Evaluation of the potential value of artificial intelligence (AI) in public health using fluoride intake as the example.

Abstract

Highlights

ChatGPT4o can summarize complex epidemiological questions related to fluoride.
ChatGPT4o can generate innovative and viable fluoride research hypotheses.
ChatGPT4o has potential value in public health education and research.
AI still has limitations, therefore human supervision is necessary.

Aim

We aimed to test whether and how ChatGPT understood the epidemiological problems related to fluoride intake and whether ChatGPT could produce novel and feasible hypotheses to tackle the challenges in the research for the disorders caused by a deficient or excessive fluoride intake.

Material and methods

We designed a set of questions to evaluate the knowledge of ChatGPT version 4o on the epidemiological problems related to fluoride intake. Three evaluators then reviewed these answers. We then requested ChatGPT4o to produce hypotheses for the eight disorders related to insufficient or excessive fluoride intake. These hypotheses were then evaluated independently by three evaluators. Finally, summaries were made through group discussions among all the authors.

Results

For the three questions on basic knowledge about the effect of fluoride on public health, the answers from ChatGPT were rated as excellent or good. For the 12 answers from ChatGPT to the epidemiological questions, 8 out of 12 answers were graded A, as excellent. Four answers were rated as B for good. The descriptions provided by ChatGPT on the effects of fluoride intake were comprehensive and well-structured. Six out of 8 answers were graded as excellent and the other 2 as good. ChatGPT proposed a hypothesis for each of the 8 disorders that are caused by either a deficiency or excess level of fluoride. Four hypotheses were rated as novel and feasible. Three hypotheses were considered relatively new and feasible. Only one hypothesis was regarded as an established hypothesis.

Conclusion

As AI technology develops, it can assist health professionals in understanding the disorders and researchers in their work on the mechanisms behind the disorders caused by insufficient or excessive fluoride intake.

Graphical Abstract

https://ars.els-cdn.com/content/image/1-s2.0-S0147651325001411-ga1_lrg.jpg

1. Introduction

Adequate fluoride intake is an important public health outcome due to the role of fluoride in the prevention of dental caries, a widespread chronic disease affecting people from all age groups (Martignon et al., 2017, Marshall et al., 2005). Tooth decay can lead to pain, and infection, and, if untreated, can lead to tooth loss. By incorporating fluoride into drinking water, toothpaste, and other dental products, the incidence of tooth decay can be effectively reduced (Hiraishi et al., 2011, Fejerskov, 2004). Systemic and topical fluoride strengthens tooth enamel and enhances remineralization, making teeth more resistant to demineralizing acid produced by plaque bacteria and sugars in the mouth (Vasisth et al., 2024). Additionally, the cost-effectiveness of water fluoridation and widespread distribution through drinking water to all populations, regardless of age, education, or socioeconomic status, make it a highly impactful public health intervention (Cronin et al., 2021, Ramasubbu et al., 2024).

Though fluoride intake is beneficial for dental health, excessive exposure is associated with several adverse effects. Chronic high fluoride consumption can cause dental fluorosis, a condition that results in discoloration and mottling of the tooth enamel (Abba et al., 2023, Chandrajith et al., 2022). More severe exposure can lead to skeletal fluorosis, characterized by pain and damage to bones and joints (Gerges et al., 2023, Meena and Gupta, 2021). Emerging research also suggests potential associations between high fluoride intake and negative impacts on neurological development which may affect cognitive function and IQ, particularly among children (Veneri et al., 2023, Xiang et al., 2024, Kumar et al., 2023 Jun). Additionally, there are concerns about fluoride’s role in thyroid dysfunction and other endocrine disruptions (Elghareeb et al., 2024, Taher et al., 2024). These risks underscore the importance of carefully regulating fluoride intake in public health policy to optimize the benefits with safety.

Studies on the application of AI in health dissemination on fluoride intake have not been reported. However, AI has been explored broadly for health education and could be effectively applied to fluoride intake education (Nagi et al., 2023, Gilbert, 2024, Jarab et al., 2024). ChatGPT can process collected text data using advanced natural language processing techniques. ChatGPT can analyze current knowledge databases by processing a large number of coherent texts using a large number of nodes and layers, so it can understand and answer a very wide range of instructions and questions (Sallam, 2023, Li et al., 2024). Currently, several studies also illustrate examples of ChatGPT’s use of machine learning, knowledge graphs, and other techniques to analyze data for application to the healthcare domain (?irkovi? and Katz, 2023, Blanchard et al., 2024). We speculate that AI-powered tools can analyze large datasets to identify patterns and trends in fluoride-related health outcomes, tailor educational information to meet individual needs and disseminate research findings through various digital platforms. Fluoride toxicity research needs to cross multiple disciplines and fields such as chemistry, medicine, and environmental science, etc. ChatGPT’s powerful and efficient ability to parse databases can provide researchers with more comprehensive knowledge on the toxic effects of fluoride on various organs in the human body, and on the transport and transformation of fluoride in ecosystems in environmental science. This integration of interdisciplinary knowledge helps to identify new research directions and potential research questions. In this study, we intend to ask several questions on the application of AI for public education on fluoride intake using ChatGPT as an example. These queries include, whether AI can provide personalized advice on fluoride use based on local water fluoride levels, individual health histories, and age-specific guidelines. Additionally, we explore whether the machine learning algorithms of ChatGPT can help public health officials identify areas at risk for either fluoride deficiency or overdose, thereby allowing for targeted educational campaigns. Overall, we investigate 1). whether the integration of AI can enhance public understanding of fluoride’s benefits and risks, thus, enabling the public to make better decisions about fluoride intake; and 2).Whether ChatGPT can be used to generate research hypotheses to assist investigators in the study of disorders caused by the fluoride deficiency or overdose.

2. Materials and methods

2.1. Tools and source for ChatGPT test

ChatGPT version 4o was used to for questions and answers as the AI language program for the study. Version 4o was the most updated version of ChatGPT at the time of the study (K?l?nç and Mans?z, 2024). Publication databases, such as PubMed, Google search, and website resources searched with Google Chrome were used as the main sources to search for articles and confirming information.

2.1.1. Test 1 – Information summary of fluoride in public health

To determine the capability of ChatGPT4o to efficiently collect information on fluoride intake, we examined whether ChatGPT4o could accurately generate knowledge on all aspects of fluoride intake, including benefits, risks, guidelines, and common FAQs (Fig. 1). For each of these answers, we also asked ChatGPT4o to provide articles as references to support its statement in its answers.

2.1.2. Test 2 – Data gathering on population and management in fluoride intake

To test the capability of ChatGPT4o in data collection, we asked ChatGPT4o a series of comprehensive questions that cover various aspects of fluoride, including, its presence, its sources, and the factors affecting its distribution in water and food sources (Fig. 1). These questions queried global regions with high natural fluoride concentrations in water, geological factors that influence the natural proportion of fluoride in water, population sizes that are affected by high fluoride concentrations in water, methods, and frequency that fluoride levels are monitored in different regions, countries that have implemented water fluoridation programs, and strategies that are employed by stakeholders to manage and mitigate high fluoride levels. We also asked ChatGPT4o to collect data related to fluoride intake from different geographic locations, based on race, ethnicity, and age.

