Medical claims are widely used in workforce estimation and health services research but identifying primary care (PC) clinicians is challenging. Existing methods relying on specialty or activity-based metrics are imprecise and hard to duplicate. We seek to develop a simple, accurate decision tree classifier using claims data, trained on survey responses from family physicians (FPs). And to use this classifier to estimate the PC workforce in Virginia. We linked 2016-2023 Virginia All-Payer Claims Database (APCD) data with responses from American Board of Family Medicine surveys to infer whether FPs provided PC. Using claims-derived features, we trained a decision tree to classify clinicians providing PC. We developed an enhanced version of the tree, adding exclusion criteria, and applied both classifiers to the entire APCD, including other PC clinician types. Virginia clinicians. Estimated classifier accuracy by clinician type and workforce estimates with bootstrapping used to estimate 95% confidence intervals. The base decision tree correctly classified 93% of the training sample using three clinician-level features: percent of claims with diagnosis category of Z00, percent with place of service 19-23 or 31-32, and percent of patients having multiple visits with the clinician. We estimate the enhanced tree achieved at least 89% accuracy for every clinician type. We found that from 2016 to 2023, the percentage of Virginia PC clinicians who were nurse practitioners (NPs) grew from 17 to 32%, with NPs passing FPs as the most common PC clinician type. This decision tree approach overcomes shortcomings of existing methods and offers a straightforward, scalable, interpretable tool for classifying PC clinicians, with applications in workforce planning and health services research.
To evaluate the association between nasal procedures and long-term continuous positive airway pressure (CPAP) treatment persistence in a large US administrative claims database. Retrospective cohort study using the Komodo Health Sentinel Database (2020-2026). Adults with sleep apnea (ICD-10 G47.3x) who initiated CPAP were classified into six cohorts: temperature-controlled radiofrequency (TCRF) nasal valve remodeling (n = 920), septoplasty (n = 24,575), functional rhinoplasty (n = 758), combined functional rhinoplasty and septoplasty (n = 3,910), other nasal surgery (n = 12,995), and no nasal surgery (n = 1,297,292). The primary outcome was CPAP discontinuation, defined as a ≥180-day gap in positive airway pressure claims. Covariate-adjusted Cox regression and coarsened exact matching (CEM) on six variables assessed the association between nasal procedures and discontinuation. Sensitivity analyses included inverse probability of censoring weighting (IPCW), alternative gap thresholds (90-365 days), E-values, and insurance stratification. Among 1,340,450 CPAP initiators, the overall discontinuation rate was 29.0%. TCRF was associated with the lowest discontinuation (HR 0.62, 95% CI 0.53-0.72; p < 0.001); IPCW addressing informative censoring strengthened this association (HR 0.48, 0.41-0.56). Other nasal procedures showed more modest associations: combined FR and septoplasty (HR 0.84, 0.79-0.90), other nasal surgery (HR 0.86, 0.83-0.89), septoplasty (HR 0.87, 0.84-0.89), and functional rhinoplasty (HR 0.90, 0.78-1.02). CEM confirmed these findings, with TCRF showing HR 0.57-0.61 across model specifications. One-year persistence was 84.2% for TCRF versus 76.1% for no surgery. The TCRF association was consistent across insurance types and all sensitivity analyses. The E-value was 2.61 (CI bound 2.12). Nasal procedures were associated with higher CPAP persistence compared with no surgery, with TCRF showing the strongest and most consistent association. These findings suggest that treating nasal obstruction may help sustain long-term CPAP use, though prospective studies are needed to confirm causality.
The authors examined the distribution of behavioral health episodes of care (EOCs) by modality-telehealth only, in-person only, and both (hybrid)-to identify factors associated with telehealth use. Using 2021-2022 MarketScan Commercial Database claims, the authors categorized EOCs by modality. Descriptive analyses were followed by logistic regression to identify independent associations of patient and provider characteristics with telehealth use. Among 937,711 EOCs, 41.1% were in-person only, 32.9% telehealth only, and 26.0% hybrid. Compared with individuals <18 years, all adult age groups had higher odds of telehealth-only care. Female sex and urban residence were also associated with greater telehealth use, whereas residence outside the Northeast and nonpsychiatrist providers were associated with lower use. Among hybrid EOCs, telehealth initiation was more likely among adults ages 36-55, urban residents, and those treated by nonpsychiatrists. Telehealth constituted a substantial proportion of behavioral health care, highlighting the importance of maintaining flexible telehealth policies.
