“AI is aimed at augmenting rather than replacing oral health practitioners’ judgement and skills”

Anisha Hall Hoppe, Dental Tribune International

Tue. 3 September 2024

share

save

A new review article by US researchers explores the inequities and biases in oral healthcare and highlights the potential of artificial intelligence (AI) in addressing the challenge of disparities in the provision of oral healthcare to low-income and minority populations, supporting care that is personalised, unbiased and transparent. Dental Tribune International spoke with two of the researchers, Drs Ahmed S. Sultan and Zaid H. Khoury, about their findings in bringing AI to the fore to help populations in need.

Dr Ahmed S. Sultan. (Image: Ahmed S. Sultan)

Dr Khoury, can the use of AI in oral healthcare help in identifying and reducing biases that affect treatment outcomes for diverse patient populations?
Dr Khoury: AI can aid the oral healthcare provider by generating a treatment plan with alternative approaches that is patient-centred and considers the linguistic and health background of that particular patient. This way, the patient would have a better understanding of the cost-effectiveness of each personalised treatment modality as it fits the patient’s needs and thus will be more involved in the informed consent process. Real-time sharing of the tentative treatment plan with other dental specialists can further enhance the definitive treatment plan through interdisciplinary collaboration. Therefore, AI can achieve better oral healthcare accessibility for diverse populations through multidisciplinary care and treatment options that suit the patient’s financial affordability.

What practical steps can dentists take to incorporate interpretable and explainable AI models into their practice to enhance treatment transparency and patient trust?
Dr Khoury: Dental providers may feel uncertain about incorporating AI technologies into their dental office and may perceive using these to be complex so I think that the first step is the oral healthcare provider understanding the basics and the value of explainable AI models. This can be achieved by continuing education through webinars and searching the relevant literature.

Patients and providers must recognise that explainable AI is aimed at augmenting rather than replacing oral health practitioners’ judgement and skills. With the help of software programmers, providers can then employ deep learning and other machine learning apps to analyse various data sets within their practice, whether derived from the electronic health record or imaging, to generate explainable predictions that better achieve personalised oral care for each patient, augmenting the provider’s skills and acting as a second opinion for the patient and ultimately reducing bias.

“AI can achieve better oral healthcare accessibility for diverse populations.”—Dr Khoury

Dr Zaid H. Khoury. (Image: Zaid H. Khoury)

How does the integration of AI in dental diagnostics and treatment planning improve access to quality care for underserved communities?
Dr Khoury: There is a lack of dental providers, whether general practitioners or specialists, in underserved and remote areas. Although the majority of AI-assisted diagnostic apps are still in the early research phase, they have shown promise, particularly when employed in outreach environments, where the alternative sometimes is a total absence of diagnostic input from an oral healthcare provider. With the presence of a basic internet connection, explainable AI models can generate predictions based on analysis of captured imaging data of the soft tissue or radiographs, directing the provider on what to do next or indicating the urgency of referral to a specialist. There is an unmet need for this service, and it will ensure that patients seek appropriate care in a timely manner for dental caries, periodontal disease or oral cancer, enhancing the quality of care in these communities.

What specific challenges do AI technologies address in managing oral disease, and how can these technologies enhance preventive care in routine dental visits?
Dr Sultan: For oral mucosal disease, some AI technologies assist the general dentist in identifying the clinical characteristics of precancerous conditions; however, most of these technologies have been validated on previously acquired imaging data as opposed to real-time intra-oral scanning and instant prediction. In this example, a patient can consent to the provider taking an intra-oral picture with a camera that can be analysed in a separate step using one of the AI-assisted diagnostic algorithms, generating a prediction on the likelihood of cancerous transformation. This will aid the general practitioner in the decision on whether a biopsy is required. Similar AI-powered systems to detect dental caries and periodontal disease exist and can be used to detect early disease for proper preventive or other therapeutic measures.

“The language barrier is a big factor that can negatively affect patients’ accessibility, understanding and comfort during oral healthcare visits.”—Dr Sultan

In what ways can AI-driven tools assist dentists in overcoming language barriers and improving communication with patients of various linguistic backgrounds?
Dr Sultan: Let’s take the US. The US has the largest population of foreign-born nationals of any country, making it increasingly likely for the provider to encounter diverse patient populations that speak various languages. The language barrier is a big factor that can negatively affect patients’ accessibility, understanding and comfort during oral healthcare visits. Employing AI interpretation language apps such as Google’s Translator AI GPT, whether offered at the clinic or through the patient’s smartphone, will ultimately improve the patient’s understanding and active involvement in treatment planning and improved treatment outcomes. This way patients can not only translate spoken words from a different language but also simplify medical and dental terminology, making it a less overwhelming experience for the patient.

How can dental practitioners ensure the responsible use of AI in their practices to avoid exacerbating existing inequities in oral healthcare?
Dr Sultan: This must start at the development stage, and AI models should be trained on diverse data sets. Once validation for the AI technologies is finalised and gets approved by the US Food and Drug Administration, there will be a need to formulate a standardised protocol for application across various healthcare domains, including dentistry. Providing such diagnostic AI aids to all patients at a minimal cost (if any or through insurance) would be one avenue to avoid exacerbating existing inequities. To achieve this, a concerted effort from governmental and corporate stakeholders is required.

“Our core focus is ethical AI and striving to maintain the highest standards for the responsible use of AI.”—Dr Sultan

Is there anything else you would like our readers to know about the present paper or any other ongoing research you’re involved in?
Dr Sultan: Yes, the Division of Artificial Intelligence Research at the University of Maryland School of Dentistry is focused on exploring the incorporation of both explainable and interpretable AI models in AI research. This involves unravelling the opaque black box of AI models and designing rationale-based decision-making models to build trust for researchers and patients. Our core focus is ethical AI and striving to maintain the highest standards for the responsible use of AI, including reducing biases and oral healthcare inequities that affect treatment outcomes for diverse patient populations.

Topics:

Tags: