Category: 3. Business

The sequential intracerebroventricular (ICV) and intraventricular (IV) administration of CR7-GD2 CAR T-cell therapy was found to be better tolerated vs concurrent administration in pediatric patients with central nervous system (CNS) tumors, according to data from a phase 1 study (NCT04099797) presented at the 2025 Society of Neuro-Oncology Annual Meeting.¹

During the study, investigators at Baylor College of Medicine in Houston, Texas, evaluated 3 administration techniques for the CR7-GD2 CAR T-cell therapy: IV only; concurrent ICV and IV; and sequential ICV and IV.

Findings showed that in patients treated with IV-only administration (n = 8) at 10 million cells/m² (n = 3) or 30 million cells/m² (n = 6), no dose-limiting toxicities (DLTs) were reported, and instances of tumor inflammation–associated neurotoxicity (TIAN) were low grade, occurring in 7 patients. One patient experienced transient grade 4 cytokine release syndrome (CRS); 5 experienced grade 1 CRS. Among patients treated via IV alone who had preexisting neurological defects (n = 7), 6 achieved clinical improvement, including 2 partial responses on imaging, which 1 patient maintained for more than 2 years.

ICV/IV Administration of CAR T-Cell Therapy in Pediatric CNS Tumors

• Sequential ICV and IV administration was found to be better tolerated compared with combined ICV and IV treatment.

• Two DLTs were reported in the combined administration group vs none in patients given sequential treatment.

• In the sequential cohort, patients experienced only low-grade TIAN/CRS and had clinical stability.

In patients who received concurrent ICV/IV administration (n = 3), with doses given 1 week apart, 2 experienced DLTs secondary to grade 3 TIAN, including 1 patient with concurrent grade 3 CRS. These DLTs persisted for several weeks and required intervention with anakinra and dexamethasone. In this group, 1 patient achieved a mixed response with transient clinical improvement and stable disease burden. The remaining 2 patients had clinical and radiographic progression.

The neurotoxicity observed with concurrent administration prompted investigators to adopt a sequential strategy, with ICV administration given 4 weeks after IV treatment. In patients treated with this strategy (n = 3), treatment was well tolerated for at least 3 cycles each, with no DLTs reported; these patients experienced low-grade TIAN/CRS and clinical stability.

“While back-to-back ICV/IV dosing resulted in increased toxicity compared [with] IV-only therapy, sequential dosing was better tolerated, potentially indicative of the negative sequelae of compounded inflammation,” lead study author Jasia Mahdi, MD, and colleagues wrote in a poster presentation of the data. Mahdi is a pediatric neurologist and neuro-oncologist, and director of Neuro-Oncology for the Division of Pediatric Neurology at Texas Children’s Hospital, as well as an assistant professor of pediatrics-neurology at Baylor College of Medicine.

How was the phase 1 study conducted?

The single-center trial enrolled 2 cohorts of patients between 12 months and 22 years of age.² The first included patients with histologically confirmed, GD2-expressing or H3K27-altered, diffuse midline glioma (DMG) or high-grade glioma (HGG) that was newly diagnosed, defined as prior to radiographic progression or recurrence; patients with histologically confirmed, GD2-expressing or H3K27-altered, recurrent, refractory, or progressive DMG/HGG; and those with recurrent, refractory, or progressive high-grade CNS tumors with confirmed GD2-expression. Cohort 2 included patients with recurrent, refractory, or progressive pontine HGG with confirmed GD2-expression or H3K27-altered DMG.

All patients needed to have tumors less than 5 cm in maximum dimension, measurable disease on at least 2 dimensions per MRI, and a Karnofsky/Lansky performance status of at least 50.

The C7R-GD2 CAR T-cell therapy was administered after lymphodepleting chemotherapy in all arms and cohorts, comprising 2 days of cyclophosphamide and 3 days of fludarabine.¹ In patients who received combined dosing, the CAR T-cell therapy was given at 5 x 10⁶ cells/m² ICV, followed by 15 x 10⁶ cells/m² IV, then another IV dose at the same level. In the sequential cohort, patients received an IV dose at 10 million cells/m², followed by an ICV dosing in cycle 2 and beyond, starting at 2 million cells and escalating to 5 million cells.

In both ICV arms, patients remained for inpatient monitoring for 6 to 10 days, followed by close outpatient follow-up. Disease evaluation occurred at week 6.

