U.S. healthcare spending has remained persistently high despite repeated efforts at correction. This essay offers a structural explanation. Waste, excess prices, and administrative complexity matter, but much spending growth reflects durable features of the sector that cannot be readily eliminated. Baumol's cost disease provides the core framework: in labor-intensive services with limited productivity gains, costs rise because wages in them must keep pace with more productive sectors. Medical technology more often expands capacity, utilization, and clinical expectations than it reduces labor inputs. The U.S. physician training pathway is unusually long and expensive, and federal residency caps have artificially constrained physician supply, reinforcing a high compensation floor. The healthcare and social assistance sector functions as a de facto industrial policy, as it is the nation's largest employment sector and the top employer in 38 states, making aggregate cost compression politically costly in ways that are structural, not incidental. Domestic multiplier effects deepen that political durability. Five distinctively American features further limit centralized cost control: population scale and decentralization, higher per capita income, a heavier chronic disease burden, the absence of a national health technology assessment authority, and weaker redistributive institutions. Given the constraints, the aspiration to make American healthcare dramatically cheaper without major disruption is unrealistic. A more credible agenda is to foster local stewardship within a structurally high-cost system.
The precise identification of intersegmental planes is critical in lung segmentectomy. Indocyanine green (ICG) is commonly used, but it requires specialized equipment and it also has a short fluorescence duration. This study explored the use of intravenous vitamin B2 (riboflavin) with a blacklight (ultraviolet A, UV-A) as a simpler, cost-effective alternative. In a porcine model (n = 4), both ICG and vitamin B2 accurately delineated the same intersegmental plane. However, vitamin B2 fluorescence lasted over 30 min -significantly longer than ICG's 7.3 min -and it was visible to the naked eye without specialized imaging systems. The method was safe and practical, utilizing readily available UV light. This approach may be particularly useful for visualizing lung segments during surgery without the need for expensive fluorescence cameras. Although the results in healthy animals are promising, further validation in humans is required. Overall, vitamin B2 with blacklight shows potential as a reliable, long-lasting, and accessible tool for pulmonary segment identification during anatomical resections.
Canada's food guide (CFG), updated in 2019, recommends choosing lower fat milk, fortified plant-based milk alternatives, and unsweetened beverages. This study examined trends in household purchases of dairy and plant-based beverages (PBBs), following CFG release, as well as price differences between beverage categories. NielsenIQ HomeScan data from 2018 to 2020 (∼4500 households/year) in Québec and Ontario were analyzed for annual purchases of total dairy beverages (milk, drinkable yogurt/kefir, milkshakes) and total PBBs (milk alternatives, plant-based drinkable yogurt/kefir). Product descriptions/Universal product codes (n=1451) were matched to online nutrition information. Mixed linear models examined changes in volume purchased over time, adjusting for household sociodemographic characteristics, and linear regressions examined price differences. On average, households purchased 97.0L of total dairy beverages and 8.8L of total PBBs annually, with 2% milk (45.7L) and almond beverages (4.9L) being the most purchased in these categories. Purchases of total dairy beverages, dairy beverages with added sugars, skim, 1%, and 2% milk were lower in 2019 than in 2018. However, purchases were higher in 2020 than in 2019 for total dairy beverages, 2% and whole milk, total PBBs, almond beverages, and "other" types of milk alternatives. Soy beverage purchases remained unchanged. Households with children purchased more dairy with added sugars and whole milk, and fewer PBBs. Differences by province, education, and ethnicity were also found. PBBs were slightly more expensive than dairy beverages ($0.23CAD/100ml vs $0.21CAD/100ml, p<0.01), with higher prices in Québec than Ontario. Overall, results suggest partial alignment with 2019 CFG recommendations.
