This study is a secondary analysis of an RCT (3), aimed at assessing the association between patients and PTs, and factors that they consider while rating GPE.
Aim 1. we performed a correlation analysis, then assessed agreement between patients’ and PT’s ratings using contingency table and Bland-Altman plot. GPE scores were categorized as 1=worsened, 2=no change, and 3=improved (Table 1).
Aim 2: we built two regression models for each time point, with dependent variable being GPE rated by patients and PTs, and independent variables being outcome measures.
Aim 3: We compared model fit statistics at two time points to examine the potential GPE risk of bias using model fit statistics.
A total of 220 participants were included in this study, of whom 142 (64%) were women with mean age of 51 years.
Aim 1: The correlation analysis showed a moderate (ρ=0.62) positive relationship between the GPE rated by patients and PTs. There was higher agreement between the improved group and PTs.
Bland-Altman plot: Overall, for scores higher than 1, the Bland-Altman plot showed stronger agreement as points are less scattered and generally within 95% confidence interval (CI).
Aim 2: Patient’s model: The final model with change score included pain severity, disability, depression, kinesophobia, pain catastrophizing, coping strategies, central sensitization, and negative beliefs about LBP. In the regression model for PT, GPE was associated with disability, pain, and function.
Aim 3: Overall, both the models performed better with follow-up scores as opposed to change scores, however, the differences were not large.
4. Conclusion: This study showed a moderate agreement between GPE rated by patients and PTs. Pain, disability, and function were associated with GPE response variance. Additional, patient’s GPE rating has association with psychosocial factors. However, these factors explained a small proportion of the response variance. Therefore, further qualitative studies are required to identify other influential factors.