teachers, and schools are. This new component was named “engagement with the World”. It not only included moments to teach and co-construct knowledge and tools but also included interludes between training meetings. The Wikimedia Colombia team and Wikimedistas Wayuú would visit the schools and accompany teachers in the classroom. This presentation illustrates a new way of thinking about Wikimedia platforms in direct relation to the educational contexts and realities in which they are used.

The EduWiki Conference 2025 was a platform for shared experiences from working in the education space of Wikimedia. The conference was organized by the Wikipedia & Education User Group in collaboration with host Wikimedia Colombia and counted with partial funding from the Wikimedia Foundation. It took place in Bogotá, Colombia, May 30-June 1, 2025 and brought together 90 individuals—educational leaders from Wikimedia and organizations sharing its mission. This conference followed the Wikimedia+Education Conference 2019 held in Donostia, Spain and and EduWiki Conference 2023 held in Belgrade, Serbia.

This session presents early experiments using a Retrieval-Augmented Generation (RAG) approach to prepare wiki edits within an open knowledge ecosystem. These experiments are part of a larger project on AI-assisted collaborative editing and represent the initial stages of the work outlined in this publication. This pilot phase was conducted with pedagogy students during the first semester of 2025. The session will explore how RAG was integrated into wiki edit preparation, highlighting methodological choices, tools used, and key challenges encountered in an educational setting. Preliminary findings suggest that RAG can enhance content preparation by improving the review of academic sources, breaking language barriers, and enabling a multimedia approach. These early insights contribute to the broader conversation on AI-enhanced open knowledge practices, shaping future developments in post digital editing environments and informing the ethical, technical, and pedagogical dimensions of AI-assisted knowledge production in education.

Ototoxic chemicals in the workplace can pose a risk to hearing and balance functions. Our objective was to identify evidence-based practices for occupational health settings in managing ototoxicity. This resulted in the document, Health Management of Workers Exposed to Ototoxic Chemicals, created by the International Ototoxicity Management Group. To develop a practical approach for any workplace, we reviewed a variety of sources and used an international panel of interdisciplinary experts. Evidence included data from experimental, observational, and review studies. Thirty-two subject matter experts were invited to review the document; twenty-two completed the review and unanimously endorsed the ototoxicity management system as proposed. Six key action steps were proposed to: (1) identify workers exposed to ototoxic chemicals, (2) perform auditory and vestibular assessments, (3) follow-up after monitoring health, (4) document worker data, (5) maintain healthy safety culture, and (6) review ototoxicity management approach. These steps focus on the management of workers who are at-risk for workplace ototoxic chemical exposure at any level (with or without concurrent noise exposures). Early identification strategies include self-report questionnaires; auditory testing; vestibular screening; referrals for diagnosis; management of cases; and monitoring of exposure scenarios to prevent further cases.

Objective: This study estimated the prevalence of hearing difficulty and workplace exposures to ototoxic chemicals and noise within US industries and occupations. Methods: This cross-sectional study examined 2023 National Health Interview Survey data for 17,726 workers. Weighted prevalence and adjusted prevalence ratios of self-reported hearing difficulty and self-reported workplace exposures to ototoxic chemicals, noise and a combination of these exposures were estimated. Results: Eleven percent of workers reported hearing difficulty. Within the past year, 13% were exposed to ototoxic chemicals, 16% were exposed to noise, and 7% were exposed to both. Twenty-eight percent had a history of noise exposure. Conclusions: Workplace hearing hazards are common. This study was the first to report the overall prevalence of ototoxic chemical exposure among US workers. This exposure cannot be overlooked when protecting worker hearing.

Background: Several health institutions developed strategies to improve health content on Wikimedia platforms given their unparalleled reach. The objective of this study was to compare an online volunteer-based Wikimedia outreach campaign with university course Wikipedia assignments (both focused on improving hearing health content in Wikimedia’s public digital knowledge archives), in terms of the reach of the contributions and the extent of the participants’ input. A secondary objective was to examine the feasibility and the implementation of the different strategies. Methods: The research team partnered for the (1) coordination of improvements in hearing and healthcare content through educational programs using Wikimedia platforms, (2) participation in the global campaign Wiki4WorldHearingDay2023 and (3) evaluation of the proposed strategies. Metrics used in the comparison of the two strategies included the number of articles edited, number of views of the edited articles (as reach) and the extent of edits, captured as the number of words. The feasibility evaluation included assessing recruitment success and the implementation of the proposed plan among faculty, students from various university programs, and volunteers representing different constituencies. Results: The effort increased the availability of quality plain language information on hearing conditions and hearing care. Both strategies demonstrated to be feasible by their success in recruiting participants who contributed to the effort and by measurable outputs as edits. The contribution of content to Wikimedia platforms as part of education activities provided a more robust result. Wiki4WorldHearingDay2023 145 participants (78 from educational programs) contributed 167,000 words, 258 + references and 140 images to 322 Wikipedia articles (283 existing and 39 new ones), which were viewed 16.5 million times. Contributions occurred in six languages. Edits in Portuguese, mainly by those involved in educational programs, led the number of articles (226 or 70.2%) that were expanded or created during the 5-month tracking period. Conclusions: The elements that contributed to the success of the studied strategies include an impact topic, coordination with educational programs, international multidisciplinary collaborations, the dissemination of the initiative in several platforms, connection with a robust local Wikimedia affiliate, and the use of a technical infrastructure that provides metrics and coordination mechanisms. Graphical