Jadbinder Seehra

Ambient voice technologies (AVT) utilise automatic speech recognition (ASR) and generative artificial intelligence for speech-to-text conversion, and are advocated as efficient, accurate tools for clinical documentation. We investigated ( n = 8) AVT systems using surgical case-reports as narratives in a controlled single-speaker evaluation prioritising safety outcomes: specifically 3 commercial transcription-systems (Dragon-Medical-One, Heidi-Health, Tortus); 4 ASR speech-to-text Application Programming Interfaces (Speechmatics-Enhanced, Amazon-Medical-Transcribe, Whisper, GPT4oTranscribe); and an experimental two-stage ASR–Large Language Model (LLM) pipeline incorporating GPT4oTranscribe with GPT-5-LLM generative error correction (GPT4oTranscribe-Corrected-5). Reference-transcripts ( n = 100; 32,897 words, range = 44–449, mean = 329/per-transcript) derived from surgical case-reports were recorded and input into AVT systems as audio-recordings for transcription-output generation. Primary outcome was proportion of transcription-outputs containing at least one clinically significant Class 3 error graded for potential harm. Secondary outcomes included transcription accuracy: Domain-Word-Error-Rate (DWER) against SNOMED-CT, lexical-accuracy (ROUGE score) and semantic similarity (BERT, BART scores). To investigate impact of the experimental pipeline on clinically significant errors, raw ASR and LLM-processed transcription-outputs were compared and errors classified as resolved, remaining or newly introduced. Across ( n = 800) transcript-outputs, 30–68% (GPT4oTranscribe-Corrected-5; Amazon-Medical-Transcribe and Dragon-Medical-One, respectively) contained at least one clinically significant Class 3 error. LLM-processing reduced the proportion of transcript-outputs affected (53% to 30%), amongst 89 clinically significant errors in ASR-output, 51 were resolved, 38 remained significant and 4 were newly introduced. Reference-transcript length significantly increased odds of Class 3 errors for 4 systems (GPT4oTranscribe, Tortus, Amazon-Medical-Transcribe, Dragon-Medical-One; odds ratio 2.65–3.34/additional 100 reference-words). Errors with potential to cause severe harm or death (NHS England Level 3–4) totalled ( n = 53) across systems, concentrated in the domains of medication-type/dose (39.6%), and investigations/laboratory results (30.2%). Performance varied across systems for transcription metrics ( P < 0.001). GPT4oTranscribe-Corrected-5 (DWER = 3.60%) and Heidi-Health (DWER = 5.67%) performed best for medical terminology; Amazon-Medical-Transcribe worst (DWER = 24.03%). For lexical and semantic similarity GPT4oTranscribe-Corrected-5 performed best, followed by Heidi-Health. The ASR–LLM pipeline reduced clinically significant errors but did introduce new ones in 3% of transcript-outputs. The safety and effectiveness of these systems in clinical practice requires further evaluation.

Objective: The aim of this study was to assess the aesthetic judgements made by lay people and professionals relating to an asymmetric maxillary central incisors' gingival margin position. Setting: Multi‐center institutional study. Materials and Methods: A high‐quality intra‐oral photograph of a previously treated case was manipulated (six images). Parents of children undergoing conventional orthodontic treatment (PCT), parents of children undergoing combined surgical‐orthodontic treatment for a unilateral impacted maxillary central incisor (PI) and professionals (dental and orthodontic specialists) were asked to evaluate and rank the images in terms of aesthetics and provide free‐text responses to support their rankings (least and most attractive). All data was analysed statistically with regressions at 5% and a thematic analysis of the free‐text responses was performed. Results: The responses from 120 participants (mean age 42.2 years; Standard Deviation [SD] 8.6 years; 62.2% female) were analysed (30/group). Image E (symmetrical gingival margins with the lateral incisor gingival margins 0.5–1.0 mm below both the gingival margin of the maxillary canine and central incisor which are at the same level in relation to each other) was ranked the most attractive (60.8%). The presence of an asymmetric gingival margin between the maxillary central incisors was rated the least attractive (Images D and F; 30.9% and 39.2%, respectively). The highest and lowest mean ranking scores (adjusted for gender) were Image E (5.12 points; 95% CI: 4.76–5.48 points) and F (1.83 points; 95% CI: 1.57–2.10 points) A similar trend was evident within groups. The PCT group scored Image D the lowest (2.17 points; 95% CI: 1.72–2.62 points). Gender influenced the ranking of the images. Themes to support the ranking of the most attractive image were related to the importance of overall symmetry. Themes to support the ranking of the least attractive image were based on overall asymmetrical (uneven) appearance ( n = 117). Conclusion: An asymmetric gingival margin position between the maxillary central incisors negatively influences the rating/perception of attractiveness by lay persons and professionals.

