Identifying self-reported health-related problems in home-based rehabilitation of older patients after hip replacement in China: a machine learning study based on Omaha system theory

BACKGROUND: With the aging of the population, the number of total hip replacement surgeries is increasing globally. Hip replacement has undergone revolutionary advancements in surgical methods and materials. Due to the short length of hospitalization, rehabilitation care is mainly home-based. The needs and concerns about such home-based rehabilitation are constantly changing, requiring continuous attention. OBJECTIVE: To explore effective methods for comprehensively identifying older patients’ self-reported outcomes after home-based rehabilitation for hip replacement, in order to develop appropriate intervention strategies for patient rehabilitation care in the future. METHODS: This study constructed a corpus of patients’ self-reported rehabilitation care problems after hip replacement, based on the Omaha classification system. This study used the Python development language and implemented artificial intelligence to match the corpus data on the cooperation platform, to identify the main health-related problems reported by the patients, and to perform statistical analyses. RESULTS: Most patients had physical health-related problems. More than 80% of these problems were related to neuromusculoskeletal function, interpersonal relationships, pain, health care supervision, physical activity, vision, nutrition, and residential environment. The most common period in which patients’ self-reported problems arose was 6 months post-surgery. The relevant labels that were moderately related to these problems were: Physiology-Speech and Language and Physiology-Mind (r = 0.45), Health-Related Behaviors-Nutrition and Health-Related Behaviors-Compliance with Doctors’ Prescription (r = 0.40). CONCLUSION: Physiological issues remain the main health-related issues for home-based rehabilitation after hip replacement in older patients. Precision care has become an important principle of rehabilitation care. This study used a machine learning method to obtain the largest quantitative network data possible. The artificial intelligence capture was fully automated, which greatly improved efficiency, as compared to manual data entering. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12911-023-02353-7.