The incidence of DLTs was the trial’s primary end point.² Response rate was a secondary end point.

In the combined ICV/IV arm, 1 male patient and 2 female patients had a median age of 15 years (range, 4-17).¹ Tumor locations included thalamus and midbrain (n = 1), pons (n = 1), and C1-C7/T1 (n = 1). All had H3K27-altered DMG. Patients had a median time since diagnosis before enrollment of 9.4 months, and they had a median time from the end of radiation to enrollment of 6.5 months. These patients received a median of 1 ICV infusion.

In the sequential cohort, all 3 patients were male with a median age of 16 years (range, 14-17). Tumor locations comprised the thalamus (n = 2) and the temporal lobe (n = 1). Tumor types included H3K27-altered DMG (n = 2) and H3 wild-type, IDH wild-type diffuse HGG (n = 1). Patients had a median time from diagnosis to enrollment of 7.4 months, and a median time from the end of radiation to enrollment of 2.3 months. This group received a median of 3 ICV infusions.

What were the toxicity profiles of combined and sequential ICV/IV CAR T-cell therapy administration?

In the combined ICV/IV arm, adverse effects included CRS (grade 1, n = 2 infusions; grade 3, n = 1 infusion), immune effector cell–associated neurotoxicity syndrome (ICANS; grade 0, n = 3 infusions), and TIAN (grade 1, n = 1 infusion; grade 3, n = 2 infusions). Two patients received tocilizumab (Actemra), and all 3 were given dexamethasone, which continued upon hospital discharge.

In the sequential group, toxicities included CRS (grade 0, n = 13 infusions; grade 1, n = 3 infusions), ICANS (grade 0, n = 16), and TIAN (grade 0, n = 2 infusions; grade 1, n = 14 infusions). All patients required tocilizumab, but no patients received dexamethasone.

What was reported from a trial case study?

Investigators highlighted a 16-year-old male patient with H3K27-altered thalamic and periventricular occipital lobe DMG who underwent 9 cycles of treatment in the sequential ICV/IV arm, including 1 IV infusion and 8 ICV infusions. This patient remained on treatment as of data cutoff, and he has experienced CRS and TIAN at a maximum grade of 1. MRI revealed stable disease, and the patient was also stable on a neurological exam.

References

1. Mahdi J, Stuckert A, Tat C, et al. Phase I study of intravenous and intracerebroventricular C7R-GD2.CAR T cell therapy for pediatric central nervous system (CNS) tumors. Presented at 2025 Society of Neuro-Oncology Annual Meeting; November 19-23, 2025; Honolulu, HI. Abstract CTP-07.
2. C7R-GD2.CAR T cells for patients with GD2-expressing brain tumors (GAIL-B). ClinicalTrials.gov. Updated September 4, 2025. Accessed November 23, 2025. https://clinicaltrials.gov/study/NCT04099797

News Highlights:

• Zyphra ZAYA1 becomes the first large-scale Mixture-of-Experts model trained entirely on AMD Instinct™ MI300X GPUs, AMD Pensando™ networking and ROCm open software.
• ZAYA1-base outperforms Llama-3-8B and OLMoE across multiple benchmarks and rivals the performance of Qwen3-4B and Gemma3-12B.
• Memory capacity of AMD Instinct MI300X helped Zyphra simplify its training capabilities, while achieving 10x faster model save times.

SANTA CLARA, Calif., Nov. 24, 2025 (GLOBE NEWSWIRE) — AMD (NASDAQ: AMD) announced that Zyphra has achieved a major milestone in large-scale AI model training with the development of ZAYA1, the first large-scale Mixture-of-Experts (MoE) foundation model trained using an AMD GPU and networking platform. Using AMD Instinct™ MI300X GPUs and AMD Pensando™ networking and enabled by the AMD ROCm™ open software stack, the achievement is detailed in a Zyphra technical report published today.

Results from Zyphra show that the model delivers competitive or superior performance to leading open models across reasoning, mathematics, and coding benchmarks—demonstrating the scalability and efficiency of AMD Instinct GPUs for production-scale AI workloads.

“AMD leadership in accelerated computing is empowering innovators like Zyphra to push the boundaries of what’s possible in AI,” said Emad Barsoum, corporate vice president of AI and engineering, Artificial Intelligence Group, AMD. “This milestone showcases the power and flexibility of AMD Instinct GPUs and Pensando networking for training complex, large-scale models.”

