Nature Human Behaviour

This study developed a large language model (LLM)-based solution to identify people at HIV risk using electronic health records. We transformed structured EHR data, including demographics, diagnoses, and medications, into narrative descriptions ordered by visit date and applied GatorTron, a widely used clinical LLM trained on 82 billion words of de-identified clinical text. We compared GatorTron with traditional machine learning models, including LASSO and XGBoost. We identified a cohort with 54,265 individuals, where only 3,342 (6%) had new HIV diagnoses. Our LLM solution, based on GatorTron, achieved excellent performance, reaching an F1 score of 53.5% and an AUC of 0.88, comparable to traditional machine learning approaches. Subgroup analysis showed that, across age, sex, and race/ethnicity groups, both LLM and traditional models achieved AUCs above 0.82. Interpretability analyses showed broadly consistent patterns across LLM models and traditional machine learning models.

Objective. Menopause is a significant physiological transition with implications for health outcomes (e.g., cardiometabolic), yet gaps remain in understanding the menopause transition, including how menopause timing and type influence health outcomes. Large-scale cohort studies in midlife (age~40-60) females, including the All of Us Research Program (AoURP), provide opportunities to study menopause across diverse populations and data modalities. We characterized menopause-related data in AoURP, focusing on age distributions and concordance between EHR diagnosis codes and self-reported survey responses. Methods. We analyzed menopause-related survey, EHR diagnostic code, and genomic data among ~396,000 participants in AoURP with female sex. We summarized menopause data across modalities, overlap between survey, EHR, and genomic data, and age distributions overall and across sociodemographic characteristics. Results. Among ~396,000 females, surveys captured ~193,000 menopause observations, nearly seven times more than structured EHR diagnoses (~28,000), suggesting under- ascertainement in EHR data. Nearly all females (~99%) with an EHR menopause diagnosis also reported menopause in the survey. Approximately 22,000 participants had intersected EHR, survey, and genomic menopause-related data. Survey-based age patterns matched expectations, with participants <40 years predominantly reporting pre-menopausal status and those >60 years predominantly reporting post-menopausal status. A small subset (N{approx}1,700; 4%) (age>70 years) reported no menopause, suggesting response or recall bias. EHR menopause codes were concentrated after age>45 years, with a notable spike at age 65. Modest differences in survey-based menopause age distributions were observed by sociodemographic characteristics (e.g., race, ancestry). Conclusions. These findings inform sampling strategies, power calculations, phenotype definition, and study design for menopause research using AoURP.

Sex differences are commonly observed in neuroimaging phenotypes and in the risk of brain diseases, yet the underlying genetic mechanisms remain poorly understood. We investigated sex differences in the genetic architecture of 805 neuroimaging phenotypes in 22,950 males and 22,950 females matched for sample size and covariates, and systematically compared sex-stratified with sex-combined genetic analyses. We found eight variant-trait associations with significant sex differences, 235 fine-mapped sex-dominant causal associations, 457 sex-dominant colocalizations with sex hormones, and 96 sex-dominant colocalizations with schizophrenia. Compared with sex-combined analysis, sex-stratified analysis identified 47 new genetic associations, 170 new fine-mapped causal associations, 1,019 new colocalizations with sex hormones, and 191 new colocalizations with schizophrenia. Additionally, sex-stratified analysis improved global heritability and genetic-correlation estimates and enhanced polygenic prediction for certain phenotypes. This work highlights the need to routinely perform sex-stratified genetic association analyses to elucidate sex-specific and sex-shared genetic control of neuroimaging phenotypes and related disorders.