BackgroundUnsafe abortion remains a critical public health challenge among young women in Ghana, persisting despite the provisions of PNDC Law 102 permitting abortion.AimThis study examined the determinants of unsafe abortion among women aged 15-25 years in the Bongo District, Upper East Region, with particular focus on contraceptive use, health-system accessibility, and socio-economic pressures.MethodsUsing a qualitative study, thirty-four women from five communities were recruited through purposive sampling and interviewed using a semi-structured guide.ResultsFour main themes emerged firstly, the idea of having an unsafe abortion was so ingrained in the society that it had become a kind of a routine for people in the district; Secondly, not using contraceptives was more a result of the women's worries about becoming infertile, side effects, and their partners disapproval than simply not knowing about them; Thirdly, formal health services were so expensive, far, and the providers were so judgmental that many just gave up on them; Lastly, social stigma namely, the women's fear of disgracing their families and being expelled from school. The findings are that unsafe abortion in Bongo District is a systemic failure, not a lack of awareness.ConclusionThe findings of this study highlighted that unsafe abortion in Bongo District is a structural failure of the reproductive health system, of social support networks, and of policy implementation rather than a deficit of individual knowledge or moral resolve. Young women in the district are neither uninformed about the risks of unsafe procedures nor indifferent to their health.
Xylitol is a sugar alcohol that has low-glycemic sweetener suitable for diabetics. It prevents cavities, functions as a prebiotic, and has emollient and moisturizing properties. Xylitol uses as a food additive, sweetener, and dental products. Although the global xylitol market is growing, driven mainly by the chewing gum industry, its production depends on expensive and unsustainable chemical methods, which has motivated the development of biotechnological alternatives. This review examines biotechnological xylitol production in Candida species. These yeasts convert efficiently xylose-containing biomass into xylitol, presenting a more economical and eco-friendly alternative to traditional extraction methods. The review will discuss xylitol applications, microbial production, advantages, production parameters, specific strains used, advantages and challenges. The online version contains supplementary material available at 10.1007/s12088-025-01456-1.
Individual glycemic responses to foods vary and can be predicted using microbiome, activity, and dietary data. However, these data are expensive and invasive to collect, and it is not known how much each modality contributes to accuracy. We aim to quantify the contributions of dietary, sleep, continuous glucose monitor (CGM), and microbiome features for glycemic response prediction; understand how much personal data are required for training; and evaluate how microbiome sample timing impacts model accuracy. We used data from 8334 participants in the Human Phenotype Project cohort study who provided demographic, anthropometric, dietary, and CGM data. Participants self-reported meals in a dietary tracking application for a mean of 10.78 d, during which they wore CGMs. We trained CatBoost models to predict postprandial glycemic response (PPGR) using 2-h incremental area under the curve and peak 2-h postprandial glucose rise (Glumax). We conducted ablation studies with varied feature combinations to assess the contribution of each data modality. We used 3 train/test splits (split-by-meal, 5-d personal training, and split-by-person) to assess the impact of personal training data. Lastly, we evaluated accuracy as a function of microbiome sample timing (from before meal logs to ≤60 d after). The model combining all features performed best, and CGM was the most informative feature. Models trained with more personal data had the best performance (PPGR split-by-meal R = 0.731; split-by-person R = 0.590), and personal training data had a larger effect on accuracy than microbiome. Microbiome features improved predictions most when collected within 7 d of meal logs and did not improve performance without personal training data or for samples collected >14 d after meal logs. Although CGM was the most important feature group, combining it with personal training data and timely microbiome samples led to the most accurate models in our analysis. These findings can help researchers understand the tradeoffs between the time and effort of data collection and how data types impact model performance.
This study presents an experimental investigation into the development of a broadly accessible, research-scale geopolymer concrete (GPC) 3D printing platform and formulation. Concrete 3D printing is a rapidly emerging production method that remains underutilized, largely due to legal and regulatory restrictions and, to a degree, the high barrier to entry in the associated research stemming from expensive specialized hardware. In this work, a low-cost tabletop clay printer was modified for GPC production, demonstrating that a dedicated geopolymer formulation can be effectively utilized for additive manufacturing as a low-carbon alternative to conventional concrete. The mix design was engineered to favor geopolymerization while maintaining printability, employing sodium carboxymethyl cellulose (Na-CMC) as a viscosity-modifying additive to enable a low water-to-solid ratio while maintaining workability. The influence of formulation and processing on extrusion behavior and defect formation is discussed. The printer modification protocol and recipe mix design are described in detail to ensure reproducibility and interstudy comparability. The resulting material demonstrated successful polymerization and solid mechanical performance, confirmed by 7-day direct tensile testing of 3D-printed dog-bone specimens (1.1 ± 0.2 MPa), multiscale defect analysis, and attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy.
