Reliability and accuracy of telemedicine-based shoulder examinations.

BACKGROUND: Telemedicine has become a critical component in the evaluation and management of patients with shoulder pathology. However, the interobserver reliability of findings on virtual physical examination relative to in-person evaluation is unclear. The purpose of this study was to determine the reliability of prerecorded video telemedicine for the evaluation of shoulder pathology compared with traditional in-person physical examination. METHODS: New patients with unilateral shoulder pain presenting to a single-surgeon shoulder clinic were recruited between July and November 2020. In 1 visit, patients were evaluated with standardized in-person and video-guided telemedicine physical examinations in randomized order. Patients were evaluated for range of motion (ROM) and symptoms including pain, weakness, and apprehension. The telemedicine examination was recorded and consisted of a video guide displaying self-directed shoulder examination maneuvers that patients performed during remote coaching by an independent non-physician observer. The in-person physical examination was performed by the treating physician. The telemedicine videos were evaluated by two independent observers for interobserver reliability. The treating physician subsequently evaluated the telemedicine videos after a minimum two-month washout period for intraobserver reliability and intra-platform reliability. Interobserver and intraobserver reliability analyses were conducted using Kuder-Richardson formula 20 (KR-20). Specificity and likelihood ratios were calculated with P < .05 representing statistical significance. RESULTS: A total of 32 patients (17 male and 15 female patients; average age, 50.2 ± 16.2 years) were included in the analysis. Overall Kuder-Richardson formula 20 (KR-20) reliability across 40 physical examination maneuvers was 0.391 ± 0.332 (76.4% ± 15.4% agreement) between the in-person and telemedicine examinations. Telemedicine maneuvers examining ROM limitations had the highest degree of reliability, sensitivity, specificity, and likelihood of also producing a positive finding on the in-person examination (0.700 ± 0.114, 66.5%, 81.0%, and 6.06, respectively). Telemedicine maneuvers identifying apprehension associated with glenohumeral instability were found to have the lowest reliability, sensitivity, and likelihood of producing a positive finding on the in-person examination (0.170 ± 0.440, 23.5%, and 0.518, respectively). All patients were satisfied with their telemedicine experience. CONCLUSION: The overall reliability of a non-physician-directed video-guided telemedicine examination ranged from unacceptable to good. Shoulder ROM limitations identified during the telemedicine examinations were found to be the most reliable, whereas evaluations of instability were found to be the least reliable. Although initial telemedicine evaluation by a non-physician may be appropriate for ROM evaluation, in-person physician evaluation is recommended to confirm suspected diagnoses, especially if clinical concern for shoulder instability exists. Alternative methods of telemedicine delivery should be explored to improve the reliability of self-directed physical examination maneuvers.

Generative Pre-trained Transformer for Pediatric Stroke Research: A Pilot Study.

BACKGROUND: Pediatric stroke is an important cause of morbidity in children. Although research can be challenging, large amounts of data have been captured through collaborative efforts in the International Pediatric Stroke Study (IPSS). This study explores the use of an advanced artificial intelligence program, the Generative Pre-trained Transformer (GPT), to enter pediatric stroke data into the IPSS. METHODS: The most recent 50 clinical notes of patients with ischemic stroke or cerebral venous sinus thrombosis at the UTHealth Pediatric Stroke Clinic were deidentified. Domain-specific prompts were engineered for an offline artificial intelligence program (GPT) to answer IPSS questions. Responses from GPT were compared with the human rater. Percent agreement was assessed across 50 patients for each of the 114 queries developed from the IPSS database outcome questionnaire. RESULTS: GPT demonstrated strong performance on several questions but showed variability overall. In its early iterations it was able to match human judgment occasionally with an accuracy score of 1.00 (n = 20, 17.5%), but it scored as low as 0.26 in some patients. Prompts were adjusted in four subsequent iterations to increase accuracy. In its fourth iteration, agreement was 93.6%, with a maximum agreement of 100% and minimum of 62%. Of 2400 individual items assessed, our model entered 2247 (93.6%) correctly and 153 (6.4%) incorrectly. CONCLUSIONS: Although our tailored generative model with domain-specific prompt engineering and ontological guidance shows promise for research applications, further refinement is needed to enhance its accuracy. It cannot enter data entirely independently, but it can be employed in tandem with human oversight contributing to a collaborative approach that reduces overall effort.

Inferring population dynamics from single-cell RNA-sequencing time series data.

Recent single-cell RNA-sequencing studies have suggested that cells follow continuous transcriptomic trajectories in an asynchronous fashion during development. However, observations of cell flux along trajectories are confounded with population size effects in snapshot experiments and are therefore hard to interpret. In particular, changes in proliferation and death rates can be mistaken for cell flux. Here we present pseudodynamics, a mathematical framework that reconciles population dynamics with the concepts underlying developmental trajectories inferred from time-series single-cell data. Pseudodynamics models population distribution shifts across trajectories to quantify selection pressure, population expansion, and developmental potentials. Applying this model to time-resolved single-cell RNA-sequencing of T-cell and pancreatic beta cell maturation, we characterize proliferation and apoptosis rates and identify key developmental checkpoints, data inaccessible to existing approaches.

Oxidative Degradation of Polypropylene Mesh in E. coli Environment.

Medical implants of polypropylene (PP) are commonly used in many surgical procedures to support tissues. Previous studies on polypropylene meshes removed from patients demonstrated biodegradation relative to the amount of time after implantation. Among the many possible factors, bacterial colonization is believed to be one of the causes for the biodegradation of PP. To gain insights on this hypothesis, PP mesh samples were tested in Luria-Bertani broth (LB) media containing Escherichia coli (E. coli) to observe possible degradation in a controlled single-organism environment. Mesh samples were immersed in either an LB media with E. coli or a control solution, and the biodegradation was measured at 1-, 2-, and 3-month intervals. The samples were then harvested from both LB media with E. coli and the control media for analysis, and results were then compared with pristine polypropylene mesh. The experimental results demonstrate qualitative and quantitative bioerosion, increased oxygen content, and enhanced hydrophilicity over the surface of the mesh structure, thus confirming the oxidative degradation in vitro.

Visitation by wild and managed bees (Hymenoptera: Apoidea) to eastern U.S. native plants for use in conservation programs.

Addition of floral resources to agricultural field margins has been shown to increase abundance of beneficial insects in crop fields, but most plants recommended for this use are non-native annuals. Native perennial plants with different bloom periods can provide floral resources for bees throughout the growing season for use in pollinator conservation projects. To identify the most suitable plants for this use, we examined the relative attractiveness to wild and managed bees of 43 eastern U.S. native perennial plants, grown in a common garden setting. Floral characteristics were evaluated for their ability to predict bee abundance and taxa richness. Of the wild bees collected, the most common species (62%) was Bombus impatiens Cresson. Five other wild bee species were present between 3 and 6% of the total: Lasioglossum admirandum (Sandhouse), Hylaeus affinis (Smith), Agapostemon virescens (F.), Halictus ligatus Say, and Ceratina calcarata/dupla Robertson/Say. The remaining wild bee species were present at <2% of the total. Abundance of honey bees (Apis mellifera L.) was nearly identical to that of B. impatiens. All plant species were visited at least once by wild bees; 9 were highly attractive, and 20 were moderately attractive. Honey bees visited 24 of the 43 plant species at least once. Floral area was the only measured factor accounting for variation in abundance and richness of wild bees but did not explain variation in honey bee abundance. Results of this study can be used to guide selection of flowering plants to provide season-long forage for conservation of wild bees.