暂无摘要（点击查看详情）
When an adult with rotator cuff tendinopathy reports feeling or hearing a pop along with new pain, both the clinician and the patient may inaccurately and unhelpfully expect (framing heuristic) an injury and consider it a compensable work claim. We retrospectively reviewed medical records of 118 people that filed work claims for new shoulder pain where aspects of their care or recovery trajectory triggered a peer review. We collected data on age, sex, reports of hearing or feeling a pop (25%; 29 of 118 patients), and reports of new numbness or tingling (11%; 13 of 118 patients). Five (4%) patients had a possible acute rotator cuff rupture (relatively large defect with good muscle) and 11 (9%) had a long head of biceps rupture, age indeterminate. Accounting for confounding variables using logistic regression, possibly acute rotator cuff defects and age-indeterminate long head of biceps ruptures were associated with older age (odds ratio [OR], 1.10; 95% CI, 1.02-1.18) and the presence of a degenerative rotator cuff defect (OR, 6.0; 95% CI, 1.5-24) but not with sensation of a "pop." Based on this evidence, among people claiming injury at work, when a "pop" is reported the clinician should not expect new pathophysiology. III.
暂无摘要（点击查看详情）
Croup is a common pediatric respiratory illness primarily affecting infants and toddlers. While understanding the geographic distribution and spatial clustering of croup healthcare encounters is crucial for public health planning and resource allocation, this area remains understudied. This study aimed to identify geographic clusters of croup healthcare encounters and areas of higher healthcare demand in Alberta, Canada, and compare univariate and multivariate spatial scan statistics using emergency department (ED) visits and physician claims data. We analyzed administrative health data for children aged ≤2 years in 70 sub-regional health authorities (sRHAs) in Alberta from 2017/18 to 2022/23. Data were aggregated by sRHA and year. Kulldorff's spatial scan statistics were used to identify croup clusters adjusted for sex distribution for each year separately. Analyses were conducted on each data source separately (univariate) and for both data sources simultaneously (multivariate). During the study period, there were 32,740 ED visits and 49,389 physician claims for croup, with males accounting for 63.5% and 62.2% of the encounters, respectively. Overall, 59.8% of patients with physician claims and 82.8% of ED patients (n=21,688) accessed both healthcare settings during the study period. Significant spatial clustering was consistently identified, with 1-4 clusters annually in univariate, and 2-5 clusters in multivariate analyses. Certain northern areas appeared most consistently in all years and methods. The COVID-19 pandemic year (2020/21) showed unique patterns with the highest relative risks. ED visits data demonstrated wider geographic coverage with consistent major metropolitan area involvement, while physician claims data showed frequent clustering in different metropolitan areas. Multivariate and univariate analyses showed overlapping but distinct findings, with multivariate analysis identifying two clusters where only physician claims data showed significantly higher case numbers than expected. Significant geographic clustering of croup healthcare encounters exists in Alberta, with northern regions most consistently identified as areas of higher healthcare utilization. Both univariate and multivariate spatial scan statistics detected significant clusters, with multivariate analysis providing additional insights by simultaneously analyzing ED visits and physician claims data. The distinct clustering patterns between data sources indicate different healthcare utilization behaviours and demonstrate the value of multivariate approaches for comprehensive spatial epidemiological analysis.
