Impact of medical technologies may be predicted using constructed graph bibliometrics.

Scientific research is driven by allocation of funding to different research projects based in part on the predicted scientific impact of the work. Data-driven algorithms can inform decision-making of scarce funding resources by identifying likely high-impact studies using bibliometrics. Compared to standardized citation-based metrics alone, we utilize a machine learning pipeline that analyzes high-dimensional relationships among a range of bibliometric features to improve the accuracy of predicting high-impact research. Random forest classification models were trained using 28 bibliometric features calculated from a dataset of 1,485,958 publications in medicine to retrospectively predict whether a publication would become high-impact. For each random forest model, the balanced accuracy score was above 0.95 and the area under the receiver operating characteristic curve was above 0.99. The high performance of high impact research prediction using our proposed models show that machine learning technologies are promising algorithms that can support funding decision-making for medical research.

Complex decision making in an intensive care environment: Perceived practice versus observed reality.

RATIONALE: Advancing our understanding of how decisions are made in cognitively, socially and technologically complex hospital environments may reveal opportunities to improve healthcare delivery, medical education and the experience of patients, families and clinicians. AIMS AND OBJECTIVES: Explore factors impacting clinician decision making in the Boston Children's Hospital Cardiac Intensive Care Unit. METHODS: A convergent mixed methods design was used. Quantitative and qualitative data sources consisted of a faculty survey, direct observations of clinical rounds in a specific patient population identified by a clinical decision support system (CDSS) and semistructured interviews (SSIs). Deductive and inductive coding was used for qualitative data. Qualitative data were translated into images using social network analysis which illustrate the frequency and connectivity of the codes in each data set. RESULTS: A total of 25 observations of eight faculty-led interprofessional teams were performed between 12 February and 31 March 2021. Individual patient characteristics were noted by faculty in SSIs to be the most important factor in their decision making, yet ethnographic observations suggested faculty cognitive traits, team expertise and value-based decisions were more heavily weighted. The development of expertise was impacted by role modeling. Decisions were perceived to be influenced by the system and environment. CONCLUSIONS: Clinician perception of decision making was not congruent with the observed behaviours in a complicated and dynamic system. This study identifies important considerations in clinical curricula as well as the design and implementation of CDSS. Our method of using social network analysis to visualize components of decision making could be adopted to explore other complex environments.

Using social network analysis to confirm the 'gambit of the group' hypothesis for a small cetacean.

Social network analysis (SNA) can be used to explore a population's social structure and how individuals contribute to social cohesion. Quantifying relationships between individuals in a network can vary depending on the data available or the relationship of interest. Studies of readily visible species can use direct interaction measures in SNA, while studies of cryptic species usually rely on the 'gambit of the group'; individuals observed in a group are considered associates. This study compared the association and pectoral fin contact (PFC) networks of Atlantic spotted dolphins around Bimini to test the 'gambit of the group' hypothesis. The association network had nearly three times as many edges than the PFC network. Still, the two networks were correlated; individuals with a relationship in one network had a comparable relationship in the other. Many network measures were also correlated across networks, suggesting association is an acceptable substitute for physical interaction in certain cases. The current study supports the 'gambit of the group', but also highlights the importance of considering what types of relationships are used in the analysis of the social system of the focal species.

The partial merger of two dolphin societies.

Interactions between mammalian social groups are generally antagonistic as individuals in groups cooperate to defend resources from non-members. Members of the family Delphinidae inhabit a three-dimensional habitat where resource defence is usually impractical. Here, we describe a long-term partial fusion of two communities of Atlantic spotted dolphins (Stenella frontalis). The northern community, studied for 30 years, immigrated 160 km to the range of the southern community, observed for 20 years. Both communities featured fission-fusion grouping patterns, strongest associations between adult males, and frequent affiliative contact between individuals. For the 5-year period following the immigration, we found members of all age classes and both sexes in mixed groups, but there was a strong bias toward finding immigrant males in mixed groups. Some association levels between males, and males and females, from different communities were as high as the highest within-community associations. Affiliative contacts indicate that these individuals were forming social relationships. The mixing of two separate social groups with new bond formation is rare in terrestrial mammal groups. Such mixing between spotted dolphin groups suggests that adaptations to respond aggressively to 'outsiders' are diminished in this species and possibly other ecologically similar dolphins.