2.1.3. Test 3 – Detailed knowledge of research on the effect of insufficient and excessive levels of fluoride

To test the depth and detail of the knowledge that ChatGPT4o provided on fluoride intake, we asked ChatGPT to describe detailed information on the disorders and diseases that are caused by excessive fluoride intake (Fig. 1). We started with the most common health problems and step by step from the common known issues to the most recently discovered health issues associated with fluoride intake. Through this approach, we intended to see how well the ChatGPT understood the influence of fluoride intake on public health and its potential to act as an assistant for education in public health. The order of the questions on the health issues is as follows: 1). Tooth decay; 2). Dental fluorosis; (Fejerskov, 2004) 3). Skeletal fluorosis; (Gerges et al., 2023) 4). Neurological developmental problems; (Veneri et al., 2023, Xiang et al., 2024, Kumar et al., 2023 Jun) 5). Thyroid dysfunction; (Elghareeb et al., 2024) 6). Kidney Damage; (Tian et al., 2019, Zhu et al., 2023) 7). GI symptoms; (Noureldin et al., 2024, Sampaio et al., 2024) and 8). Low-renin hypertension (LRH) (Zhu et al., 2023).

2.1.4. Test 4. – Capability of ChatGPT to generate new hypotheses to study the effect of deficient and excess levels of fluoride

We asked ChatGPT as many hypotheses as it could generate for each of the eight disorders that were caused by deficiency and excessive levels of fluoride (Fig. 1). We then asked about the novelty of these hypotheses. After we obtained the degrees of novelty among all these hypotheses, we asked ChatGPT to select the ones that were the most novel and easiest to test.

2.2. Evaluations of outputs from ChatGPT

To evaluate ChatGPT’s answers to the above queries and requests, three authors (E1, E2, and E3) were selected to independently provide an evaluation based on the evaluation criteria. The final evaluation then was provided through discussion and agreement among the authors.

Criteria for selection of experts: working on the field of endemic diseases, including fluoride, for more than 20 years, holding positions in relevant fluoride research institutes, and having published authoritative fluoride-related research results.

Evaluation criteria: 1)We graded the content of Test 1, Test 2, and Test 3 as excellent ‘A’ responses that provided exhaustive details, accurate information. Good ‘B’ responses that covered most of the necessary aspects, accurate information. Satisfactory ‘C’ responses that provided a fair amount of detail, and the information is mostly accurate. Poor ‘D’ refers to a response that provides little but basic detail; the information is partially accurate. Failure ‘F’ means the response is severely lacking in detail, and the information is inaccurate. 2) Detailed criteria for evaluating Test 4 content are in Supplementary Material 1.

3. Results

3.1. Test 1 – capacity of ChatGPT4o to provide correct information on fluoride in the public healthChatGPT4o demonstrated a robust capacity to synthesize information about the role of fluoride role in public health, highlighting its importance (Table 1 and Supplementary Material 2). For the three questions on basic knowledge about the effect of fluoride on public health, the answers from ChatGPT were rated as excellent or good, while one answer from AI- Google AI overview was rated as good and two as satisfactory. ChatGPT correctly determined that fluoride was essential for dental health, aided in the remineralization of tooth enamel, prevented dental caries, and strengthened developing teeth. It also supported the public health measure of community water fluoridation, as the measure is a cost-effective measure for preventing cavities. ChatGPT synthesized knowledge to report that insufficient fluoride leads to increased dental caries, impairment of tooth enamel, higher dental treatment costs, and exacerbation of oral health inequities. ChatGPT also realizes that excessive fluoride intake causes dental fluorosis, skeletal fluorosis, neurological effects, thyroid dysfunction, kidney damage, and gastrointestinal issues.

Table 1. Summary of questions and answers from ChatGPT, Google AI, with evaluation grade by experts and Group regarding the information about fluoride in public health*?.

Question	E1??			Group
Q1. Why it is important to have fluoride in the human body?
A1-ChatGPT. Fluoride is an essential mineral for the human body, primarily due to its significant role in maintaining dental health. The key benefits include: 1.Prevention of Dental Caries-?*?. 2.Strengthening Tooth Enamel-. 3.Reduction of Dental Sensitivity-. 4. Support in Public Health-.	B	A	A	A
A1- Google AI overview. Fluoride is a mineral that can be found in many foods and as a dietary supplement. It’s important for the human body because it helps protect teeth and stimulate bone formation: •Teeth, •Bones,	C	B	C	C
Q2. What is the risk when there is not enough fluoride for the human body?
A2-ChatGPT. Insufficient fluoride intake can lead to several dental and potentially broader health issues: 1.Increased Risk of Dental Caries-. 2.Weakened Tooth Enamel-. 3.Higher Dental Treatment Costs-. 4.Worsening Oral Health Inequities-. 5.Potential Impact on Overall Health-.	C	B	A	B
A2- Google AI overview. Not getting enough fluoride can lead to tooth decay and possibly osteoporosis. Fluoride is a naturally occurring element that protects teeth from decay by inhibiting bacteria in plaque and tooth enamel, and by enhancing tooth remineralization. When there’s not enough fluoride, bacteria can produce acids that wear away at tooth enamel, creating an environment for cavities to form.	D	C	B	C
Q3. What are the risks of having too much fluoride intake?
A3-ChatGPT. Excessive fluoride intake can lead to several health issues, some of which can be quite serious: 1.Dental Fluorosis-. 2.Skeletal Fluorosis-. 3.Neurological Effects-. 4.Thyroid Dysfunction-. 5.Kidney Damage-. 6.Gastrointestinal Issues-	B	B	A	B
A3- Google AI overview. Consuming too much fluoride can have a number of negative effects, including: •Dental fluorosis-. •Skeletal fluorosis-. •Other health effects-	D	D	B	C

*?*E?=?Evaluator. For each of these answers, we also asked ChatGPT4o to give articles as references.

*?**?More explanation and/or data in Supplementary Material 2

*: Detailed information in Supplementary Material 2.

The limitations of ChatGPT4o in providing information on fluoride in public health include its reliance on pre-2023 data, which may exclude the latest research and developments. ChatGPT cannot access proprietary databases or subscription-based journals, limiting comprehensive literature reviews. Responses from ChatGPT may lack context or nuance for complex public health issues, and it may inadvertently propagate outdated or incorrect information due to the absence of independent verification. Additionally, ChatGPT4o may reflect biases from its training data, and its inability to securely handle personal health information poses ethical and privacy concerns. These limitations underscore the need for human expertise to supplement and verify the information provided by ChatGPT4o. ChatGPT did well on answering all questions in all aspects of fluoride intake, including benefits, risks, guidelines, and common FAQs.

3.2. Test 2 – Capability for data gathering on population and management of fluoride intake

ChatGPT4o demonstrated substantial capabilities for gathering data on the fluoride intake of a population and managing public health, although it faces notable limitations (Table 2 and Supplementary Material 3). For the 12 answers to the epidemiological questions, eight out of 12 answers from ChatGPT were graded A, as excellent. Four answers were rated as B for good. Four answers from A1- Google AI overview were graded as B, good; Two were graded as C for Satisfactory, Two as D, for poor, and four as F for failure. This data again indicates that ChatGPT provides much more complete information than Google’s web-based information. Globally, regions like East Africa’s Rift Valley, parts of South Asia, China, Mexico, the Middle East, and some areas of North America exhibit high natural fluoride concentrations due to specific geological conditions such as volcanic activity and the presence of fluoride-rich minerals. Many countries, including the United States, Canada, Australia, and the United Kingdom, have implemented water fluoridation programs to maintain fluoride levels within optimal ranges for dental health. However, the methods used to measure fluoride levels, such as ion-selective electrodes and colorimetric methods, though globally consistent, can face challenges like sampling variability, technical interferences, and resource limitations. The influence of soil type, climatic conditions, and agricultural practices further complicates the accurate assessment of fluoride distribution. Despite these limitations, ChatGPT4o can process vast amounts of data to identify patterns and provide insights into the relationship between fluoride exposure and public health outcomes. However, the model’s reliance on pre-2023 data, lack of access to proprietary databases, potential biases, and inability to handle personal health information securely highlight the need for human oversight and the integration of up-to-date, verified information to ensure the reliability and relevance of its outputs in managing fluoride intake and its associated health risks.

