Development of a care plan recommendation system for physical therapy patients

Abstract

Physical therapy includes many different treatment methods for patients' musculoskeletal problems. However, each patient has different complaints, lifestyles, and recovery rates. The importance of personalized care in physical therapy practices has therefore increased. The aim of this study is to develop a personalized recommendation system for patients' physical therapy processes. A dataset of 12,393 patients was collected from a private hospital and analyzed using machine learning and deep learning algorithms. Prediction was performed with Multi-Layer Perceptron (MLP) Classifier, Artificial Neural Network (ANN), Convolutional Neural Network (CNN), and Feedforward Neural Network (FNN) models using 5-fold cross-validation. Among single models, the ANN achieved the best performance with Precision, Recall, F1-Score, and Accuracy values of 0.8733, 0.8666, 0.8600, and 0.8666, respectively. In the ensemble stage, a Voting Classifier combining Random Forest and LightGBM achieved 96% accuracy for predicting treatment session numbers, while the stacked-ensemble pipeline integrating ANN and CNN achieved 86% success in personalized treatment plan recommendation. The novelty of this study lies in the design of a stacked ensemble pipeline that integrates treatment session prediction with personalized care plan recommendation, enhancing both effectiveness and patient-centeredness in physical therapy. Each patient can thus receive the most appropriate treatment for their health status and needs. Personalized care can lead to faster recovery, greater participation in treatment processes, and more positive long-term outcomes. This study contributes to making physical therapy more effective and patient-centered by emphasizing the importance of personalized care in healthcare.

Keywords

• Artificial Intelligence
• Deep Learning
• Physical Therapy
• Clinical Decision Support

Fizik tedavi hastalari için bir bakim plani öneri sisteminin geliştirilmesi

Öz

Fizik tedavi, hastaların kas-iskelet sistemi sorunları için birçok farklı tedavi yöntemini içerir. Ancak her hastanın şikayetleri, yaşam tarzları ve iyileşme oranları farklıdır. Bu nedenle, fizik tedavi uygulamalarında kişiselleştirilmiş bakımın önemi artmıştır. Bu çalışmanın amacı, hastaların fizik tedavi süreçleri için kişiselleştirilmiş bir öneri sistemi geliştirmektir. Özel bir hastaneden 12.393 hastadan oluşan bir veri seti toplanmış ve makine öğrenimi ve derin öğrenme algoritmaları kullanılarak analiz edilmiştir.Tahmin, 5 katlı çapraz doğrulama kullanılarak Çok Katmanlı Algılayıcı (MLP) Sınıflandırıcı, Yapay Sinir Ağı (YSA), Evrişimli Sinir Ağı (CNN) ve İleri Beslemeli Sinir Ağı (FNN) modelleri ile gerçekleştirilmiştir. Tek modeller arasında YSA, sırasıyla 0,8733, 0,8666, 0,8600 ve 0,8666'lık Hassasiyet, Geri Çağırma, F1 Puanı ve Doğruluk değerleriyle en iyi performansı elde etmiştir. Topluluk aşamasında, Random Forest ve LightGBM'yi birleştiren bir Oylama Sınıflandırıcısı, tedavi seansı sayılarını tahmin etmede %96 doğruluk oranına ulaşırken, ANN ve CNN'i entegre eden yığılmış topluluk hattı, kişiselleştirilmiş tedavi planı önerisinde %86 başarı elde etti. Bu çalışmanın yeniliği, tedavi seansı tahminini kişiselleştirilmiş bakım planı önerisiyle entegre eden ve fizyoterapide hem etkinliği hem de hasta merkezliliği artıran bir yığılmış topluluk hattının tasarımında yatmaktadır. Böylece her hasta, sağlık durumu ve ihtiyaçlarına en uygun tedaviyi alabilir. Kişiselleştirilmiş bakım, daha hızlı iyileşmeye, tedavi süreçlerine daha fazla katılıma ve daha olumlu uzun vadeli sonuçlara yol açabilir. Bu çalışma, sağlık hizmetlerinde kişiselleştirilmiş bakımın önemini vurgulayarak fizyoterapiyi daha etkili ve hasta merkezli hale getirmeye katkıda bulunmaktadır.

