Harnessing large language models' zero-shot and few-shot learning capabilities for regulatory research.

Large language models (LLMs) are sophisticated AI-driven models trained on vast sources of natural language data. They are adept at generating responses that closely mimic human conversational patterns. One of the most notable examples is OpenAI's ChatGPT, which has been extensively used across diverse sectors. Despite their flexibility, a significant challenge arises as most users must transmit their data to the servers of companies operating these models. Utilizing ChatGPT or similar models online may inadvertently expose sensitive information to the risk of data breaches. Therefore, implementing LLMs that are open source and smaller in scale within a secure local network becomes a crucial step for organizations where ensuring data privacy and protection has the highest priority, such as regulatory agencies. As a feasibility evaluation, we implemented a series of open-source LLMs within a regulatory agency's local network and assessed their performance on specific tasks involving extracting relevant clinical pharmacology information from regulatory drug labels. Our research shows that some models work well in the context of few- or zero-shot learning, achieving performance comparable, or even better than, neural network models that needed thousands of training samples. One of the models was selected to address a real-world issue of finding intrinsic factors that affect drugs' clinical exposure without any training or fine-tuning. In a dataset of over 700 000 sentences, the model showed a 78.5% accuracy rate. Our work pointed to the possibility of implementing open-source LLMs within a secure local network and using these models to perform various natural language processing tasks when large numbers of training examples are unavailable.

Consensus clustering with missing labels (ccml): a consensus clustering tool for multi-omics integrative prediction in cohorts with unequal sample coverage.

Multi-omics data integration is a complex and challenging task in biomedical research. Consensus clustering, also known as meta-clustering or cluster ensembles, has become an increasingly popular downstream tool for phenotyping and endotyping using multiple omics and clinical data. However, current consensus clustering methods typically rely on ensembling clustering outputs with similar sample coverages (mathematical replicates), which may not reflect real-world data with varying sample coverages (biological replicates). To address this issue, we propose a new consensus clustering with missing labels (ccml) strategy termed ccml, an R protocol for two-step consensus clustering that can handle unequal missing labels (i.e. multiple predictive labels with different sample coverages). Initially, the regular consensus weights are adjusted (normalized) by sample coverage, then a regular consensus clustering is performed to predict the optimal final cluster. We applied the ccml method to predict molecularly distinct groups based on 9-omics integration in the Karolinska COSMIC cohort, which investigates chronic obstructive pulmonary disease, and 24-omics handprint integrative subgrouping of adult asthma patients of the U-BIOPRED cohort. We propose ccml as a downstream toolkit for multi-omics integration analysis algorithms such as Similarity Network Fusion and robust clustering of clinical data to overcome the limitations posed by missing data, which is inevitable in human cohorts consisting of multiple data modalities. The ccml tool is available in the R language (https://CRAN.R-project.org/package=ccml, https://github.com/pulmonomics-lab/ccml, or https://github.com/ZhoulabCPH/ccml).

Warning Labels as a Public Health Intervention: Effects and Challenges for Tobacco, Cannabis, and Opioid Medications.

Warning labels help consumers understand product risks, enabling informed decisions. Since the 1966 introduction of cigarette warning labels in the United States, research has determined the most effective message content (health effects information) and format (brand-free packaging with pictures). However, new challenges have emerged. This article reviews the current state of tobacco warning labels in the United States, where legal battles have stalled pictorial cigarette warnings and new products such as electronic cigarettes and synthetic nicotine products pose unknown health risks. This article describes the emerging research on cannabis warnings; as more places legalize recreational cannabis, they are adopting lessons from tobacco warnings. However, its uncertain legal status and widespread underestimation of harms impede strict warning standards. The article also reviews opioid medication warning labels, suggesting that lessons from tobacco could help in the development of effective and culturally appropriate FDA-compliant opioid warning labels that promote safe medication use and increased co-dispensing of naloxone.

Self-reported decreases in the purchases of selected unhealthy foods resulting from the implementation of warning labels in Mexican youth and adult population.