The occupational burden of injury caused by exposure to cleaning chemicals during the pandemic is not well described. The objective of this study was to determine whether the COVID-19 pandemic was associated with a change in the rate of toxic inhalation (TI) injuries following occupational exposure to cleaning chemicals. This retrospective study spanned from July 1, 2017, to June 30, 2022, with January 1, 2020, designated as the pandemic start. TI cases were identified among workers' compensation claims filed in Washington State, USA using keyword text search and diagnostic and insurance codes. Qualitative claim data were reviewed to determine the frequency of cases with exposure to aerosolized disinfectants following spray or fog application. Rates of TI claims per full-time equivalent (FTE) workers were compared in the pre- and pandemic periods. Out of the 440 TI claims identified, 30% involved exposure to disinfectants and 13% had a diagnosis of work-related asthma. There was a decline in the overall rate of TI claims in the pre- vs. pandemic period (8.9/100,000 FTE vs. 7.4/100,000 FTE, respectively). However, the proportion of exposures to disinfectants increased during the pandemic period. Four industries, particularly "Transportation and Warehousing," experienced an increase in TI injury rate. Fog or spray application of disinfectants was involved in 20 cases (including 14 in bus transportation) during the pandemic, compared to zero cases in the pre-pandemic period. In Washington State, the COVID-19 pandemic was associated with industry-specific increased TI injury rates due to cleaning chemicals, but not an overall increased rate. A case series emphasizes that aerosolized disinfectants are a toxic inhalation risk for workers. Studies to further characterize this practice and these injuries are warranted. Prevention efforts should utilize the hierarchy of controls as well as address hazard awareness, product over-application, and the need to estimate and uphold a safe re-occupation time after fog or spray application of disinfectants.
We examined changes in smoking cessation service utilization and prescribing patterns under the universal health insurance system during the varenicline supply suspension in Japan. We analyzed publicly available aggregated data from the NDB Open Database, which covers more than 95% of all health insurance claims. Monthly claims for nicotine dependence management fees from April 2019 to March 2024 were evaluated using interrupted time-series analysis, and annual prescription trends for varenicline and nicotine patches were assessed. The varenicline supply suspension was associated with a significant immediate decline in service claims (-6443.5; 95% confidence interval: -8360.2 to -4526.9; p < 0.001). Claims showed a modest downward trend before the suspension, followed by a gradual increase thereafter. Varenicline use declined after 2021, accompanied by increased nicotine patch use. These findings indicate that the supply suspension disrupted smoking cessation services and altered prescribing patterns in Japan, highlighting vulnerability in systems with limited reimbursed therapies.
Treatment-resistant depression (TRD) represents a significant clinical challenge, yet data on its epidemiology and real-world therapy patterns remain limited. This study aimed to estimate TRD prevalence and analyze treatment patterns using claims data. This retrospective analysis used the German Analysis Database for Evaluation and Health Services Research (DADB), encompassing 2.47 million statutory health insurance beneficiaries in 2019. TRD was operationally defined as a claims-based proxy for patients with moderate-to-severe depression receiving at least three consecutive antidepressant prescriptions involving two or more changes of therapeutic strategy (switch, combination, augmentation, or ECT) using claims data. Among 360,344 individuals diagnosed with depressive episodes, 98,577 (27.36%) had moderate-to-severe depression and received pharmacological treatment. Of these, 12,054 patients met our operational TRD definition, representing 6.46% of all moderate/severe depression cases and 17.49% of those receiving continuous pharmacological treatment. Treatment patterns revealed suboptimal guideline adherence: only 27.20% of patients received combination or augmentation therapy after failure of first-line therapies, and 38.35% in third-line treatment. SSRIs dominated across all therapy lines, with frequent class-level cycling back to previously failed medication groups. ECT remained critically underused. This study provides the first large-scale, real-world evidence on epidemiology and treatment patterns of TRD in Germany. TRD affects a substantial proportion of patients with moderate-to-severe depression. As first large-scale, real-world evidence on TRD in Germany, these findings highlight urgent needs for improved clinical decision-making and better access to specialized care to address the substantial burden and suboptimal treatment trajectories of TRD in Germany.