Accurate clinical records are fundamental to dental practice. Automatic speech recognition (ASR) has the capacity to convert spoken clinical language into written text within the electronic health record; however, the accuracy of ASR in natural language processing for clinical dentistry remains uncertain. The aim of this study was to investigate the transcriptional accuracy of ASR systems using orthodontic clinical records as the experimental model. Specifically, we used 4 commercial ASR systems (Heidi Health, DigitalTCO, Dragon Medical One, Dragon Professional Anywhere), 5 application programming interfaces (Amazon, Google, Speechmatics, Whisper, GPT4oTranscribe), and a 2-stage pipeline coupling GPT4oTranscribe with the GPT4o large language model (LLM) for generative error correction (GPT4oTranscribeCorrected). Orthodontic diagnostic and treatment planning summaries ( n = 200; 10 subject domains; 43,408 words; 6 h of audio) were narrated and recorded for analysis. The primary outcome was domain word error rate (DWER), which investigates clinical terminological transcription errors against the Systematized Nomenclature of Medicine Clinical Terms (SNOMED-CT) database. Secondary outcomes included nondomain WER (N-DWER), lexical accuracy (Recall-Oriented Understudy for Gisting Evaluation [ROUGE] score), semantic similarity (Bidirectional Encoder Representations from Transformers [BERT] and Bidirectional and Auto-Regressive Transformer [BART] scores), hallucinations (transcribed text not in the spoken input), and qualitative error analysis. GPT4oTranscribeCorrected was transcriptionally most accurate (DWER = 3.5%; WER = 3.7%), with DWER decreasing by 54.9% versus GPT4oTranscribe. Heidi Health was the highest-performing commercial system (DWER = 6.2%; WER = 5.4%), with Dragon Professional Anywhere being the worst (WER = 33.9%). All systems were less accurate with technical vocabulary (DWER > N-DWER; P < 0.001), except GPT4oTranscribeCorrected. Significant differences were seen across systems for ROUGE, BERT, and BART scores ( P < 0.001). Based on post hoc pairwise comparisons, GPT4oTranscribeCorrected performed best and Dragon Professional Anywhere was consistently worst for lexical and semantic errors. Hallucinations were absent except for Whisper ( n = 57) and DigitalTCO ( n = 1). Across systems, background noise increased DWER and WER ( P < 0.001). Importantly, clinically significant errors were seen with all systems, ranging from 2% to 66% (GPT4oTranscribeCorrected clean; Dragon Medical One background noise, respectively). Variation in narrator accent had no effect in clean conditions ( P = 0.65) and a small effect with background noise ( P = 0.001). ASR systems deliver single-digit transcription error rates, particularly when coupled with LLM-based correction, but clinically significant errors persist. The verification of clinical records is essential when using current ASR systems.

Background: P-values convey statistical significance while effect estimates and confidence intervals (CIs) place emphasis on the clinical significance. The aim of this study was to determine the frequency of reporting clinical or statistical significance in European Orthodontic Society (EOS) Congress scientific poster abstracts and to ascertain whether the conclusions drawn are based on either clinical or statistical significance. Materials and methods: Abstracts published between 2014 and 2024 were included. Pre-piloting and calibration were undertaken prior to data collection. Abstract characteristics were extracted independently by two reviewers. Descriptive statistics and frequency distributions were calculated. Results: A total of 3654 abstracts were analysed. The highest number of abstracts were presented in 2018 (13.0%). Epidemiological studies (cross-sectional, case-control, cohort, survey) (54.7%) were frequently presented, followed by laboratorial studies (17.4%) and systematic reviews (7.7%). No inferential statistics were commonly reported (58.5%). Within the abstracts reporting statistical significance (41.1%), typically only P-values were reported (32.5%), followed by the term ‘statistically significant’ stated only (3.7%), P-values in conjunction with estimates and 95% CIs (1.7%), P-values and 95% CIs (1.4%), 95% CIs (0.8%), estimates and 95% CIs (0.6%), and P-values and estimates (0.4%). When interpreting the reported results in the conclusion section, these were typically based on P-values (31.2%) or the term ‘statistically significant’ stated only, without consideration of outcomes between groups (14.7%). 95% CIs (0.7%), P-values and 95% CIs (1.3%), P-values in conjunction with estimates and 95% CIs (1.6%) and estimates and 95% CIs (0.5%) were infrequently considered. Across the study timeframe (per year), the results and author conclusions tended to be based on P-values primarily. Limitations: Only one society congress was assessed which may impact the generalizability of the results. Conclusions: Clinical significance is often under-reported in abstracts presented at EOS. Where applicable the reporting of clinical significance (effect size with CIs) and their interpretation in poster abstracts should be stipulated as this allows clinicians to gauge both the size and range of the observed differences between groups and the relevance to their clinical practice.