BACKGROUND: Front-of-package nutritional warning labels (WLs) are designed to facilitate identification and selection of healthier food choices. We assessed self-reported changes in purchasing different types of unhealthy foods due to WLs in Mexico and the association between the self-reported reductions in purchases of sugary beverages and intake of water and sugar-sweetened beverages. METHODS: Data came from 14 to 17 year old youth (n = 1,696) and adults ≥ 18 (n = 7,775) who participated in the Mexican arm of the 2020-2021 International Food Policy Study, an annual repeat cross-sectional online survey. Participants self-reported whether the WLs had influenced them to purchase less of each of nine unhealthy food categories due to WLs. Among adults, a 23-item Beverage Frequency Questionnaire was used derive past 7-day intake of water and sugary beverages analyzed to determine the relationship between self-reported reductions in purchasing sugary drinks due to the WLs. Multilevel mixed-effects logistic regression models were fitted to estimate the percentage of participants who self-reported reducing purchases within each food group, and overall. Sociodemographic characteristics associated with this reduction were investigated as well. RESULTS: Overall, 44.8% of adults and 38.7% of youth reported buying less of unhealthy food categories due to the implementation of WL, with the largest proportion reporting decreased purchases of cola, regular and diet soda. A greater impact of WLs on the reported purchase of unhealthy foods was observed among the following socio-demographic characteristics: females, individuals who self-identified as indigenous, those who were overweight, individuals with lower educational levels, those with higher nutrition knowledge, households with children, and those with a significant role in household food purchases. In addition, adults who reported higher water intake and lower consumption of sugary beverages were more likely to report reduced purchases of sugary drinks due to the WLs. Adults who reported greater water intake and lower sugary beverages intake were significantly more likely to report buying fewer sugary drinks due to the WLs. CONCLUSION: Our findings suggest that implementation of WLs has reduced perceived purchases of unhealthy foods in Mexico. These results underscore the potential positive impact of the labeling policy particularly in subpopulations with lower levels of education and among indigenous adults.

Countering online marketing and user endorsements with enhanced cannabis warning labels: An online experiment among at-risk youth and young adults.

As cannabis legalization expands and online marketing intensifies, this study examines whether online social cues can amplify youth-targeted cannabis advertising and whether cannabis warning labels (CWLs) can counteract these influences. A U.S. online sample of 970 adolescents and 1776 young adults susceptible to cannabis use were recruited from Qualtrics in summer 2022. Each participant was randomly assigned to one of the 3 (CWLs: none vs. textual vs. pictorial) by 3 (comments: none vs. anti-cannabis vs. pro-cannabis) conditions in an online experiment. Participants were exposed to three online marketing posts promoting marijuana edibles (randomly selected from a large pool, N = 1260), each with either no warning label, a textual warning, or a pictorial warning (text and picture), and with either five comments (pro- or anti-cannabis in valence) or none. Results showed that among adolescents, pro-cannabis comments increased product appeal (vs. anti-cannabis comments: b = 0.18, p = .025; vs. no comments: b = 0.21, p = .021), and did so more than young adults. For adolescents, only pictorial warnings reduced product appeal (b = -0.20, p = .028). For young adults, both pictorial (b = -0.18, p = .002) and textual warnings (b = -0.12, p = .029) reduced product appeal. Furthermore, both textual (adolescents: b = -0.20, p = .004; young adults: b = -0.15, p = .005) and pictorial (adolescents: b = -0.30, p < .001; young adults: b = -0.18, p = .001) warnings reduced cannabis use intentions. Findings support requiring enhanced CWLs accompany online marketing ads.

Assessing transnational spillover effects of Mexico's front-of-package nutritional labeling system among Mexican Americans in the US.

OBJECTIVE: In 2020, Mexico implemented innovative front-of-package nutrition warning labels (FoPWLs) for packaged foods to increase the salience and understanding of nutrition information. This study evaluated Mexican Americans' self-reported exposure to Mexican FoPWLs and self-reported effects of FoPWLs on purchasing behavior. METHODS: The 2021 International Food Policy Study surveyed online panels of adult Mexican Americans in the US (n = 3361) to self-report on buying food at Mexican-oriented stores, noticing Mexican FoPWLs, and being influenced by FoPWLs to purchase less of eight different unhealthy foods (each assessed separately). After recoding the frequency of buying foods in Mexican stores and noticing FoPWLs (i.e., "often" or "very often" vs. less often), logistic models regressed these outcomes on sociodemographics, adjusting for post-stratification weights. RESULTS: Most participants (88.0%) purchased foods in Mexican stores. Of these, 64.1% reported noticing FoPWLs, among whom many reported that FoPWLs influenced them to buy fewer unhealthy foods (range = 32% [snacks like chips] - 44% [colas]). Participants were more likely to buy foods in Mexican stores and notice FoPWLs if they were younger, had ≥two children at home vs no children (AOR = 1.40, 95%CI = 1.15-1.71; AOR = 1.37, 95%CI = 1.03-1.80, respectively), and more frequently used Spanish (AOR = 1.91, 95%CI = 1.77-2.07; AOR = 1.87, 95%CI = 1.69-2.07). Also, high vs. low education (AOR = 1.51, 95%CI = 1.17-1.94) and higher income adequacy (AOR = 1.37, 95%CI = 1.25-1.51) were positively associated with noticing FoPWLs. Being female and more frequent Spanish use were consistently associated with reporting purchase of fewer unhealthy foods because of FoPWLs. CONCLUSIONS: Many Mexican Americans report both exposure to Mexican FOPWLs and reducing purchases of unhealthy foods because of them.