AO_SCPLOWBSTRACTC_SCPLOWThe volume of scientific manuscripts is rising faster than the available pool of expert reviewers, and AI tools are emerging as a possible response, ranging from frontier large language models applied directly to peer review to purpose-built multi-agent systems. Scalable, standardized benchmarks are needed to regularly evaluate how these tools compare to one another and to human reviewers. We present ReviewBench, an open-source, venue-agnostic framework that compares human and AI reviews across structure, alignment with a papers major claims, impact, and critique category. We apply ReviewBench to 145,021 review comments from human reviewers, frontier large language models (GPT-5.2, and Gemini 3 Pro), and Reviewer3.com (R3), a multi-agent peer review system. The dataset spans papers in computer science (ICLR 2025, n = 1,000), social science (Nature Human Behaviour, n = 142), and life science (eLife, n = 1,000). Across disciplines, AI reviews are more structured and engage more directly with a papers major claims, with R3 more often surfacing consequential comments, defined as comments capable of undermining those claims. When restricting to critical comments, however, human reviewers rank first on consequential rate on more individual papers than any AI source, despite a lower average. We identify a bimodal reviewer distribution with peaks near 0% and 100%, indicating that many reviewers outperform AI on this metric, but a substantial fraction of reviewers near 0% brings the average down. Critique typing demonstrates systematic differences, where humans emphasize contribution and clarity, while AI emphasizes validity, sufficiency, and transparency. Together, these findings argue against framing AI as a replacement for human review and instead support a complementary model in which AI scales technical verification of major claims while human judgment remains essential for evaluating contribution and shaping editorial decisions.

Evidence accumulation is a fundamental aspect of human decision-making. However, how the precise temporal structure of evidence shapes the accumulation process has not been systematically studied. As a result, current understanding of evidence accumulation remains largely limited to its time-averaged behavior. We tested human subjects in a visual estimation task in which they inferred the angular position of an unknown source from a noisy stimulus sequence. Introducing systematic temporal perturbations, i.e., breaks of different durations and at different positions in the otherwise regular evidence sequence, revealed that subjects actively compensated for the memory loss endured during the break by dynamically enhancing evidence integration and memory maintenance immediately after the break. We derived a new time-continuous Bayesian updating model that is dynamically constrained by optimal performance-effort trade-offs. With two free parameters determining the overall resource-efficiencies of encoding and memory maintenance, the model accurately predicts the rich dependencies of subjects accumulation behavior on the evidence schedule, including subjects individual tendencies to emphasize either early (primacy) or late (recency) samples in the evidence sequence. Our results suggest that evidence accumulation is a non-stationary, dynamically controlled process that optimally balances the information gained from incoming evidence against the cognitive effort required to acquire and maintain it. The proposed model is general and should apply broadly across many task domains.

A distinctive feature of bodily experience is its transparency. During skilled action, our limbs recede from awareness and function as the medium of interaction rather than perceptual objects1. This is reflected in systematic perceptual biases: humans reliably underestimate the weight of their own hands2, potentially reflecting predictive motor processes that modulate self-generated sensory signals. Wearable technologies may test the limits of this perceptual transparency. Exoskeletons and other augmentative devices attach directly to the body, adding mass that must be integrated into sensorimotor control3; yet little is known about how such devices are experienced as they become integrated into the sensorimotor system. Here, we tested whether training with finger-extending exoskeletons alters their perceived weight and whether such changes depend on active use. We developed a Bayesian analytic framework combining individual psychometric modelling with a regression-based decomposition of perceived weight, to partition contributions of the biological hand and attached exoskeletal device. Thirty-four right-handed adults completed a weight-perception task before and after 20 minutes of training with either finger-extending or non-augmenting control devices. Participants compared the perceived weight of their right hand, with or without the exoskeleton, to reference weights suspended from the opposite wrist. Before training, the weight of both the biological hand and the exoskeleton were underestimated to a similar degree ([~]25- 30%), suggesting rapid perceptual integration following attachment. Training selectively increased attenuation of the perceived weight of the finger-extending exoskeleton, with no corresponding change for the biological hand and little evidence for a general training effect. These findings support a two-stage embodiment process in which passive attachment initiates perceptual updating, while sensorimotor training consolidates integration through functional interaction with the device. Perceived weight thus provides a behavioral marker of embodiment, offering insight into how the sensorimotor system integrates wearable augmentative technologies.