Lung cancer remains the leading cause of cancer-related mortality worldwide, with poor survival rates due to late-stage detection. Current low-dose computed tomography screening faces barriers including high costs and false-positive rates reaching 24%, while artificial intelligence offers opportunities to enhance early detection through longitudinal clinical data analysis. This study developed a Multi-Channel Convolutional Neural Network (MCNN) for lung cancer risk prediction using Taiwan's National Health Insurance Research Database, encompassing 523,539 patients (2,809 lung cancer, 23,783 other cancer, and 496,947 non-cancer). The MCNN was designed as a lightweight model processing nine channels of diagnostic codes, medications, and medical orders over a three-year observation period. Systematic feature selection reduced estimated feature storage requirements by 99.8%, from approximately 1,184 GB for the full ICD feature space to approximately 2.11 GB for the selected features, while retaining clinical relevance. Model performance was assessed using stratified 10-fold cross-validation against seven machine learning baselines, and interpretability was examined through SHAP analysis. The MCNN achieved an F₁-score of 66.91%, precision of 84.47%, and recall of 59.79%. Ablation studies confirmed multi-modal integration benefits, with diagnostic codes providing primary predictive power. SHAP analysis revealed distinct temporal patterns validating the model's ability to identify pre-diagnostic phases through healthcare engagement patterns. Findings are based on internal validation within a single national database, and key risk factors such as smoking history are not captured in administrative claims data; future evaluation in independent external cohorts is therefore warranted to confirm these findings. The model's high precision minimizes false-positive rates while its computational efficiency and clinical interpretability support practical implementation as a complementary claims-based screening support tool for early cancer detection.
Diagnostic stewardship-performing the right test for the right patient at the right time-improves diagnostic accuracy and reduces healthcare resource waste. Various stewardship interventions have been introduced, yet their effects dissipate without systematic, sustained monitoring. One-time education and isolated system controls consistently show effect attenuation over time, and inappropriate ordering practices rapidly rebound once active oversight ceases. To date, the literature has focused on monitoring within single institutions; integration with population-level surveillance systems remains underexplored. This review proposes a framework linking institutional-level and system-level monitoring. At the institutional level, key performance indicators such as tests per patient-day and test-to-test ratios, combined with dashboard visualization, statistical process control charts, and cyclical audit-and-feedback structures, enable continuous surveillance of ordering patterns and drive behavioral change. Root cause analysis of monitoring data can identify specific drivers of over-ordering, and machine learning approaches show promise for predicting both overutilization and underutilization. These institutional tools, however, cannot track patients across facilities or assess population-level test appropriateness. At the system level, health insurance claims and administrative data enable macroscopic monitoring across entire populations. National experiences from Korea, Canada, and the United States demonstrate that ordering code pattern analysis can systematically identify inappropriate utilization-and that the structural design of reimbursement systems is more effective than voluntary recommendations in controlling low-value testing. The inherent limitation of claims data-the absence of test result values-can be partially overcome through linkage with other administrative databases. Bridging these two levels requires healthcare data standards and interoperability infrastructure, including Logical Observation Identifiers Names and Codes (LOINC), Nomenclature for Properties and Units (NPU), Systematized Nomenclature of Medicine-Clinical Terms (SNOMED CT), and Fast Healthcare Interoperability Resources (FHIR), yet practical barriers such as mapping quality variability, privacy constraints, and standardization costs persist. The scope of monitoring should extend beyond test ordering to encompass the entire total testing process, engaging diverse stakeholders across the testing continuum. With multidisciplinary workforce development, artificial intelligence-based clinical decision support, value-based reimbursement models, and rigorous multi-center studies, diagnostic stewardship monitoring can evolve into sustainable healthcare infrastructure that serves both individual patient safety and population health.