Effectiveness of alcohol warning labels for at-risk groups and the general public: A policy-informing randomized experiment in Chile.

OBJECTIVE: The World Health Organization recommends using health-risk warnings on alcoholic beverages. This study examines the impact of separate or combined warning labels for at-risk groups and the general population on alcohol purchase decisions. METHODS: In 2022, 7758 adults who consumed alcohol or were pregnant/lactating women (54.0 % female, mean age = 40.6 years) were presented with an online store's beverage section and randomly assigned to one of six warning labels in a between-subjects experimental design: no-warning, pregnant/lactating, drinking-driving, general cancer risk, combined warnings, and assorted warnings across bottles. The main outcome, the intention to purchase an alcoholic vs. non-alcoholic beverage, was examined with adjusted risk differences using logistic regressions. RESULTS: Participants exposed to the general cancer risk warning decreased their alcoholic choices by 10.4 percentage points (pp.) (95 % CI [-0.139, -0.069], p < 0.001, OR = 0.561), while those in the pregnancy/lactation warning condition did it by 3.8 pp. (95 % CI [-0.071, -0.005], p = 0.025, OR = 0.806). The driving-drinking warning had no significant effect. Participants exposed to the combined warnings label, or the assorted warnings reduced alcohol purchase decisions by 6.1 pp. (95 % CI [-0.095, -0.028], p < 0.001, OR = 0.708) and 4.3 pp. (95 % CI [-0.076, -0.010], p = 0.011, OR = 0.782), respectively. Cancer warning outperformed other labels and was effective for subgroups such as pregnant/lactating women, young adults, and low-income individuals. CONCLUSIONS: General cancer risk warnings are more effective at reducing alcohol purchase decisions compared to warning labels for specific groups or labels using multiple warnings. In addition to warning labels, other policies should be considered for addressing well-known alcohol-related risks (e.g., drinking and driving).

Perception is reality: qualitative insights into how consumers perceive alcohol warning labels.

AIMS: This study explores perspectives of on-pack alcohol warning labels, and how they might influence alcohol purchase and/or consumption behavior to inform culturally appropriate label design for effective behavior change. METHODS: New Zealand participants ≥18 years, who reported having purchased and consumed alcoholic beverages in the last month were recruited via a market research panel and grouped into 10 focus groups (n = 53) by ethnicity (general population, Maori, and Pacific peoples), age group, and level of alcohol consumption. Participants were shown six potential alcohol health warning labels, with design informed by relevant literature, label framework, and stakeholder feedback. Interviews were transcribed and analyzed via qualitative (directed) content analysis. RESULTS: Effective alcohol labels should be prominent, featuring large red and/or black text with a red border, combining text with visuals, and words like "WARNING" in capitals. Labels should contrast with bottle color, be easily understood, and avoid excessive text and confusing imagery. Participants preferred specific health outcomes, such as heart disease and cancer, increasing message urgency and relevance. Anticipated behavior change included reduced drinking and increased awareness of harms, but some may attempt to mitigate warnings by covering or removing labels. Contextual factors, including consistent design and targeted labels for different beverages and populations, are crucial. There was a strong emphasis on collective health impacts, particularly among Maori and Pacific participants. CONCLUSIONS: Our findings indicate that implementing alcohol warning labels, combined with comprehensive strategies like retail and social marketing campaigns, could effectively inform and influence the behavior of New Zealand's varied drinkers.