“Efficiency has always been a core guiding principle at Zyphra. It shapes how we design model architectures, develop algorithms for training and inference, and choose the hardware with the best price-performance to deliver frontier intelligence to our customers,” said Krithik Puthalath, CEO of Zyphra. “ZAYA1 reflects this philosophy and we are thrilled to be the first company to demonstrate large-scale training on an AMD platform. Our results highlight the power of co-designing model architectures with silicon and systems, and we’re excited to deepen our collaboration with AMD and IBM as we build the next generation of advanced multimodal foundation models.”

Efficient Training at Scale, Powered by AMD Instinct GPUs
The AMD Instinct MI300X GPU’s 192 GB of high-bandwidth memory enabled efficient large-scale training, avoiding costly expert or tensor sharding, which reduced complexity and improving throughput across the full model stack. Zyphra also reported more than 10x faster model save times using AMD optimized distributed I/O, further enhancing training reliability and efficiency. With only a fraction of the active parameters, ZAYA1-Base (8.3B total, 760M active) matches or exceeds the performance of models such as Qwen3-4B (Alibaba), Gemma3-12B (Google), Llama-3-8B (Meta), and OLMoE.¹

Building on prior collaborative work, Zyphra worked closely with AMD and IBM to design and deploy a large-scale training cluster powered by AMD Instinct™ GPUs with AMD Pensando™ networking interconnect. The jointly engineered AMD and IBM system, announced earlier this quarter, combines AMD Instinct™ MI300X GPUs with IBM Cloud’s high-performance fabric and storage architecture, providing the foundation for ZAYA1’s large-scale pretraining.

For further details on the results, read the Zyphra technical report, the Zyphra blog, and the AMD blog, for comprehensive overviews of the ZAYA1 model architecture, training methodology, and the AMD technologies that enabled its development.

Supporting Resources

About AMD
For more than 50 years AMD has driven innovation in high-performance computing, graphics, and visualization technologies. Billions of people, leading Fortune 500 businesses, and cutting-edge scientific research institutions around the world rely on AMD technology daily to improve how they live, work, and play. AMD employees are focused on building leadership high-performance and adaptive products that push the boundaries of what is possible. For more information about how AMD is enabling today and inspiring tomorrow, visit the AMD (NASDAQ: AMD) website, blog, LinkedIn, and X pages.

Contact:
David Szabados
 AMD Communications
+1 408-472-2439
david.szabados@amd.com

Liz Stine
AMD Investor Relations
+1 720-652-3965 
liz.stine@amd.com

_________________________
¹ Testing by Zyphra as of November 14, 2025, measuring the aggregate throughput of training iterations across the full Zyphra cluster measured in quadrillion floating point operations per second (PFLOPs). The workload was training a model comprised of a set of subsequent MLPs in BFLOAT16 across the full cluster of (128) compute nodes, each containing (8) AMD Instinct™ MI300X GPUs and (8) Pensando™ Pollara 400 Interconnects running a proprietary training stack created by Zyphra. Server manufacturers may vary configurations, yielding different results. Performance may vary based on use of the latest drivers and optimizations. This benchmark was collected with AMD ROCm 6.4.

Source: Advanced Micro Devices, Inc.

Released November 24, 2025

Intuit SMB MediaLabs’ latest expansion enables more advertisers to reach small and mid-market businesses with greater precision and relevance

MOUNTAIN VIEW, Calif.–(BUSINESS WIRE)–
Intuit Inc., the global financial technology platform that makes TurboTax, Credit Karma, QuickBooks, and Mailchimp, today announced that its SMB MediaLabs audiences are now available on The Trade Desk, providing advertisers with access to Intuit’s first-party small and mid-market business (SMB) audience segments. The partnership facilitates advertisers’ access to Intuit’s SMB MediaLabs, a first-of-its-kind advertising network powered by Intuit’s unmatched first-party business data, through The Trade Desk platform. With this integration, advertisers can seamlessly activate Intuit’s unique small and mid-market business insights to reach SMBs with greater precision and at scale, providing them with highly relevant advertising that connects them with products and services that can help optimize and grow their business. Advertisers will continue to abide by Intuit’s Advertising Guidelines, allowing campaigns to deliver value to Intuit’s SMB customers while adhering to responsible, privacy-conscious standards.