Perinatal mood and anxiety disorders (PMADs) represent a common comorbidity of pregnancy and postpartum with serious consequences. Despite high disease prevalence and burden, PMADs often remain undiagnosed. In recent years, universal screening for perinatal mental health disorders has become standard care and most patients are screened at least once during pregnancy or postpartum. Whether racial differences in PMAD diagnosis persist remains unknown. We aimed to measure and assess temporal trends in racial differences in PMAD diagnosis among symptomatic pregnant and postpartum Medicaid enrollees in Michigan. This serial cross-sectional observational study analyzed data from the Michigan Pregnancy Risk Assessment Monitoring System (MI-PRAMS) and Michigan Medicaid administrative claims for deliveries spanning 2016-2020. We linked MI-PRAMS responses to Medicaid claims from one year before through one year after delivery to determine whether enrollees who reported PMAD symptoms to MI-PRAMS received a PMAD diagnosis in Medicaid claims. We used weighted logistic regression to determine the association between race and PMAD diagnosis among Michigan Medicaid enrollees who reported PMAD symptoms. Out of 60,583 Medicaid enrollees with a live birth in Michigan from 2016-2020, 33% were Black and 61% were White. About 70% were between 19-29 years old and 18% were between 30-34 years old. During the perinatal period, 39% had 2 or more co-morbid conditions, and 60% had 2 or more major life stressors. In unadjusted weighted logistic regression, White enrollees had 2.146 times greater odds of PMAD diagnosis than Black enrollees (95% CI: 1.570-2.933). This association remained significant after adjusting for age, ethnicity, co-morbidities, major life stressors, and delivery year. Adjusted weighted logistic regression found that White enrollees had 2.326 times greater odds of PMAD diagnosis than Black enrollees (95% CI: 1.669-3.236). White enrollees who reported PMAD symptoms were more than twice as likely to receive a PMAD diagnosis as Black enrollees reporting PMAD symptoms. The magnitude of racial difference in PMAD diagnosis did not improve over the course of the study period. These persistent racial differences are consistent with structural barriers, provider biases, and cultural stigmas that differentially and disproportionately impact non-White patients. Tailored interventions to improve mental health care among Black Medicaid enrollees may be needed to improve PMAD diagnosis, a critical first step in mental health management. Not applicable.
Islamic bioethics is a recent, albeit growing, academic discipline. Despite commendable contributions, the field remains critically limited. Most notably, its methodology of strict application of Islamic law to ethical analyses and recommendations often lacks sufficient moral analysis, intellectual engagement, or social context. The practice's emphasis on religio-legal rulings- without an investigation of their underpinning moral values- has resulted in a field of inquiry devoid of robust normative foundations and dependent upon ineffective and unsubstantiated claims. This paper calls for a revival of Islamic philosophical discourse to enrich Islamic bioethical practice. Although once popularized by Medieval Muslim philosophers like Ibn Sina (Avicenna) and Ibn Rushd (Averroes), philosophical discourse has fallen out of favor in the Muslim world, largely due to a perceived tension with religion. This work highlights the rich tradition of philosophical discourse in the Medieval Muslim world, disproving claims of an inherent conflict between philosophy and Islam. Following an Islamic philosophical framework, three goals for Islamic bioethics are established. First, theoretical rigor aimed at continually re-assessing and re-understanding concepts integral to the practice of bioethics such as personhood, dignity, futility, autonomy, and justice. Second, a shift from essentialist understandings of the Quran- and other sources of Islamic law- to more contextual examinations in the formulation of ethical opinions. Third, an active and interdisciplinary collaboration between Muslim scholars in the determination of Islamic rulings on medical matters. Only when these goals are met is the practice of Islamic bioethics capable of meeting the needs of Muslim patients and clinicians.
