PhD defence Y. (Yanyan) Fu

Below is a brief summary of the dissertation:

Low back pain (LBP) is one of the leading causes of disability worldwide and is characterized by a heterogeneous clinical course, particularly among older adults. This thesis aimed to improve understanding of the long-term prognosis, prognostic factors, prognostic models, and methodological challenges related to LBP. Specifically, it investigated long-term disability trajectories in older adults, evaluated the role of comorbidities and existing prognostic models, externally validated the STarT Back Tool (SBT), and explored the feasibility of conducting an individual participant data meta-analysis (IPD-MA) of non-steroidal anti-inflammatory drugs (NSAIDs) for acute LBP.

Four distinct five-year disability trajectories were identified, demonstrating substantial heterogeneity in long-term outcomes, with most recovery occurring during the first six months. Less favorable trajectories were associated with older age, higher body mass index, greater baseline pain and disability, multimorbidity, radiating pain, depressive symptoms, previous back pain, and negative recovery expectations. Musculoskeletal comorbidities were associated with poorer pain and disability outcomes. A systematic review found that although the SBT showed acceptable discrimination in previous studies, most published prognostic models were at high risk of bias because of methodological limitations and insufficient validation. In contrast, external validation showed that the SBT had poor prognostic performance for predicting disability and pain outcomes in older adults receiving chiropractic care.

The planned IPD-MA of NSAIDs for acute LBP could not be completed because only one eligible trial shared individual participant data. Instead, this thesis reports the practical barriers encountered, including unavailable datasets, obsolete data storage, contractual restrictions, and limited responses from original investigators. Overall, this thesis highlights the heterogeneous prognosis of LBP in older adults, identifies multimorbidity as an important prognostic factor, questions the generalizability of existing prognostic models to older chiropractic populations, and emphasizes the need for more rigorous model validation and stronger data-sharing practices to advance personalized LBP care.