Trends (2014-2018) in the healthiness of packaged food purchases of Australian consumers before and after the introduction of voluntary Health Star Rating nutrition labels.

OBJECTIVE: To evaluate the trends in the healthiness of packaged food purchases of Australian consumers before and after the introduction of the Health Star Rating (HSR) nutrition labels. DESIGN: Panel data analysis and difference-in-differences analysis. SETTING: The Australian Government endorsed HSR nutrition labels for voluntary implementation on packaged foods in June 2014. We analyse the packaged food purchases of households across all major supermarkets before (January 2014 to June 2014) and after (June 2014-Dec 2018) the introduction of HSR. PARTICIPANTS: 6284 members of a panel of households across Australia reporting their grocery purchases to a market research company (Nielsen Homescan panel). RESULTS: The healthiness of household food purchases exhibited a U-shaped trend - decreasing from 2014 to 2017, and then increasing from 2018, corresponding to the time when a higher proportion of products were HSR-labelled. Households that purchased a higher proportion of HSR-labelled products had healthier household purchases overall. Further, the healthiness of households' category-specific food purchases was positively associated with the proportion of HSR-labelled products in categories where HSR was adopted, relative to control categories where HSR was not adopted. CONCLUSIONS: In Australia, once a substantial number of packaged food products adopted the voluntary HSR summary indicator, we observed an increasing trend in the healthiness of household food purchases. Widespread adoption of a nutrition summary indicator, such as HSR, on packaged food is likely to be beneficial for population health.

"R" you getting this? Factors contributing to the public's understanding, evaluation, and use of basic reproduction numbers for infectious diseases.

BACKGROUND: We (1) examined the effects of evaluative labels and visual aids on people's understanding, evaluation, and use of the COVID-19 reproduction number (or "r-number"), (2) examined whether people's perceived susceptibility and (intended) adherence to preventive measures changed after being exposed to the r-number, and (3) explored whether these effects and changes depended on people's numeracy skills. METHODS: In an online experiment, participants from a large Dutch representative sample (N = 1,168) received information about the COVID-19 r-number displayed on the corona dashboard of the Dutch Ministry of Health, Welfare and Sport. The r-number was either presented with or without a categorical line display (i.e., evaluative label) and with or without an icon-based tree diagram (i.e., visual aid) explaining how the number works. Regarding people's use of the statistic, we measured perceived susceptibility to COVID-19 and adherence (intention) to five preventive measures before and after exposure to the r-number. After exposure, we also measured participants' understanding, perceived usefulness, affective and cognitive evaluation, and objective numeracy. RESULTS: About 56% of participants correctly interpreted the r-number, with highly numerate people having better understanding than less numerate people. Information about the r-number was perceived as more useful when presented with a visual aid. There were no differences across experimental conditions in people's understanding, affective, and cognitive evaluations. Finally, independent of experimental conditions, intention to adhere to preventive measures was higher after seeing the r-number, but only among highly numerate people. CONCLUSIONS: Although evaluative labels and visual aids did not facilitate people's understanding and evaluation of the r-number, our results show that the statistic is perceived as useful and may be used to stimulate adherence to preventive measures. Policy makers and public health communicators are advised to clearly explain why they are giving these numbers to - especially - the less numerate people, but also how people could use them for behavior change to combat the spread of virus during a pandemic.

The impact of healthy nutrition education based on traffic light labels on food selection, preference, and consumption in patients with acute coronary syndrome: a randomized clinical trial.

BACKGROUND: Acute Coronary Syndrome is the most common heart disease and the most significant cause of death and disability-adjusted life years worldwide. Teaching a healthy eating style is one preventive measure to prevent the disease's recurrence. This study aimed to determine the effect of healthy nutrition education with the help of traffic light labels on food selection, preference, and consumption in patients with acute coronary syndrome. METHODS: This randomized, single-blinded clinical trial was conducted with 139 participants (66 in the intervention group and 73 in the control group) from January 2021 to August 2021 in Shaheed Rajaie Hospital, Tehran, Iran. The control group received standard training. The intervention group, besides this, received additional bedside training with an educational poster on traffic light labels from the research team during their final hospitalization days. Data were collected using a researcher-made questionnaire on food selection, preference, and consumption. RESULTS: The Brunner-Munzel test showed no significant difference between the two groups in terms of selection (P = 0.127), preference (P = 0.852), and food consumption (P = 0.846) in the baseline, while after the intervention, there were significant differences in selection (P > 0.001), preference (P > 0.001), and consumption (p < 0.004). Comparing the difference between the two groups in the difference between the before and after scores for selection (p < 0.001), preference (p < 0.001), and food consumption (p = 0.011) with the Brunner-Munzel test indicated a significant difference in all outcome variables. CONCLUSIONS: Teaching healthy eating styles with the help of traffic light labels affected food selection, preference, and consumption and led to healthier diets in these patients. CLINICAL TRIAL REGISTRATION NUMBER: Clinical trial registration: It was prospectively registered in the Iran Clinical Trials Registration Center on this date 30/10/2020 (IRCT20200927048857N1).

