ORCA AI in Dental Research
ORCA Meeting, Heraklion, Crete, Greece
Last updated
ORCA Meeting, Heraklion, Crete, Greece
Last updated
Early childhood caries (ECC) is a significant public health issue that affects the primary teeth of young children. The World Health Organization (WHO) defines ECC as the presence of one or more decayed, missing, or filled teeth (dmft) in children under six years of age. The aim of this study was to estimate the global prevalence of ECC using the WHO criteria.
To achieve this aim, the study design involved a systematic review of studies published from 1960 to 2019. The data sources included PubMed, Google Scholar, SciELO, and LILACS. The inclusion criteria were articles that utilized the dmft index according to WHO diagnostic criteria, had calibrated examiners, used probability sampling, and had adequate sample sizes.
Two reviewers independently searched, screened, and extracted information from the selected articles. All pooled analyses were conducted using random-effects models. The review was registered with PROSPERO (registration number: CRD42014009578).
From 472 reports identified, 214 used WHO criteria, and 125 met the inclusion criteria. Sixty-four reports from 67 countries published between 1992 and 2019 provided adequate data for the meta-analysis, covering 29 countries and 59,018 children. The global random-effects pooled prevalence of ECC was 48% (95% CI: 43% to 53%).
The study found that the prevalence of ECC varied by continent, with Africa having the lowest prevalence at 30% (95% CI: 19% to 45%), followed by Europe at 43% (95% CI: 24% to 66%), the Americas at 48% (95% CI: 42% to 54%), Asia at 52% (95% CI: 43% to 61%), and Oceania having the highest prevalence at 82% (95% CI: 73% to 89%). Differences between countries explained 21.2% of the observed variance.
The study concluded that ECC is a significant global health issue, affecting nearly half of preschool children worldwide. However, the results represent data from only 29 out of 195 countries. The findings indicated substantial variability in ECC prevalence, with more variance attributable to differences between countries rather than between continents or over time.
This study highlights the need for global efforts to prevent and manage ECC in young children. The high prevalence of ECC in many countries underscores the importance of implementing effective oral health promotion programs and policies. The study also emphasizes the need for further research to understand the factors contributing to the variability in ECC prevalence between countries and to develop targeted interventions to reduce the burden of ECC in high-risk populations.
| AI Term | Definition | Example |
|---|---|---|
Artificial Intelligence (AI)
A computer system that can imitate human intelligence, such as understanding speech, making decisions, translating languages, and learning from experiences.
AI in dentistry can analyze radiographs to detect caries
Machine Learning
A field under AI, where a computer system learns to identify patterns and make predictions based on them. Uses tabular or structured datam such as databases
Machine learning is used in dentistry to analyze historical patient data and predict the probability of oral diseases based on the data available.
Deep Learning
Similar to machine learning, but uses non structured data, such as images, audio or text
Large Language Models (LLMs)
LLMs use deep learning techniques to process language and mimic human communication.
These systems can be utilized in dentistry to create conversational virtual dental assistants, guiding patients with scheduling appointments, offering reminders, or answering common oral care questions. ChatGPT is a LLM
Generative AI
Generative AI learns patterns and structures and then generates something that’s similar but new. It can make things like images, text, etc.
By applying generative AI, it is possible to create clinical scenarios, images of teeth or any other clinical material based on patients' dental records.
Hallucinations
Errors in AI generated outputs, producing results that don't align with reality.
GenAI can invent references or situations. An LLMs always responds with confidence, and sometimes gets it right.
Responsible AI
Principles and practices directing the design of AI systems to be safe and fair.
Responsible AI in dentistry implies the ethical use of AI, ensuring that it does not compromise patient privacy, and its recommendations are fair and non-discriminatory. For instance, an AI model trained on data from a specific type of patients can give innacturate or biased outputs in another kind of patient (race, SES, sex, age, etc).
Multimodal Models
A model that can work with different types, or modes, of data simultaneously.
In dentistry, multimodal models can combine visual data from radiographs and textual data from patient records to deliver predictions about a patient's oral health condition.
Prompts
An instruction entered into a system that tells the AI what task to perform.
Prompts are used to guide an AI system to analyze a radiograph and identify areas of concern. For example, a dentist could give a prompt like "identify caries".
Copilots
A personal assistant that works alongside in various digital applications.
A dental research copilot could assist researchers by searching literature, summarizing reports, and even drafting parts of research papers.
Plugins
Small software add-ons that enhance the functionality of a larger application.
Plugins can be developed for dental software to interface with AI models, enabling features like automated image analysis or patient data prediction.