Table 2. Summaries of ChatGPT and Google AI-generated data gathered from populations and management of fluoride intake with individual evaluation grades and overall group scores*?.

Question	E1	E2		E3		Group
Q1. What are the global regions with high natural fluoride concentrations in water and soil?
A1-ChatGPT. Regions with high natural fluoride concentrations in water and soil are typically associated with specific geological and environmental conditions. Here are some notable regions: 1.East Africa-?. 2.South Asia: Parts of India- particularly in the states of Andhra Pradesh, Rajasthan, Gujarat, and Punjab?.3.China: Certain regions in China-. 4.Mexico-. 5.Middle East-. 6.North America-	B		A		A	A
A1- Google AI overview. According to a 2022 global fluoride hazard map, areas with high natural fluoride concentrations in water and soil include: •Africa: 15?% of Africa has a greater than 50?% chance of groundwater fluoride concentrations exceeding 1.5?mg/L- •Australia and Oceania-. •South America-. •Asia-. •North America-. •Other areas: –	B		B		B	B
Q2 Which countries have implemented water fluoridation programs, and what are the average fluoride levels in these regions?
A2-ChatGPT. Many countries around the world have implemented water fluoridation programs to prevent dental caries. Here are some notable examples and their average fluoride levels (Supplementary Materials 3): United States-. Canada-. Australia-. United Kingdom-. Ireland-. New Zealand-. Singapore-. Malaysia-. Hong Kong-. Brazil-	B	B		B		B
A2- Google AI overview. Around 24 countries have water fluoridation programs, providing artificially fluoridated water to about 5.7?% of the world’s population. Some countries have more widespread programs, while others have more limited ones. •Widespread programs: Australia (90?%), Brunei (95?%), Chile (70?%), Ireland (70?%), Malaysia (66?%), New Zealand (52?%), Singapore (100?%), and the United States (70?%). •Limited programs: Vietnam, Fiji, Papua New Guinea, and South Korea	B	B		B		B
Q3. What geological factors influence the natural occurrence of fluoride in water and soil?
A3-ChatGPT. The natural occurrence of fluoride in water and soil is influenced by several geological factors. Here are the key factors: 1. Type of Bedrock and Minerals-. 2. Geothermal Activity-. 3. Climate and Weathering-. 4. Soil Composition and Properties-. 5. Hydrogeology-. 6. Anthropogenic Factors-. Industrial Activities-	A	A		A		A
A3- Google AI overview. Many geological factors can influence the natural occurrence of fluoride in water and soil, including: •Rock types-. •Aquifer conditions-. •Groundwater age-. •pH •Temperature	B	B		B		B
Q4. What are the primary natural sources of fluoride in water and soil (e.g., volcanic activity, mineral deposits)?
A4-ChatGPT. The primary natural sources of fluoride in water and soil include: 1. Mineral Deposits-. 2. Volcanic Activity-. 3. Weathering of Fluoride-Containing Rocks-. 4. Soil Composition-. 5. Geothermal and Hydrothermal Activities	B	A		A		A
A4- Google AI overview. Fluorides occur naturally in the earth’s crust where they are found in rocks, coal, clay, and soil. They are released into the air in wind-blown soil. Hydrogen fluoride is released to the air from fluoride-containing substances, including coal, minerals, and clays, when they are heated to high temperatures.	D	B		D		D
Q5. How do agricultural practices and industrial activities contribute to fluoride levels in water and soil?
A5-ChatGPT. Agricultural practices and industrial activities can significantly contribute to fluoride levels in water and soil. Here’s how: Agricultural Practices: 1.Use of Phosphate Fertilizers-. 2.Irrigation Practices-. Industrial Activities: 1.Aluminum Production-. 2.Phosphate Processing- 3.Glass and Ceramic Manufacturing-. 4.Coal Burning-. 5.Brick and Tile Manufacturing- Environmental and Health Impacts: •Soil Contamination-. •Water Contamination-. •Airborne Fluoride-	B	A		A		A
A5- Google AI overview. Not yet have an answer. (As Aug 1, 2024)	F	F		F		F
Q6. What are the typical fluoride concentration ranges in groundwater and surface water in different regions?
A6-ChatGPT. Fluoride concentrations in groundwater and surface water vary significantly depending on geological and environmental factors. Groundwater concentration ranges were provided for 5 sample areas (East Africa – Rift Valley, South Asia, China, North America, Middle East. Surface water concentrations were provided for 4 areas, including North America, Europe, Asia, Australia and Africa. (Ranges from these different regions are found in Supplementary Material 3):	B	B		C		B
A6- Google (not an AI Overview). Because of the large number of variables, the fluoride concentrations in groundwater can range from well under 1?ppm to more than 35?ppm. In Kenya and South Africa, the levels can exceed 25?ppm.	D	D		B		D
Q7. How do climate and weather patterns affect fluoride levels in water and soil?
A7-ChatGPT. Climate and weather patterns significantly influence fluoride levels in water and soil through several mechanisms: 1. Evaporation and Concentration: •Arid and Semi-Arid Climates-. •Seasonal Variations: -. 2. Water Flow and Dilution: •Precipitation and Runoff: -. •Flooding: -. 3. Soil Moisture and Leaching: •Soil Moisture: -. •Dry Climates: -. 4. Temperature and Mineral Solubility: •Temperature Effects: -. •Cold Climates: -. 5. Climate Change Impacts: •Changing Precipitation Patterns:-. •Rising Temperatures: -.	B	A		A		A
A7- Google(not an AI Overview). Fluorine uptake is increasing proportionally with the increasing of the content of fluorine in water, as well as with air temperature in the area, since summer temperatures are higher than winter temperatures. Drinking-water is more often the largest single part to daily fluoride intake	C	C		C		C
Q8. What are the seasonal variations in fluoride concentrations in water sources?
A8-ChatGPT. Seasonal variations in fluoride concentrations in water sources can be influenced by a range of environmental and hydrological factors. Here are some key patterns observed globally: 1. Dry Seasons-•Concentration Increase-. •GivenExample Regions (Supplementary Material 3):-. 2. Rainy Seasons. •Dilution Effect-. •Runoff and Leaching: 3. Seasonal Groundwater Recharge: •Aquifer Recharge: -. •Fluctuations in Wells 4. Agricultural Practices: •Irrigation Cycles:-. •Fertilizer Application: 5. Temperature and Mineral Solubility: •Temperature-Dependent Solubility-. Regional Examples	B	B		A		B
A8- Google AI overview. -Not yet have answer (As Aug 1, 2024).	F	F		F		F
Q9. How does soil type and composition affect fluoride retention and leaching? Measurement and Data Collection
A9-ChatGPT. Soil Type and Texture 1.Clay Soils: oRetention:-. oLeaching-. 2.Sandy Soils: oRetention:-. oLeaching: -. 3.Loamy Soils:-. oRetention:-. oLeaching:-. Soil Composition 1.pH Levels: oAcidic Soils (Low pH):-. oAlkaline Soils (High pH):-. 2.Organic Matter: oRetention-. oLeaching:-. 3.Calcium Content: oRetention:-. oLeaching:-. Measurement and Data Collection 1.Sampling: oSoil Sampling:-. oWater Sampling:-. 2.Analytical Techniques: oIon Selective Electrode (ISE):-. oSpectrophotometry: -. 3.Data Analysis: oStatistical Methods:-. oGeospatial Analysis:-.	B	A		A		A
A9- Google AI overview. Not yet have an answer (As Aug 1, 2024)	F	F		F		F
Q10. What methods are used to measure fluoride levels in water and soil, and how consistent are these methods globally?
A10-ChatGPT. Methods to Measure Fluoride Levels in Water and Soil 1. Ion-Selective Electrode (ISE) Method: •Principle: •Application: •Consistency: 2. Colorimetric Method: •Principle: •Common Reagent: •Application: •Consistency: 3. Ion Chromatography (IC): •Principle: •Application: •Consistency: 4. Fluoride Ion-Selective Electrode with TISAB Buffer •Principle •Application: •Consistency: •Standardization: •Quality Control: •Adaptations	B	A		A		A
A10- Google AI overview. There are multiple methods for measuring fluoride levels in water and soil, including colorimetric, potentiometric, gas chromatographic (GC), and spectrometric methods: •Colorimetric:-. •Potentiometric:-. •GC: Used to determine trace levels of fluoride in biological media. •Spectrometric: Can be used to estimate fluoride levels in water, soil, and vegetables.	C	C		B		C
Q11. How frequently are fluoride levels monitored in different regions, and what agencies are responsible for this monitoring?
A11-ChatGPT. The frequency of monitoring fluoride levels in water and soil varies by region, depending on local regulations, environmental conditions, and public health priorities. Here are some typical monitoring practices (Supplementary Material 3): 1. United States: •Frequency:-. •Responsible Agencies:-. 2. Canada: •Frequency:-. •Responsible Agencies:-. 3. Australia: •Frequency:-. •Responsible Agencies:-. 4. United Kingdom: •Frequency:-. •Responsible Agencies:-. 5. India: •Frequency:-. •Responsible Agencies:-. 6. China: •Frequency:-. •Responsible Agencies-. 7. European Union: •Frequency:-. •Responsible Agencies-.	B	B		A		B
A11- Google AI overview. The frequency and agencies responsible for monitoring fluoride levels vary depending on the type of water system: •Community water systems-. •Private wells-.	D	D		C		D
Q12. What are the common challenges in collecting accurate data on fluoride distribution?
A12-ChatGPT. Collecting accurate data on fluoride distribution in water and soil involves several challenges that can impact the reliability and consistency of the results. Here are some common challenges: 1. Sampling Variability: •Spatial Variability-. •Temporal Variability: -. 2. Technical and Methodological Issues: •Instrumentation and Calibration: -. •Sample Handling and Preservation: -. 3. Analytical Interferences: •Presence of Interfering Ions: -. •Matrix Effects: -. 4. Resource Limitations: •Financial Constraints: -. •Technical Expertise: -. 5. Data Management and Sharing: •Data Integration: -. •Data Accessibility: -. 6. Environmental and Human Factors: •Environmental Conditions: -. •Human Activities: -.	A	A		A		A
A12- Google AI overview. No response given as of Aug 1, 2024)	F	F		F		F