Small businesses make up 99% of companies in the U.S.¹, making the owners and decision-makers of these businesses a high-value market segment with significant spending power. However, this audience has historically been difficult for advertisers to reach accurately, with campaigns often reliant on fragmented or outdated third-party data, resulting in both wasted ad spend for brands and irrelevant experiences for SMBs. The launch of this integration significantly changes this dynamic, making Intuit’s network of unique audiences, which spans millions of SMBs, more readily accessible to advertisers on The Trade Desk platform. Using aggregated, de-identified insights from the Intuit platform, advertisers can more efficiently connect with verified SMB decision-makers, helping brands improve campaign performance while also delivering more relevant ad experiences for SMBs, to help their businesses grow and thrive.

“This partnership marks a fundamental shift in how B2B marketers will be able to engage small and mid-market businesses. For too long, the industry has struggled with accuracy and relevance in targeting the SMB audience—a critical gap we are now closing,” said Christopher Moneta, Director, SMB MediaLabs, Intuit. “By fusing Intuit’s unique, deterministic SMB insights with the powerful execution capabilities of The Trade Desk, we are setting a new standard. Brands can now confidently deliver highly relevant advertising that reaches the right decision-makers across every channel, while SMBs can more easily discover the products and solutions they need to succeed.”

The Trade Desk is the latest demand-side platform (DSP) to directly partner with the Intuit SMB MediaLabs network, and the first DSP where this first-party SMB data will be discoverable for advertisers, providing them with efficient campaign management capabilities and improved cross-channel measurement. Available through the SMB MediaLabs self-service offerings as an endpoint on the LiveRamp Data Marketplace, the integration also significantly expands the reach of Intuit’s SMB MediaLabs across connected TV, audio, display, and digital out-of-home channels.

“As the first media-buying platform to bring Intuit’s SMB MediaLabs audiences to market discoverably, we’re giving advertisers direct access to one of the most trusted sources of small business intelligence,” said Matthew Fantazier, VP, Data Partnerships, The Trade Desk. “This partnership enables us to help brands connect their messages to real decision-makers, with more precision, transparency, and scale.”

Launched in 2023, Intuit’s SMB MediaLabs allows advertisers to create targeted campaigns that have the potential to reach millions of small and mid-market businesses. Access to this key audience has proven to be game changer for marketers seeking to connect with decision-makers and drive measurable business outcomes.

For more information, visit medialabs.intuit.com and thetradedesk.com.

About Intuit

Intuit is the global financial technology platform that powers prosperity for the people and communities we serve. With approximately 100 million customers worldwide using products such as TurboTax, Credit Karma, QuickBooks, and Mailchimp, we believe that everyone should have the opportunity to prosper. We never stop working to find new, innovative ways to make that possible. Please visit us at Intuit.com and find us on social for the latest information about Intuit and our products and services.

¹ According to a report from the U.S. Small Business Administration, published in July 2024.

View source version on businesswire.com: https://www.businesswire.com/news/home/20251124125255/en/

Intuit Media Contact:

Jaymie Sinlao

jaymie_sinlao@intuit.com

Source: Intuit Inc.

Released November 24, 2025

Proving the durability of Porsche’s first-generation E1 Cayenne, Moussaoui set off with a simple plan: enter Kuwait into his GPS, pack a bag, and hit the road. What followed was an extraordinary four-month journey that showcased the resilience of a showroom-spec Cayenne.

The 22-year-old adventurer covered 28 countries and more than 34,000 km in his stock Cayenne V6 – a car that had already clocked 276,000 km before the trip began. “The whole point of my trip was to do it with a completely stock car,” Moussaoui explained.

Unmodified yet unstoppable: The Cayenne’s true test

Aside from upgrading the headlights with OEM Porsche parts, the Cayenne remained untouched – running on factory 21-inch wheels with low-profile tyres. Despite the risk of potholes, Moussaoui was never stranded.
 

Speaking at Icons of Porsche in Dubai Design District, Moussaoui shared his motivation: a lifelong passion for Porsche. “Porsche is my forever brand. I wouldn’t have chosen another car for this trip. My parents put me in a 911 (993) when I left the hospital as a newborn – my dad said there was just enough room in the back.”

The only maintenance required for the four-month journey that took him from the UAE deserts to the snow fields of Switzerland was regular oil changes and new spark plugs in London.

“Oil changes are the cheapest insurance to ensure the engine stays strong. It’s the ultimate testament to its durability and it didn’t skip a beat.”