Food for thought: Reinforced learning and recall of physical activity calorie equivalent (PACE) and numerical calorie content in an associative learning task.

Calorie overconsumption has been proposed as a critical contributing factor to rising obesity rates. To combat this health issue, governments and policymakers have suggested implementing numerical caloric content labels. Alternatively, physical activity calorie equivalent (PACE) labels are being proposed as an easier-to-understand metric, representing the amount of physical activity required to burn off calorie content. This study examined individuals' ability to correctly estimate either the numerical caloric content or the PACE values of food images in an associative learning task. Moreover, it assessed whether this knowledge was learned and retained over time. One hundred and ninety-one participants were instructed to estimate either the numerical caloric content or PACE values of thirty food images. To facilitate learning, feedback on the correct number of calories or PACE values was provided during the first session (Time 1). To assess retention, people re-estimated numerical caloric content or PACE values of the same food pictures three days later (Time 2) and seven days later (Time 3), where feedback was not provided. Results showed that participants in both groups improved their estimations using feedback, with people being consistently more accurate when estimating numerical calorie content. Yet, our results also suggest that participants consolidated their knowledge of PACE values over time. Finally, our findings show that hunger moderates individuals' estimation ability, where hungrier people are less accurate than satiated ones. The results contribute to our understanding of how consumers process, estimate, and learn PACE labels versus numerical caloric content, and provide valuable information for researchers and policymakers to develop and implement nutritional labels as a health strategy.

Can I still eat this? Using implicit and explicit measures to explore consumer behavior toward food products with date labels.

This study investigates implicit and explicit attitudes toward products before and beyond the best-before date (BBD) using an Implicit Association Test and an online questionnaire. Moreover, we test whether consumer perception of and behavior toward products beyond the BBD can be manipulated using a priming task. We use a three-group between-subjects design where respondents had to recall either a frugal, a wasteful, or an unrelated behavior. Results show that consumers have negative implicit associations with products beyond the BBD. Reduced health and safety perceptions, consumers' strategies to determine edibility, and general risk perception of products beyond the BBD predict consumption of these products. While recalling a frugal behavior does not have significant effects, recalling a wasteful behavior prior to evaluating products beyond the BBD leads to a decrease in the perceived safety and healthfulness of these products.

How a food scanner app influences healthy food choice.

The use of mobile applications to assist with food decision making has increased significantly. Although food scanner applications provide nutritional information to consumers in the marketplace, little is known about their effects on users' intentions and behavior. This research investigates whether a mobile food scanner app can influence consumers toward healthier food choices. Four studies tested whether information displayed through a food scanner app (as opposed to no information or front-of-packaging label information) influenced purchase intentions for food products (Studies 1-3) or led consumers to make healthier food choices (Study 4). Application-provided information enhanced hypothetical choice and purchase intentions of healthy products in comparison no information, but it did not influence real behavior when participants made choices in an experimental supermarket. Information provided through a food scanner app was systematically outperformed by front-of-packaging label information.

Fruitful changes? Exploring household fruit purchase decisions following comprehensive food policy regulations in Chile.

Changes in unprocessed healthy food purchases associated with the implementation of comprehensive food policy remain understudied. This study analyzes whether, following the announcement, modification, and implementation of Chile's Food Labeling and Advertising law targeting highly processed food (occurring in 2012, 2015, and 2016, respectively), households improved their fruit purchase decisions: purchase participation (i.e., buying likelihood) and purchase quantity. Expenditure data from a representative sample of Chilean households were employed, covering two consecutive survey waves conducted in 2011/2012 and 2016/2017. After controlling for socioeconomic factors (e.g., prices and income), results indicate that only purchase participation increased, providing weak support for positive spillover effects of a comprehensive food policy on fruit purchases. Subsample analyses reveal that this increase was driven by college-educated, childless, and low-income households and was stronger for sweeter and more convenient fruits. Considering that households in Chile do not meet health recommendations for daily fruit intake, additional policy efforts targeting healthy, unprocessed food consumption could be considered.

