Prognostic factors identification using machine learning in temporomandibular disorder treatment responders

Background: Temporomandibular disorders (TMD) are complex chronic conditions that significantly impair quality of life and impose a considerable social and economic burden. Although various treatment strategies have been developed, the prognostic factors influencing therapeutic outcomes remain poorly defined. This study aimed to identify relevant prognostic factors for treatment response using machine learning methods, with patient readiness for discharge serving as the primary outcome. Methods: A total of 1050 medical records from patients diagnosed with TMD and treated at the Orofacial Pain and Dental Sleep Medicine clinic, between January 2018 and June 2023, were retrospectively analyzed with a follow-up period of one year. Twenty-six clinical and demographic variables were initially extracted and preprocessed using one-hot encoding. After the removal of highly correlated variables, the final dataset comprised 36 features derived from encoded categorical variables. Seven machine learning algorithms, namely Decision Tree, Random Forest, Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Classifier (SVC), Gradient Boosting, and Extreme Gradient Boosting (XGBoost), were trained to predict discharge readiness. Model performance was evaluated using accuracy, precision, recall, F1 score, and the receiver operating characteristic-area under the curve (ROC-AUC). Feature importance was assessed using information gain and Shapley Additive exPlanations (SHAP) to interpret model predictions. Results: The Random Forest model demonstrated superior performance in predicting readiness for discharge, with an accuracy of 0.7901, precision of 0.8611, recall of 0.8943, F1 score of 0.7901, and ROC-AUC of 0.6442. SHAP analysis identified onset duration, pain severity, anxiety level, presence of neck pain, and body mass index (BMI) as the most influential predictors. Conclusions: The identified prognostic factors highlight the multidimensional nature of TMD and support their relevance in guiding patient-specific management strategies. The integration of machine learning approaches may enhance clinical decision-making and contribute to the development of more personalized and effective treatment pathways for TMD.

Machine learning; Model selection; Prognostic factors; Temporomandibular disorders; Treatment outcome

[1] Asquini G, Devecchi V, Viscuso D, Bucci R, Michelotti A, Liew BX, et al. An exploratory data-driven approach to classify subgroups of patients with temporomandibular disorders based on pain mechanisms. The Journal of Pain. 2025; 26: 104721

[2] Herrero Babiloni A, Exposto FG, Peck CM, Lindgren BR, Martel MO, Lenglet C, et al. Temporomandibular disorders cases with high-impact pain are more likely to experience short-term pain fluctuations. Scientific Reports. 2022; 12: 1657.

[3] Manfredini D, Favero L, Gregorini G, Cocilovo F, Guarda-Nardini L. Natural course of temporomandibular disorders with low pain-related impairment: a 2-to-3-year follow-up study. Journal of Oral Rehabilitation. 2013; 40: 436–442.

[4] Smith T, Singh G, Mcnamee G, Newton C. Musculoskeletal physiotherapists’ discharge practices for people treated with low back pain: a United Kingdom survey. Musculoskeletal Care. 2024; 22: e1851.

[5] Staniszewski K, Willassen L, Berge T, Johansson A, Rosen A, Schjødt B. High pain intensity is a risk factor of non-resolving TMD: a three-year follow-up of a patient group in a Norwegian interdisciplinary evaluation program. Journal of Pain Research. 2022; 15: 1283–1296.

[6] Kassam S, Wong E, Thompson M, Tran T, Bosma R, Sheffe S. “Discharge doesn’t mean the end” exploring success in discharge to community self-management for young adults living with chronic pain: a qualitative study. Canadian Journal of Pain. 2024; 8: 2346943.

[7] Meloto CB, Slade GD, Lichtenwalter RN, Bair E, Rathnayaka N, Diatchenko L, et al. Clinical predictors of persistent temporomandibular disorder in people with first-onset temporomandibular disorder. The Journal of the American Dental Association. 2019; 150: 572–581.e10.

[8] Jha N, Lee KS, Kim YJ. Diagnosis of temporomandibular disorders using artificial intelligence technologies: a systematic review and meta-analysis. PLOS ONE. 2022; 17: e0272715.

