Altmetric: o impacto social dos artigos científicos

• Ano de defesa: 2019
• Autor(a) principal: Araujo, Amanda Costa
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Cidade de São Paulo; Brasil; Pós-Graduação; Programa de Mestrado e Doutorado em Fisioterapia; UNICID
• Palavras-chave em Português: Bibliometria; Publicações periódicas; Mudança social; Mídias sociais; Altmetric; Social impact; Social media; CNPQ::CIENCIAS DA SAUDE::FISIOTERAPIA E TERAPIA OCUPACIONAL
• Link de acesso: https://repositorio.cruzeirodosul.edu.br/handle/123456789/249

Background: There is great interest from authors and publishers in sharing the results of their studies over the internet in order to increase the dissemination of research findings. In this way, articles tend to be discussed and the impact of these articles tends to increase. This novel concept of impact can be measured by a tool named Altmetric. Altmetric aims to understand the individual impact of each article through the attention attracted online. Objectives: The first objective of this thesis was to describe the characteristics and associated factors with Altmetric in randomized controlled trials, systematic reviews and clinical practice guidelines in the field of physiotherapy interventions for patients with low back pain. The second objective was to describe the characteristics of scientific journals and scientific articles that are associated with Altmetric in research fields. The third objective was to describe the characteristics of the 100 articles with the highest Altmetric scores in 2017. Methods: For answering the first objective, a random sample of 200 randomized controlled trials, 66 systematic reviews and 5 clinical practice guidelines were selected from the Physiotherapy Evidence Database (www.pedro.org.au). Data were extracted from the articles referring to the methodological quality of the eligible studies, individual and total scores of the Altmetric mentioned and Altmetric reader. Descriptive analyses and linear regression models were used. For answering the second objective, a methodological review in all areas of research was conducted. Data were extracted from the articles referring to: journal’s impact factor, number of citations, access type, time since publication and dissemination of the article by the journal. Descriptive analyses were used for this study. For answering the third objective, the 100 articles with the highest Altmetric scores were downloaded from the Altmetric's website. Data were extracted from the articles referring to: journal’s impact factor, access type, type of title, dissemination of the article by the journal, if the study hypothesis was confirmed or refuted, number of research centers where the study was conducted, number of months since publication and number of citations normalized by publication time. Descriptive analyses were used for this study. Results: Most randomized clinical trials were published in English, with descriptive title type (title describing the aim, but does not reveal the main conclusions) and published in open access journals that endorse CONSORT (Consolidated Standards of Reporting Trials). A total of four variables were associated with randomized controlled trials and the Altmetric score: journal’s impact factor [β= 3.4 points, (95% CI 0.98 to 5.86)], number of years since publication [β= -4.9 points, (95% CI -8.50 to -1.47)], number of citations divided by the years since publication [β= 5.2 points, (95% CI 2.49 to 7.88)] and descriptive title [β= -29.4 points, (95% CI -46.48 to -12.23)]. Only one variable was associated with Altmetric reader: number of citations divided by the years since publication [β= 10.1 points, (95% CI 7.74 to 12.46)]. Most of the systematic reviews and clinical practice guidelines were published in English, with descriptive title, published in open access journals, multicenter studies and have dissemination of the article by the journal. One variable was associated with systematic reviews and the Altmetric score mentioned: journal impact factor [β= 20.2 points, (95% CI 12.82 to 27.59)]. One variable was associated with Altmetric reader: number of citations divided by the years since publication [β= 5.6 points, (95% CI 4.25 to 6.98)]. The methodological review found a weak to moderate association for the variables journal’s impact factor, number of citations, type of access, time since publication and dissemination of the article by the journal. Besides that, the 100 articles with the highest Altmetric scores were published in English, most published in journals with a high impact factor and with dissemination of the article by the journal. Conclusion: Researchers should preferrably select high impact factor journals for submission and use declarative (title expressing the results of the trial) or interrogative titles (title introducing the trial in the form of a question). These factors increase the likelihood of a larger visibility in online media. We also found that the number of citations and journal’s impact factor are often associated with Altmetric in several areas of research. Finally, the articles with the highest visibility in 2017 were published in English, mostly in high impact scientific journals and used media release produced by the publishing journal.