JAMA Network Open

BackgroundArtificial Intelligence (AI) will have unintended consequences for radiology. When a radiologist misses an abnormality on an image, their liability may differ according to whether or not AI also missed the abnormality. MethodsU.S. adults viewed a vignette describing a radiologist being sued for missing a brain bleed (N=652) or cancer (N=682). Participants were randomized to one of five conditions. In four conditions, they were told an AI system was used. Either AI agreed with the radiologist, also failing to find pathology (AI agree) or did find pathology (AI disagree). In the AI agree+FOR condition, AI agreed with the radiologist and an AI false omission rate (FOR) of 1% was presented. In the AI disagree+FDR condition, AI disagreed and an AI false discovery rate (FDR) of 50% was presented. There was also a no AI control condition. Otherwise, vignettes were identical. Participants indicated whether the radiologist met their duty of care as a proxy for whether they would side with defense (radiologist) or plaintiff in trial. ResultsParticipants were more likely to side with the plaintiff in the AI disagree vs. AI agree condition (brain bleed: 72.9% vs. 50.0%, p=0.0054; cancer: 78.7% vs. 63.5%, p=0.00365) and in the AI disagree vs. no AI condition (brain bleed: 72.9% vs. 56.3%, p=0.0054; cancer: 78.7% vs. 65.2%, p=0.00895). Participants were less likely to side with the plaintiff when FDR or FOR were provided: AI disagree vs AI disagree+FDR (brain bleed: 72.9% vs. 48.8%, p=0.00005; cancer: 78.7% vs. 73.1%, p=0.1507), and AI agree vs. AI agree+FOR (brain bleed: 50.0% vs. 34.0%, p=0.0044; cancer: 63.5% vs. 56.4%, p=0.1085). DiscussionRadiologists who failed to find an abnormality are viewed as more culpable when they used an AI system that detected the abnormality. Presenting participants with AI accuracy data decreased perceived liability. These findings have relevance for courtroom proceedings.

ObjectiveTo determine the SARS-CoV-2 seroprevalence among school workers in the setting of full in-person schooling and the highly transmissible Omicron variants of concern. DesignCross-sectional study among school staff, comparing to period-, age-, sex- and postal code-weighted data from Canadian blood donors from the same community. SettingThree large school districts in the greater Vancouver metropolitan area, British Columbia, Canada, with serology sampling done between January 26, 2022 and April 8, 2022. ParticipantsSchool staff actively working in the Vancouver, Richmond and Delta School Districts. Main outcome measureSARS-CoV-2 seroprevalence based on nucleocapsid (N)-protein testing, adjusted for the sensitivity and specificity of the assay. ResultsA majority (65.8%) of the 1845 school staff enrolled reported close contact with a COVID-19 case outside the household. Of those, about half reported close contact with a COVID-19 case at school either in a student (51.5%) or co-worker (54.9%). In a representative sample of 1620 (87.8%) school staff, the adjusted seroprevalence was 26.5% [95%CrI: 23.9 - 29.3%]. This compared to an age, sex and residency area-weighted seroprevalence of 32.4% [95%CrI: 30.6 - 34.5%] among 7164 blood donors. ConclusionDespite frequent COVID-19 exposures, the prevalence of SARS-CoV-2 infections among the staff of three main school districts in the Vancouver metropolitan area was no greater than a reference group of blood donors, even after the emergence of the more transmissible Omicron variant. What is already known on this subject?O_LIEarlier studies indicate that COVID-19 infection rates are not increased among school staff at previous stages of the pandemic compared to the community, yet controversy remains whether this will remain true after the emergence of the highly transmissible Omicron variant. C_LI What this study adds?O_LIDespite frequent COVID-19 exposures, this study identified no detectable increase in SARS-CoV-2 seroprevalence among school staff working in three metro Vancouver public school districts after the first Omicron wave in British Columbia, compared to a reference group of blood donors from the same age, sex and community area. C_LI

ImportanceMany patients will develop more than one skin cancer, however most research to date has examined only case status. ObjectiveDescribe the frequency and timing of the treatment of multiple skin cancers in individual patients over time DesignLongitudinal claims and electronic health record-based cohort study SettingVanderbilt University Medical Center database called the Synthetic Derivative, VA, Medicare, Optum Clinformatics(R) Data Mart Database, IBM Marketscan ParticipantsAll patients with a Current Procedural Terminology code for the surgical management of a skin cancer in each of five cohorts. ExposuresNone. Main Outcomes and MeasuresThe number of CPT codes for skin cancer treatment in each individual occurring on the same day as an ICD code for skin cancer over time ResultsOur cohort included 5,508,374 patients and 13,102,123 total skin cancers treated. Conclusions and RelevanceNearly half of patients treated for skin cancer were treated for more than one skin cancer. Patients who have not developed a second skin cancer by 2 years after the first are unlikely to develop multiple skin cancers within the following 5 years. Better data formatting will allow for improved granularity in identifying individuals at high risk for multiple skin cancers and those unlikely to benefit from continued annual surveillance. Resource planning should take into account not just the number of skin cancer cases, but the individual burden of disease. Key pointsQuestion: How many skin cancer patients are treated for more than one skin cancer and how soon after the first skin cancer do they occur? Findings: 43% of patients were treated for more than one skin cancer, the majority of which occurred within two years after the initial skin cancer. Just 3% of patients were treated for 10 or more skin cancers, but these patients accounted for 22% of all of the skin cancer treatments in the cohort Meaning: Nearly half of all skin cancer patients were treated for multiple skin cancers, while those without a second skin cancer after two years were less likely to be treated for a subsequent skin cancer within the next five years.