*?* More explanation and/or data in Supplementary Material 3

*: Detailed information is explained in Supplementary Material 3.

Collecting data on population fluoride intake and managing its impacts on public health by ChatGPT faces several specific limitations. First, sampling variability is a significant issue, with fluoride concentrations differing widely due to local geology and environmental conditions, necessitating extensive and frequent sampling to obtain representative data. Technical challenges such as ensuring proper calibration of analytical instruments, preventing sample contamination, and managing interferences from other ions complicate accurate measurement. Additionally, resource constraints, including financial limitations and lack of trained personnel, hinder comprehensive monitoring and data collection efforts. Data management also poses difficulties, with challenges in integrating data from various sources, maintaining data accessibility, and ensuring consistent methodologies across different monitoring agencies. Lastly, environmental factors like extreme weather and human activities such as agriculture and industrial emissions can unpredictably alter fluoride levels, further complicating data collection and analysis. These limitations underscore the need for coordinated efforts and advanced technologies to ensure reliable data for effective public health management.

3.2.1. Test 3 – Detailed knowledge on research on effect of deficiency and excessive levels of fluoride

The descriptions provided by ChatGPT on the effects of fluoride intake are comprehensive and well-structured, covering a wide range of fluoride-related health impacts (Detailed results reported in Supplementary Material 4). The evaluation scales in Table 3 show that 6 out of 8 answers are graded as “A”, the best grade in our evaluation system (Supplementary Material 1).

Table 3. Evaluation of Answers from ChatGPT regarding effects of deficiency and excessive levels of fluoride.

Effect	E1	E2	E3	Group
Tooth Decay	A	A	A	A
Dental Fluorosis	A	A	A	A
Skeletal Fluorosis	A	A	A	A
Neurological problems	A	B	B	B
Thyroid Dysfunction	A	A	A	A
Dysfunction of the kidneys	A	A	A	A
GI symptoms	A	A	A	A
Low-renin hypertension	B	B	B	B

3.3. Tooth decay

The description in this section accurately reflected the well-established benefits of fluoride in preventing dental caries (Supplementary Material 4). The explanation of how fluoride works through both pre-eruptive and post-eruptive mechanisms, including enamel remineralization and antibacterial properties, was well-articulated. The reference to epidemiological studies and the importance of regulated fluoride intake was crucial in supporting the claims. The description effectively highlighted the balance needed between adequate and excessive fluoride intake.(Table 4)

Table 4. Evaluation of hypotheses generated by ChatGPT.

Issue	Hypothesis	E1	E2	E3	Group
Tooth decay caused by insufficient level of fluoride	Individual genetic variations in enamel formation genes modify the susceptibility to dental caries in populations with insufficient fluoride intake	Relatively New Hypothesis Partially Feasible	Relatively new, feasible	Relatively new, feasible	Relatively new, feasible
Dental fluorosis caused by excessive amounts of fluoride	High dietary intake of calcium and vitamin D moderates the risk of developing dental fluorosis in children exposed to excessive fluoride levels during critical periods of tooth development	Relatively New Hypothesis Partially Feasible	Relatively new, feasible	Relatively new, feasible	Relatively new, feasible
Skeletal fluorosis caused by excessive amounts of fluoride	Individuals with genetic variants in collagen-related genes have a higher susceptibility to skeletal fluorosis under conditions of excessive fluoride exposure compared to those without these variants.	Relatively New Hypothesis Partially Feasible	Relatively Novel, feasible	Novel, feasible	Relatively new, feasible
Neurological problems caused by excessive amounts of fluoride	Chronic exposure to high levels of fluoride induces neuroinflammation through activation of microglia, leading to neurological impairments in cognitive function and behavior in adolescents	Relatively New Hypothesis Partially Feasible	Relatively new, partially feasible	Relatively new, partially feasible	Relatively new, partially feasible
Thyroid Dysfunction caused by excessive amounts of fluoride	Excessive fluoride exposure disrupts gut microbiota balance, leading to altered gut-thyroid axis communication and subsequent thyroid dysfunction, characterized by changes in thyroid hormone levels and increased autoimmune thyroiditis incidence	Novel Hypothesis Partially Feasible	Novel, feasible	Novel, feasible	Novel, feasible
Thyroid Dysfunction caused by excessive amounts of fluoride	Excessive fluoride exposure leads to epigenetic modifications, such as DNA methylation changes in genes involved in thyroid hormone synthesis and regulation, contributing to the development of thyroid dysfunction	Relatively New Hypothesis.	Novel, feasible	Novel, feasible	Novel, feasible
Dysfunction of the kidneys caused by excessive amounts of fluoride	Excessive fluoride intake leads to renal dysfunction through oxidative stress and inflammatory pathways, resulting in altered kidney structure and function	Established Hypothesis Partially Feasible	Established, feasible	Relatively new, feasible	Established hypothesis, feasible
GI symptoms caused by excessive amounts of fluoride	Chronic exposure to high levels of fluoride disrupts serotonin signaling pathways in the gastrointestinal tract, leading to altered gut motility and increased incidence of GI symptoms such as nausea, bloating, and diarrhea	Novel Hypothesis High Risk, High Reward	Novel, feasible	Novel, feasible	Novel, feasible
Low-renin hypertension caused by excessive amount of fluoride	Excessive fluoride exposure leads to low-renin hypertension by inducing hyperactivity in the sympathetic nervous system, which disrupts the normal renin-angiotensin-aldosterone system (RAAS) feedback mechanism	Novel Hypothesis Partially Feasible	Novel, feasible	Novel, feasible	Novel, feasible

3.4. Dental fluorosis

The description of dental fluorosis was clear and precise, effectively detailing the impact of excessive fluoride intake during tooth development (Supplementary Material 4). The discussion of the range of severity based on fluoride exposure levels and the emphasis on regulated fluoride levels in community water fluoridation was well-presented. The connection between fluoride intake, hypo mineralization, and the resulting dental fluorosis was accurately portrayed.