In 25-40% of patients with thromboembolic stroke, current diagnostics fail to identify a cause. It has been shown that prolonged heart rhythm monitoring can detect atrial fibrillation (AF) as the underlying cause in one-third of these patients. Even though an implantable loop recorder detects AF in 37% of cases up to four years after stroke, several studies show no reduction in the number of recurrent strokes with anticoagulation treatment. It seems that a high degree of AF around the stroke, so-called high burden, has a more likely causal link, than AF detected months or years after stroke. Anticoagulation treatment may be more effective in high AF burden stroke patients. A less expensive, non-invasive detection method, applicable immediately after stroke, may be more suitable in practice. Furthermore, recent research shows promising results in the importance of determining electrocardiographic, anatomical and blood biomarkers indicative for an atrial myopathy and better detection AF.
Pseudomonas aeruginosa (PA) is isolated in up to 12% of post-operative head and neck cancer patients, capable of surgical site tissue invasion triggering acute inflammation, thrombus formation, and localized microcirculatory collapse. This may result in flap failure and delay post-operative radiotherapy due to extended hospital stays. This report reviews literature on acetic acid washouts as adjunctive treatment for post-surgical PA infection and presents a single-centre case series. A systematic review protocol was registered with PROSPERO, and a comprehensive literature search on acetic acid use in PA infection was conducted. A retrospective single-centre review of head and neck patients undergoing ablative and reconstructive surgery with microvascular free tissue transfer (2020-2023) was performed. Patients with clinical and biochemical evidence of PA infection received 2-3 daily bedside washouts with 0.05% acetic acid for 7-14 days, alongside intravenous antibiotics. Six patients with post-operative PA infection were identified. Acetic acid washouts were administered for 1-2 weeks alongside IV antibiotics based on susceptibility testing. Four patients required return to theatre (mean days post-op=20.5; range=5-44). Post-operative inpatient stays ranged from 9 to 38 days. No flap failures or metal-work removals occurred. Our literature search found no comparable studies. Acetic acid is an easily available, non-toxic, inexpensive topical agent, with bedside washout courses costing £28-£84 per patient. This modest expense is outweighed by the substantial health and psychological benefits of avoiding flap failure. Its low side effect profile supports its use in managing PA-infected surgical sites post-operatively in head and neck cases.
Heavy-metal-free metal selenide nanocrystals have attracted great interest as low-cost, environmentally friendly, and solution-processable materials for future optoelectronics and thermoelectrics. Despite the widespread prospects of these materials, their preparation and processing remain a challenge, as the fabrication is laborious and may require harsh environments and toxic chemicals, and, more importantly, most widely used selenium precursors for the synthesis of colloidal NCs are not only toxic and expensive but also highly unstable, which makes them extremely difficult to handle and leads to unreproducible results. In this work, we present a facile synthetic approach for obtaining colloidally stable ternary silver bismuth selenide (AgBiSe2) and quaternary silver bismuth sulfide selenide (AgBi-(S1-x Se x )2) nanocrystals by utilizing bis-(acyl) selenides as an alternative precursor in the hot-injection synthesis method. The employed bis-(acyl) selenide precursors are stable under ambient conditions, easily processable, and allow great composition control on the anion side and, consequently, band gap fine-tuning in the case of AgBi-(S1-x Se x )2 nanocrystals, making them highly attractive for solution-processed optoelectronic devices.