IntroductionThe SARS-CoV-2 secondary attack rate (SAR) in schools is low when mitigation measures are adopted, Data on the relative impact of such strategies are limited. We evaluated the SARS-CoV-2 SAR in Massachusetts schools during 2020-21 and factors associated with transmission risk. MethodsIn a convenience sample of 25 Massachusetts public K-12 school districts, de-identified information about SARS-CoV-2 cases and their school-based contacts was reported using a standardized contact-tracing tool. Index cases were included if they were in school while infectious. SAR was defined as the proportion of in-school contacts acquiring SARS-CoV-2 infection and designated as possible or probable in-school transmission by school-based teams. We compared exposure-specific SAR using unadjusted risk ratios (RR) with 95% confidence intervals (CI); p-values were calculated using Fishers exact tests. ResultsEight districts (70 schools with >33,000 enrolled students) participated. There were 435 index cases and 1,771 school-based contacts (Table 1). Most contacts (1327/1771 [75%]) underwent SARS-CoV-2 testing and 39/1327 (2.9%) contacts tested positive. Of 39 positive contacts, 10 (25.6%) had clear out-of-school exposures and were deemed not in-school transmissions, so were excluded from further calculations. Twenty-nine (74.4%) contacts were deemed possible or probable in-school transmissions, resulting in an in-school SAR of 2.2%. Of the 29 in-school transmissions, 6 (20.7%) were staff-to-staff, 7 (24.1%) were staff-to-student, 3 (10.3%) were student-to-staff, and 13 (44.8%) were student-to-student; 6 (20.7%) occurred from index cases attending work/school while symptomatic. The unadjusted SAR (Table 2) was significantly higher if the index case was a staff member versus a student (RR 2.18, 95% CI 1.06-4.49; p=0.030), if the index case was identified via in-school contact tracing versus via school-based asymptomatic testing (RR 8.44, 95% CI 1.98-36.06; p=0.001), if the exposure occurred at lunch versus elsewhere (RR 5.74, 95% CI 2.11-15.63; p<0.001; all lunch transmissions were staff-to-staff), and if both parties were unmasked versus both masked (RR 6.98, 95% CI 3.09-15.77; p<0.001). For students, SAR did not differ by grade level. O_TBL View this table: org.highwire.dtl.DTLVardef@1b1f706org.highwire.dtl.DTLVardef@d0cea9org.highwire.dtl.DTLVardef@afcd27org.highwire.dtl.DTLVardef@1c5e7e7org.highwire.dtl.DTLVardef@1fc0731_HPS_FORMAT_FIGEXP M_TBL O_FLOATNOTable 1:C_FLOATNO O_TABLECAPTIONNumber of SARS-CoV-2 index cases, possible and probable in-school transmissions, secondary attack rates, and number of contacts per index case in 8 Massachusetts K-12 public school districts, 2020-21 C_TABLECAPTION C_TBL O_TBL View this table: org.highwire.dtl.DTLVardef@13c5853org.highwire.dtl.DTLVardef@cac231org.highwire.dtl.DTLVardef@1931f5eorg.highwire.dtl.DTLVardef@1f095c1org.highwire.dtl.DTLVardef@1f7bc21_HPS_FORMAT_FIGEXP M_TBL O_FLOATNOTable 2:C_FLOATNO O_TABLECAPTIONNumber of index cases and contacts and secondary attack rate by type of exposure: 8 public MA K-12 districts, 2020-21 C_TABLECAPTION C_TBL ConclusionsSecondary attack rates for SARS-CoV-2 were low in public school settings with comprehensive mitigation measures in place before the emergence of the delta variant; lack of masking and staff-to-staff dining were associated with increased risk.

BackgroundThe unprecedented context of the COVID-19 pandemic poses the opportunity to study several questions in circumstances that would probably not otherwise occur. We sought to determine the dynamics of pharmaceutical company drug sales revenue, market capitalization and payments to physicians during the pandemic, focusing on payments to so-called key opinion leaders (KOLs). MethodsWe analyzed the CMS Open Payments data of 15 top pharmaceutical company general payments to US physicians. We calculated total payments per year for all physicians, KOLs and 2018 KOLs in subsequent years. Drug-related fold changes in payments, drug revenues and company market capitation were calculated using Q1-2018 as reference. Yearly differences in payments, drug sales revenue and market capitalization were tested using generalized estimation equations (GEE). A double-sided p<0.05 was considered significant. ResultsThe analyzed dataset comprised 8,563,872 payments to 382,779 physicians. In 2020, we observed a reduction in payments to physicians and KOLs compared to prior years. The total amount per KOL physician per company also decreased for each year for KOLs and the 2018 KOLs in the subsequent years. Payments per drug, but neither drug revenues nor pharmaceutical company market capitalization, followed a downward trend in 2020 compared to prior years. GEE analysis confirmed that, compared to 2018, the decrease in payments to KOLs overall and for the top drugs of each company was statistically significant. Yet, no significant differences in drug sales revenue and market capitalization was observed. ConclusionsA substantial and significant reduction in payments to KOLs during the first fiscal year of the COVID-19 pandemic was not associated with a reduction in drug sales revenue of blockbuster drug products and the market capitalization of 15 top pharmaceutical companies. Overall, these findings suggest that a substantial part of pharmaceutical payments to KOLs do not appear to impact top drug sales revenues.