Across deserts and mountains: The route to Icons of Porsche

From Dubai, Mateo drove to Kuwait via Saudi Arabia. From there he continued on to Iraq, Turkey and across to Europe crossing through England, Scotland and Wales.

“The most challenging part of the journey was negotiating the border entry from Kuwait into Iraq around Basra, which was an experience. I spent four hours at the border and there were 20 kiosks for documents and fees.”

After its return to Dubai, Mateo gave it another oil change but it’s still running on the same tyres after giving them a rotation in Monaco at the 13,000 km mark.

Despite conquering Saudi Arabian deserts, jagged mountains of Iraq, snow-capped peaks of Italy, fording lakes in Switzerland and cruising German Autobahns, Mateo decided his E1 Cayenne needed to complete a lap of the legendary Nürburgring Nordschleife.

 

“It’s insane to think that after completing this massive loop of the world in my Cayenne with over 300,000 km on it, the 20.8 km of the Nürburgring Nordschleife were still the toughest kilometres that car has ever done. But again, it survived without a problem.”

What’s next? Adventures still to come

After some rest for both car and driver, a new expedition is already in the pipeline with Mateo planning to take his Cayenne to visit Jordan, Syria and Lebanon as well as a few more Scandinavian countries and later plans to drive it through Asia, America and Africa.

“I’ve covered 28 countries without a hitch in a car that is absolutely as standard as the day it left the factory, so why would I stop now?”

Study design

This study utilized PC and PHE methods, which are comparable to those used in GBD 2010 and European surveys measuring adult DWs [17, 18]. However, we differed in our survey approach and survey instruments. We used a paper version of the questionnaire, tailored to the situation, as the survey instrument instead of an electronic questionnaire. Furthermore, we conducted a face-to-face survey instead of an online survey from March 2021 to October 2022.

Research setting and participants

The primary target of the sample for this study was people who had some knowledge or awareness of a particular health state in children. Participants were preferably close contacts of children or children themselves. Due to the face-to-face nature of the survey, the large amount of human, material, and financial resources required, and the limited time available due to the epidemic control during the survey period, the respondents of this study were selected from the parents of children in the birth cohort already established by the research group; the parents of children attending and being hospitalised in Hunan Provincial Children’s Hospital, Xiangya Hospital, Xiangya No. 2 Hospital, Xiangya No. 3 Hospital; and the paediatricians of the community and the general hospitals; The general population living in other urban areas of Hunan Province and the administrative districts of Changsha, as well as university students.

The inclusion criteria for the respondents were: (1) be 18 years of age or older; (2) possess normal intelligence, a certain level of literacy, and the ability to comprehend the questionnaire; and (3) have an understanding of the content and purpose of the study, agree to participate, and voluntarily sign an informed consent form. Exclusion criteria include: (1) participants who did not meet the inclusion criteria; (2) incomplete questionnaire responses; and (3) participants who refused to cooperate or did not sign the informed consent form.

The sample size was estimated in consultation with experts in the field of statistics and computer simulations revealed that pairwise comparisons, with more than five comparisons using Probit regression analysis, were able to identify differences, and we also referred to the published literature [19], where 206 illnesses and injuries were included in face-to-face surveys of 5,750 people. The number of disease and injury categories investigated in our study was 148. The number of pairs for two-by-two comparison is 148*147/2 = 10,878, each questionnaire in this study incorporates 16 PC method pairs, a total of 10,878/16 = 680 different questionnaires are needed for one round of survey, each questionnaire needs to be completed by one person independently, 8 rounds of survey are planned in this study, the sample size to be surveyed is 8*680 = 5440, the PC method of our study investigates a total of 5455 respondents in 8 rounds of survey.

Since it was difficult for us to access high schools or middle schools to survey children under the age of 18 in person. Therefore, we surveyed a random sample of students in colleges and universities to make it more representative of the health preferences of this age group. Our survey involved health state descriptions and the anchoring tool, the PHE method, was more difficult to understand, requiring respondents to have some knowledge and equivalent measures of a certain health state, so we surveyed almost exclusively children’s parents in hospitals and neighbourhoods and required parents to have a certain level of cognition, resulting in an overall high level of education in the included population.