3.5. Skeletal fluorosis

This section provided a detailed explanation of the causes and progression of skeletal fluorosis, emphasizing the role of prolonged fluoride exposure (Supplementary Material 4). The discussion on how fluoride accumulated in bones and disrupted bone metabolism was well-supported by epidemiological evidence. The focus on regions with high natural fluoride levels and the need for preventive measures was appropriate.

3.6. Neurological problems

The description was comprehensive, covering the potential cognitive and neurodevelopmental effects of excessive fluoride intake. The mention of oxidative stress and alterations in neurotransmitter function provided a plausible mechanism for the observed neurological issues. However, given that the evidence in this area is still evolving, the cautious rating by some evaluators was understandable. The emphasis on preventive measures and public health policies was well-placed. Therefore, the answer was graded as “B”.

3.7. Thyroid dysfunction

This section effectively highlighted the relationship between high fluoride intake and thyroid dysfunction, particularly in iodine-deficient populations (Supplementary Material 4). The explanation of how fluoride competes with iodine and disrupts thyroid hormone synthesis was accurate and supported by relevant studies. The emphasis on the importance of monitoring fluoride levels to prevent thyroid issues while still benefiting from its dental health properties was well-balanced.

3.8. Dysfunction of the kidneys

The description here was thorough, detailing the direct and indirect effects of high fluoride intake on kidney function (Supplementary Material 4). The discussion of mechanisms such as oxidative stress, inflammation, and fibrosis provided a clear understanding of how fluoride could lead to renal dysfunction. The mention of studies linking high fluoride levels to chronic kidney disease added credibility to the claims.

3.9. GI symptoms

This section accurately described the gastrointestinal symptoms associated with high fluoride intake, both in acute and chronic scenarios (Supplementary Material 4). The discussion of how fluoride irritated the mucosal lining and alters digestive enzyme activity was well-explained. The emphasis on managing fluoride levels to prevent GI issues was appropriate and supported by evidence.

3.10. Low-renin hypertension

The connection between high fluoride intake and low-renin hypertension was presented with reasonable evidence (Supplementary Material 4). The description of how fluoride impacted the renin-angiotensin system and kidney function provided a plausible explanation for the observed hypertension in high-fluoride areas. However, the rating as “B” by some evaluators suggests that this area might require further research to establish stronger causal links.

Overall, the descriptions provided by ChatGPT are scientifically sound and well-referenced and cover the key aspects of fluoride’s effects on various health conditions. The emphasis on both the benefits and potential risks of fluoride intake demonstrates a balanced approach to the topic. However, we note that most of the evaluation of answers provided by AI come from review articles, possibly leading to those secondary-cited conclusions that may have been influenced by the views of review article authors.

4. Test 4 – Capacity in the generation of research hypotheses to study the effect of deficiency and exceed levels of fluoride

ChatGPT proposed a hypothesis for each of the eight disorders that were caused by deficiency and excessive levels of fluoride ((Supplementary Material 5–12). Four hypotheses were rated as novel and feasible, including those that suggested that genetic variations influence susceptibility to dental caries and that high calcium and vitamin D intake moderated dental fluorosis risk. The other three hypotheses, such as those that linked excessive fluoride to neuroinflammation and thyroid dysfunction via gut microbiota disruption or epigenetic modifications, were also considered to be relatively new and feasible. These hypotheses offer insights for mechanisms and interactions that could guide future research on fluoride-related health impacts. Only one hypothesis was regarded as an established hypothesis.

4.1. Hypothesis for investigating tooth decay

The hypothesis “Individual genetic variations in enamel formation genes modify the susceptibility to dental caries in populations with insufficient fluoride intake” is rated as relatively new, and feasible. It proposes a personalized approach to dental caries prevention by integrating genetic and environmental factors (Supplementary Material 5). By focusing on diverse populations with low fluoride levels, the study can identify specific genetic variants in genes like AMELX, ENAM, and MMP20 (Wright, 2006) that influence enamel strength and formation, potentially revealing genetic subgroups at higher risk for dental caries. This research could identify targeted interventions and more precise public health guidelines for dental care. The study’s design includes comprehensive fluoride exposure assessment and advanced dental examination techniques, ensuring robust and nuanced data collection. Statistical interaction models will elucidate the interplay between genetic predisposition and fluoride exposure, highlighting the importance of personalized dental health practice.

4.2. Hypothesis for dental fluorosis

The hypothesis “High dietary intake of calcium and vitamin D moderates the risk of developing dental fluorosis in children exposed to excessive fluoride levels during critical periods of tooth development” is considered relatively new and feasible (Supplementary Material 6). The hypothesis explores the capacity of nutritional intervention to mitigate the prevalent dental condition. By focusing on children living in areas with naturally high fluoride levels and assessing their dietary intake of calcium and vitamin D, the study aims to examine whether these nutrients can reduce the severity of dental fluorosis. This approach could offer a simple and accessible preventive strategy through dietary optimization and reduce the health burden of dental fluorosis. The use of regression models to analyze interactions between dietary factors and fluoride exposure further enhances the study’s relevance in uncovering valuable insights into the protective role of nutrition in dental health.

4.3. Hypothesis for skeletal fluorosis

The hypothesis “Individuals with genetic variants in collagen-related genes have a higher susceptibility to skeletal fluorosis under conditions of excessive fluoride exposure compared to those without these variants” is graded as relatively new and feasible (Supplementary Material 7). This hypothesis explores the interplay between genetic predisposition and environmental factors in the development of skeletal fluorosis. By focusing on regions with endemic fluorosis and screening for variants in critical collagen-related genes like COL1A1 and COL1A2, the study aims to identify genetic subgroups at increased risk. Employing interaction models further strengthens the investigation of genetic and environmental interactions as well as the targeted interventions for those most at risk.

4.4. Neurological problems

The hypothesis “Chronic exposure to high levels of fluoride induces neuroinflammation through activation of microglia, leading to neurological impairments in cognitive function and behavior in adolescents” is considered as relatively new and feasible (Supplementary Material 8). It discovers the neurotoxic effects of fluoride, a commonly encountered environmental factor. By targeting adolescents in high-fluoride areas, the research will establish a direct link between fluoride exposure and neuroinflammation, potentially contributing to cognitive and behavioral impairments. This research could have substantial public health implications by highlighting the need for revised fluoride safety guidelines. The study’s comprehensive design includes detailed fluoride exposure assessment, neuropsychological, behavioral evaluations, biomarker analysis and ensures robust data collection.