ObjectiveTo estimate the effectiveness of 2-dose and 3-dose mRNA vaccination (BNT162b2 and mRNA-1273) against general SARS-CoV-2 infection (asymptomatic or symptomatic) caused by the omicron variant. DesignPropensity-score matched retrospective Cohort Study. SettingLarge public university undergoing weekly Covid-19 testing in South Carolina, USA. ParticipantsPopulation consists of 24,145 university students and employees undergoing weekly Covid-19 testing between January 3rd and January 31st, 2022. The analytic sample was constructed via propensity score matching on vaccination status: Unvaccinated, completion of 2-dose mRNA series within previous 5 months, and receipt of mRNA booster dose within previous 5 months. The resulting analytic sample consists of 1,944 university students and 658 university employees. InterventionVaccination with a two dose or 3 dose regimen of the BNT162b2 or mRNA-1273 vaccine. ResultsBooster protection against any SARS-CoV-2 infection was 66.4% among employees (95% CI: 46.1-79.0%; P<.001) and 45.4% among students (95% CI: 30.0-57.4%; P<.001). Compared to the 2-dose mRNA series, estimated increase in protection from the booster dose was 40.8% among employees (P=.024) and 37.7% among students (P=.001). We did not have enough evidence to conclude a statistically significant protective effect of the 2-dose mRNA vaccination series, nor did we have enough evidence to conclude that protection waned in the 5-month period after receipt of the 2nd or 3rd mRNA dose. Furthermore, we did not find evidence that protection varied by manufacturer. ConclusionsCovid-19 mRNA booster doses offer moderate protection against any SARS-CoV-2 infection caused by the omicron variant and provide a substantial increase in protection relative to the 2-dose mRNA vaccination series.

ImportanceThe Affordable Care Act expanded Medicaid eligibility in participating states to individuals with incomes up to 138% of the federal poverty line. The effects of this policy on the diagnosis and treatment of melanoma in low-income populations has yet to be described. ObjectiveTo evaluate the effect of Medicaid expansion on changes in insurance status and clinical characteristics of low-income patients with newly diagnosed melanoma. Design, Setting, and ParticipantsThis cross-sectional study included patients younger than 65 with a new diagnosis of malignant melanoma from January 1, 2011 to December 31, 2016, in the US National Cancer Institutes Surveillance Epidemiology and End Results database. ExposuresResidence in a state that expanded Medicaid on January 1, 2014. Main Outcomes and MeasuresThe primary outcomes were insurance status, melanoma staging, and overall survival. ResultsIn Medicaid expansion states, there were 1,719 low-income patients with newly diagnosed melanoma during the pre-expansion time period and 1,984 (15% increase) during the post-expansion time period. In nonexpansion states, there were 326 low-income patients with newly diagnosed melanoma during the pre-expansion time period, and 288 during the post-expansion time period (12% decrease). Compared with nonexpansion states, expansion states had a significantly greater reduction in percentage of uninsured patients following Medicaid expansion (adjusted odds ratio, 6.27 [95% CI, 4.83 to 8.14]). Overall survival was not statistically different between expansion and nonexpansion states (HR, 0.89 [95% CI, 0.74 to 1.06]). There were no statistically significant differences in melanoma staging at diagnosis between the expansion and nonexpansion groups (p = 0.05). Conclusions and RelevanceMedicaid expansion was associated with increased melanoma diagnoses in low-income patients and a decreased proportion of uninsured patients. However, our study did not identify differences in clinical outcomes associated with Medicaid expansion. Key PointsO_ST_ABSQuestionC_ST_ABSWas Medicaid expansion associated with changes in insurance status and clinical characteristics of low-income melanoma patients? FindingsMedicaid expansion was associated with increased diagnoses of melanoma in low-income populations and reductions in the proportion of uninsured melanoma patients. However, there were no statistically significant changes in staging at diagnosis or overall survival associated with Medicaid expansion. MeaningIncreased health insurance coverage associated with Medicaid expansion could potentially improve timely detection and treatment of melanoma for low-income populations.