BackgroundSustained retention in care supports continuous access to antiretroviral therapy, routine clinical monitoring, and long-term viral suppression. ObjectiveTo compare the effectiveness of interventions for improving retention in care among people living with HIV (PLHIV). DesignSystematic review and network meta-analysis Data sourcesPubMed, Embase, CINAHL, PsycINFO, Web of Science, and the Cochrane Library from 1995 to December 2024. Eligibility criteriaRandomised controlled trials (RCTs) evaluating interventions to improve retention in care, viral load suppression, or quality of life (QoL) among PLHIV, compared with standard of care (SoC) or other interventions. Data extraction and synthesisPairs of reviewers independently screened studies, extracted data, and assessed risk of bias using ROBUST-RCT. We conducted a fixed-effect frequentist network meta-analysis and rated interventions categories relative to SoC based on effect estimates effects and the certainty of evidence.. Dichotomous outcomes were summarized as odds ratios (ORs) with 95% confidence intervals (CIs), and continuous outcomes as mean differences (MDs) with 95% CI. ResultsEighty-four trials enrolling 107 137 PLHIV evaluated 13 intervention categories. For retention in care, five interventions supported by moderate or high certainty evidence proved superior to SoC: multi-month dispensing (OR 2.02, 95% CI 1.32 to 3.09), task shifting (OR 1.94, 95% CI 1.42 to 2.66), differentiated service delivery (OR 1.47, 95% CI 1.22 to 1.76), behavioural counselling (OR 1.36, 95% CI 1.21 to 1.54), and supportive interventions (OR 1.31, 95% CI 1.11 to 1.55). For viral load suppression, two interventions supported by moderate or high certainty evidence proved superior to SoC: task shifting (OR 2.07, 95% CI 1.25 to 3.43) and behavioural counselling (OR 1.34, 95% CI 1.11 to 1.67). Across outcomes, no intervention demonstrated convincing superiority over other active interventions. ConclusionsAmong 13 intervention categories, only a subset provided moderate or high-certainty evidence of superiority to the standard of care, and no superiority to other interventions. Persistent evidence gaps for key populations, diverse settings, and long-term outcomes support the need for context-sensitive and patient-centred interventions. RegistrationPROSPERO CRD42024589177 Strengths and limitations of this study[tpltrtarr] This systematic review followed Cochrane methods and was reported in accordance with PRISMA-NMA guidelines. [tpltrtarr]The network meta-analysis integrated direct and indirect evidence to compare multiple intervention categories within a single framework. [tpltrtarr]Risk of bias and certainty of evidence were assessed using ROBUST-RCT and the GRADE approach for network meta-analysis, respectively. [tpltrtarr]Some networks were sparse, and limited representation of key populations and long-term follow-up constrained the strength and generalisability of inferences.

Social determinants of health (SDoH) are important for clinical care, but it remains unclear how much AI-captured social context is preserved after clinician editing in ambient documentation workflows. We retrospectively analyzed 75,133 paired ambient AI-drafted and clinician-finalized note sections from ambulatory care at a large academic health system. Using a rule-based NLP pipeline, we extracted 21 SDoH categories and quantified retention, deletion, and addition. SDoH appeared in 25.2% of AI drafts versus 17.2% of final notes. At the mention level, AI captured 29,991 SDoH mentions, of which 45.1% were deleted, 54.9% were retained with clinicians adding 3,583 new mentions. Insurance and marital status were most often deleted, whereas substance use and physical activity were more often retained. Deletion patterns also varied by specialty, supporting the need for specialty-aware ambient AI systems.

IntroductionOn January 20, 2025, the U.S. government froze foreign assistance including for PEPFAR, though a limited waiver for "life-saving" interventions was subsequently granted. PEPFARs 2025 monitoring results, released April 17, 2026, covered only quarter 4 while an earlier inadvertent release included all four quarters. Combining both data sets, we systematically assess facility-level programmatic performance and reporting trends to quantify service disruptions accounting for reporting discrepancies. MethodsWe categorized facilities by reporting continuity across Q1 2024 and Q4 2025 (e.g. continuous, intermittent, dropped, or new) and assessed changes in service delivery by the category of health facility for key HIV treatment, testing, PMTCT, and prevention programming. We additionally analyze changes in employed human resources for health (HRH) reported by PEPFAR. ResultsPEPFAR data included 31,746 facilities and community service sites. 71.3% were classified as continuous reporters, 16.9% intermittent reporters, 2.5% community services, 3.9% dropped in 2025, and 3.1% new in 2025. Total number of people accessing HIV treatment declined modestly by -0.3%, but differed by facility category. Continuous facilities saw a 0.5% increase in people on treatment, while intermittent facilities saw a -1.7% decrease. HIV testing declined -17%. HIV diagnoses declined -13% in continuous facilities, -35% in community services, and -29% in intermittent facilities. PMTCT infant testing and diagnoses declined by -6% and -12% in continuous facilities, respectively, and -60% and -31% in intermittent facilities, respectively. PrEP initiations declined -33%. Total direct service delivery HCWs reduced -62,541 (-24%) ConclusionThese findings reveal substantial disruptions across PEPFAR service areas, with the steepest declines among intermittent and community-based delivery sites, alongside a 24% reduction in direct service delivery healthcare workers. As potentially the final data set PEPFAR will ever release, these findings represent a troubling inflection point. The dismantling of public data systems and accountability structures undermine progress and enable programmatic gaps to develop and go unnoticed that risk allowing HIV resurgence to occur over the coming years.