Can environmental traffic light warning labels reduce meat meal selection? A randomised experimental study with UK meat consumers.

An important area for tackling climate change and health improvement is reducing population meat consumption. Traffic light labelling has successfully been implemented to reduce the consumption of unhealthy foods and sugary drinks. The present research extends this work to meat selection. We tested 1,300 adult UK meat consumers (with quotas for age and gender to approximate a nationally representative sample). Participants were randomised into one of four experimental groups: (1) a red traffic light label with the text 'High Climate Impact' displayed on meat meal options only; (2) a green traffic light label with the text 'Low Climate Impact' displayed on vegetarian and vegan meal options only; (3) red/orange/green (ROG) traffic light labels displayed on relevant meals; and (4) control (no label present). Participants made meal selections within their randomised group across 20 meal trials. A beta-regression was performed to ascertain the change in primary outcome (proportion of meat meals selected across the 20 trials) across the different groups. The red-only label and ROG labels significantly reduced the proportion of meat meals selected compared to the unlabelled control group, by 9.2% and 9.8% respectively. The green-only label did not differ from control. Negatively framed traffic light labels seem to be effective at discouraging meat selection. The labels appeared to be moderately acceptable to meat eaters, who did not think the labels impacted the appeal of the products. These encouraging findings require replication in real-life settings.

Gamified Crowdsourcing as a Novel Approach to Lung Ultrasound Data Set Labeling: Prospective Analysis.

BACKGROUND: Machine learning (ML) models can yield faster and more accurate medical diagnoses; however, developing ML models is limited by a lack of high-quality labeled training data. Crowdsourced labeling is a potential solution but can be constrained by concerns about label quality. OBJECTIVE: This study aims to examine whether a gamified crowdsourcing platform with continuous performance assessment, user feedback, and performance-based incentives could produce expert-quality labels on medical imaging data. METHODS: In this diagnostic comparison study, 2384 lung ultrasound clips were retrospectively collected from 203 emergency department patients. A total of 6 lung ultrasound experts classified 393 of these clips as having no B-lines, one or more discrete B-lines, or confluent B-lines to create 2 sets of reference standard data sets (195 training clips and 198 test clips). Sets were respectively used to (1) train users on a gamified crowdsourcing platform and (2) compare the concordance of the resulting crowd labels to the concordance of individual experts to reference standards. Crowd opinions were sourced from DiagnosUs (Centaur Labs) iOS app users over 8 days, filtered based on past performance, aggregated using majority rule, and analyzed for label concordance compared with a hold-out test set of expert-labeled clips. The primary outcome was comparing the labeling concordance of collated crowd opinions to trained experts in classifying B-lines on lung ultrasound clips. RESULTS: Our clinical data set included patients with a mean age of 60.0 (SD 19.0) years; 105 (51.7%) patients were female and 114 (56.1%) patients were White. Over the 195 training clips, the expert-consensus label distribution was 114 (58%) no B-lines, 56 (29%) discrete B-lines, and 25 (13%) confluent B-lines. Over the 198 test clips, expert-consensus label distribution was 138 (70%) no B-lines, 36 (18%) discrete B-lines, and 24 (12%) confluent B-lines. In total, 99,238 opinions were collected from 426 unique users. On a test set of 198 clips, the mean labeling concordance of individual experts relative to the reference standard was 85.0% (SE 2.0), compared with 87.9% crowdsourced label concordance (P=.15). When individual experts' opinions were compared with reference standard labels created by majority vote excluding their own opinion, crowd concordance was higher than the mean concordance of individual experts to reference standards (87.4% vs 80.8%, SE 1.6 for expert concordance; P<.001). Clips with discrete B-lines had the most disagreement from both the crowd consensus and individual experts with the expert consensus. Using randomly sampled subsets of crowd opinions, 7 quality-filtered opinions were sufficient to achieve near the maximum crowd concordance. CONCLUSIONS: Crowdsourced labels for B-line classification on lung ultrasound clips via a gamified approach achieved expert-level accuracy. This suggests a strategic role for gamified crowdsourcing in efficiently generating labeled image data sets for training ML systems.