ImportanceCigarette smoking during pregnancy increases the risks of miscarriage, ectopic pregnancy, placental complications, hypertensive disorders of pregnancy, and infant mortality. Reducing smoking in pregnancy remains a pressing public health priority. ObjectiveTo project the impact of a proposed nicotine product standard on maternal complications and infant mortality using the Smoking, E-cigarette use, and Pregnancy (SEP) microsimulation model. Design, Settings, and ParticipantsIndividual-level, annual-cycle microsimulation of U.S. females of reproductive age that tracks smoking, vaping, pregnancy, and pregnancy outcomes was constructed. Model inputs were drawn from national surveillance and vital statistics, including National Health Interview Survey (NHIS) for general population, Behavioral Risk Factor Surveillance System (BRFSS) (2016-2023) for smoking/vaping among pregnant women and NVSS Natality/Linked Birth-Infant Death files for late-pregnancy morbidities and infant mortality. Policy effects on tobacco and e-cig usage transitions are based on FDAs expert-elicitation and used to simulate outcomes under status quo vs. the new policy from 2027-2100. ExposureSmoking and vaping. Main Outcome(s)Maternal outcomes during pregnancy (ectopic pregnancy, miscarriage, placenta previa, placental abruption, hypertensive disorders of pregnancy/pre-eclampsia, eclampsia), infant mortality, pregnancy-related costs, and maternal Quality Adjusted Life Years (QALYs). ResultsUnder the nicotine product standard (policy start 2027), smoking in pregnancy falls sharply from 6.0% in 2027 to 1.2% by 2040. These behavioral shifts translate into large perinatal gains through 2100: approximately 167,000 ectopic pregnancies, 950,000 miscarriages, 15,000 placenta previa, 62,000 placental abruptions, 167,000 hypertensive disorders of pregnancy/pre-eclampsia, 9,000 eclampsia cases, and 64,000 infant deaths are averted cumulatively. Maternal health improves as well, with 103,000 pregnancy QALYs gained. Health system spending falls despite conservative costing, with $4.9 billion in pregnancy-related medical costs avoided. Across uncertainty bounds, direction and magnitude of benefit remain favorable for all maternal morbidities and infant deaths, indicating that new policy yields substantial and durable health gains alongside meaningful cost offsets. ConclusionsA proposed nicotine product standard is projected to improve maternal and infant outcomes and yield sizable pregnancy-related health gains and cost offsets. The SEP model complements prior tobacco policy evaluation frameworks while focusing on maternal and infant health.

BackgroundIntentional inhalation of 1,1-difluoroethane (DFE), the propellant in aerosol duster products, is a leading cause of inhalant-abuse death in the United States. The CPSC has cited death counts from its Clearinghouse in regulatory proceedings, yet no peer-reviewed publication has described the methods used to identify these cases. ObjectivesTo estimate DFE- and duster-related deaths in the CPSC Clearinghouse for 2011-2021, characterize reporting patterns, and assess classification reliability against an independently coded dataset. MethodsDeath records (N = 6,316) were identified from 261,076 Clearinghouse records using CPSC product codes for chemicals, aerosols, gases, and related products. Each record was classified through narrative review and substance coding. Inter-rater reliability was assessed against an independently coded dataset from Families United Against Inhalant Abuse (FUAIA) using Cohens kappa and Gwets AC1. ResultsOf 2,451 inhalant-abuse deaths identified (70.8% male; mean age 36.9 years), 2,097 (85.6%) involved DFE or aerosol duster products. DFE/duster deaths rose from 110 (2011) to 266 (2016). Only 17% of cases were received in the same calendar year as the incident. Prior to reconciliation, comparison with the FUAIA dataset yielded Cohens kappa of 0.90 (95% CI [0.89, 0.91]); all discrepancies were subsequently resolved through joint review. ConclusionAerosol duster products account for approximately 86% of inhalant-abuse deaths reported to the CPSC Clearinghouse; however, these counts significantly underestimate true prevalence. The concentration of mortality in a single, widely available product class supports targeted product-level interventions and provides the first peer-reviewed baseline for evaluating the impact of regulatory and prevention efforts.

BackgroundLow-value services are common in cancer care. The onset of the COVID-19 pandemic caused a dramatic decrease in health care utilization, leading many to suspect that low-value cancer services may decrease. MethodsIn this retrospective cohort study, we used administrative claims from the HealthCore Integrated Research Environment, a repository of medical and pharmacy data from US health plans representing over 80 million members, to identify 204,581 patients diagnosed with breast, colorectal, and/or lung cancer between January 1, 2015, and March 31, 2021. We used linear probability models to investigate the relation between the onset of COVID-19 pandemic and 5 guideline-based metrics of low-value cancer care: 1) Positron Emission Tomography/Computed Tomography (PET/CT) instead of conventional CT imaging for initial staging; 2) conventional fractionation instead of hypofractionation for early-stage breast cancer; 3) non-guideline-based antiemetic use for minimal-, low-, or moderate-to-high-risk chemotherapies; 4) off-pathway systemic therapy; and 5) aggressive end-of-life care. ResultsAmong 204,581 patients, the mean [SD] age was 63.1 [13.2], 68.1% were female, 83,593 (40.8%) had breast cancer, 56,373 (27.5%) had colon cancer, and 64,615 (31.5%) had lung cancer. Rates of low-value cancer services did not exhibit meaningful declines during the pandemic: PET/CT imaging, adjusted percentage point difference 1.87 (95% CI -0.13 to 3.87); conventional radiotherapy, adjusted percentage point difference 3.93 (95% CI 1.50 to 6.36); off-pathway systemic therapy, adjusted percentage point difference 0.82 (95% CI -0.62 to 2.25); non-guideline-based antiemetics, adjusted percentage point difference -3.62 (95% CI -4.97 to -2.27); aggressive end-of-life care, adjusted percentage point difference 2.71 (95% CI -0.59 to 6.02). DiscussionLow-value cancer care remained prevalent through the pandemic. Policymakers should consider changes to payment and incentive design to turn the tide toward higher-value cancer care.

