**Background:** Prisoners face significant barriers to accessing oral healthcare due to security constraints, limited dental personnel, and logistical challenges. In Thailand, the 'Good Health Good Heart' project aims to improve prisoner health, but dental services remain limited, especially outside Bangkok. Teledentistry using intraoral cameras (IOCs) has been proposed as a way to increase access to oral health screening in prisons, but its diagnostic accuracy compared to direct examination by a dentist had not been rigorously evaluated.
**Methods:** This crossover study was conducted at Sisaket Provincial prison in Thailand. Eight of 45 male prisoner health volunteers (PHVs) underwent a 1-day, 3-hour teledentistry training program, including calibration against a dentist (requiring ≥0.80 Kappa agreement). One trained PHV was randomly selected for data collection. From 1621 male prisoners, 309 volunteered for screening; 157 reported dental problems and were eligible. Five were excluded (2 for health issues, 3 for blurred images), leaving 152 prisoners with 215 symptomatic teeth. In Phase II, the PHV used an IOC (2-megapixel resolution) to capture images of symptomatic areas, which were independently assessed by the PHV and a dentist for treatment needs (dental filling, scaling, extraction, surgical removal of impacted tooth). In Phase III, a second dentist performed direct oral examinations to determine actual treatment needs. Sensitivity, specificity, PPV, and NPV were calculated using direct examination as the reference standard.
**Key Results:** The mean age of participants was 33.6 years (SD 9.1, range 19–67). Chief complaints included toothache (38.2%), tooth cavity (27.6%), broken tooth (23.0%), tooth mobility (8.5%), and tooth sensitivity (2.0%). Most symptomatic teeth were in the mandibular posterior region (51.6%). For teledentistry examination by the dentist compared to direct examination, sensitivity and specificity were above 90% for all treatment categories: dental filling (sensitivity 97.1%, specificity 97.2%), dental scaling (100%, 100%), simple extraction (93.9%, 98.0%), and surgical removal (100%, 98.6%). PPV ranged from 81.8% to 100% and NPV from 83.1% to 100%. For teledentistry examination by the PHV, sensitivity was 97.1% for dental filling but only 33.3% for dental scaling and 33.3% for surgical removal. Specificity was above 80% for all categories. PPV was 50.0% for dental filling, 40.0% for dental scaling, 94.7% for simple extraction, and 60.0% for surgical removal. NPV was 99.3% for filling, 98.1% for scaling, 51.8% for extraction, and 97.1% for surgical removal.
**Clinical Implications:** Teledentistry using an IOC can facilitate dental disease screening in prisons with acceptable diagnostic accuracy when performed by a dentist, but accuracy is substantially lower when performed by a PHV, particularly for detecting periodontal treatment needs and distinguishing between restorable and non-restorable teeth. The study highlights that intraoral photographs alone cannot capture subgingival calculus, bleeding on probing, tooth mobility, or pulpal health, limiting definitive diagnosis. The authors suggest that PHVs may need additional training, and that real-time teledentistry consultation between PHVs and dentists could improve accuracy. Despite limitations, teledentistry screening could help prioritize treatment needs, reduce the burden on mobile dental units, and be integrated into prison admission protocols, especially in settings with high turnover and limited dental personnel.