How early visual cortex (EVC) represents working memory (WM) content while continuing to process incoming sensory input remains unclear. Using fMRI with prolonged delays to isolate mnemonic activity, together with analyses of cross-decoding, low-dimensional subspace structure, and representational geometry, we examined the relationship between sensory and mnemonic orientation representations in human EVC. Cross-decoding generalized poorly between sensory and mnemonic epochs, but this did not imply unrelated codes. Rather, the two occupied separable low-dimensional subspaces while preserving representational geometry across epochs. During discrimination and estimation, sensory- and mnemonic-trained decoders yielded dissociable readouts of concurrent sensory and mnemonic information from the same EVC measurements. Mnemonic coding showed little dependence on the retinotopic radial bias that characterized sensory coding, and trial-by-trial variability in mnemonic representation predicted both discrimination choices and estimation reports. Our findings support a population-level account in which mnemonic information in EVC is re-expressed in a separable but geometrically preserved format.

Genome-guided dietary advice is a goal of precision nutrition. However, the contribution of gene-diet interactions (GxDs) to disease risk remains unclear, hindering the identification of diet-outcome pairs more likely amenable to genetic-based recommendations. We thus implemented a two-step approach: first, we comprehensively assessed the contributions of genome-wide GxDs to cardiometabolic outcomes across a broad array of dietary exposures in UK Biobank participants (N = 141,144 to 325,989). Second, we selected the 20 significant diet-outcome pairs from the 713 pairs tested (p < 7.0 x 10-5) and derived GxD polygenic scores. In an independent sample, all scores were nominally associated with their corresponding outcomes, with 12 of 20 polygenic scores Bonferroni significant (p < 0.0025). Further analyses revealed GxD polygenic scores were associated with clinical outcomes such as incident gout, suggesting translational potential. Altogether, these results showcase the promise of GxD scores to inform precision nutrition.

To minimise health disparities, equitable access to medical treatment is paramount. In a pioneering intervention, National Health Service Englands Hepatitis C virus (HCV) programme has implemented country-wide peer support to boost treatment access. Peer support workers (peers) are individuals with relevant lived experience, who promote testing and treatment in marginalised populations underserved by traditional health services. We evaluated the English peers intervention, exploiting its staggered rollout and rich surveillance data between June 2016 and May 2021. Peers increased HCV cases identified by 13{middle dot}9% (95% credible interval (95% CrI) [5{middle dot}3, 21{middle dot}7]), sustained viral responses by 8{middle dot}0% (95% CrI [-4{middle dot}4, 18{middle dot}6]), and drug services referrals by 8{middle dot}8% (95% CrI [-12{middle dot}5, 22{middle dot}6]). The interventions effectiveness was magnified during the first COVID-19 lockdown and individuals supported by peers typically belonged to populations with poor treatment access. Our findings indicate that peers can boost equity in treatment access on a national scale.

Non-pharmaceutical interventions (NPIs) have been important for controlling SARS-CoV-2 transmission, particularly before and during initial vaccine rollout. During the pandemic, the US Centers for Disease Control and Prevention issued isolation and masking guidance in case of COVID-19-like illness, a positive SARS-CoV-2 test, or known exposure to SARS-CoV-2. However, the impact of this guidance on mitigating transmission in office workplaces is unclear. We used a network-based mathematical model to estimate the impact of this guidance on SARS-CoV-2 transmission among office workers and their communities. The model represented social contacts in the home, office, and community. We used data from the CorporateMix study to parametrize social contacts among office workers and calibrated the model to represent the COVID-19 epidemic in Georgia, USA from January 2021 through August 2022. In the reference scenario (58% adherence to guidance among office workers and the broader population), workplace transmission accounted for a small fraction of total infections. Reducing adherence among office workers to 0% increased workplace transmissions by 27.1% and increasing adherence to 75% reduced workplace transmission by 7.0%. Increasing adherence to 75% among office workers had minimal impact on symptomatic cases and deaths; increasing it among the broader population was more effective in reducing office worker cases and deaths. In our model, moderate adherence to recommended NPIs in workplaces was effective in reducing transmission, but increasing adherence had limited benefit given workplaces that have low contact intensity and hybrid work arrangements. These results underscore the public health benefits of community-wide adoption of recommended NPIs.