IntroductionCoronavirus disease-19 (COVID-19) has caused a marked increase in all-cause deaths in the United States, mostly among adults aged 65 and older. Because younger adults have far lower infection fatality rates, less attention has been focused on the mortality burden of COVID-19 in this demographic. MethodsWe performed an observational cohort study using public data from the National Center for Health Statistics at the United States Centers for Disease Control and Prevention, and CDC Wonder. We analyzed all-cause mortality among adults ages 25-44 during the COVID-19 pandemic in the United States. Further, we compared COVID-19-related deaths in this age group during the pandemic period to all drug overdose deaths and opioid-specific overdose deaths in each of the ten Health and Human Services (HHS) regions during the corresponding period of 2018, the most recent year for which data are available. ResultsAs of September 6, 2020, 74,027 all-cause deaths occurred among persons ages 25-44 years during the period from March 1st to July 31st, 2020, 14,155 more than during the same period of 2019, a 23% relative increase (incident rate ratio 1.23; 95% CI 1.21-1.24), with a peak of 30% occurring in May (IRR 1.30; 95% CI 1.27-1.33). In HHS Region 2 (New York, New Jersey), HHS Region 6 (Arkansas, Louisiana, New Mexico, Oklahoma, Texas), and HHS Region 9 (Arizona, California, Hawaii, Nevada), COVID-19 deaths exceeded 2018 unintentional opioid overdose deaths during at least one month. Combined, 2,450 COVID-19 deaths were recorded in these three regions during the pandemic period, compared to 2,445 opioid deaths during the same period of 2018. MeaningWe find that COVID-19 has likely become the leading cause of death--surpassing unintentional overdoses--among young adults aged 25-44 in some areas of the United States during substantial COVID-19 outbreaks. NoteThe data presented here have since been updated. As a result, an additional 1,902 all-cause deaths occurring among US adults ages 25-44 during the period of interest are not accounted for in this manuscript.

Key PointsO_LIIn a randomized trial, an educational booklet and video did not increase evaluation completion or living donor kidney transplant receipt. C_LIO_LIFor patients who received the booklet and video intervention, experiencing discrimination in healthcare reduced evaluation completion. C_LIO_LILong-term follow-up and a large sample size yielded sufficient power to validate a true null effect of the intervention on key outcomes. C_LI BackgroundKidney transplantation (KT) evaluation is a complex, lengthy process; and living donor KT (LDKT) is the optimal treatment for kidney failure. Interventions at the start of evaluation may improve evaluation completion and LDKT rates. This study tested whether (a) an educational booklet and video (the "Talking About Living Kidney donation" [TALK] intervention) increased evaluation completion and LDKT when delivered under a streamlined KT evaluation program; and (b) if no effects found, explore differential effects by psychosocial/sociocultural factors (e.g., healthcare-related discrimination). MethodsWe conducted a randomized-controlled trial of the TALK intervention using permuted block randomization at an urban transplant center. Participants were enrolled 05/2015-06/2018; follow-up through 08/2022. Staff were blinded to block size, not allocation. Fine-Gray proportional hazards models examined intent-to-treat and per-protocol approaches. Primary outcomes were the cumulative incidence of evaluation completion and LDKT receipt. We explored interaction analyses by psychosocial/sociocultural factors and TALK-assignment. ResultsAmong 1108 participants (574 [52%] TALK, 534 [48%] No-TALK; median age: 59.13 [IQR: 48.92-67.10]; 243 [22%] Black, 783 [71%] White, 82 [7%] Other; 695 [63%] male), TALK did not significantly improve evaluation completion (sub-distribution hazard [SHR]=1.06; 95% CI: 0.92-1.22) or LDKT receipt (SHR=0.83; 95% CI: 0.55-1.25) in intent-to-treat and per-protocol analyses. In exploratory per-protocol analyses, discrimination significantly modified the effect of TALK on evaluation completion (SHR=0.42; 95% CI: 0.29-0.61). The "No-Discrimination" TALK participants had greater evaluation completion than No-TALK (SHR=1.32; 95% CI: 1.10-1.58), but the "Discrimination" TALK participants had lower evaluation completion than No-TALK (SHR=0.56; 95% CI: 0.41-0.77). ConclusionsDespite streamlined care, TALK did not improve evaluation completion or LDKT rates. A significant interaction in the per-protocol analyses for evaluation completion suggests prior healthcare-related discrimination may limit educational intervention effectiveness. Future studies should explore approaches that address systemic barriers and complement, rather than rely on, educational strategies to promote LDKT (ClinicalTrials.gov Identifier: NCT02342119).