Intravenous thrombolysis with recombinant tissue plasminogen activator (rt-PA) and urokinase (UK) are both recommended for acute ischemic stroke (AIS) in China. Compared with rt-PA, UK is less expensive and more widely available in clinical practice. The Treatment of Acute Ischemic Stroke with Intravenous Urokinase (TASK-UK) study aims to investigate clinical outcomes among patients with AIS who receive UK thrombolysis, and to compare efficacy and cost-effectiveness between UK and rt-PA. TASK-UK is a multicenter, real-world observational study, with a planned sample size of 1000 patients in the UK group and 800 in the rt-PA group. This study enrolls patients with AIS who are eligible for intravenous thrombolysis from 20 participating centers and conducts follow-up until 90 days after thrombolysis. The primary efficacy outcome is functional independence at 90 days, defined as a modified Rankin scale (mRS) score of 0-2. The primary safety outcome is symptomatic intracranial hemorrhage (sICH) occurring within 22-36 h after thrombolysis. The exploratory outcome is defined as an intergroup comparison of direct medical costs during hospitalization. The TASK-UK study will provide high-quality real-world evidence on the effectiveness and cost-effectiveness of UK thrombolysis for AIS. NCT06194968 (ClinicalTrials.gov).
Bladder cancer is the 11th most common cancer in the United Kingdom, with approximately 10,500 new cases annually. Diagnosis and surveillance typically involve cystoscopy, an expensive, time-consuming, and uncomfortable procedure which has encouraged efforts to identify biomarkers, particularly in urine, given its direct contact with malignant tissue. Urine collected from 100 participants (50 bladder cancer patients, 50 controls) was subjected to solvent extraction followed by gas chromatography-mass spectrometry (GC-MS) to determine potential volatile and semi-volatile biomarkers. The results were analysed using classical univariate statistics and machine learning methods. Five machine learning algorithms were evaluated, with recursive feature elimination (RFE) identifying optimal biomarker panels. Machine learning with XGBoost achieved area under the receiver operating characteristic curve (AUROC) of 0.869 (95% CI: 0.740-0.988), representing a significant improvement over the classical statistical approach (AUROC 0.752). An 8-metabolite panel achieved balanced sensitivity and specificity of 85%, or 95% sensitivity with 70% specificity when optimised for screening. The findings indicate that solvent extraction of urine shows promise for isolating putative biomarkers of bladder cancer. Employing machine learning achieved diagnostic accuracy potentially suitable for clinical deployment as a non-invasive bladder cancer detection tool.
Automated solid-phase peptide synthesis (SPPS) is the premier methodology for producing high-purity peptides across a broad range of applications. However, despite its efficiency, automated SPPS often requires expensive instrumentation and generates significant chemical waste. To overcome these limitations, we present a robust SPPS protocol that competes the thermal acceleration of microwave-assisted systems using standard laboratory equipment, offering a highly accessible and cost-effective alternative. The methodology enables the rapid assembly of a diverse array of linear and cyclic peptides (10-100 mg scale), including two biochemical active sequences, with reaction times reduced to minutes and a decrease in chemical waste. Given that elevated temperatures could compromise stereochemical integrity, we performed an in-depth investigation into epimerization risks using a multi-modal analytical suite, including high-performance liquid chromatography, nuclear magnetic resonance, and ion mobility mass spectrometry. Our results demonstrate that this instrumental approach not only provides satisfactory crude yields and purities but also maintains high stereochemical fidelity and native biochemical function, offering a robust alternative to high-cost automated systems.
Rapid, non-invasive detection of intracranial hemorrhage at the point of care is crucial for improving patient outcomes after head trauma. To overcome the limitations of bulky and expensive gold-standard imaging systems like CT and MRI, this study introduces a portable detection system based on electromagnetic induction. The core of our method is a simplified single-coil sensor that detects hemorrhage-induced changes in tissue conductivity by monitoring variations in coil impedance, facilitated by a high-precision inductance-to-digital converter (LDC1101). We rigorously evaluated the system through electromagnetic simulation, customized coil design, hardware implementation, and software development featuring an RGB localization algorithm. Experiments using a 3D-printed head phantom demonstrated that the system effectively distinguishes simulated ICH lesions (30 mL NaCl solution) from normal brain tissue and successfully localizes the hemorrhage. This work establishes a novel and practical technological pathway for developing portable, low-cost ICH monitoring devices suitable for bedside and point-of-care applications.
