Table 3.

Effect Estimates for Comparison of Cohorts (SPLIT vs Usual Care) for Primary and Secondary Outcomes Over the 6-Month Follow-Up

Crude ß/OR (95% CI)aP ValueAdjusted ß/OR (95% CI)aP Value
Back-related disability (RMDQ, 0 to 24)
    Raw score−5.19 (−6.09 to −4.29)≤.001−2.94 (− 3.63 to −2.24)≤.001
    Achieved MIC, yes vs nob  7.07 (3.31 to 15.08)≤.001  8.67 (3.98 to 18.89)≤.001
    Poor disability, yes vs noc  0.07 (0.003 to 0.13)≤.001  0.24 (0.15 to 0.38)≤.001
Pain intensity (NPRS, 0 to 10)
    Raw score−1.76 (−2.18 to −1.35)≤.001−0.88 (−1.18 to −0.57)≤.001
    Achieved MIC, yes vs nob  3.55 (1.94 to 6.48)≤.001  3.50 (1.88 to 6.49)≤.001
Perceived effect of treatment (GPES, −5 to +5)
    Raw score  1.50 (1.08 to 1.91)≤.001  1.40 (0.97 to 1.82)≤.001
    Achieved MIC, yes vs nob  4.66 (2.39 to 9.12)≤.001  4.40 (2.22 to 8.71)≤.001
HRQoL (EQ-5D-3L, 0 to 1)
    Raw score  0.19 (0.16 to 0.23)≤.001  0.11 (0.08 to 0.14)≤.001
  • EQ-5D-3L = EuroQoL 5 dimensions 3 levels; GPES = Global Perceived Effect Scale; HRQoL = health-related quality of life; MIC = minimal important change; NPRS = Numeric Pain Rating Scale; OR = odds ratio; RMDQ = Roland-Morris Disability Questionnaire; SPLIT = stratified primary care for low back pain; UC = usual care.

  • a For continuous outcomes (RMDQ, NPRS, GPES, and EQ-5D-3L raw scores over time), crude/adjusted ß values were derived via a linear mixed-effects model for the comparison of UC vs SPLIT. For categorical outcomes (achieved MIC of RMDQ, NPRS, and GPES, along with poor disability), crude/adjusted OR values were derived via a logistic mixed-effects model for the comparison of UC vs SPLIT. Adjusted ß/OR and 95% CI values were adjusted for age, duration of low back pain episode, referred leg pain, and baseline STarT Back Screening Tool psychosocial subscale, NPRS, RMDQ, and EQ-5D-3L values.

  • b Based on established MIC criteria: ≥30% decrease from baseline for RMDQ and NPRS and a GPES score ≥3.

  • c Based on a cutoff RMDQ score ≥7.