BackgroundArtificial intelligence (AI)-based mobile health (mHealth) smartphone apps for skin cancer detection are increasingly available to the general population, but their impact on care is unclear. MethodsThe SPOT study is an investigator-initiated and -designed, unblinded, randomized controlled trial. Participants from a Dutch non-profit health insurance living in and around region Rotterdam the Netherlands, were recruited between August and December 2021. Participants were randomly assigned (3:2) to either free access to an AI-based mHealth app for skin cancer risk detection or care-as-usual. The primary endpoint was the difference in healthcare consumption for (pre)malignant and benign skin lesions at 12-months follow-up in the intention-to-treat population. Secondary endpoints included differences in the proportion of surgical interventions, overall use of dermatological care, and costs. FindingsAmong the 19,009 participants, the incidence of claims for (pre)malignant skin lesions was 2{middle dot}8-fold higher than among non-responders. Within the group of study participants, the skin cancer incidence was higher among the intervention group compared to the control group at 12 months follow-up (2{middle dot}7% (n=305) vs. 2{middle dot}3% (n=171); risk difference (RD) 0{middle dot}4% (95% confidence interval (CI) -0{middle dot}07 to 0{middle dot}85), p = 0{middle dot}10), though this difference was not statistically significant. Furthermore, participants in the intervention group had significantly more claims for benign skin lesions (3{middle dot}9% (n=443) vs. 2{middle dot}6% (n=198), RD 1{middle dot}3 (95% CI 0{middle dot}7 to 1{middle dot}7), p < 0{middle dot}001), underwent more surgical interventions, and had higher mean costs per participant ({euro}63 (95% CI 58 to -67), vs. {euro}47 (41 to -52); p<0{middle dot}001) compared to controls. InterpretationIn the first 12 months of this study, access to an AI-based mHealth app for skin cancer risk detection showed a modest trend toward a higher rate of skin cancer detection compared to care-as-usual. However, it also resulted in significantly more dermatological care for benign skin lesions. FundingDSW and SkinVision(R) Research in context Evidence before this studyPrior to the start of this study, we conducted a PubMed search for articles published between January 1, 2011, and December 31, 2021, using the search terms artificial intelligence AND skin cancer. This search resulted in 809 articles which were screened for relevance. We also included the results of one prospective validation study for which we had conducted the analyses ourselves, but which had not yet been published at that time. Several commercial companies have implemented such algorithms in smartphone-based mobile health (mHealth) applications, making them available to the general public. A systematic review and meta-analysis reported that AI-based apps assessing skin cancer risk from macroscopic images achieved varying sensitivity and specificity, depending on the algorithm and the type of skin cancers detected. Four studies had prospectively validated the specific mHealth app investigated in this study against histopathology, with reported sensitivities ranging from 57% to 87% and specificities from 27% to 83%. Despite promising indications, no randomised controlled trials have yet evaluated the effectiveness of AI-based mHealth apps for skin cancer screening in the general population. Added value of this studyThe SPOT study is, to our knowledge, the first randomised controlled trial to investigate how implementing an AI-based skin cancer risk detection app in the general population affects skin cancer detection and healthcare use for benign skin lesions. We found a modestly higher skin cancer incidence amongst those who were offered to use the app, though this difference was not statistically significant. However we also found those who were offered to use the app had a significantly larger increase in healthcare visits and procedures for benign skin lesions. Suggesting that implementation in the general population may involve a possible trade-off between increased skin cancer detection and unnecessary care due to overdiagnosis. Implications of all the available evidenceEven though research in a sterile setting shows potential for implementation of AI-based mHealth apps, the results from this study suggests that nationwide implementation of an mHealth with its current accuracy is not the most optimal strategy. Targeted implementation in higher-risk populations may offer a more favourable balance between benefits and harms.

ImportancePrior studies using large registries suggested a modest increase in risk for neurodevelopmental diagnoses among children of mothers with immune activation during pregnancy, and such risk may be sex-specific. ObjectiveTo determine whether in utero exposure to the novel coronavirus SARS-CoV-2 is associated with sex-specific risk for neurodevelopmental disorders up to 18 months after birth, compared to unexposed offspring born during or prior to the pandemic period. DesignRetrospective cohort. ParticipantsLive offspring of all mothers who delivered between March 2018 and May 2021 at any of eight hospitals across two health systems in Massachusetts. ExposurePCR evidence of maternal SARS-CoV-2 infection during pregnancy. Main Outcome and MeasuresElectronic health record documentation of ICD-10 diagnostic codes corresponding to neurodevelopmental disorders. ResultsThe pandemic cohort included 18,323 live births, including 877 (4.8%) to individuals with SARS-CoV-2 positivity during pregnancy. The cohort included 1806 (9.9%) Asian individuals, 1634 (8.9%) Black individuals, 1711 (9.3%) individuals of another race, and 12,694 (69%) White individuals; 2614 (14%) were of Hispanic ethnicity. Mean maternal age was 33.0 years (IQR 30.0-36.0). In adjusted regression models accounting for race, ethnicity, insurance status, hospital type (academic center vs. community), maternal age, and preterm status, SARS-CoV-2 positivity was associated with statistically significant elevation in risk for neurodevelopmental diagnoses among male offspring (adjusted OR 1.99, 95% CI 1.19-3.34; p=0.009) but not female offspring (adjusted OR 0.90, 95% CI 0.43-1.88; p=0.8). Similar effects were identified using matched analyses in lieu of regression. Conclusion and RelevanceSARS-CoV-2 exposure in utero was associated with greater magnitude of risk for neurodevelopmental diagnoses among male offspring in the 12 months following birth. As with prior studies of maternal infection, substantially larger cohorts and longer follow-up will be required to reliably estimate or refute risk. Trial RegistrationNA Key PointsO_ST_ABSQuestionC_ST_ABSAre rates of neurodevelopmental disorder diagnoses greater among male or female children with COVID-19 exposure in utero compared to those with no such exposure? FindingsIn a cohort of 18,323 infants delivered after February 2020, males but not females born to mothers with a positive SARS-CoV-2 PCR test during pregnancy were more likely to receive a neurodevelopmental diagnosis in the first 12 months after delivery, even after accounting for preterm delivery. MeaningThese findings suggest that male offspring exposed to COVID-19 in utero may be at increased risk for neurodevelopmental disorders.