4.5. Hypothesis for thyroid dysfunction

The hypothesis “Excessive fluoride exposure disrupts gut microbiota balance, leading to altered gut-thyroid axis communication and subsequent thyroid dysfunction, characterized by changes in thyroid hormone levels and increased autoimmune thyroiditis incidence” is considered as significant and novel because it integrates the fields of endocrinology and microbiome research to explore an indirect pathway through which fluoride might affect thyroid health (Supplementary Material 9). By focusing on adults in high-fluoride areas and assessing fluoride intake, thyroid function, and gut microbiota composition, the research intends to elucidate a complex interaction that has not been widely studied. This approach could reveal a previously unrecognized mechanism linking environmental fluoride exposure to thyroid dysfunction, and inform public health policies and interventions. The study’s design includes comprehensive fluoride exposure assessment, regular thyroid function tests, advanced gut microbiota analysis.

4.6. Hypothesis for dysfunction of the kidneys

The hypothesis “Excessive fluoride intake leads to renal dysfunction through oxidative stress and inflammatory pathways, resulting in altered kidney structure and function” is regarded as an established hypothesis, yet also feasible (Tian et al., 2019, Hu et al., 2023) (Supplementary Material 10). The proposal seeks to elucidate the specific mechanisms underlying fluoride-induced kidney damage, which are not well understood. Several studies on this aspect have been reported (Hu et al., 2023, Ajuwon et al., 2024, Saber et al., 2020). These studies have revealed the similar results to what the ChatGPT predicted. Although this hypothesis itself is not novel, it also suggests that the other hypotheses could achieve the expected results if they are novel.

4.7. Hypothesis for GI symptoms

The hypothesis “Chronic exposure to high levels of fluoride disrupts serotonin signaling pathways in the gastrointestinal tract, leading to altered gut motility and increased incidence of GI symptoms such as nausea, bloating, and diarrhea” is novel and feasible because it explores a biochemical mechanism linking fluoride exposure to gastrointestinal health (Supplementary Material 11). By focusing on serotonin signaling, the study aims to uncover how fluoride affects gut motility and contributes to common GI symptoms. This research could provide insightful thoughts into the systemic effects of fluoride and identify possible therapeutic targets for managing fluoride-induced GI disturbances. The study’s comprehensive design, including detailed fluoride exposure assessment, standardized GI symptom documentation, and measurement of serotonin levels in both blood and stool samples, ensures robust data collection. By considering control variables and using regression analyses, the research aims to offer a nuanced understanding of the relationship between fluoride exposure and gastrointestinal health.

4.8. Hypothesis for low-renin hypertension

The hypothesis “Excessive fluoride exposure leads to low-renin hypertension by inducing hyperactivity in the sympathetic nervous system, which disrupts the normal renin-angiotensin-aldosterone system (RAAS) feedback mechanism” is novel and feasible as it explores a neuroendocrine pathway through which fluoride might influence blood pressure regulation (Supplementary Material 12). By focusing on adults in high-fluoride regions and assessing comprehensive fluoride exposure, blood pressure, renin activity, aldosterone levels, and sympathetic nervous system markers, the research will elucidate how fluoride impacts the RAAS via sympathetic hyperactivity. This method offers original ideas about the systemic effects of fluoride, highlighting targets for preventing and managing fluoride-related hypertension. The study’s robust design, control of the confounding variables and use of multivariable regression models, ensure a thorough examination of the proposed mechanisms, potentially informing therapeutic strategies for hypertension management.

In summary, the hypothesis proposed by ChatGPT is based on a certain fluoride research background, the basic chemical properties of fluoride, and common toxicity mechanisms. These fundamentals tend to have a degree of continuity and universality. There are limitations of ChatGPT in accessing articles after 2023 and in subscribing to journals. We searched for the research published after 2023 as well as the subscribed journals (non-ChatGPT sources) and found that ChatGPT proposed novel hypotheses in the research areas of fluoride-related tooth decay, skeletal fluorosis, thyroid dysfunction, dysfunction of the kidneys, GI symptoms, and low-renin hypertension. However, a few similar research reports have been on hypotheses regarding dental fluorosis and neurological problems (Supplementary Material 13). Based on the above results, it is found that the hypothesis proposed by ChatGPT still has certain novelty.

5. Discussion

The most significant accomplishment of our investigation is the demonstration of AI’s capability, specifically using ChatGPT4o, to provide accurate and comprehensive information on fluoride intake and its implications for public health. This original study highlights the ability of AI to understand the complex epidemiological problems associated with fluoride intake, generate novel and feasible hypotheses for research, and provide detailed knowledge on the disorders caused by fluoride deficiency or excess. This study establishes AI as a promising tool in public health education and research, particularly in addressing both the beneficial and adverse effects of fluoride intake.

Our study shows the ability of ChatGPT4o to collect and synthesize data on population fluoride intake, identify patterns and trends in fluoride-related health outcomes, and provide tailored educational content. Additionally, our study effectively compares the AI’s responses to those from Google Chrome, showcasing the superiority of AI in providing more complete and accurate information on fluoride intake and its health impacts.

In terms of generating new hypotheses, we needed ChatGPT to provide relevant literature when generating hypotheses, and in conjunction with our search for literature from non-ChatGPT sources, we analysed the proportion of non-ChatGPT sources used in validating hypotheses (Supplementary Material 14). We found a higher proportion of non-ChatGPT sources in studies of dental fluorosis, skeletal fluorosis, neurological problems and renal dysfunction 72.7?%, 42.9?%, 33.3?%, 14.3?% respectively. Expert assessment found three hypotheses to be relatively novel and feasible, while the hypothesis on renal dysfunction was considered an established and previously researched hypothesis. However, for the four hypotheses with 0?% non-ChatGPT sources, their expert evaluations were overall more novel and feasible than the others. Thus, our study found that ChatGPT has the potential to propose new hypotheses.

The impact of this study for the application of AI in the study of health issues of fluoride is multifaceted. By showcasing the potential of AI to enhance public health education on fluoride intake, it opens new avenues for leveraging AI in other areas of public health. The study’s findings can inform public health policies and practice, particularly in regions with high fluoride exposure. Moreover, the novel hypotheses generated by the AI can guide future research, new discoveries in the health effects of fluoride. This work underscores the importance of integrating AI into public health strategies to improve health education, data collection, and research.

Despite its accomplishments, the study has several limitations. One significant limitation is the reliance on pre-2023 data, which may exclude the latest research and developments in the field. The inability of AI to access proprietary databases or subscription-based journals limits the comprehensiveness of its literature reviews. Despite the significant potential of ChatGPT for data analysis, its application requires caution. The following limitations may exist: ChatGPT’s training data comes from a variety of published and unpublished textual sources with varying data quality. ChatGPT is inconsistent in its ability to provide in-depth professional understanding when dealing with large amounts of fluoride-related textual data information. Additionally, the AI may lack context or nuance in addressing complex public health issues and could propagate outdated or incorrect information due to the absence of independent verification. The study also notes potential biases from the AI’s training data and ethical and privacy concerns related to handling personal health information.

Future studies may include updating the AI’s training data with the latest research to enhance the accuracy and relevance of its outputs. Expanding the AI’s access to a broader range of databases and journals would improve the comprehensiveness of its literature reviews. Further research should focus on addressing the ethical and privacy concerns associated with AI in public health applications. Meanwhile, AI-driven knowledge graphs and semantic networks are used to discover conceptual relationships to generate hypotheses, which are screened by experts or AI tools for valuable research and optimised with reinforcement learning and critical review techniques. Additionally, validating the novel hypotheses generated by AI through empirical research could collect new evidence and improve public health interventions. The study could also explore the integration of AI with other technologies to enhance its capabilities in public health education and research.

6. Conclusion

In conclusion, our study demonstrates the potential value of artificial intelligence, specifically ChatGPT4o, in public health education and research regarding fluoride intake. The AI showcased its ability to provide comprehensive and accurate information, generate novel research hypotheses, and offer tailored educational content. These findings highlight AI’s capacity to enhance public understanding of fluoride’s benefits and risks, so that the public can make more informed decisions about fluoride intake.