Early behavioral and temperamental differences are important indicators of later socioemotional development and psychopathology risk, yet their neural bases near birth remain incompletely understood. Using resting-state fMRI data from the Developing Human Connectome Project, we examined whether neonatal functional connectivity predicts 18-month behavioral and temperament outcomes in 397 infants (277 term-born, 120 preterm-born). Outcomes were assessed using the Child Behavior Checklist (CBCL) and the Early Childhood Behavior Questionnaire (ECBQ). We applied a stability-driven, ROI-constrained connectome-based predictive modeling framework to identify robust whole-brain connectivity features associated with later externalizing, internalizing, surgency, negative affect, and effortful control. Significant predictive models were observed for multiple outcomes across the whole cohort as well as within term-born and preterm-born groups, with clear differences in predictive architecture between cohorts. Across analyses, prefrontal and temporoparietal systems were repeatedly implicated, alongside medial temporal, fusiform, parahippocampal, and orbitofrontal-related regions. These findings indicate that large-scale neonatal functional organization is meaningfully related to later socioemotional and behavioral variation, and that preterm birth is associated with partly distinct predictive connectivity patterns.

Visual surface perception is a fundamental aspect of vision, yet its neural implementation remains poorly understood. Troxlers perceptual filling-in paradigm provides a tractable illusion for studying surface perception, in which a peripheral figure becomes perceptually assimilated into the surrounding background after a period of sustained fixation. Although neural correlates of this phenomenon have been reported in early visual cortex, the underlying mechanisms, particularly the contribution of feedback signaling, remain unresolved. Here we use ultra-high-field (7T) layer-fMRI to investigate perceptual filling-in in the human visual cortex. While experimentally controlling perceptual filling-in, we measured GE-BOLD responses in ten participants. Analyses across cortical depth in the independently localized figure representation in primary visual cortex (V1) revealed neural correlates of filling-in in deep cortical layers, which are associated with feedback input. These findings provide evidence that perceptual filling-in and visual surface perception in general are supported by feedback signals to early visual cortex.

BackgroundChlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis are curable sexually transmitted infections (STIs) associated with adverse birth outcomes. Most infections are asymptomatic. Whether antenatal STI screening improves birth outcomes remains uncertain. MethodsIn a randomized three-group trial in South Africa, pregnant women aged 18 years or older were assigned before 27 weeks gestation to: (1) screening and treatment for Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis at enrollment, with tests-of-cure (One-Time Screening); (2) screening and treatment at enrollment, repeated at 30 to 34 weeks (Two-Time Screening); or (3) Standard-of-Care (Syndromic management). The primary outcome was a composite of preterm birth (<37 weeks gestation) or low birthweight (<2500 g), analyzed in the modified intention-to-treat population of participants with live births. Components of the composite outcome were evaluated individually as the main secondary outcomes. The study was registered with ClinicalTrials.gov, NCT04446611. FindingsOf 2247 enrolled participants, 1910 had live births. The composite outcome occurred in 22{middle dot}9% of the One-Time Screening group (risk ratio [RR] 0{middle dot}99; 95% confidence interval [CI] 0{middle dot}81-1{middle dot}21), 20{middle dot}6% of the Two-Time Screening group (RR 0{middle dot}89; 95% CI 0{middle dot}72-1{middle dot}09), compared with 23{middle dot}2% of the Standard-of-Care group. Preterm birth occurred in 18{middle dot}9% of the One-Time Screening group (RR 1{middle dot}00; 95% CI 0{middle dot}80-1{middle dot}26), 14{middle dot}5% of the Two-Time Screening group (RR 0{middle dot}77; 95% CI 0{middle dot}60-0{middle dot}99), and 18{middle dot}8% of the Standard-of-Care group. Low birthweight occurred in 14{middle dot}1% of the One-Time Screening group (RR 1{middle dot}10; 95% CI 0{middle dot}83-1{middle dot}46), 12{middle dot}9% of the Two-Time Screening group (RR 1{middle dot}01; 95% CI 0{middle dot}76-1{middle dot}35), and 12{middle dot}8% of the Standard-of-Care group. InterpretationNeither screening strategy for Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis reduced the primary composite outcome of preterm birth or low birthweight, or low birthweight alone. The Two-Time antenatal STI screening strategy, however, reduced preterm birth by 23%.

