Arthur Sá Ferreira

Health locus of control and its relationship with #kinesiophobia, #disability, and prognosis in chronic low back #pain. Brazilian Journal of Pain, 9, e202670 doi.org/10.63231/2595-… @lulunkes

🆕 C5–C6 and thoracic spine mobilization with postural correction exercise compared with conventional therapy in patients with adhesive capsulitis: a two-group, parallel-arm, single-blinded, randomized clinical trial—study protocol @BMC_series doi.org/10.1186/s13063…

“The next unit of science: Is the scientific paper due to be replaced?” thetransmitter.org/from-bench-to-…

Many researchers overlook or misinterpret statistical assumptions. Misunderstanding these assumptions can lead to using the wrong statistical tests, drawing inaccurate conclusions, and weakening a study’s credibility. Even common analyses can produce misleading results if their assumptions are not properly checked. I recently came across an R Shiny app called AssumpSure, created by Ahmed Bargheet, designed to address these problems. It offers an intuitive way to check assumptions, interpret results, and choose the right statistical approach. Here’s what it can do: ✔️ Automatically check the assumptions for your chosen test ✔️ Clearly explain what the results mean ✔️ Recommend the right statistical method if assumptions are not met ✔️ Run the analysis and create publication-ready plots ✔️ Download all results and figures for your paper or report Want to explore this free, easy-to-use Shiny app? Take a look: github.com/Ahmedbargheet/… Subscribe to my email newsletter for more practical tips on statistics, data science, R, and Python. Learn more by visiting this link: statisticsglobe.com/newsletter #RStats #database #statisticians #DataAnalytics

Hypothesis generation is fundamental to science. What happens when we outsource this task to a machine? In this month's Editorial we explore the potentially disruptive shift of using AI for hypothesis generation. nature.com/articles/s4159…

For foundation models in radiology, the answer matters; but the pathway to that answer may matter even more. 🧠🤖 This opinion paper argues that post-deployment monitoring of foundation models in radiology must go far beyond checking final outputs, because model behaviour is shaped by invisible deployment settings, user prompts, and dynamic retrieval contexts. The authors make a strong case for institutional deployment with true observability, so health systems can detect behavioural drift before it affects patient care. A sharp and timely read for anyone thinking seriously about moving LLMs and multimodal AI from exciting demos into safe clinical infrastructure. "This paper reframes AI governance beautifully: for foundation models, safety is not just about whether the final answer “looks fine,” but whether we can understand, monitor, and learn from the entire system pathway that produced it." - Amit Gupta buff.ly/RSDPqor (Vera Sorin et al.)

Goodness-of-fit tests or measuring contamination fractions? @Statistics_Jour tandfonline.com/doi/full/10.10…

#J_Epidemi Most Accessed: Understanding Marginal Structural Models for Time-Varying Exposures: Pitfalls and Tips Tomohiro Shinozaki & Etsuji Suzuki jeaweb.jp/journal/mostAc… @J_Epidemi

Researchers propose a "transparent universal credit system to incentivize peer review." pnas.org/doi/10.1073/pn…

Experienced researchers are less likely to produce ‘disruptive’ science than are those just starting their careers, finds an analysis of scientific papers published by 12.5 million researchers over 60 years go.nature.com/42jgSHL

This is a genuinely useful read for anyone trying to modernize their R workflow around reproducibility, Quarto, and large-scale analysis.#rstats" target="_blank" rel="nofollow noopener">hbiostat.org/rflow/#rstats #reproducibility #quartopub

Uso da Inteligência Artificial na Academia doi.org/10.17605/OSF.I… via @OsfFramework

Immediate Effects of Self-Massage Compared to Static Stretching on Neuromuscular and Functional Performance of Physically Active Healthy Adults: A Crossover Study doi.org/10.1016/j.jore…

‘Seeing pain differently’: the impact of physiotherapists’ attitudes and treatment beliefs on postural assessment in neck pain – an eye-tracking study @neymeziat DOI: 10.2478/advrehab-2026-0002 reference-global.com/article/10.247… #pain #neckpain #msk #physiotherapy

AI is, at its core, statistics. Better AI starts with better questions about bias, assumptions, and accuracy. A sharp read from Aimée Hall and the RSS 👇 select-statistics.co.uk/blog/think-lik… #AI #DataScience #Statistics #RStats

📚Research Note: Distinguishing Mixed Methods from Multi-Methods Research – Why Integration Matters🔎 Read here: doi.org/10.1016/j.jphy… @sherbino

#J_Epidemi Most viewed on J-Stage (Mar. 2026): Methodological Tutorial Series: Confounder Selection and Sensitivity Analyses to Unmeasured Confounding From Epidemiological and Statistical Perspectives Kosuke Inoue et al. doi.org/10.2188/jea.JE… @J_Epidemi

New #OpenAccess research from Archives of Rehabilitation Research & Clinical Translation (ARRCT). This study characterizes changes in physical function from acute hospitalization to home health care across different diagnostic categories, highlighting recovery patterns during care transitions. Read the article: archives-rrct.org/article/S2590-… #ACRM #ARRCT #OpenAccess #RehabilitationResearch #CareTransitions #PhysicalFunction #HomeHealth #HealthServices #ClinicalResearch

🤖 🧘 Between Automation and Alienation: Rethinking AI in Radiology As artificial intelligence becomes increasingly embedded in radiology workflows, an important question emerges: is AI reducing workload — or simply redistributing it? This European Radiology article explores the evolving relationship between radiologists and AI, highlighting a growing tension between automation and professional autonomy. AI promises efficiency, but the reality is more nuanced. Gains in speed and throughput may be counterbalanced by new demands, like increased oversight, alert fatigue, documentation burden, and lingering ambiguity around accountability when errors occur. In some settings, improved productivity risks being reframed not as relief, but as expectation. 💬 Do you think AI will ultimately reduce or increase your workload? #AI #Radiology #FutureOfRadiology #DigitalHealth buff.ly/CbvFpei (Elif Can & Jawed Nawabi)

“A questão central, portanto, não é se a inteligência artificial deve integrar a pesquisa científica, mas como fazê-lo sem comprometer os fundamentos que sustentam a produção do conhecimento.” ✅