Real-world evidence (RWE) has emerged as an essential complement to randomized controlled trial data, providing insights into the effectiveness, safety, and tolerability of healthcare interventions in routine practice. The increasing digitalization of healthcare systems has enabled the generation of vast amounts of real-world data (RWD) from electronic medical records, claims databases, and digital health technologies. When appropriately analyzed, these data can inform clinical decision-making, regulatory evaluations, payer assessments, and healthcare resource optimization. Despite its growing importance, the generation of high-quality RWE presents significant methodological and operational challenges. Common challenges such as misclassification bias, confounding, incomplete longitudinal follow-up, poor data linkage, and risks of selective reporting. Addressing these challenges requires rigorous study designs, transparent methodologies, pre-specified and publicly registered protocols, and adherence to standard reporting guidelines. Advances in artificial intelligence (AI) and natural language processing hold promise for improving data accuracy, addressing time-varying confounding, scalability, efficiency, and enhancing study reproducibility. However, AI-driven approaches currently serve as supportive tools rather than autonomous solutions and require robust scientific oversight to ensure appropriate study design, causal reasoning, and clinical contextualization. This article highlights key principles, challenges, and emerging trends in RWE generation and dissemination, with a focus on the role of methodological rigor and expert scientific communication. Agencies and medical communication experts play a crucial role in ensuring that RWE generation aligns with stakeholder needs, regulatory frameworks, and scientific integrity. By leveraging innovative methodologies, robust protocols, transparent reporting, and stakeholder-focused dissemination strategies, RWE can reach its full potential in transforming patient care and driving evidence-based decision-making across healthcare systems. What is this article about?This article explores the use of real-world evidence (RWE) in healthcare. RWE is generated from data collected during routine patient care, such as electronic medical records, insurance claims, and digital health tools. RWE provides insights into how treatments perform in everyday clinical practice and complement traditional clinical trial data.What were the results?/What methodology/protocol is described?This article outlines how RWE is generated and analyzed using large healthcare databases and advanced analytics. Based on a targeted overview of RWE in healthcare, it discusses key challenges including data quality, accuracy, bias, incomplete follow-up, and need for transparent study-design and reporting. It also highlights the potential role of artificial intelligence in improving data extraction and study reliability while emphasizing the importance of human scientific oversight.What do the results mean?/Why is this important?Well-designed and clearly communicated RWE studies provide insights into treatment patterns, patient outcomes, and safety in real-world settings. These insights support better decision-making by healthcare professionals, regulators, and payers. By combining rigorous methods, emerging technologies, and expert scientific communication, RWE can improve patient care, guide policy decisions, and support efficient healthcare resource use.
Background: Explainable artificial intelligence (XAI) is increasingly proposed to improve trust in mammography-based artificial-intelligence systems, but the validity and clinical readiness of published explanations remain unclear. We aim to systematically review XAI methods applied to mammography and synthesize how explanations are evaluated for validity, robustness, and clinical usefulness. Methods: We conducted a systematic review according to PRISMA 2020. MEDLINE/PubMed, Embase, Scopus, Web of Science Core Collection, and the Cochrane Library were searched from 1 January 2015 to 15 January 2026. Two reviewers independently screened records and extracted data; disagreements were resolved by consensus with a third reviewer. Included studies used mammography as the primary input and reported an explicit explanation or interpretability mechanism. Because the literature was methodologically heterogeneous, we performed a structured narrative synthesis and an adapted XAI-specific appraisal of explanation claims, quantitative evaluation, external validation, human-factor assessment, and reporting transparency. Results: Fourteen studies were included. Ten studies addressed detection or lesion classification and four addressed risk or outcome prediction. Primary XAI families were interpretable-by-design architectures (6/14), post hoc saliency or attribution methods (5/14), and feature-level explanation methods (3/14). Five studies remained at tier-1 qualitative plausibility only, seven reached tier-2 internal quantitative explanation evaluation, two reached tier-3 external or cross-dataset interpretability assessment, and none reported reader or workflow studies. In the dedicated mammography saliency benchmark, Pointing Game scores for Grad-CAM, Grad-CAM++, and Eigen-CAM ranged from 0.30 to 0.41, indicating only modest lesion-pointing reliability despite acceptable classifier performance. Conclusions: Mammography XAI remains dominated by visually plausible explanations that are inconsistently validated. The literature is moving toward task-aligned and intrinsically interpretable designs, yet external validation and clinician-centered evaluation remain rare. Future studies should pre-specify explanation claims, use task-appropriate quantitative metrics, report explanation robustness under distribution shift, and test whether explanations improve human decision-making.
The skin serves as the primary physical barrier and plays a critical role in the aesthetic appearance, particularly through its pigmentation and texture. Collagenprash and Collagen builder are developed to support skin integrity by promoting collagen-associated protein expression. This study systematically evaluated these claims through a variety of biochemical and cellular assays. The formulations were assessed for biochemical and cellular antioxidant activity, collagenase inhibition, interleukin-mediated anti-inflammatory activity, melanogenesis inhibition, and proteins associated with collagen expression by using immunofluorescence analysis. The presence of a collagen blend, ascorbic acid, and botanical extracts enhanced the expression of prolyl 4-hydroxylase (P4H) and lysyl hydroxylase, key enzymes involved in post-translational modification of collagen. Upregulation of P4H facilitates hydroxylation of proline residues, a critical step for stabilization of the collagen triple helix, whereas increased lysyl hydroxylase expression promotes hydroxylysine formation and subsequent intermolecular cross-linking. Therefore, these formulations may have the potential to support collagen biosynthesis and maintain skin texture.
