Towards reducing repeated hospital attendance and admissions for patients with low back pain using a machine learning approach

Dr Kwan, Kenny Yat Hong   (Principal Investigator (PI))

Dr Wang Qingchen   (Co-Investigator)

24

2021-09-01

1062880

Towards reducing repeated hospital attendance and admissions for patients with low back pain using a machine learning approach

Big data, Hospital readmission, Low back pain, Machine learning, Repeated attendance, Risk stratification

Others - Medicine, Dentistry and Health

18192271

Health and Medical Research Fund - Full Grant

2020

On-going

Objectives: Low back pain (LBP) is one of the leading causes of disease burdens globally, leading to significant functional limitations. Repeated hospital attendance and readmissions for LBP patients can increase long-term back-related disability paradoxically and healthcare costs. The public health challenge is for clinicians to identify such patients early and implement targeted multidisciplinary rehabilitation. Hypothesis: Big data analytics can create accurate models to predict repeated hospital attendance for LBP patients(AUC0.9). Design and subjects: A retrospective study based on a territory-wide cohort of 20,000 patients with LBP from 2009 to 2020. Study instruments: Demographics, blood test and bacterial cultures results, psychiatric history and drug prescription records will be retrieved through electronic patient record and analyzed by super computer. Main outcome measures: The primary outcome will be the predictive ability of any repeated medical attendance after the index episode of hospital encounter for LBP. Secondary outcomes will be models based on different time periods of readmissions. Data analysis: Logistic regression and machine learning models. Expected results: This project will develop a highly predictive model for repeated hospital attendance and readmission for patients will LBP available for routine clinical use. Early identification of such patients can result in more effective use of healthcare resources and improve patients’ outcomes.