Nevin Manimala

J Med Econ. 2026 Dec;29(1):654-671. doi: 10.1080/13696998.2026.2630598. Epub 2026 Mar 4.

ABSTRACT

AIMS: This study compares deep learning architectures with traditional regression and tree-based models for individual-level healthcare cost prediction, with particular attention to performance differences across data contexts.

METHODS: We conducted a preregistered systematic review (PROSPERO CRD420251129440). Web of Science, PubMed, Embase, and Scopus were searched through August 2025. Eligible studies used real-world individual-level data (claims, electronic health records, or registries) to predict cost-related outcomes with at least one deep learning architecture and one classical regression comparator, and reported quantitative performance. Data were extracted on population, predictors, outcome horizon, model type, validation strategy, performance metrics, calibration, and interpretability.

RESULTS: Eight studies met inclusion criteria, spanning the United States, Europe, and Asia. In longitudinal designs-such as multi-year claims prediction and medication or hospitalization time-series forecasting-sequential deep learning models, particularly LSTM and CNN-LSTM hybrids, consistently outperformed regression and tree-based algorithms. Reported gains included approximately 10-20% reductions in RMSE/MAE, R² improvements of 0.01-0.15, and AUROC values up to 0.78 for high-risk classification. Across studies, prior costs and utilization were consistently the strongest predictors, while social determinants and free-text features were rarely incorporated. In contrast, for low-dimensional, structured, cross-sectional medical data, generalized linear models and tree-based approaches remain robust baseline models due to their interpretability and calibration stability.

LIMITATIONS: Evidence is based on a small and heterogeneous set of eight studies, with limited external or temporal validation, short prediction horizons, and sparse assessment of calibration, economic interpretability, and fairness, warranting cautious interpretation.

CONCLUSIONS: Deep learning offers clear gains for longitudinal, sequence-rich cost forecasting, whereas tree-based methods remain highly competitive for cross-sectional tabular prediction. Overall, these findings are consistent with the proposed Complexity-Performance Hypothesis, which posits that the predictive advantages of deep learning emerge primarily when model capacity is well matched to data complexity.

PMID:41779998 | DOI:10.1080/13696998.2026.2630598

J Am Acad Orthop Surg Glob Res Rev. 2026 Mar 3;10(3). doi: 10.5435/JAAOSGlobal-D-25-00432. eCollection 2026 Mar 1.

ABSTRACT

INTRODUCTION: Hospitals contribute to substantial environmental waste and greenhouse gas emissions, with operating rooms accounting for 50% to 70% of hospital waste. Arthroscopic rotator cuff repair (RCR), a commonly performed procedure, typically uses disposable instruments to minimize infection risk. There is limited evidence regarding the clinical safety and effectiveness of disposable instruments compared with reusable instruments. We aimed to evaluate whether reusable instrumentation for arthroscopic RCR affects clinical outcomes.

METHODS: This was a retrospective cohort study involving 191 patients undergoing primary arthroscopic RCR. Patients were divided into reusable (N = 89) and disposable (N = 102) instrumentation cohorts. Primary outcomes included rates of postoperative soft-tissue infection and septic revision within 1 year postoperatively. Data were analyzed using frequentist and Bayesian statistical methods.

RESULTS: Infection rates and septic revisions were similar between reusable and disposable instrumentation groups, with one septic revision in each cohort (P = 1.0). Aseptic revision rates were also similar (P = 0.50). Surgical times did not significantly differ between groups (reusable: 1.50 ± 0.33 hours; disposable: 1.61 ± 0.41 hours; P = 0.076). Bayesian analysis supported these findings, demonstrating no meaningful difference in infection risks between groups, with median odds ratios close to 1.0 and credible intervals including 1.0.

CONCLUSION: Observed proportions of revision and infection were similar in magnitude. These findings suggest that reusable instruments have the potential to be a safe and sustainable alternative to single-use instruments in arthroscopic RCR. However, owing to the rarity of infection and revision, future multisite studies are necessary to assess whether risk of these outcomes is noninferior to nonreusable instruments.

PMID:41779932 | DOI:10.5435/JAAOSGlobal-D-25-00432

J Contin Educ Nurs. 2026 Mar;57(3):130-136. doi: 10.3928/00220124-20251204-01. Epub 2026 Mar 1.