Atherosclerotic peripheral arterial disease (PAD) can lead to chronic limb-threatening ischemia (CLTI) and limb loss. Treatments include lifestyle modifications, medications, and both open and minimally invasive operative approaches, including balloon angioplasty. Drug-eluting balloon (DEB) angioplasty is a promising alternative to uncoated balloon angioplasty for treating PAD. Ballooning and coating the inside of atherosclerotic vessels with cytotoxic agents inhibits cellular mechanisms responsible for atherosclerosis and neointimal hyperplasia, thereby preventing or postponing its complications. Although economic analyses may have demonstrated the cost-effectiveness of DEB angioplasty, they remain considerably more expensive than uncoated balloons, and there is uncertainty around their effectiveness. This is an update of our previously published 2016 review. To evaluate the benefits and harms of DEB angioplasty compared with uncoated, plain old balloon angioplasty (POBA) in people with symptomatic lower-limb PAD. We systematically searched the following databases for randomized controlled trials and controlled clinical trials: Cochrane Vascular Specialised Register, Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase Ovid, and CINAHL EBSCO. We also searched the WHO International Clinical Trials Registry Platform and ClinicalTrials.gov. We used the bibliographies of relevant papers to identify other studies. The most recent searches were carried out on 13 March 2023. We included randomized controlled trials comparing DEB angioplasty with POBA for intermittent claudication or critical limb ischemia (CLI). We used standard Cochrane methods. Our primary outcome was amputation. Our secondary outcomes included amputation-free survival, secondary vessel patency, change in ankle-brachial index (ABI), change in quality of life (QoL), change in functional walking ability, and all-cause mortality. We used GRADE to assess the certainty of evidence for selected outcomes (amputation, secondary vessel patency including target lesion revascularization and binary restenosis, change in ABI, and all-cause mortality). A total of 31 trials randomizing 5292 participants met the inclusion criteria. Nineteen trials included femoropopliteal arterial lesions, nine included tibial arterial lesions, and three included both. The trials were carried out in Europe, the US, China, Singapore, Jordan, Japan, and New Zealand. All trials used paclitaxel. All but four trials were industry-sponsored. There was heterogeneity in the frequency of stent deployment and antiplatelet regimens between trials. Participants were followed up for up to five years. There were better outcomes with DEB angioplasty for target lesion revascularization at one year, from 204 per 1000 lesions with POBA to 80 per 1000 lesions with DEB angioplasty (odds ratio (OR) 0.34, 95% confidence interval (CI) 0.28 to 0.41; 23 studies, 4172 participants; P < 0.00001; low-certainty evidence). DEB angioplasty was also superior for binary restenosis at one year, from 443 per 1000 vessels with POBA to 187 per 1000 vessels with DEB angioplasty (OR 0.29, 95% CI 0.23 to 0.38; 8 studies, 1288 participants; P < 0.00001; moderate-certainty evidence). There was no difference between DEB angioplasty and POBA in amputation at one year, from 16 per 1000 participants with POBA to 22 per 1000 participants with DEB angioplasty (OR 1.45, 95% CI 0.91 to 2.29; 27 studies, 4469 participants; P = 0.12; moderate-certainty evidence); all-cause mortality, from 45 per 1000 participants with POBA to 44 per 1000 participants with DEB angioplasty (OR 0.97, 95% CI 0.71 to 1.32; 25 studies, 4312 participants; P = 0.83; moderate-certainty evidence); or change in ABI, from 0.1 to 0.35 higher with POBA and 0.03 higher to 0.03 lower with DEB angioplasty (mean difference (MD) 0, 95% CI -0.03 to 0.03; 5 studies, 1156 participants; P = 0.96; moderate-certainty evidence), although none of the studies were powered to detect a significant difference in these clinical endpoints. Meta-analysis of 31 trials with 5292 participants demonstrated that there may be evidence of an advantage of DEB angioplasty compared with POBA in several anatomic endpoints including late lumen loss, target lesion revascularization (low-certainty evidence), and binary restenosis (moderate-certainty evidence). Conversely, there may be little to no evidence of advantage with DEB angioplasty for clinical endpoints such as amputation (moderate-quality evidence), amputation-free survival, death (moderate-quality evidence), change in ABI (moderate-quality evidence), QoL, or functional walking ability. Well-designed randomized trials with long-term follow-up are needed to further compare DEB angioplasty with POBA adequately for both anatomic and clinical study endpoints.