ImportanceReopening of universities in the U.S. has been controversial in the setting of the coronavirus disease 2019 (COVID-19) pandemic. ObjectiveTo investigate (1) the association between new COVID-19 cases since September 1st with the number of students returning to campus in each county across the U.S. and (2) how different reopening policies at universities correlated with new COVID-19 cases. DesignObservational cohort study using publicly available data sources. Multivariable regression models estimated both effects of university reopening and different reopening policies. Settings and ParticipantsPopulations in U.S. counties reporting new confirmed COVID-19 cases from August 1st to October 22nd. Exposures(1) total enrollment of students under the in-person or hybrid policies per county population and (2) proportion of online and hybrid enrollment within each county. Main Outcomes and MeasuresMean number of daily new confirmed COVID-19 cases per 10,000 county population from September 1st to October 22nd. ResultsFor 2,893 counties included in the study, mean number of daily confirmed cases per 10,000 county population rose from 1.51 from August 1st to August 31st to 1.98 from September 1st to October 22nd. Mean number of students returning to universities was 2.1% (95% CI, 1.8% to 2.3%) of the county population. The number of students returning to campus had an increased association ({beta} = 2.006, P < 0.001) with new confirmed COVID-19 cases within the local county region where the institution resided. For 1,069 U.S. counties with universities, the mean proportion of online enrollment within each county was 40.1% (95% CI, 37.4% to 42.8%), with most students enrolling in-person or hybrid mode. In comparison to holding classes in-person, reopening universities online ({beta} = -0.329, P < 0.001) or in a hybrid mode ({beta} = -0.272, P = 0.012) had a decreased association with new confirmed COVID-19 cases. Conclusions and RelevanceA higher number of students returning to campus in U.S. counties was associated with an increase in new confirmed COVID-19 cases; reopening online or partially online was associated with slower spread of the virus, in comparison to in-person reopening. Key PointsO_ST_ABSQuestionC_ST_ABSAre students returning to universities and specific reopening policies associated with new confirmed coronavirus cases in United States? FindingsIn this cohort study of 2,893 U.S. counties, the number of students returning to campus was significantly associated with a higher number of new confirmed COVID-19 cases. In 1,069 U.S. counties with universities, online or hybrid reopening was significantly associated with a lower risk of new cases compared with in-person reopening. MeaningAn increased risk of coronavirus infection was seen in surrounding regions after universities reopened last fall, and this effect was largest in those holding in-person classes.

BackgroundBetween 1979 and 2016, US overdose death rates rose in a smooth fashion, described by Jalal and Burke using an exponential growth curve that fit observed data nearly perfectly. Fluctuations above this curve have subsequently been seen during shocks related to drug supply and the COVID-19 pandemic. However, large-magnitude dips below the curve have never been demonstrated. Given that overdose mortality began sharply falling during 2023-2024, we assess updated overdose trends against the Jalal-Burke curve. MethodsWe examined US overdose deaths from the National Vital Statistics System between January 1979-December 2024. We recreated the Jalal-Burke curve, fitting an exponential growth curve to overdose rates from 1979 to 2016, linearly projecting through 2024, with 95% confidence intervals. We also examined trends by specific substance involvement. ResultsAfter precipitously surpassing exponential growth predictions in 2020-2023, overdose deaths decreased sharply from approximately 32 per 100,000 in 2021-2023 to 23.7 in 2024, falling below the lower bound of Jalal-Burke curve (24.98 per 100,000) for the first time since 2001. These decreases reflected declining illicit fentanyl-involved deaths (with and without stimulants); however, deaths involving stimulants without fentanyl, and those involving xylazine, represent an increasing share of deaths in 2024. ConclusionsRather than simply representing a return to the Jalal-Burke exponential growth curve, recent decreases in overdose deaths represent the first significant, large-magnitude deviation below exponential growth projections. These trends represent a very positive development. However, challenges in the US drug crisis are shifting, requiring a tailored public health response.

BackgroundCancer patients treatment is often dispersed across multiple physician communities that may not exchange information. We measured the association between this type of discoordinated care and healthcare utilization, cost, and mortality in stage IV non-small cell lung cancer (NSCLC). MethodsStage IV NSCLC patients [&ge;]65 years were identified from the Surveillance, Epidemiology, and End Results database attached to Medicare claims. A discoordinated care subgroup was selected using a novel index that approximated how many physician communities each patient encountered. Communities were defined by (1) using claims data to create a nationwide social network of healthcare providers and then (2) dividing that network into communities using cluster label propagation. Associations between discoordinated care and overall survival (OS), cancer-specific survival (CSS), hospitalizations, the burden of diagnostic imaging, and cost were assessed. ResultsOf the 11,417 patients in our cohort, 5,855 received discoordinated care. Discoordination was associated with younger age, higher socioeconomic status, higher physician density, and lack of a home health aide. Discoordinated care was associated with improved OS and CSS (HR = 0.92, 95% CI 0.88 - 0.95 for OS). However, discoordinated patients also received 32% more MRIs (p = 0.007) and paid $494.02 more for imaging (p = 0.004). There was no association with other kinds of imaging, rates of hospitalization, or other healthcare costs, including total cost. ConclusionsDiscoordinated care was associated with additional MRIs, but also improved survival. The reason is unclear, but discoordinated patients may be seeking the best care at the expense of continuity. Key PointsO_ST_ABSQuestionC_ST_ABSHow does dispersing healthcare across multiple physician communities impact healthcare utilization and survival in patients with advanced stage non-small cell lung cancer? FindingsIn this retrospective cohort study, we found that discoordinated care was associated with increased utilization of MRIs and total cost of imaging. Surprisingly, it was also associated with improved survival. MeaningCancer patients with dispersed healthcare may be seeking care through a tertiary care center or clinical trial. This may lead to increased healthcare utilization but also improved survival.

