Diagnostic accuracy of teledentistry for dental caries among children and adolescents (6-18 years): a systematic review

Teledentistry can be used to enhance access to dental care by reducing treatment and travel costs. It can also improve the quality of dental care by sharing patient information remotely with dental specialists and getting their advice in diagnosis and treatment planning. This systematic review is aimed to assess the diagnostic accuracy of teledentistry compared to the standard visual inspection for dental caries among children and adolescents. We developed search strategies for different databases (PubMed, CINAHL, Dentistry and oral sciences source) using MeSH terms and keywords. Two reviewers independently screened and extracted the studies against the eligibility criteria. The quality assessment for diagnostic accuracy studies-2 tool was used to assess the risk of bias of the included studies. In case of any disagreement, the third reviewer was consulted. We included seven studies that met the inclusion criteria. six studies reported on validity and one study reported on the feasibility of teledentistry in diagnosing dental caries. The pooled sensitivity of teledentistry, when compared to visual examination was 82% (95% CI: 81-84%), and specificity was 90% (95% CI: 89-92). In one study from Australia, it was reported that the total estimated cost for teledentistry and visual examination was 50 million dollars and 135 million dollars, respectively. According to the study findings, teledentistry was 2.7 times less expensive than the visual examination method. Based on the limited findings, it is concluded that teledentistry has the potential for diagnosing dental caries in children and adolecents, but further studies in varying contexts are needed to ascertain its benefits especially in low- and middle- income countries.

Methods
Search strategy and study selection

We developed search strategies for different databases (PubMed, CINAHL, Dentistry and oral sciences source) using medical subheading (MESH) terms and keywords. The references of the systematic reviews and other identified articles were also searched manually to find any relevant article that might have been missed in the database search. To manage the records of the studies imported from the electronic databases, a citation management system (Endnote X9 software) was used. The studies were later imported into Covidence (a web-based systematic review program) for screening. After the de-duplication of studies in Covidence, two reviewers (YA and MKR) carried out the title and abstract screening to identify relevant studies. The same reviewers carried out the full-text screening of the screened studies against the eligibility criteria for final inclusion of studies for data extraction. Two reviewers carried out the whole screening process, and any disagreement between the two reviewers was resolved by discussion. A third reviewer (JKD) was consulted for a final decision in case of any disagreement. We followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines.

Data extraction

We used double data extraction whereby two reviewers (YA and MKR) independently extracted data from the included studies on a designed data extraction form on Microsoft Excel. Discrepancies were resolved through discussion. The data extraction sheet included information about the study (first author, year of publication, journal, study title, country, income region of the country according to the world bank, aim of the study, study design, study setting, sample size, the participants (age range), the index test, the reference standard, the results (sensitivity and specificity), study limitations, funding source of the study, conflict of interest, and the comments by the reviewer).

远程牙科（Teledentistry）可通过降低治疗与出行成本，提升牙科诊疗服务的可及性。其还可通过与牙科专科医师远程共享患者信息、获取其诊断与治疗计划建议，优化牙科诊疗质量。本系统综述旨在评估远程牙科相较于标准目视检查，在儿童与青少年龋齿诊断中的诊断准确性。我们利用医学主题词（MeSH）与关键词，为PubMed、CINAHL、牙科与口腔科学数据库（Dentistry and oral sciences source）构建了检索策略。两名评价者独立按照纳入排除标准对研究进行筛选与数据提取。本研究采用诊断准确性研究质量评估工具-2（quality assessment for diagnostic accuracy studies-2 tool）评价纳入研究的偏倚风险。当两名评价者出现意见分歧时，由第三名评价者进行裁决。本综述共纳入7项符合纳入标准的研究：6项报告了远程牙科诊断龋齿的有效性，1项报告了其临床应用可行性。相较于目视检查，远程牙科的合并灵敏度为82%（95%置信区间：81%~84%），合并特异度为90%（95%置信区间：89%~92%）。澳大利亚的一项研究显示，远程牙科与目视检查的总预估成本分别为5000万美元与1.35亿美元。根据该研究结果，远程牙科的成本仅为目视检查法的约37%，即较目视检查法便宜约2.7倍。基于现有有限的研究结果，本综述得出结论：远程牙科具备应用于儿童与青少年龋齿诊断的潜力，但仍需在不同场景下开展进一步研究，以明确其应用价值，尤其在低收入与中等收入国家。

方法
检索策略与研究筛选
我们利用医学主题词（MeSH）与关键词，为PubMed、CINAHL、牙科与口腔科学数据库构建了检索策略。此外，我们手动检索系统综述与其他已识别文献的参考文献列表，以补充数据库检索中可能遗漏的相关文献。为管理电子数据库导入的研究记录，我们采用文献管理系统Endnote X9软件。随后将研究记录导入Covidence（一款基于网页的系统综述管理程序）进行筛选。在Covidence中完成研究去重后，两名评价者（YA与MKR）先后开展标题与摘要筛选，以识别相关研究。随后，两名评价者按照纳入排除标准对筛选出的文献进行全文筛查，以最终确定纳入数据提取的研究。整个筛选流程由两名评价者独立完成，若二者间出现意见分歧，可通过讨论解决；若仍无法达成一致，则由第三名评价者（JKD）作出最终裁决。本综述遵循系统综述与Meta分析首选报告条目（PRISMA）指南。

数据提取
我们采用双人独立数据提取法：两名评价者（YA与MKR）通过预先设计的Microsoft Excel数据提取表，独立从纳入研究中提取相关数据。数据提取过程中出现的分歧通过讨论解决。数据提取表涵盖以下信息：研究第一作者、发表年份、期刊、研究标题、研究开展国家、世界银行划分的该国收入层级、研究目的、研究设计、研究场景、样本量、研究对象（年龄范围）、目标检测方法、参照标准、研究结果（灵敏度与特异度）、研究局限性、研究资助来源、利益冲突情况，以及评价者备注。