Large dental data repository created in the U.S.

A team of U.S. researchers has created a dental data repository that can be used by faculty, residents, and students to answer many questions about the nation's oral health and how best to continue improving it.

The aptly named BigMouth database has already digested detailed information from more than 1 million patients, thanks to the team members' collation of data from the electronic health records (EHRs) at several U.S. dental schools.

The data have provided insights into the differences and similarities in the institutions' dental patient populations and the academics' approaches to treating these populations. For example, during the period of 2010-2011, patients at the participating universities had high rates of caries and periodontitis, while prophylactic procedure rates were lower and varied widely among institutions.

"Having completed this successful limited launch, we look forward to expanding both the types of data to include information on medications and other parameters, and the number of institutions that contribute to BigMouth," project leader Muhammad Walji, PhD, told attendees of the recent American Medical Informatics Association's 2013 annual meeting in Washington, DC. Walji is the director of informatics for the Office of Technology Services and Informatics at the University of Texas Health Science Center at Houston (UTHealth) School of Dentistry.

Bite-size steps used to create BigMouth

Almost all U.S. dental schools have adopted EHRs, and most use the axiUm (Exan Group) platform. In addition, 32 dental institutions belong to the Consortium for Oral Health Research and Informatics (COHRI), which was created to share data and tools for achieving best practices. Walji and other informatics and dentistry experts at four of those schools -- UTHealth, Tufts and Harvard universities in Boston, and the University of California, San Francisco -- created BigMouth with a grant from the National Library of Medicine.

The group built the database using an open-source and National Institutes of Health-funded software suite called Informatics for Integrating Biology & the Bedside (i2b2). The researchers first laid out the ground rules governing the project. These stipulate, for example, that participating institutions must adhere to privacy and security requirements with respect to the data, and that a committee will be formed to review all research proposals that seek to use repository data. Institutions that are COHRI members are eligible to contribute to BigMouth, and they need to complete a data use agreement and obtain approval by an institutional review board to proceed.

The team members then proceeded to develop scripts to extract data from each institution's EHR systems and generate a file that could be loaded into the i2b2 software, which is housed on a server within the UTHealth network. The focus is on structured data: demographics, diagnoses, medical and dental histories, procedures, odontograms, periodontal charts, and treating providers. The data are deidentified except for the patients' ZIP codes and dates such as the date of birth and dental visit dates.

Next came the gargantuan job of harmonizing the terminology and content in the captured data to make it easy for people to search the repository data without being able to identify the individual institutions.

The biggest challenge was the lack of standardized forms and technologies across the participating dental schools, Walji said. They accomplished this goal thanks to such tools as the EZCodes Dental Diagnostic Terminology, but still point to the "pressing need for standardization in dentistry," he said. To this end, experts have clustered into COHRI work groups to standardize the medical and dental history forms for adults and children, and also caries risk assessment.

Database gaining traction

To access BigMouth, contributing institutions sign up at the data repository's website. Two types of access are allowed. Level 1 access enables perusal of the information in the data repository using the i2b2 Workbench -- a task made easier because the data are presented in the same hierarchy and order as they are in axiUm. Level 2 access is for people conducting research. Researchers are only granted this access after their study proposal has been approved by the review committee.

BigMouth was formally rolled out in August 2012. Walji and his colleagues created a snapshot of the data, showing variations in rates of periodontitis, preventive procedures, and therapeutic procedures among institutions.

Recently, the University of Pittsburgh School of Dental Medicine started to contribute data to BigMouth. Information from more than 1 million patients is now included in the database.

Walji and his colleagues hope members of the dentistry community will harness the data and work to both improve the quality of the information provided and probe the factors behind variances in parameters, with a goal of improving oral care across the country.