Addressing label noise for electronic health records: insights from computer vision for tabular data.

The analysis of extensive electronic health records (EHR) datasets often calls for automated solutions, with machine learning (ML) techniques, including deep learning (DL), taking a lead role. One common task involves categorizing EHR data into predefined groups. However, the vulnerability of EHRs to noise and errors stemming from data collection processes, as well as potential human labeling errors, poses a significant risk. This risk is particularly prominent during the training of DL models, where the possibility of overfitting to noisy labels can have serious repercussions in healthcare. Despite the well-documented existence of label noise in EHR data, few studies have tackled this challenge within the EHR domain. Our work addresses this gap by adapting computer vision (CV) algorithms to mitigate the impact of label noise in DL models trained on EHR data. Notably, it remains uncertain whether CV methods, when applied to the EHR domain, will prove effective, given the substantial divergence between the two domains. We present empirical evidence demonstrating that these methods, whether used individually or in combination, can substantially enhance model performance when applied to EHR data, especially in the presence of noisy/incorrect labels. We validate our methods and underscore their practical utility in real-world EHR data, specifically in the context of COVID-19 diagnosis. Our study highlights the effectiveness of CV methods in the EHR domain, making a valuable contribution to the advancement of healthcare analytics and research.

Evaluation of marketing authorization and labels of medicines in 2021 WHO Model List of Essential Medicines for Children in China, the Russian Federation and Brazil.

OBJECTIVE: This work compares the marketing authorization, labels and dosage forms of medicines in the WHO Model List of Essential Medicines for Children (EMLc) in China, the Russian Federation and Brazil to urge policymakers to pay more attention to paediatric medication. METHODS: Medicines were selected from the 8th EMLc. By searching relevant databases, which include different types of medical information in China, the Russian Federation and Brazil, the marketing authorization, labels and dosage forms of paediatric medicines in the three countries were evaluated. RESULTS: A total of 485 drug products containing 312 active pharmaceutical ingredients listed in the WHO EMLc were evaluated. Among them, 344 products were approved for use in China, 286 in the Russian Federation and 264 in Brazil. Out of the 344 approved medicines, 317 (92.15%) were authorized for paediatric use in China, 224 (78.32%) in the Russian Federation and 218 (82.58%) in Brazil. In terms of guidance information labelling on drug labels, 75.08%, 83.04% and 88.07% of paediatric drugs approved in China, the Russian Federation and Brazil, respectively, clearly indicated the usage and dosage for paediatric use. Additionally, injections and tablets were the most prevalent dosage forms in these three countries. CONCLUSION: There is still scope for enhancing the marketing authorization and development of dosage forms for paediatric medicines in the three countries. Furthermore, additional measures are being implemented to enhance the information provided on drug labels for children, particularly in China.

Warning labels for sugar-sweetened beverages and fruit juice: evaluation of 27 different labels on health effects, sugar content, energy and exercise equivalency.

OBJECTIVES: Front-of-pack warning labels may reduce consumption of sugar-sweetened beverages, potentially mitigating negative health outcomes. Comparisons between different warning label types to inform future research and policy directions are lacking. This study compared 27 warning labels across six message types for their potential to reduce sugar-sweetened beverage consumption. DESIGN AND METHODS: A national sample of regular soda (n = 2578) and juice (n = 1048) consumers aged 14-60 years participated in an online survey. Participants evaluated randomly allocated labels; one from each of six warning label sets (health-graphic, sugar-pictogram, sugar-text, exercise equivalents, health-text, energy information) on four measures of perceived effectiveness (PE: overall effectiveness, discourage from drinking, emotional response, persuasive potential). Participants could also provide open comments. A general linear model compared differences in mean scores across label sets for each measure of PE. RESULTS: PE ratings differed significantly between label sets. Labels clearly quantifying sugar content (sugar-teaspoons) received consistently high PE ratings, whereas 'high in sugar' labels did not. Health-graphic labels were rated highly across all PE measures except persuasive potential. Exercise labels only rated highly on persuasive potential. Health-text results were mixed, and energy labels were consistently low. CONCLUSIONS: Simple, factual labels were easily interpreted and perceived as most effective. Labels quantifying sugar content were consistently high performers and should be advanced into policy to help decrease overconsumption of sugar-sweetened beverages.