Determination of health statuses

After reviewing relevant literature, we identified 148 childhood health statuses based on the disease spectrum of Chinese children’s outpatient and inpatient services, the Global Burden of Childhood Diseases list, the WHO Children’s Disease Statistical List, and major childhood health statuses gained through interviews with children’s parents, as well as the disease spectrum of pediatric outpatient and inpatient services at comprehensive tertiary hospitals and children’s specialty hospitals (See Supplementary documents 1). The text describes six categories of children’s health statuses: birth defects and congenital disability diseases (24), acute infectious disease (31), chronic diseases and injuries (34), accidental injuries (36), mental and behavioral disorders diseases (14), and malignant neoplasm diseases (9). (See Supplement Table S1).

Lay description of health statuses

The principles for describing children’s health statuses are to use concise, non-clinical vocabulary, to highlight the main functional consequences and symptoms associated with the health statuses, and to keep the description to 50 words or less. The same health statuses assessed in adult DWs were identified by referring to lay descriptions of adult health statuses and incorporating them for children [20,21,22]. In the first phase of lay disease descriptions, we measured the functional and symptomatic dimensions of the 148 childhood health statuses included, using the “International Classification of Functioning, Disability, and Health, (ICF)” (https://apps.who.int/iris/bitstream/handle/10665/42407/9241545429.pdf?sequence=1) assessment scale (See Supplement Table S2). This helps to characterise the specific health statuses of these manifestations. Where possible, descriptions were determined based on standard clinical professional classification systems to accurately reflect the severity of a particular condition. The research team extensively discussed and revised the functional health and symptom presentation of the 148 childhood health statuses before finalising a preliminary version. Pediatric experts and doctors from community health service centers were consulted to review and modify the preliminary textual descriptions. This was done to ensure that they accurately reflected the characteristics, common presentations, and duration of associated symptoms involving the sequelae of impaired functioning. Finally, we obtained versions of the textual descriptions of childhood health statuses suitable for use in general population surveys using the PC and PHE methods (See Supplement Table S1).

Health status valuation

A comprehensive evaluation of the 148 children’s health statuses was conducted using the PC and PHE methodologies. The PC technique is an ordinal measure that assesses relative differences in individual functioning and health by comparing pairs of children’s health statuses. It also captures the assessor’s preferred choices for these health statuses. The PHE technique is a group health benefit transformation method that requires the assessor to retrospectively assess two hypothetical health items. The first health item is to prevent 1,000 people from developing a disease that leads to rapid death, while the second health item is to prevent 1,500, 2,000, 3,000, 5,000, or 10,000 people (based on randomly selected bids for each question) from developing a disease that is not fatal (i.e., one of the 148 health statuses for children) but would experience the symptoms and durations mentioned in the descriptions. Evaluators are asked to choose which health program they believe produces greater overall population health benefits. The PHE method is utilised for the purpose of evaluating and comparing health statuses affecting entire populations. This is achieved by estimating the propensity for children to experience a loss of welfare due to different health statuses, and subsequently translating this into an equivalent value of welfare loss.

The 148 childhood diseases were first paired two-by-two, for a total of 148*147/2 = 10,878 pairs. We arranged 16 pairs and 3 PHE questions per questionnaire, each questionnaire needed pairs without put back randomly selected from 10,878 pairs and 3 PHE questions without put back randomly selected from 148 childhood diseases until all pairs were included in the questionnaire. Each round of the survey required the completion of 10,878/16 = 680 different questionnaires (See Supplementary documents 2).

Data collection procedures and instruments

In this study, a paper-based questionnaire served as the primary survey instrument. The questionnaire was developed by the research team in consultation with experts and was finalized based on the pre-survey results. The questionnaire mainly consisted of basic socio-demographic information of the respondents (e.g., age, gender, usual address, type of household registration, marital status, education level, annual household income level, type of occupation, presence of children in the household, presence of medical background, etc.), as well as 16 PC questions and 3 PHE questions.

Trained investigators conducted face-to-face interviews. Respondents were informed of the survey’s purpose and questionnaire content and asked to sign an informed consent form. The enumerator supervised respondents while they completed the questionnaire and answered any questions. The questionnaire was completed in full. Respondents did not receive payment for their participation in the survey. The investigator prompted respondents to imagine two children with the disease and to weigh who was healthier between the two. Two investigators worked in pairs to enter the completed questionnaires into a pre-developed computer program.