Anahtar Kelimeler

• Yapay Zeka
• Derin Öğrenme Fizik Tedavi
• Klinik Karar Desteği

Kaynakça

1. [1] Secinaro S, Calandra D, Secinaro A, Muthurangu V, Biancone P. “The role of artificial intelligence in healthcare: a structured literature review”. BMC Med. Inf. Decis. Making, 21, 1–23, 2021.
2. [2] Bi Q, Goodman KE, Kaminsky J, Lessler J. “What is machine learning? A primer for the epidemiologist”. Am. J. Epidemiol. 188 (12), 2222–2239, 2019.
3. [3] Collins GS, Moons KGM. “Reporting of artificial intelligence prediction models”. Lancet, 393, (10181), 1577–1579, 2019.
4. [4] Mahmoud H, Aljaldi F, El-Fiky A, Battecha K, Thabet A, Alayat M, et al. “Artificial Intelligence machine learning and conventional physical therapy for upper limb outcome in patients with stroke: a systematic review and meta-analysis”. Eur. Rev. Med. Pharmacol. Sci., 27 (11), 2023.
5. [5] Puyuelo-Quintana, G, Cano-De-La-Cuerda, R, Plaza-Flores A, Garces-Castellote E, Sanz-Merodio D, Goni-Arana A, et al. “A new lower limb portable exoskeleton for gait assistance in neurological patients: a proof of concept study”. J. NeuroEng. Rehabil. 17, 1–16, 2020.
6. [6] Lyu M, Chen WH, Ding X, & Wang J. “Knee exoskeleton enhanced with artificial intelligence to provide assistance-as-needed”. Review of Scientific Instruments, 90(9), 2019.
7. [7] Lopez-Meyer P, George DF, Edward SS. "Automatic detection of temporal gait parameters in poststroke individuals." IEEE Transactions on Information Technology in Biomedicine, 15(4), 594-601, 2011.
8. [8] Khera P, Kumar N. “Role of machine learning in gait analysis: a review”. Journal of Medical Engineering & Technology, 44(8), 441-467, 2020.