Cryo-electron tomography (cryo-ET) enables structural characterization of biomolecules under near-native conditions. Existing approaches for interpreting the resulting three-dimensional volumes are computationally expensive and have difficulty interpreting density associated with small proteins/complexes. To explore alternate approaches for identifying proteins in cryo-ET data, we pursued a Graph Network and topologically invariant approach. Here, we report on a fast algorithm that distinguishes volumes containing protein density from noise by searching for nuances of evolutionarily conserved motifs and the geometric characteristics of protein structure. Graph Identification of Proteins in Tomograms (GRIP-Tomo) 2.0 is a machine-learning pipeline that extracts interpretable topological features of protein structures within noisy experimental backgrounds. Compared to version 1.0, the new pipeline includes three upgrades that significantly improve performance, including synthetic tomogram generation simulating realistic noise, graph-based persistent feature extraction as protein fingerprints, and High Performance Computing acceleration. GRIP-Tomo 2.0 achieves over 90% accuracy in distinguishing proteins from noise for synthetic datasets and over 80% accuracy for real datasets with Angstroms per pixel close to 1 from the protein mixtures of in-house samples, which represents a foundational step toward advancing cryo-ET workflows and empowering automated detection of both small and large proteins for visual proteomics. https://github.com/EMSL-Computing/grip-tomo.
Access to dialysis facilities greatly affects patients' quality of life and healthcare costs. Transportation alone represents 21% of dialysis-related expenses. Going beyond a simple focus on access, this study aimed to evaluate dialysis care accessibility in France by using a spatial analysis approach, offering insights for healthcare planning and resource allocation strategies. We developed a novel spatial accessibility indicator based on an Enhanced Two-Step Floating Catchment Area (E2SFCA) method, incorporating custom distance-decay functions calibrated to actual patient travel patterns. The resulting value represents the number of available dialysis stations per individual within a population cluster. A value close to 1 indicates a balance between supply and demand with adequate geographic proximity, values <1 suggest insufficient availability or reduced accessibility, and values >1 indicate greater availability relative to local needs. The study analyzed data for all French chronic hemodialysis patients as of December 31, 2021, excluding those on home dialysis. Using hexagonal grid mapping and the Local Indicators of Spatial Association test, we identified spatial clusters and disparities in dialysis accessibility. The median accessibility score across all modalities was 0.68 (interquartile range: 0.48-0.93) dialysis stations per patient in the facilities' catchment areas, increasing to 0.88 (0.66-1.14) in patient-occupied hexagons. Self-care units showed the highest alignment between supply and demand (median: 1.02, 0.74-1.34), and facility-based units exhibited lower overall accessibility (median: 0.62, 0.42-0.89). An east-west accessibility division emerged. Rural and mountainous areas consistently had lower accessibility scores than urban areas. Our study poses a number of challenges, particularly in terms of defining supply and interpreting the results. Our E2SFCA method revealed substantial regional disparities and modality-specific patterns in dialysis accessibility across France. This approach is an exploratory planning tool to inform resource allocation decisions while addressing geographic inequities.