Socioeconomic inequalities in health and well-being are a major public health concern, particularly in ageing populations. Education is a key determinant shaping multiple aspects of health outcomes. We used cross-sectional data from wave 9 of the German sample (n=4,148) of the Survey of Health, Ageing and Retirement in Europe (SHARE) to test whether formal education is associated with well-being in later adulthood, with health literacy, self-rated health, and preventive health behaviours as possible mediators. Our results showed that education was positively associated with greater well-being, but only via indirect pathways. Specifically, self-rated health, health literacy, and fruit and vegetable consumption mediated the relationship between education and well-being accounting for 54.7, 24.7, and 12.6 percent of the total effect, respectively. In addition, there were significant positive correlations between education and health literacy, as well as high-intensity physical activity, daily fruit and vegetable consumption, more preventive health check-ups, and less smoking. In contrast, alcohol consumption was more common among those with higher levels of education. All health behaviours and health literacy were correlated directly or indirectly (i.e., mediated by health) with well-being. These findings highlight the importance of examining indirect pathways linking education to well-being in later life. Interventions aimed at improving health literacy and promoting healthy behaviours may help reduce educational inequalities in quality of life among older adults.

Salient sensory stimuli are known to evoke neural activations across distributed brain regions. However, the temporal dynamics of these responses over sub-second timescales remain poorly understood, in part due to limitations in the temporal resolution of non-invasive neuroimaging methods. We examined the spatiotemporal dynamics of neural activations evoked by salient sensory stimuli (rare sounds) using 1,194 widely distributed intracranial electrodes in 5 neurosurgical patients. Salient stimuli preferentially activated 263 of 1,194 electrodes (22%), with responses segregating into two largely distinct spatiotemporal patterns: (1) phasic activation in sensorimotor regions, and (2) sustained activation within the salience network. Cross-correlation analysis revealed that phasic sensorimotor activation preceded sustained salience network activation on a trial-by-trial basis. These findings support an updated view of salience processing in the human brain, revealing that salient stimuli evoke two sequential stages of neural activation--phasic sensorimotor responses followed by sustained salience network activity--rather than simultaneous widespread activation.

Multisensory perception is a cornerstone paradigm for understanding how the brain constructs coherent representations of the world from noisy, fragmented sensory inputs. For decades, researchers have used the magnitude of crossmodal illusions, the width of the temporal binding window, and related behavioral indices as direct proxies for integration strength, and have leveraged these measures to compare multisensory function across developmental, clinical, and aging populations. Here we argue that this descriptive practice is fundamentally compromised: behavioral readouts of multisensory integration are composite measures jointly shaped by unisensory precision, amodal priors, and the binding process itself, and cannot be interpreted in isolation. Drawing on simulations within a Bayesian Causal Inference framework, we show how identical behavioral patterns can arise from very different underlying causes, leading to systematic misattribution of group differences to deficits or enhancements in integration. We review complementary computational frameworks, including drift diffusion, multisensory correlation detection, and statistical facilitation models, and outline their respective explanatory limits. Finally, we provide a model-based inference pipeline, from experimental design and unisensory baselines to parameter estimation and interpretation, that disentangles sensory fidelity, prior expectations, and integrative tendency. Adopting this normative approach is essential for cumulative progress in basic multisensory research and for its translation to neuropsychiatric assessment, lifespan research, and artificial perceptual systems.

Drug overdose deaths in the United States reached record levels during the fentanyl era before recently declining. A plausible hypothesis is that a sudden drop in fentanyl purity beginning in 2023 caused the downturn in overdose mortality. We evaluated this hypothesis by replicating a published analysis with regional overdose data, using models that account for time trends and autocorrelation, and negative control indicators to test for spurious correlation. When fentanyl purity was rising, the national purity series did not track overdose increases in most regions and showed only a modest association in the West. When both purity and mortality later declined, the observed associations were also seen with unrelated macroeconomic indicators that shared the same time pattern. National fentanyl purity alone does not provide a sufficient explanation for recent overdose declines.