ABSTRACT

BACKGROUND: Discrepancies between nurses’ ideal and actual roles can undermine job satisfaction, role identity, and care quality. This study explored how registered nurses in Jordan perceive their ideal versus actual roles and how these perceptions differ by demographic and organizational factors.

METHOD: A descriptive cross-sectional design was used with 357 nurses from governmental, educational, and private hospitals. Data were collected with a sociodemographic questionnaire and the Pieta Nursing Role Conception tool, which evaluates service, professional, and bureaucratic roles. Analyses included descriptive statistics, t tests, and analysis of variance.

RESULTS: A significant overall discrepancy was found (mean difference = 0.52, p < .001), with the largest gaps in service (0.96) and professional roles (0.72). Bureaucratic roles were practiced more than desired (-0.08). Role discrepancies varied by age, hospital type, education, experience, and practice area.

CONCLUSION: The findings highlight the need for continuing education, leadership development, and policy reforms to align nursing roles with professional expectations and improve workforce outcomes.

PMID:41779906 | DOI:10.3928/00220124-20251204-01

J Proteome Res. 2026 Mar 4. doi: 10.1021/acs.jproteome.5c01075. Online ahead of print.

ABSTRACT

Mass spectrometry-based single-cell proteomics emerges as the most promising method for studying cellular heterogeneity at the global proteome level with unprecedented depth and coverage. Its widespread application remains limited due to robustness, reproducibility, and throughput requirements, still difficult to meet as analyzing large cohorts of single cells is necessary to ensure statistical confidence. In this context, we conducted method optimizations at three levels. First, we benchmarked three distinct workflows compatible with the nanoElute2 platform using different sample collection/preparation plate supports (EVO96 oil-free, LF48 oil-based, and LF48 oil-free, a streamlined automated sample resuspension, and direct injection protocol). Then, we compared the optimized EVO96 workflow on nanoElute2 with Evosep-based separations operating at two analytical throughputs (80 and 120 samples per day). Subsequently, we evaluated digestion efficiency using a range of enzyme/protein ratios (1:1; 10:1; 20:1; 50:1) to maximize peptide recovery. Finally, the chromatographic setup was refined to determine the best compromise between throughput and robustness. Altogether, these optimizations allowed to establish a robust workflow quantifying up to 5000 proteins in 10 min gradient time per single HeLa cell at a 55 samples-per-day throughput.

PMID:41779902 | DOI:10.1021/acs.jproteome.5c01075

Elife. 2026 Mar 4;13:RP100623. doi: 10.7554/eLife.100623.

ABSTRACT

Cells must adopt flexible regulatory strategies to make decisions regarding their fate, including differentiation, apoptosis, or survival in the face of various external stimuli. One key cellular strategy that enables these functions is stochastic gene expression programs. However, understanding how transcriptional bursting, and consequently, cell fate, responds to DNA damage on a genome-wide scale poses a challenge. In this study, we propose an interpretable and scalable inference framework, DeepTX, that leverages deep learning methods to connect mechanistic models and single-cell RNA sequencing (scRNA-seq) data, thereby revealing genome-wide transcriptional burst kinetics. This framework enables rapid and accurate solutions to transcription models and the inference of transcriptional burst kinetics from scRNA-seq data. Applying this framework to several scRNA-seq datasets of DNA-damaging drug treatments, we observed that fluctuations in transcriptional bursting induced by different drugs were associated with distinct fate decisions: 5′-iodo-2′-deoxyuridine treatment was associated with differentiation in mouse embryonic stem cells by increasing the burst size of gene expression, while low- and high-dose 5-fluorouracil treatments in human colon cancer cells were associated with changes in burst frequency that corresponded to apoptosis- and survival-related fate, respectively. Together, these results show that DeepTX enables genome-wide inference of transcriptional bursting from single-cell transcriptomics data and can generate hypotheses about how bursting dynamics relate to cell fate decisions.

PMID:41779826 | DOI:10.7554/eLife.100623

PLoS One. 2026 Mar 4;21(3):e0340876. doi: 10.1371/journal.pone.0340876. eCollection 2026.

