Table 2.
Multivariate analysis of overall predictors of back related dysfunction (Roland score)
Roland score* | |||||
8-week | 52-week | ||||
Variable | Categories | β ** (SE) | P-value | β ** (SE) | P-value |
Baseline Roland score | Continuous | 0.43 (0.05) | <.0001 | 0.56 (0.05) | <.0001 |
Baseline bothersomeness score | Continuous | 0.10 (0.10) | 0.30 | 0.001 (0.10) | 0.99 |
Any disability | Yes vs. No | 0.78 (0.46) | 0.09 | 0.01 (0.49) | 0.99 |
SF-36 Mental Health score | Continuous | -0.02 (0.01) | 0.17 | -0.01 (0.01) | 0.58 |
Age | Continuous | -0.01 (0.02) | 0.51 | 0.02 (0.02) | 0.33 |
Gender | Female vs. Male | 0.42 (0.41) | 0.32 | 0.37 (0.44) | 0.39 |
Education level | College graduate, Yes vs. No | -0.19 (0.42) | 0.65 | -0.81 (0.44) | 0.07 |
Employment | Heavy lifting job vs. Unemployed | -0.16 (0.70) | 0.81 | -1.76 (0.74) | 0.16 |
Light/Medium lifting job vs. Unemployed | -0.53 (0.57) | -1.22 (0.60) | |||
Sedentary job vs. Unemployed | -0.61 (0.59) | -0.86 (0.63) | |||
Employment unknown vs. Unemployed | -1.07 (1.84) | -0.93 (1.90) | |||
Medication use | Narcotics vs. None | 0.43 (0.69) | 0.81 | 1.77 (0.73) | 0.04 |
Other vs. None | 0.02 (0.43) | 0.12 (0.46) | |||
Self-efficacy | High vs. Low | 0.61 (0.93) | 0.51 | -0.20 (1.00) | 0.84 |
Expectation of acup. help | Top tertile vs. Lower two tertiles | -0.58 (0.47) | 0.42 | -0.23 (0.50) | 0.55 |
Could not rate vs. Lower two tertiles | -0.44 (0.53) | 0.45 (0.56) | |||
Duration of chronic LBP | >= 1 y vs. <1 y | 0.24 (0.45) | 0.60 | 0.49 (0.48) | 0.30 |
Pain travels below knee | Yes vs. No | -0.02 (0.51) | 0.98 | -0.37 (0.54) | 0.50 |
Days of LBP in last 3 mo | Continuous | 0.02 (0.01) | 0.10 | 0.01 (0.01) | 0.46 |
Intense LBP treatment | Yes vs. No | 1.20 (0.63) | 0.06 | 0.51 (0.66) | 0.44 |
Treatment group | Individualized acupuncture vs. Usual care | -2.38 (0.56) | <.0001 | -2.21 (0.59) | 0.001 |
Standardized acupuncture vs. Usual care | -2.44 (0.55) | -1.81 (0.58) | |||
Simulated acupuncture vs. Usual care | -2.86 (0.55) | -1.03 (0.58) |
* For ease of identification, all independent variables with P < 0.5 are in boldface type
** The parameter estimates β refer to the amount of change in the outcome that is based on a one unit change in that covariate (continuous variables) or a change in category (categorical variables)