Background: Throughout a trial report, clear and accurate reporting is essential. The aim of this study was to assess the reporting quality of RCT abstracts presented at the European Orthodontic Society (EOS) Congress between 2015–2024. Associations between reporting quality and abstract characteristics were explored. Material and methods: All EOS Congress scientific abstracts published between 2015–2024 (2020 excluded) were included. Descriptive statistics and frequency distributions were calculated. Mean values for adequate reporting per CONSORT item and sum score were calculated. On an exploratory basis, univariable linear regression between summary score and abstract characteristics was undertaken. Results: 139 RCT congress abstracts were analysed. The most frequent years of RCT abstract presentation were 2018 (14.4%) and 2023 (19.4%). RCT abstracts were more likely to be poster types (62.6%), with corresponding authors based in Europe (77.7%), single centre (64.7%) and reporting a non-significant result for the primary outcome (54.0%). The mean overall total quality reporting score was 17.6 (SD 5.6, min 15 and max 22.8) out of a maximum score of 50. Items that tended not to be reported include authors contact details, participant (settings), randomization procedures and trial registration. Posters achieved lower scores compared to oral abstracts (−2.96; 95% CI −4.86, −1.07; P < 0.01). Additionally, an abstract word count greater than 251 words was associated with higher total score (3.28; 95% CI 0.74, 5.82; P = 0.012). A weak association (0.29, 95% CI: −0.03–0.62, p = 0.07) between year of abstract publication and an improvement in overall reporting score over time was evident. Limitations: Only one society congress were assessed which may impact the generalisability of the results. Conclusions: In relation to the CONSORT reporting Randomized Controlled Trials in Journal and Conference Abstracts checklist, RCT abstracts presented at EOS congress between 2015–2024 remain sub-optimal with reporting of key items lacking. However, reporting quality scores have shown an association with type of presentation (oral or poster), abstract word count and some evidence of improvement over time. Measures to encourage clear and consistent RCT abstract reporting are required.

Background: Issues regarding research transparency have been highlighted in the literature. The aim of this investigation was to assess the trend of adhering to the reporting of key transparency practices (data sharing, code sharing, COI disclosure, funding disclosure, and protocol registration) by articles published in three leading orthodontic journals. Material and methods: A Scopus database search was undertaken to identify articles published in 2013, 2018 and 2023 in three orthodontic journals. Independent data extraction was performed. Frequency distributions and cross tabulations (COI not disclosed vs COI disclosed) were calculated. Univariable logistic regression was undertaken to detect associations between COI disclosure and publication characteristics, journal type and reporting of transparency practices. Results: 1498 articles were analysed. No data sharing statement (60.7%) was commonly cited. The intention to share data was stated in approx. 15% articles. No article stated a code sharing statement. Almost 70% articles had not registered a protocol. A third of articles disclosed the funder or sponsor (32.7.%), while approximately an equal number of articles disclosed (49.1%) or did not disclose (50.9%) any conflict of interest. Articles published in the EJO had higher odds of reporting protocol registration (OR 7.42; 95% CI: 4.55, 12.10; P < .01) and funding disclosure (OR 3.44; 95% CI: 2.52, 4.69; P < .01) compared to AO and AJODO. A COI disclosure statement was more likely to be in articles published in AJODO. The odds of reporting a data sharing statement were higher in articles published in EJO (OR 8.72; 95% CI: 5.72, 13.29; P < .001). Apart from code sharing, improvements over the 5-year intervals in reporting of transparency indicators were evident. Limitations: Only three journals were assessed which may impact the generalisability of the results. Conclusions: Despite improvements over the 5-year intervals, the reporting of research transparency indicators requires improvement in articles published in orthodontic journals. To encourage transparent research conduct and open science practices, journals and their editors should promote reporting of research transparency indicators.