ABSTRACT

BACKGROUND: Serum calcium, a key biochemical marker in the body, plays a crucial role in maintaining bone health. Nevertheless, research exploring the link between preoperative serum calcium levels and the occurrence of postoperative complications in elderly orthopedic patients is currently lacking.

AIMS: This study sought to assess the ability of preoperative serum calcium levels to predict the occurrence of postoperative complications in geriatric orthopedic surgery.

METHODS: We utilized multivariate logistic regression to identify correlations between serum calcium levels and complications. Generalized additive models to analyze the dose-response relationship with curve fitting and threshold effect evaluation. Subgroup analyses further evaluated the impact of other covariates.

RESULTS: This study included 690 elderly patients undergoing orthopedic surgery. Common postoperative complications primarily included infection, hypoalbuminemia, and electrolyte imbalance, etc. The study demonstrated that preoperative serum calcium levels were an independent protective factor against postoperative complications (OR: 0.24, CI: 0.07-0.76, P = 0.036). When comparing groups based on serum calcium tertiles, patients in the low calcium group exhibited a 79% higher risk of complications compared to the high calcium group (OR = 1.79, 95% CI: 1.12-2.78). Further nonlinear relationship analysis revealed a threshold effect between serum calcium and postoperative complication risk, with a turning point at 2.4 mmol/L. The association was statistically significant below this value but not above it. Subgroup analyses and interaction tests showed that age, gender, comorbidities, and medications, cognitive function, cardiac function, and surgical complexity were not significantly associated with this correlation (P > 0.05 for interaction).

CONCLUSION: Preoperative calcium screening and correction may represent a simple, low-cost strategy to reduce postoperative complications in elderly orthopedic patients. This study provides evidence for the importance of actively correcting calcium levels before surgery and establishes the value of serum calcium as an early warning indicator for poor prognosis.

PMID:41779825 | DOI:10.1371/journal.pone.0340876

PLOS Glob Public Health. 2026 Mar 4;6(3):e0004659. doi: 10.1371/journal.pgph.0004659. eCollection 2026.

ABSTRACT

Despite progress in malaria control, malaria remains a major public health burden in sub-Saharan Africa, particularly among children under five. The introduction of malaria vaccines, including RTS,S/AS01 (Mosquirix) and the recently WHO-recommended R21/Matrix-M, offers renewed hope for reducing malaria morbidity and mortality. The effectiveness of these vaccines, however, depends largely on caregivers’ awareness, knowledge, and attitudes. This study assessed caregivers’ knowledge and attitudes, and the factors associated with awareness of the malaria vaccine in the Tiko Health District of Cameroon. A community-based cross-sectional study was conducted among 410 caregivers of children aged 0-5 years who were selected using a multistage sampling technique. Data were collected using a structured pre-tested questionnaire. Descriptive statistics summarized participants’ characteristics, and knowledge and attitude scores were generated using a structured scoring system with a 60% cut-off defining adequate knowledge and positive attitudes. Logistic regression analysis identified factors independently associated with malaria vaccine awareness with statistical significance set at p < 0.05. The median age of participants was 32 years(IQR:27-40), and most were female(83.2%). Although 60.7% of caregivers had heard of the malaria vaccine, only 26.6% demonstrated adequate knowledge and 25.1% had positive attitudes. Healthcare workers were the primary source of vaccine information(35.4%). Caregivers whose children had a previous malaria episode were less likely to be aware of the vaccine(AOR:0.55; 95% CI:0.28-0.97). Conversely, caregivers who trusted health workers (AOR:3.02; 95% CI:1.83-4.99) and those who routinely attended childhood immunization services (AOR:3.57; 95% CI:2.27-5.60) were more likely to be aware of the vaccine. Caregivers in the Tiko Health District exhibited limited knowledge and generally negative attitudes toward the malaria vaccine. Strengthening health-worker engagement, improving communication during routine immunization services, and addressing gaps in caregivers’ understanding may enhance malaria vaccine uptake in the district.

PMID:41779824 | DOI:10.1371/journal.pgph.0004659

Proc Natl Acad Sci U S A. 2026 Mar 10;123(10):e2526432123. doi: 10.1073/pnas.2526432123. Epub 2026 Mar 4.