Statistical analysis

The statistical analyses for this study were conducted using R software (version 4.3.1), Stata MP software (version 17.0), and Microsoft Excel 2021. Probit regression models were used to estimate relative outcomes for childhood health statuses based on pooled PC data. The result reflects the relative differences in severity between childhood health statuses on a quantitative scale, and also shows participants’ choice preferences for each childhood health statuses [23]. The probit regression model was used to determine the selection of the first disease and injure as the healthier state in pairwise comparisons. A response variable of 1 was assigned to this selection, while a value of 0 was assigned to the alternative selection. The probit regression model incorporated indicator variables for each disease and injure, with a value of 1 assigned to the first state in a PC, −1 to the second state in a PC, and 0 to all states other than the pair being considered. Additionally, the interval regression model was used to analyze the pooled group health equivalence data. Finally, a linear regression model was used to anchor the estimates from the probit regression. Logit transformations based on the PHE responses were performed to map to a DW scale of 0 to 1. Finally, 1000 bootstrap iterations were used to calculate 95% uncertainty intervals (UIs) [17, 18].

In addition, we conducted a trend analysis of the DWs of the childhood health statuses included in this study, to validate the logical soundness of the study and the reliability of the DW values. Trend analysis i.e. by plotting a trend line on the DW values of diseases with severity ratings to see if they logically fit the intuition. R_{Spearman’s correlation coefficient} (rs) was used to test for correlations between the DW values obtained for different subgroups and to identify overall differences in DW values measured under different population characteristics. These factors included gender (male and female), age (≥ 35 and < 35 years), education level (bachelor’s degree/above and below), annual household income (high income ≥ 100,000 yuan and low income < 100,000 yuan), type of household registration (urban and rural), presence of a medical background, and physical labor status (manual and non-manual), whether there are children in the family (with child and without child), and whether there is a medical background (with medical background (MB) and without MB). Using the results of the ICF assessment of 148 children’s health statuses (See Supplement Table S3), we investigated the impact of various functional attributes and symptomatic manifestations on children’s DW values.

US senators Josh Hawley and Richard Blumenthal have asked the heads of the Federal Trade Commission (FTC) and the Securities and Exchange Commission (SEC) to investigate revenue from ads on Facebook and Instagram that promote scams and banned goods.

“The FTC and SEC should immediately open investigations and, if the reporting is accurate, pursue vigorous enforcement action where appropriate” to force Meta to disgorge profits, pay penalties and agree to cease running such advertisements, Hawley and Blumenthal wrote in a letter to the federal agencies.

Earlier this month, Reuters reported that internal documents from late 2024 stated that that year – about $16bn – from illicit advertising. One document noted Meta, which owns Facebook and Instagram, earns $3.5bn in revenue from “higher risk” scam ads every six months. Other documents stated that Meta’s anti-fraud rules didn’t appear to apply to many ads that regulators and the company’s own staff believed “violated the spirit” of its rules against scam advertising.

In response to the Reuters report, Meta said it had reduced user reports of scams by 58% over the last 18 months.

The Hawley-Blumenthal letter “makes claims that are exaggerated and wrong”, Meta spokesman Andy Stone said. “We aggressively fight fraud and scams because people on our platforms don’t want this content, legitimate advertisers don’t want it and we don’t want it either.”

Hawley, a Republican, and Blumenthal, a Democrat, expressed skepticism about Meta’s efforts to combat illicit advertising. They pointed to the company’s “ad library”, a publicly accessible database of advertising that appears on Meta’s social-media platforms.

“Even a short review of Meta’s Ad Library at the time of this letter shows clearly identifiable advertisements for illicit gambling, payment scams, crypto scams, AI deepfake sex services, and fake offers of federal benefits,” they wrote.

The senators cited Reuters reporting that Meta itself estimated its platforms were involved in a third of all scams in the US, and went on to note that the FTC estimates Americans lost $158.3bn to scams last year.

“Scams have been allowed to take over Facebook and Instagram as Meta has drastically cut its safety staff, including for FTC mandated reviews, even as it dumps unimaginable sums into its generative AI projects.“

Blumenthal and Hawley expressed particular concern about fake ads purporting to represent the US government or political figures. They cited an example of a bogus ad that claimed Donald Trump was offering $1,000 to recipients of food assistance.

“While Meta has been warned about advertisement deepfakes impersonating politicians, it still continues to run fraudulent clips,” their letter states. “The beneficiaries of these scams are often cybercrime groups based in China, Sri Lanka, Vietnam and the Philippines.”