9. [9] Ng G, Andrysek J. “Classifying changes in amputee gait following physiotherapy using machine learning and continuous inertial sensor signals”. Sensors, 23(3), 1412, 2023
10. [10] Kianifar R, Lee A, Raina S, Kuli´c D. “Automated assessment of dynamic knee valgus and risk of knee injury during the single leg squat”. IEEE journal of translational engineering in health and medicine, 5, 1–13, 2017.
11. [11] Alfakir A, Arrowsmith C, Burns D, Razmjou H, Hardisty M, Whyne C. “Detection of low back physiotherapy exercises with inertial sensors and machine learning: algorithm development and validation”. JMIR Rehabilitation and Assistive Technologies, 9 (3), e38689, 2022.
12. [12] Arrowsmith C, Burns D, Mak T, Hardisty M, Whyne C, “Physiotherapy exercise classification with single-camera pose detection and machine learning”. Sensors, 23 (1), 363, 2022.
13. [13] Khan O, Badhiwala JH, Grasso G., Fehlings MG. “Use of machine learning and artificial intelligence to drive personalized medicine approaches for spine care”. World neurosurgery, 140, 512–518, 2020.
14. [14] Romaniszyn-Kania P, Pollak A, Bugdol MD, Bugdol MN, Kania D, Manka A. et al. “Affective state during physiotherapy and its analysis using machine learning methods”. Sensors, 21 (14), 4853, 2021.
15. [15] Burns DM, Leung N, Hardisty M, Whyne CM, Henry P. McLachlin S. “Shoulder physiotherapy exercise recognition: Machine learning the inertial signals from a smartwatch”. Physiol. Meas., 9, 75007, 2018.
16. [16] Burns D, Razmjou H, Shaw, J, Richards, R, McLachlin, S, Hardisty, M, Henry, P, Whyne, C, et al. “Adherence tracking with smart watches for shoulder physiotherapy in rotator cuff pathology: Protocol for a longitudinal cohort study”. JMIR Res. Protoc., 9, e17841, 2020.
17. [17] Alfakir A, Arrowsmith C, Burns D, Razmjou H, Hardisty M, Whyne C, et al. “Detection of Low Back Physiotherapy Exercises With Inertial Sensors and Machine Learning: Algorithm Development and Validation”. JMIR Rehabil. Assist. Technol., 9, e38689, 2022.
18. [18] Abujaber, A. A., Alkhawaldeh, I. M., Imam, Y., Nashwan, A. J., Akhtar, N., Own, A., Al-Khayyat, A., ... & Hassanat, A. B. Predicting 90-day prognosis for patients with stroke: a machine learning approach. Frontiers in Neurology, 14, 1270767, 2023.
19. [19] Sprint, G., Cook, D. J., Weeks, D. L., & Borisov, V. Predicting functional independence measure scores during rehabilitation with wearable inertial sensors. IEEE access, 3, 1350-1366, 2015.
20. [20] Cafri, G., et al. Predicting risk for adverse health events using random forest. Journal of Applied Statistics, 45(12), 2279-2294, 2018.
21. [21] Almeida, G., Brito Correia, F., Borges, A. R., & Bernardino, J. Hospital length-of-stay prediction using machine learning algorithms—a literature review. Applied Sciences, 14(22), 10523, 2024.
22. [22] Mennella, C., Maniscalco, U., De Pietro, G., & Esposito, M. A deep learning system to monitor and assess rehabilitation exercises in home-based remote and unsupervised conditions. Computers in Biology and Medicine, 166, 107485, 2023.
23. [23] Katsuki, M., et al. Deep Learning-Based Functional Independence Measure Score Prediction After Stroke in Convalescent Rehabilitation. International Journal of Environmental Research and Public Health, 18(13), 6791, 2021.
24. [24] Zhou, M., et al. Applying machine learning to predict future adherence to physical activity programs. Preventive Medicine Reports, 15, 100913, 2019.
25. [25] Zhang, S., Li, Y., Zhang, S., Shahabi, F., Xia, S., Deng, Y., & Alshurafa, N. Deep learning in human activity recognition with wearable sensors: A review on advances. Sensors, 22(4), 1476, 2022.
26. [26] Altman NS. “An introduction to kernel and nearest-neighbor nonparametric regression”. Am. Stat. 46 (3), 175–185, 1992.
27. [27] Vinutha HP, Poornima B, Sagar BM. “Detection of outliers using interquartile range technique from intrusion dataset. In Information and decision sciences: Proceedings of the 6th international conference on ficta” Springer Singapore, 511-518, 2018.
28. [28] Altinok D. “A Diverse Set of Freely Available Linguistic Resources for Turkish. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics”. Association for Computational Linguistics, vol 1, 13739–13750, 2023.