ABSTRACT

The present study examined the accuracy of conscious representations that emerge from implicit statistical learning (ISL), a fundamental cognitive process through which we extract regularities in the environment. While ISL has several characteristics of unconscious processing (e.g., it operates unintentionally, produces subjectively unconscious knowledge), participants in ISL experiments always report some fragmentary conscious knowledge. Thus, the notion that ISL is truly an unconscious process has been the subject of perpetual debates. In the present study, we challenge the assumption that these conscious reports reflect direct access to the acquired knowledge. Combining previously collected and novel data, we tested the hypothesis that participants’ conscious reports in ISL reflect a post hoc conscious model of their nonconscious knowledge. Across two experiments, participants were exposed to sequences of stimuli (letters, faces, or body movements in VR) generated by different regularities (artificial grammars and grammar-like bigram regularities). In a subsequent test, they decided whether novel strings were grammatical or not and reported their subjective awareness trial-by-trial. In both experiments, we found extreme Bayesian evidence that repetitions embedded in the testing strings made participants more aware of the knowledge driving their grammaticality decisions, above and beyond their influence on responses or accuracy. This suggests that, lacking access to the true basis of their decisions, participants attributed their responses to the most salient feature available: the repetitions. Thus, we find evidence that our conscious experience can misrepresent not only the external world but also our own unconsciously learned representations.

PMID:41779790 | DOI:10.1073/pnas.2526432123

Proc Natl Acad Sci U S A. 2026 Mar 10;123(10):e2533066123. doi: 10.1073/pnas.2533066123. Epub 2026 Mar 4.

ABSTRACT

This study presents a biocompatible, ultrasound-responsive platform for remotely activating mechanochemical reactions within live plant tissue. Fluorogenic Mechanophore-embedded silica NanoParticles (FMNPs) that are thermally stable were engineered to emit blue fluorescence at 440 nm upon mechanical activation. In Solanum lycopersicum (tomato) leaves, activation was achieved through the synergistic combination of gas vesicles (GVs) and high-frequency focused ultrasound (FUS, 550 kHz), enabling spatially localized and minimally invasive stimulation. Low-frequency ultrasound (25 kHz) triggered activation but caused extensive tissue damage, while high-frequency FUS alone was biocompatible yet insufficient to activate FMNPs. Incorporation of GVs as a cavitation amplifier significantly boosted activation efficiency under mild acoustic conditions without observable tissue disruption. In planta fluorescence imaging confirmed that FMNPs retained their functionality after injection into leaf vasculature, and only the combination of GV and FUS produced a statistically significant fluorescence increase, indicating successful mechanochemical activation. This represents a demonstration of noninvasive and biocompatible ultrasound-induced mechanophore activation in live plants. This modular and noninvasive strategy opens possibilities for programmable release of regulatory and metabolic chemicals, biosensing, and synthetic molecular control in plant systems.

PMID:41779773 | DOI:10.1073/pnas.2533066123

PLoS Comput Biol. 2026 Mar 4;22(3):e1013985. doi: 10.1371/journal.pcbi.1013985. Online ahead of print.

ABSTRACT

Identifying the key order parameters that connect a protein’s native structure to its dynamical and evolutionary behavior remains a central challenge. We introduce topological and geometrical metrics-specifically, writhe and Local Topological Energy (LTE)-to investigate these connections. Applying these tools to both present-day and ancestral forms of thioredoxin and β-lactamase, we show that LTE strongly correlates with established dynamical measures such as the Dynamical Flexibility Index (DFI). Remarkably, LTE distributions also track the evolutionary trajectories of these proteins, suggesting that the topological geometry of the native state encodes key aspects of both dynamics and evolution. Through molecular dynamics simulations, we further reveal critical shifts in the topological landscape of proteins, providing a molecular mechanism by which functional evolution proceeds via alterations in conformational dynamics. Extending our analysis to over 100 proteins, we provide the first compelling evidence that topological descriptors derived from static structures can reliably predict dynamical behavior. In general, our findings demonstrate that simple geometrical metrics capture essential features of protein conformational landscapes, offering a powerful new approach to bridging protein structure, dynamics, and evolution.

PMID:41779772 | DOI:10.1371/journal.pcbi.1013985