CRediT authorship contribution statement

Wei Dan: Writing – review & editing, Investigation. Li Minghui: Writing – review & editing, Investigation. Sun Shuqiu: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation. Gu Tianshu: Methodology, Investigation, Formal analysis, Data curation. Cao Yanhong: Methodology, Investigation, Formal analysis, Data curation. Wei Wei: Writing – review & editing, Writing – original draft, Project administration, Data curation, Conceptualization. Yao Lan: Writing – review & editing, Methodology, Investigation, Data curation, Conceptualization. Sun Dianjun: Writing – review & editing, Methodology, Investigation, Data curation, Conceptualization. Fly Alyce D: Writing – review & editing, Methodology, Investigation. Gu Weikuan: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of Generative AI and AI-assisted technologies in the writing processDuring the preparation of this work, the author(s) used ChatGPT4o to generate descriptions and draw pictures for authors to modify to compose the manuscript and to make the figures for illustration of scientific concepts. The authors take all responsibility for these figures.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper

Acknowledgments

The findings and conclusions in this report are those of the author(s) and do not necessarily.represent the official position of any organization. This work was partially supported by funding from the University of Tennessee Health Science Center (R073290109) to WG in Memphis, TN, USA, and National Natural Science Foundation of China (Nos.82373699) to WW in China. The funders had no role in the study design, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.

Human subjects

Data from patients in this study were obtained from public websites.

Consent statement/Ethical approval

N/A

Author contributions

Conceived and designed the experiments: WW, LY, DS, WG. Performed data search and collection: TG, WG, Analyzed the data: All Authors. Contributed analysis tools: WG, ML, DS. Wrote the first draft of the manuscript: TG,WD, LY, WG, edited by LY, ADF. Revised and approved manuscript: All authors.

Appendix A. Supplementary material

What’s this?

Download: Download Word document (22KB)

Supplementary material

Download: Download Word document (308KB)

Supplementary material

Download: Download Word document (321KB)Supplementary material

Data availabilityData will be made available on request. All the data are available in the public databases as provided in the manuscript.

References

Abba et al., 2023

S.I. Abba, J.C. Egbueri, M. Benaafi, J. Usman, A.G. Usman, I.H. Aljundi

Fluoride and nitrate enrichment in coastal aquifers of the Eastern Province, Saudi Arabia: the influencing factors, toxicity, and human health risks

Chemosphere, 336 (2023), Article 139083, 10.1016/j.chemosphere.2023.139083

View PDF View article View in ScopusGoogle ScholarAjuwon et al., 2024

O.R. Ajuwon, T.A. Adeleke, B.O. Ajiboye, A.O. Lawal, I. Folorunso, B. Brai, F.A. Bamisaye, J.A. Falode, I.M. Odoh, K.I. Adegbite, O.B. Adegoke

Fermented rooibos tea (Aspalathus linearis) ameliorates sodium fluoride-induced cardiorenal toxicity, oxidative stress, and inflammation via modulation of NF-KappaB/IkappaB/IkappaKB signaling pathway in wistar rats

Cardiovasc Toxicol., 24 (3) (2024), pp. 240-257, 10.1007/s12012-024-09826-9

View in ScopusGoogle ScholarBlanchard et al., 2024

M. Blanchard, V. Venerito, P. Ming Azevedo, T. Hügle

Generative AI-based knowledge graphs for the illustration and development of mHealth self-management content

Front Digit Health, 6 (2024), Article 1466211, 10.3389/fdgth.2024.1466211

View in ScopusGoogle ScholarChandrajith et al., 2022

R. Chandrajith, S. Bhagya, S. Diyabalanage, S. Wimalasiri, M.A.B. Ranatunga, J.A.C. Barth

Exposure assessment of fluoride intake through commercially available black tea (Camellia sinensis L.) from areas with high incidences of chronic kidney disease with undetermined origin (CKDu) in Sri Lanka

Biol. Trace Elem. Res, 200 (2) (2022), pp. 526-534, 10.1007/s12011-021-02694-2

View in ScopusGoogle Scholar?irkovi? and Katz, 2023

A. ?irkovi?, T. Katz

Exploring the potential of ChatGPT-4 in predicting refractive surgery categorizations: comparative study

JMIR Form. Res., 7 (2023), Article e51798, 10.2196/51798

View in ScopusGoogle ScholarCronin et al., 2021

J. Cronin, S. Moore, M. Harding, H. Whelton, N. Woods

A cost-effectiveness analysis of community water fluoridation for schoolchildren

BMC Oral. Health, 21 (1) (2021), p. 158, 10.1186/s12903-021-01490-7

View in ScopusGoogle ScholarElghareeb et al., 2024

M.M. Elghareeb, G.E. Elshopakey, S. Rezk, A. Ateya, E.S. El-Ashry, M. Shukry, H.I. Ghamry, B.S. Alotaibi, N.M.A. Hashem

Nigella sativa oil restores hormonal levels, and endocrine signals among thyroid, ovarian, and uterine tissues of female Wistar rats following sodium fluoride toxicity

Biomed. Pharm., 170 (2024), Article 116080, 10.1016/j.biopha.2023.116080

View PDF View article View in ScopusGoogle ScholarFejerskov, 2004

O. Fejerskov

Changing paradigms in concepts on dental caries: consequences for oral health care

Caries Res, 38 (3) (2004), pp. 182-191, 10.1159/000077753

View in ScopusGoogle ScholarGerges et al., 2023

M.M. Gerges, V.T. Ngo, J. Tu, P.M. Colletti, H. Wassef

Skeletal fluorosis

Clin. Nucl. Med, 48 (12) (2023), pp. 1127-1130, 10.1097/RLU.0000000000004911

View in ScopusGoogle ScholarGilbert, 2024

T.K. Gilbert

Generative AI and generative education

Ann. N. Y Acad. Sci., 1534 (1) (2024), pp. 11-14, 10.1111/nyas.15129

View in ScopusGoogle ScholarHiraishi et al., 2011

N. Hiraishi, R. Sono, M.S. Islam, M. Otsuki, J. Tagami, T. Takatsuka

Effect of hesperidin in vitro on root dentine collagen and demineralization

J. Dent., 39 (5) (2011), pp. 391-396, 10.1016/j.jdent.2011.03.002

View PDF View article View in ScopusGoogle ScholarHu et al., 2023

Y. Hu, Z. Yan, Y. He, Y. Li, M. Li, Y. Li, D. Zhang, Y. Zhao, M.M. Ommati, J. Wang, M. Huo, J. Wang

Ameliorative effects of different doses of selenium against fluoride-triggered apoptosis and oxidative stress-mediated renal injury in rats through the activation of Nrf2/HO-1/NQO1 signaling pathway

Food Chem. Toxicol., 174 (2023), Article 113647, 10.1016/j.fct.2023.113647

View PDF View article View in ScopusGoogle ScholarJarab et al., 2024

A.S. Jarab, W. Al-Qerem, D.M. Al-Hajjeh, S. Abu Heshmeh, T.L. Mukattash, A.Y. Naser, H. Alwafi, Y.N. Al Hamarneh

Artificial intelligence utilization in the healthcare setting: perceptions of the public in the UAE

Int J. Environ. Health Res. (2024), pp. 1-9, 10.1080/09603123.2024.2363472

Google Scholar

K?l?nç and Mans?z, 2024

D.D. K?l?nç, D. Mans?z

Examination of the reliability and readability of Chatbot Generative Pretrained Transformer’s (ChatGPT) responses to questions about orthodontics and the evolution of these responses in an updated version

Am. J. Orthod. Dentofac. Orthop., 165 (5) (2024), pp. 546-555, 10.1016/j.ajodo.2023.11.012