29. [29] Zhang X, Zhang J, Zhang J & Zhang Y. “Research on the combined prediction model of residential building energy consumption based on random forest and BP neural network”. Geofluids,, 7271383, 2021(1),
30. [30] Breiman L “Random forests”. Machine learning, 45, 5-32, 2001.
31. [31] Zekić-Sušac M, Has A, Knežević M. “Predicting energy cost of public buildings by artificial neural networks, CART, and random forest”. Neurocomputing, 439, 223–233, 2021.
32. [32] Duman E. “Implementation of XGBoost Method for Healthcare Fraud Detection”. Scientific Journal of Mehmet Akif Ersoy University, 5(2), 69-75, 2022.
33. [33] Cao Y, Liu G, Sun J, Bavirisetti DP, Xiao G. “PSO-Stacking improved ensemble model for campus building energy consumption forecasting based on priority feature selection”. Journal of Building Engineering, 72, 106589, 2023.
34. [34] Arjunan P, Poolla K, Miller C. “EnergyStar++: Towards more accurate and explanatory building energy benchmarking”. Applied Energy, 276, 115413, 2020.
35. [35] Santolamazza A, Cesarotti V, Introna V. “Anomaly detection in energy consumption for Condition-Based maintenance of Compressed Air Generation systems: an approach based on artificial neural networks”. IFAC-PapersOnLine, 51(11), 1131-1136, 2018.
36. [36] Lee ZJ, Lin Y, Chen ZY, Yang Z, Fang WG, Lee CH. “Ensemble deep learning applied to predict building energy consumption”. In 2023 IEEE 6th Eurasian Conference on Educational Innovation (ECEI), 339-341, 2023.
37. [37] Mlangeni S, Ezugwu AE, Chiroma H. “Deep learning model for forecasting institutional building energy consumption”. In 2020 Conference on Information Communications Technology and Society (ICTAS). 1-8, IEEE, 2020.
38. [38] Gunn S.. Support Vector Machines for Classification and Regression. Image Speech and Intelligent Systems Technical Report. Bildiriler Kitabı: 230–67, 1998.
39. [39] Chen JF, Do QH, Hsieh HN.. “Training artificial neural networks by a hybrid PSO-CS algorithm”. Algorithms, 8(2), 292-308, 2015.
40. [40] Heidari, AA, Faris H, Mirjalili S, Aljarah I, Mafarja M. “Ant lion optimizer: theory, literature review, and application in multi-layer perceptron neural networks”. Nature-inspired optimizers: theories, literature reviews and applications, 23-46, 2020.
41. [41] Park YS, Lek S. “Artificial neural networks: Multilayer perceptron for ecological modeling”. In Developments in environmental modelling. 28, 123-140, 2016.
42. [42] Qaddoura R, Al-Zoubi A, Faris H, Almomani I. “A multi-layer classification approach for intrusion detection in iot networks based on deep learning”. Sensors, 21(9), 2987, 2021.
43. [43] Hosmer DW, Lemeshow S. Sturdivant RX. “Applied logistic regression”, Wiley Hoboken, New Jersey, ISBN: 978-0-470-58247-3, 2013.
44. [44] Usman HM, ElShatshat R, El-Hag AH., Jabr RA. “Estimation of distribution transformer kVA load using residential smart meter data”. Electric Power Systems Research, 204, 107663, 2022.
45. [45] Ke G. Meng, Q, Finley T, Wang, T, Chen W, Ma W, Ye Q, Liu TY. “LightGBM: A highly efficient gradient boosting decision tree”. Advances in Neural Information Processing Systems, 30, 2017.
46. [46] Prokhorenkova L, Gusev G, Vorobev A, Dorogush AV, Gulin A. “CatBoost: unbiased boosting with categorical features”. Advances in neural information processing systems, 31, 2018.
47. [47] Dorogush AV, Ershov V, Gulin A. CatBoost: gradient boosting with categorical features support. 2018.
48. [48] Abujayyab SKM, Kassem MM, Khan AA, Wazirali R, Coşkun M, Taşoǧlu, E Oztürk, A Toprak, F. Wildfire susceptibility mapping using five boosting machine learning algorithms: The case study of the Mediterranean region of Turkey. Genetics Research, 2022.
49. [49] Achu AL, Thomas J, Aju CD, Gopinath G, Kumar S, Reghunath R. “Machine-learning modelling of fire susceptibility in a forest-agriculture mosaic landscape of southern India”. Ecological Informatics, 64, 101348, 2021.
50. [50] Balaban ME, Kartal E. “Veri Madenciliği Süreci ve Futbol Maç Sonuçlarının Öngürülmesine İlişkin Bir Uygulama”. R ile Veri Madenciliği Uygulamaları. İstanbul: Çağlayan Kitabevi; 2016.
51. [51] Lavrač N, Motoda H, Fawcett T, Holte R, Langley P, Adriaans P. “Introduction: Lessons Learned from Data Mining Applications and Collaborative Problem Solving”. Machine Learning, 57(1), 13-34, 2004.
52. [52] Han J, Kamber M, Pei J. “Data mining concepts and techniques”, 3 th edition. Waltham: The Morgan Kaufmann; 2012.