Wastewater-based surveillance of SARS-CoV-2 has become an established approach for monitoring health status and disease outbreak at the community level. Although analytical protocols for RNA quantification have been widely implemented and optimised, the characterisation of variability introduced across all stages, such as storage, sample processing, and PCR quantification, remained limited, including its impact upon epidemiological interpretation. This study proposes an analytical framework to quantify variability across the complete workflow for SARS-CoV-2 RNA measurement in wastewater, including storage stability, sample processing, and PCR quantification. This framework was developed for a large SARS-CoV-2 dataset of two viral gene targets (N1, E-Sarbeco), which were quantified alongside human-associated CrAssphage for a total of 20,124 RT-qPCR data points derived from four water recycling centres (WRCs) over a 24-month period. This analysis revealed that sample refrigeration yielded richer epidemiological data than freezing, and that CrAssphage had minimal utility as a normalisation biomarker given its higher variability. Quantification of analytical variability for SARS-CoV-2 targets enabled development of a composite metric (Total Estimated SARS-CoV-2), improving sensitivity and reducing non-detects at the expense of reduced precision. This novel multi-stage framework had the capability to improve the sensitivity and interpretation of WBE data and the characterisation of WBE methodologies, with broad applicability for pathogen monitoring.
Persistent somatic symptoms (PSS) are prevalent across Europe. As little is known about affected patients' healthcare experience and its relationship with symptom outcomes, these were explored in a cross-sectional survey across four European countries. This cross-sectional online survey was distributed in Germany, Italy, the Netherlands and Poland (04/2023-05/2024). The survey included adults aged ≥18 years with PSS (Patient Health Questionnaire-15 ≥10) who had used healthcare services in the past year. Healthcare factors were assessed across four dimensions: availability, affordability, accessibility and adequacy. Multinomial logistic regression examined associations between healthcare factors and symptom course (improvement, persistence and deterioration). Duration of untreated illness was also compared across countries. Of 595 participants (87% female) included in regression analyses, 13% reported improvement, 70% persistence and 17% deterioration. Duration of untreated illness was longest in Poland, followed by Italy, the Netherlands and Germany (48, 24, 12 and 3 months, respectively, p<0.001). After adjusting for age, gender and country, nearly all healthcare factors were associated with higher odds of symptom persistence and deterioration across countries, including difficulty accessing services (persistence OR 1.39 (95% CI 1.12 to 1.75); deterioration OR 1.84 (95% CI 1.39 to 2.44)), financial difficulties (persistence OR 1.60 (95% CI 1.25 to 2.05); deterioration OR 2.17 (95% CI 1.59 to 2.96)), postponing care due to costs (persistence OR 1.54 (95% CI 1.07 to 2.19); deterioration OR 2.53 (95% CI 1.67 to 3.86)), dissatisfaction with insurance (persistence OR 1.26 (95% CI 1.03 to 1.54); deterioration OR 1.57 (95% CI 1.19 to 2.06)) and poor clinician-patient communication (persistence OR 1.19 (95% CI 1.11 to 1.26); deterioration OR 1.25 (95% CI 1.16 to 1.34)). Country-specific effects were found only for insurance satisfaction, with dissatisfaction linked to worse outcomes in Germany (persistence OR 1.87 (95% CI 1.19 to 2.93); deterioration OR 4.02 (95% CI 2.09 to 7.74)). Across four European countries, patients with PSS reported that their symptoms did not improve despite receiving care. This was the case even in countries with substantially shorter treatment delays, suggesting that early access alone may not be sufficient. Barriers related to access, affordability and communication were associated with worse outcomes, highlighting key areas for improvement across healthcare systems.
Foundation models (FMs) and large language models (LLMs) are transforming cancer AI by integrating heterogeneous data sources, including medical imaging, electronic health records, and molecular profiles. By learning from large-scale, unstructured, and label-free inputs, these models may support diagnosis, biomarker discovery, prognostic assessment, treatment personalization, and workflow automation. In this narrative review, we propose the paradigm of "Leave No Data Behind" to describe the promise that broad oncology data integration may generate clinically meaningful outputs. We critically assess whether this paradigm is supported by current evidence and identify the key challenges that must be addressed to harness the full potential of FMs and LLMs for clinical implementation in oncology.