[9] Cui Y, Kang F, Li X, Shi X, Zhang H, Zhu X. Predicting temporomandibular disorders in adults using interpretable machine learning methods: a model development and validation study. Frontiers in Bioengineering and Biotechnology. 2024; 12: 1459903.

[10] Hunter DJ, Holmes C. Where medical statistics meets artificial intelligence. New England Journal of Medicine. 2023; 389: 1211–1219.

[11] Pudjihartono N, Fadason T, Kempa-Liehr AW, O’Sullivan JM. A review of feature selection methods for machine learning-based disease risk prediction. Frontiers in Bioinformatics. 2022; 2: 927312.

[12] Molnar C. Interpretable machine learning: a guide for making black box models explainable. 3rd edn. Leanpub: British Columbia. 2025.

[13] Schiffman E, Ohrbach R, Truelove E, Look J, Anderson G, Goulet J, et al. Diagnostic criteria for temporomandibular disorders (DC/TMD) for clinical and research applications. Journal of Oral & Facial Pain and Headache. 2014; 28: 6–27.

[14] World Health Organization. International statistical classification of diseases and related health problems, 10th revision (ICD-10). 2016. Available at: https://icd.who.int/browse10/2016/en (Accessed: 22 January 2025).

[15] International Labour Office. The International standard classification of occupations (ISCO-08) companion guide. Geneva; 2023. Available at: https://webapps.ilo.org/ilostat-files/ISCO/newdocs-08-2021/ISCO-08/ISCO-08%20EN%20Vol%201.pdf (Accessed: 22 January 2025).

[16] Choi S, Yoon S, Lee H. Beyond measurement: a deep dive into the commonly used pain scales for postoperative pain assessment. The Korean Journal of Pain. 2024; 37: 188–200.

[17] Sornsenee P, Kongtragulsub K, Watcharajiranich K, Chantanuwat R, Aungchayakul A, Mangkhalathat K, et al. Factors associated with anxiety and depression among micro, small, and medium enterprise restaurant entrepreneurs due to Thailand’s COVID-19-related restrictions: a cross-sectional study. Risk Management and Healthcare Policy. 2022; 15: 1157–1165.

[18] Luangapichart P, Saisavoey N, Viravan N. Efficacy and feasibility of the minimal therapist-guided four-week online audio-based mindfulness program ‘mindful senses’ for burnout and stress reduction in medical personnel: a randomized controlled trial. Healthcare. 2022; 10: 2532.

[19] Wang X, Yang Y, Lin L, Yao Q, Zhang J. Obesity and temporomandibular joint disorders: a systematic review and meta-analysis. BMC Oral Health. 2023; 23: 607.

[20] Cigdem Karacay B, Sahbaz T. Investigation of the relationship between probable sleep bruxism, awake bruxism and temporomandibular disorders using the Diagnostic Criteria for Temporomandibular Disorders (DC/TMD). Dental and Medical Problems. 2023; 60: 601–608.

[21] Travers P, Lacy BE, Cangemi DJ. Irritable bowel syndrome—less irritable, or better treatments? Current Opinion in Gastroenterology. 2024; 40: 27–33.

[22] Onan D, Younis S, Wellsgatnik WD, Farham F, Andruškevičius S, Abashidze A, et al. Debate: differences and similarities between tension-type headache and migraine. Journal of Headache and Pain. 2023; 24: 92.

[23] Da-Cas CD, Valesan LF, Nascimento LPD, Denardin ACS, Januzzi E, Fernandes G, et al. Risk factors for temporomandibular disorders: a systematic review of cohort studies. Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology. 2024; 138: 502–515.

[24] Hashimoto K. Update on subclinical thyroid dysfunction. Endocrine Journal. 2022; 69: 725–738.

[25] Nelson KL, Davis JE, Corbett CF. Sleep quality: an evolutionary concept analysis. Nursing Forum. 2022; 57: 144–151.

[26] Jiang X, Xu C. Deep learning and machine learning with grid search to predict later occurrence of breast cancer metastasis using clinical data. Journal of Clinical Medicine. 2022; 11: 5772.

[27] Ponce-Bobadilla AV, Schmitt V, Maier CS, Mensing S, Stodtmann S. Practical guide to SHAP analysis: explaining supervised machine learning model predictions in drug development. Clinical and Translational Science. 2024; 17: e70056.