AimsTo characterize decreases in overdose mortality in the United States between 2023 and 2024 by substance involvement, geography, race/ethnicity, demographic, and other key dimensions. DesignPopulation-based study of national death records. SettingUnited States. Participants/casesAll individuals who died from drug overdose between January 1999 and December 2024. MeasurementsAnnual or monthly (annualized) overdose deaths per 100,000 population. Year and month of occurrence of overdose death; substance involvement; census region and division; state; county; race/ethnicity, age, and sex. FindingsAfter over two decades of mostly exponential increases, monthly data show consistent decreases in overdose deaths between June 2023 and December 2024. Decreases reflected declining illicit fentanyl-involved deaths (with and without stimulants); however, increasing trends through 2024 were still seen in deaths involving stimulants without fentanyl, and those involving xylazine. Death rates in the Northeast, South and Midwest fell to 19.5, 19.4 and 17.3 per 100,000, respectively, in December 2024, but remained elevated in the West, compared with other regions, at 27.2 per 100,000. Non-Hispanic Black and African Americans had the largest decrease in death rates in 2023-2024 falling 29.3%, but remained elevated at 36.0 per 100,000, compared to the national average of 23.7 per 100,000. Non-Hispanic American Indian and Alaska Native individuals had the highest overdose mortality rate in 2024, at 50.8 per 100,000. ConclusionsRecent decreases in overdose deaths are encouraging and unprecedented. Racial gaps remained large but shrunk by a modest margin. The geography of the overdose crisis has shifted, with the West now the most affected region, which may have implications for the targeting of funding. The nature of the crisis is also shifting, as stimulants and xylazine continue to represent increasingly important public health challenges, and renewed attention to nonfatal aspects of addiction in the US is needed.

BackgroundMillions of primary school students across the United States are about to return to in-person learning. Amidst circulation of the highly infectious Delta variant, there is danger that without the appropriate safety precautions, substantial amount of school-based spread of COVID-19 may occur. MethodsWe used an extended Susceptible-Infected-Recovered computational model to estimate the number of new infections during 1 semester among a student population under different assumptions about mask usage, routine testing, and levels of incoming protection. Our analysis considers three levels of incoming protection (30%, 40%, or 50%; denoted as "low", "mid", or "high"). Universal mask usage decreases infectivity by 50%, and weekly testing may occur among 50% of the student population; positive tests prompt quarantine until recovery, with compliance contingent on symptom status. ResultsWithout masking and testing, more than 75% of susceptible students become get infected within three months in all settings. With masking, this values decreases to 50% for "low" incoming protection settings ("mid"=35%, "high"=24%). Testing half the masked population ("testing") further drops infections to 22% (16%, 13%). ConclusionWithout interventions in place, the vast majority of susceptible students will become infected through the semester. Universal masking can reduce student infections by 26-78%, and biweekly testing along with masking reduces infections by another 50%. To prevent new infections in the community, limit school absences, and maintain in-person learning, interventions such as masking and testing must be implemented widely, especially among elementary school settings in which children are not yet eligible for the vaccine.

BackgroundOut-of-hospital cardiac arrest (OHCA) accounts for over 350,000 deaths annually in the United States, and survival depends on early bystander cardiopulmonary resuscitation (CPR). Although many cardiac arrests occur on or near school grounds, Massachusetts has no statewide CPR graduation requirement and little current data on school preparedness. MethodsWe conducted a cross-sectional electronic survey of all 413 public high schools in Massachusetts (including charter, vocational, and technical) between September 29 and November 17, 2025. The 14-item survey asked about enrollment, staff size, CPR and automated external defibrillator (AED) resources, student and staff training, and the presence of a cardiac emergency response plan (CERP). The CERP item was excluded from analysis due to inconsistent interpretation. We summarized resources per 1,000 students or staff, compared Title I and non-Title I schools, and explored geographic variation and multivariable predictors of AED availability and CPR teaching. ResultsOne hundred schools responded (24.2%), representing 13 of 14 counties, with a median enrollment of 662 students, and 33.0% were Title I schools. Overall, 72.0% reported teaching CPR. The median student training rate was 138.2 per 1,000 students, though only 10.0% of schools with non-missing data reported all students trained and 15.0% reported at least 70% of students trained. Among responding schools, Title I schools had fewer trained students than non-Title I schools (median 50.3 vs 199.8 per 1,000; Holm-Bonferroni adjusted p = 0.025), despite similar AED and manikin availability. This disparity persisted in sensitivity analyses using median imputation for missing student training data. County-level analyses suggested geographic variation in both training rates and AED density, although county-level estimates were based on small numbers of responding schools. ConclusionsAmong responding Massachusetts high schools, most reported some CPR instruction, but only a small minority achieved broad student coverage, with particularly low training rates in Title I schools. These exploratory findings underscore the need for policies ensuring equitable CPR training access, particularly in Title I schools, and support targeted investment in school-based cardiac emergency preparedness