Background/Objectives: The inclusion of a participant flow diagram in randomized clinical trials (RCTs) is a requirement of the CONSORT guidelines. The aim of this study was to assess the reporting quality of flow diagrams of RCTs published in orthodontic journals in relation to the CONSORT Flow Diagram for Parallel Group RCTs. Materials/Methods: RCTs published between January 2011 and December 2023 in five orthodontic journals were identified and trial characteristics were extracted. The reporting of the flow diagram (if included) was assessed for completeness in relation to the CONSORT flow diagram template. Descriptive statistics and cross tabulations between RCT characteristics and presence/no presence of a flow diagram were performed. On an exploratory basis, univariable associations between RCT characteristics and presence/no presence of a flow diagram were performed and univariable logistic regression to examine the effect of publication year on flow diagram reporting. Results: Three hundred and thirty-four RCTs met the inclusion criteria. The majority were published in 2021 (n = 39, 11.7%), and had 2 arms (n = 279, 83.5%). Three-hundred and seven (92.0%) RCTs were published in journals endorsing the CONSORT guidelines. Two hundred and thirty-three (69.8%) RCTs included a flow diagram and from these, 48.1% (n = 112) were fully compliant with flow diagram reporting. 121 (51.9%) omitted at least one item of the CONSORT reporting template. Significant associations between journal type, CONSORT endorsement by authors, ethical approval status, presence of a published protocol, significance of the primary outcome, involvement of a statistician, presence of conflict of interest, center type, type of analysis undertaken and the presence/ no presence of a flow diagram were present. Across the study timeframe, the odds of inclusion of RCT flow diagram increased per additional year (OR:1.47; 95%CI:1.34,1.61; p < .001). Limitations: Only five orthodontic journals. Conclusions/Implications: Despite improvements over time, the inclusion and reporting of CONSORT flow diagram for parallel group RCTs in trials published in orthodontic journals requires improvement. To mitigate potential biased interpretation of trial results, journal editors should ensure a complete CONSORT flow diagram is submitted by researchers.

Background: Maxillary lateral incisor agenesis is often managed with orthodontic space closure and canine substitution. Anatomic gingival margin heights associated with the maxillary anterior teeth are considered important for achieving aesthetic excellence with space closure, but evidence relating to the perceptions of lay people is poor. Objectives: This study investigated the influence of gingival margin height variation in the maxillary anterior teeth following orthodontic space closure and canine substitution in the absence of maxillary lateral incisor teeth on perceived aesthetics judged by a lay population. Material and Methods: Images of a case treated with space closure and bilateral canine substitution were digitally modified to create gingival margin height variation for the substituted lateral incisors and canines. Six variations were created, including one conforming to accepted norms for anatomic gingival margin heights. Lay people represented by parents of orthodontic patients ranked the images based on aesthetics, with data analysed statistically using linear models at 5%. Results: One hundred and twenty responders were included (median age 48.0 years), 50% male and 81.7% Caucasian, with 87% able to provide specific preferences. In absolute terms, the highest-ranking [image C] had substituted lateral incisor gingival margins below the central incisors (considered normal) but substituted canine margins symmetrically below the substituted lateral incisors (considered to be too low); followed by [image D] with symmetrically level gingival margins; and [image F] with asymmetric substituted lateral incisor margins and substituted canine margins symmetrically too low. The lowest-ranked [image E] had normal substituted lateral incisor margins but asymmetric substituted canine margins. Image C was most often ranked first (29.8% of responders) and image E last (22.1%). After adjusting for potential confounding effects, image E received the worse scores. Conclusions: Anatomic norms for gingival margin height in absent maxillary lateral incisor-canine substitution do not correlate with the highest-ranking aesthetic choices of a lay population. Lateral incisor gingival margins symmetrically below the central incisors and substituted canine margins symmetrically below these ranked highest. Vertical asymmetry in the substituted canine position was considered the least aesthetic.

Palatal impaction of the maxillary canine is commonly considered as a genetic condition, while the lateral incisor root has been suggested to guide maxillary permanent canine eruption. This review assessed whether agenesis of the maxillary lateral incisor is associated with eruption disturbances of the maxillary canine (displacement or impaction). Unrestricted searches in six databases for observational case–control human studies until May 2024 assessing eruption disturbances of the permanent maxillary canine in conjunction with permanent maxillary lateral incisor agenesis were performed. Study selection, data extraction and risk‐of‐bias assessment with the Joanna Briggs Institute's tool were performed in duplicate. Random‐effects meta‐analyses of odds ratios (ORs) with their 95% confidence intervals (CIs) were performed, followed by meta‐regressions and sensitivity analyses. Seventeen reports pertaining to 17 case–control studies were included covering 10,155 patients (11.6 year‐old/46% male on average). Maxillary lateral incisor agenesis was associated with eruption disturbance of the maxillary canine (17 studies; OR = 3.88; 95% CI = 2.05–7.34; p < 0.001; I ² = 58). Lateral incisor agenesis was more associated with canine displacement (7 studies; OR = 7.56; 95% CI = 2.10–27.18; p = 0.008; I ² = 57) and to a lesser degree with canine impaction (10 studies; OR = 2.39; 95% CI = 1.28–4.48; p = 0.01). No significant effects of patient age or sex were seen, and the results were robust to study design and precision. The quality of evidence was low due to the inclusion of observational case–control studies with methodological issues and residual confounding. Evidence indicates that permanent maxillary lateral incisor agenesis is associated with eruption disturbances of the permanent maxillary canine, including mainly displacement, but also impaction. Trial Registration: PROSPERO database number: CRD42024549898