[28] Bhak Y, Ahn TK, Peterson TA, Han HW, Nam SM. Machine learning models for low back pain detection and factor identification: insights from a 6-year nationwide survey. The Journal of Pain. 2024; 25: 104497.

[29] Şentürk İA, Şentürk E, Üstün I, Gökçedağ A, Yıldırım NP, İçen NK. High-impact chronic pain: evaluation of risk factors and predictors. The Korean Journal of Pain. 2023; 36: 84–97.

[30] Bond EC, Mackey S, English R, Liverman CT, Yost O. Temporomandibular disorders. National Academies Press: Washington, DC. 2020.

[31] Ballantyne JC, Sullivan MD. Intensity of chronic pain—the wrong metric? New England Journal of Medicine. 2015; 373: 2098–2099.

[32] de Oliveira-Souza AIS, do Valle Sales LR, de Fontes Coutinho AD, de Oliveira DA, Armijo-Olivo S. Effectiveness of an 8-week neck exercise training on pain, jaw function, and oral health-related quality of life in women with chronic temporomandibular disorders: a randomized controlled trial. Journal of Oral & Facial Pain and Headache. 2024; 38: 40–51.

[33] Kizek P, Pacutova V, Schwartzova V, Timkova S. Decoding chronic jaw pain: key nature of temporomandibular disorders in Slovak patients. Bratislava Medical Journal. 2025; 126: 514–523.

[34] Minervini G, Franco R, Marrapodi MM, Crimi S, Badnjević A, Cervino G, et al. Correlation between temporomandibular disorders (TMD) and posture evaluated trough the diagnostic criteria for temporomandibular disorders (DC/TMD): a systematic review with meta-analysis. Journal of Clinical Medicine. 2023; 12: 2652.

[35] Chen X, Cheng Z, Xu J, Zhao Z, Jiang Q. Causal association between body mass index and temporomandibular disorders: a bidirectional two-sample Mendelian randomization analysis. BMC Oral Health. 2023; 23: 499.

[36] Rhim E, Han K, Yun K. Association between temporomandibular disorders and obesity. Journal of Cranio-Maxillofacial Surgery. 2016; 44: 1003–1007.

[37] Yamada K, Kubota Y, Iso H, Oka H, Katsuhira J, Matsudaira K. Association of body mass index with chronic pain prevalence: a large population-based cross-sectional study in Japan. Journal of Anesthesia. 2018; 32: 360–367.

[38] Bäckström E, Wänman A, Sjöström M. The majority of patients report satisfaction more than 24 years after temporomandibular joint discectomy. Oral and Maxillofacial Surgery. 2024; 28: 1539–1545.

[39] Kang JH. Influences of decision preferences and health literacy on temporomandibular disorder treatment outcome. BMC Oral Health. 2022; 22: 385.

[40] Welvaars K, Oosterhoff JHF, van den Bekerom MPJ, Doornberg JN, van Haarst EP, van der Zee JA, et al. Implications of resampling data to address the class imbalance problem (IRCIP): an evaluation of impact on performance between classification algorithms in medical data. JAMIA Open. 2023; 6: ooad033.

[41] Richardson E, Trevizani R, Greenbaum JA, Carter H, Nielsen M, Peters B. The receiver operating characteristic curve accurately assesses imbalanced datasets. Patterns. 2024; 5: 100994.

[42] Salmi M, Atif D, Oliva D, Abraham A, Ventura S. Handling imbalanced medical datasets: review of a decade of research. Artificial Intelligence Review. 2024; 57: 273.

[43] Kaliappan J, Saravana Kumar IJ, Sundaravelan S, Anesh T, Rithik RR, Singh Y, et al. Analyzing classification and feature selection strategies for diabetes prediction across diverse diabetes datasets. Frontiers in Artificial Intelligence. 2024; 7: 1421751.

[44] Wang H, Liang Q, Hancock JT, Khoshgoftaar TM. Feature selection strategies: a comparative analysis of SHAP-value and importance-based methods. Journal of Big Data. 2024; 11: 44.