View PDF View article View in ScopusGoogle ScholarKumar et al., 2023 Jun

J.V. Kumar, M.E. Moss, H. Liu, S. Fisher-Owens

Association between low fluoride exposure and children’s intelligence: a meta-analysis relevant to community water fluoridation

Public Health, 219 (2023 Jun), pp. 73-84, 10.1016/j.puhe.2023.03.011

View PDF View article View in ScopusGoogle ScholarLi et al., 2024

Z. Li, Q. Wei, L.C. Huang, J. Li, Y. Hu, Y.S. Chuang, J. He, A. Das, V.K. Keloth, Y. Yang, C.S. Diala, K.E. Roberts, C. Tao, X. Jiang, W.J. Zheng, H. Xu

Ensemble pretrained language models to extract biomedical knowledge from literature

J. Am. Med. Inf. Assoc., 31 (9) (2024), pp. 1904-1911, 10.1093/jamia/ocae061

View in ScopusGoogle ScholarMarshall et al., 2005

T.A. Marshall, J.M. Eichenberger Gilmore, B. Broffitt, P.J. Stumbo, S.M. Levy

Diet quality in young children is influenced by beverage consumption

J. Am. Coll. Nutr., 24 (1) (2005), pp. 65-75, 10.1080/07315724.2005.10719445

View in ScopusGoogle ScholarMartignon et al., 2017

S. Martignon, M.O. Opazo-Gutiérrez, M. Velásquez-Riaño, I.R. Orjuela-Osorio, V. Avila, E.A. Martinez-Mier, M.C. González-Carrera, J.A. Ruiz-Carrizosa, B.C. Silva-Hermida

Geochemical characterization of fluoride in water, table salt, active sediment, rock and soil samples, and its possible relationship with the prevalence of enamel fluorosis in children in four municipalities of the department of Huila (Colombia)

Environ. Monit. Assess., 189 (6) (2017), p. 264, 10.1007/s10661-017-5975-6

View in ScopusGoogle ScholarMeena and Gupta, 2021

L. Meena, R. Gupta

Skeletal Fluorosis

N. Engl. J. Med., 385 (16) (2021), p. 1510, 10.1056/NEJMicm2103503

View in ScopusGoogle ScholarNagi et al., 2023

F. Nagi, R. Salih, M. Alzubaidi, H. Shah, T. Alam, Z. Shah, M. Househ

Applications of artificial intelligence (AI) in medical education: a scoping review

Stud. Health Technol. Inf., 305 (2023), pp. 648-651, 10.3233/SHTI230581

View in ScopusGoogle ScholarNoureldin et al., 2024

A. Noureldin, W. Alshehri, H. Tapias, L. Mallonee, L.M. Mancl, P. Milgrom, K. Svboda

Efficacy of 38% silver diamine fluoride in reducing gingival inflammation and plaque accumulation in older adults living in retirement-homes: a randomized controlled pilot trial

J. Dent., 143 (2024), Article 104890, 10.1016/j.jdent.2024.104890

View PDF View article View in ScopusGoogle ScholarRamasubbu et al., 2024

D. Ramasubbu, J. Lewney, B. Duane

Investigating the effectiveness of water fluoridation

Evid. Based Dent. (3) (2024), 10.1038/s41432-024-01032-4

Google Scholar

Saber et al., 2020

T.M. Saber, M.F. Mansour, A.S. Abdelaziz, R.M.S. Mohamed, R.A. Fouad, A.H. Arisha

Argan oil ameliorates sodium fluoride-induced renal damage via inhibiting oxidative damage, inflammation, and intermediate filament protein expression in male rats

Environ. Sci. Pollut. Res. Int., 27 (24) (2020), pp. 30426-30436, 10.1007/s11356-020-09366-z

View in ScopusGoogle ScholarSallam, 2023

M. Sallam

ChatGPT utility in healthcare education, research, and practice: systematic review on the promising perspectives and valid concerns

Healthcare, 11 (6) (2023), p. 887, 10.3390/healthcare11060887

View in ScopusGoogle ScholarSampaio et al., 2024

C. Sampaio, A.C.B. Delbem, T.Y. Hosida, A.V.P. Fernandes, B. do Amaral, L.A. de Morais, D.R. Monteiro, J.P. Pessan

Amount of dentifrice and fluoride concentration affect the pH and inorganic composition of dual-species biofilms of streptococcus mutans and Candida albicans

Pharmaceutics, 16 (4) (2024), p. 562, 10.3390/pharmaceutics16040562

View in ScopusGoogle ScholarTaher et al., 2024

M.K. Taher, F. Momoli, J. Go, S. Hagiwara, S. Ramoju, X. Hu, N. Jensen, R. Terrell, A. Hemmerich, D. Krewski

Systematic review of epidemiological and toxicological evidence on health effects of fluoride in drinking water

Crit. Rev. Toxicol., 54 (1) (2024), pp. 2-34, 10.1080/10408444.2023.2295338

View in ScopusGoogle ScholarTian et al., 2019

X. Tian, J. Feng, N. Dong, Y. Lyu, C. Wei, B. Li, Y. Ma, J. Xie, Y. Qiu, G. Song, X. Ren, X. Yan

Subchronic exposure to arsenite and fluoride from gestation to puberty induces oxidative stress and disrupts ultrastructure in the kidneys of rat offspring

Sci. Total Environ., 686 (2019), pp. 1229-1237, 10.1016/j.scitotenv.2019.04.409

View PDF View article View in ScopusGoogle ScholarVasisth et al., 2024

D. Vasisth, P. Mehra, L. Yadav, V. Kumari, U. Bhatia, R. Garg

Fluoride and its implications on oral health: a review

J. Pharm. Bioallied Sci., 16 (1) (2024), pp. S49-S52, 10.4103/jpbs.jpbs_929_23

Google Scholar

Veneri et al., 2023

F. Veneri, M. Vinceti, L. Generali, M.E. Giannone, E. Mazzoleni, L.S. Birnbaum, U. Consolo, T. Filippini

Fluoride exposure and cognitive neurodevelopment: systematic review and dose-response meta-analysis

Environ. Res., 221 (2023), Article 115239, 10.1016/j.envres.2023.115239

View PDF View article View in ScopusGoogle ScholarWright, 2006

J.T. Wright

The molecular etiologies and associated phenotypes of amelogenesis imperfecta

Am. J. Med. Genet A, 140 (23) (2006), pp. 2547-2555, 10.1002/ajmg.a.31358

View in ScopusGoogle ScholarXiang et al., 2024

J. Xiang, X.L. Qi, K. Cao, L.Y. Ran, X.X. Zeng, X. Xiao, W. Liao, W.W. He, W. Hong, Y. He, Z.Z. Guan

Exposure to fluoride exacerbates the cognitive deficit of diabetic patients living in areas with endemic fluorosis, as well as of rats with type 2 diabetes induced by streptozotocin via a mechanism that may involve excessive activation of the poly(ADP ribose) polymerase-1/P53 pathway

Sci. Total Environ., 912 (2024), Article 169512, 10.1016/j.scitotenv.2023.169512

View PDF View article View in ScopusGoogle ScholarZhu et al., 2023

C. Zhu, W. Gu, D. Sun, W. Wei

The mechanism underlying fluoride-induced low-renin hypertension is related to an imbalance in the circulatory and local renin-angiotensin systems

Toxicol. Lett., 381 (2023), pp. 36-47, 10.1016/j.toxlet.2023.04.010

View PDF View article View in ScopusGoogle Scholar

FULL-TEXT STUDY ONLINE AT
https://www.sciencedirect.com/science/article/pii/S0147651325001411?via%3Dihub

Research Studies

Study Tracker

Evaluation of the potential value of artificial intelligence (AI) in public health using fluoride intake as the example.

Abstract