Treatment of acromioclavicular joint (ACJ) separations continually lacks consensus between providers, and while low-grade injuries generally respond well to conservative management, high-grade injury management is more complex. For patients requiring surgery, the effect of the timing of surgery with respect to the original injury on clinical outcomes remains unclear. Clinical outcome measures would worsen as the time between ACJ dislocation injury and surgery increased. Cohort study; Level of evidence, 3. A large nationwide insurance claims database was queried for patients who had an ACJ dislocation injury using International Classification of Diseases, Ninth and Tenth Revision (ICD-9 and ICD-10) codes. Surgical management via ACJ repair or ACJ reconstruction was identified using Current Procedural Terminology codes 23550 and 23552, respectively. The time between the date of initial ACJ dislocation diagnosis and the date of surgery was used to separate patients into 4 temporal subgroups with the following parameters: 0 to 4 weeks, 4 weeks to 3 months, 3 months to 1 year, and after 1 year. Clinical outcomes-including complication rate, infection rate, and fracture rate-were compared between temporal and procedural subgroups. Outcomes were compared using chi-square tests. A total of 13,194 patient met the inclusion criteria and were included in the study, of whom 8639 received surgical management of their ACJ dislocation within 4 weeks of initial injury diagnosis, 2210 between 4 weeks and 3 months, 1264 between 3 months and 1 year, and 1081 more than 1 year after injury diagnosis. The rate of all measured adverse clinical outcomes increased over time as surgical management was delayed in the ACJ repair subgroup, and the rates of revisions and complications increased over time as surgical management was delayed in the ACJ reconstruction subgroup. The rate of all measured adverse clinical outcomes increased over time as surgical management was delayed in the ACJ repair subgroup, and the rate of revisions and complications increased over time as surgical management was delayed in the ACJ reconstruction subgroup. Further research is needed to define the role of injury severity and classification in these outcomes and identify which patients would benefit from early surgical intervention.
Autism spectrum disorder (ASD) is highly heterogeneous in symptom onset and severity, comorbidities, and treatment responsiveness, challenging the notion of a single pathogenic mechanism. Increasing evidence indicates that some individuals with ASD exhibit prominent peripheral physiological alterations, including gastrointestinal (GI) dysfunction, gut microbial dysbiosis, immune imbalance, oxidative stress, and mitochondrial/energy metabolic vulnerability. In this context, gut-derived metabolites-particularly short-chain fatty acids (SCFAs)-have emerged as plausible modulators of the neurodevelopmental milieu through the expanded gut-immune-metabolic-brain axis. This review synthesizes: (i) SCFAs' biogenesis and physiological roles, (ii) context- and developmental stage-dependent effects, (iii) the clinical heterogeneity of reported microbiome and SCFA alterations in ASD, and (iv) propionate as a frequently discussed candidate signal and the interpretive boundaries of preclinical evidence. Human studies show substantial inter-study variability in SCFA alterations (increases, decreases, or no differences), influenced by factors such as sample type (stool vs. blood), GI symptoms, diet, medication exposure, and analytical variability. Accordingly, SCFAs should not be treated as universal ASD biomarkers but rather as context-dependent metabolic signals relevant under specific clinical and biological conditions. Building on this premise, we propose the conceptual framework of "metabolic ASD" representing a metabolically informed dimension of biological variability in which peripheral metabolic-immune perturbations may contribute to neurodevelopmental vulnerability. To avoid premature causal claims, we outline design requirements for future research, including stratified study designs, longitudinal cohorts, and integrative multi-layer analyses. Ultimately, metabolic ASD should be positioned as a testable precision medicine research framework rather than a universal etiological model.