Tracking periodontal disease with electronic dental records

Search Dental Tribune

Tracking periodontal disease with electronic dental records may enhance diagnosis and treatment

Using two algorithms they developed, US researchers were able to identify the progression of periodontal disease in patients based on their electronic dental records. (Image: Visual Generation/Shutterstock)

INDIANAPOLIS, US: Despite advances in periodontal disease research and treatments, it remains a growing health issue in the US. To address this topic, researchers from Regenstrief Institute and the Indiana University School of Dentistry in Indianapolis have developed algorithms to track periodontal disease changes through electronic dental records. This method could help dental professionals follow disease progression and diagnose the disease early—when it is potentially still reversable—and thereby reduce the risk of other systemic diseases associated with periodontal disease.

For their study, the researchers used data on 28,908 patients who had received a comprehensive oral evaluation at the dental school’s clinics between 2009 and 2014. They developed two algorithms to extract periodontal disease-related information from the patients’ electronic dental records and to classify them into three groups—patients with disease progression, patients with disease improvement and patients with no disease change. The algorithms were applied to the 15 years of electronic dental record data to generate the final patient cohorts. Both algorithms showed a high accuracy of 98%, and they have been made publicly available for use by other researchers.

“Gum disease, which is typically underdiagnosed, is reversible if caught at an early stage before it has affected underlying structures and adversely impacted tooth support. Enabling dentists to track the disease using both the information in clinical notes and the periodontal charting data contained in a patient’s electronic dental record can enable diagnosis and hope,” said co-author Dr Thankam Thyvalikakath, head of the institutions’ joint Dental Informatics programme, in a press release.

She added: “We are here to develop and establish a culture of documenting and diagnosing cases in a structured manner as is done in medicine.”

The high usage of electronic dental record systems to document patient care information provides a significant opportunity to study the clinical course of periodontal disease and the influence of risk factors. “I think the advantage of our approaches is that, using routinely collected data, we can automate and monitor gum disease treatments and changes that are visible only clinically, so we can catch gum disease at an early, potentially reversible, stage. This contrasts with other approaches that leverage only radiographs, which only show advanced gum disease,” said Dr Thyvalikakath.

The authors concluded that their study demonstrated the viability of using longitudinal electronic dental record data to track periodontal disease changes and that their algorithms were successful in classifying the three different patient cohorts using the data. This approach can be used to study the clinical course of periodontal disease using artificial intelligence, including machine learning methods.

In addition, Dr Thyvalikakath commented on the importance of tracking periodontal disease for an interdisciplinary treatment approach: “There is a bidirectional relationship between certain risk factors and gum disease. For example, having diabetes increases risk of periodontal disease and having periodontal disease negatively affects the course of diabetes. A similar relationship exists between cardiovascular disease and periodontal disease. Recognising, monitoring, and treating gum disease is an important part of overall patient health.”

The study, titled “Developing automated computer algorithms to track periodontal disease change from longitudinal electronic dental records”, was published on 8 March 2023 in the special issue Advances in Biomedical and Dental Diagnostics Using Artificial Intelligence of Diagnostics.

Leave a Reply

Your email address will not be published. Required fields are marked *