Table 2

Results of sequential logistic regression analyses to evaluate factors that influence intentions to improve practice

Variable of interestLaboratory experiment (n=330)Field study (n=462)
Adjusted for confounders*OR (95% CI)p ValueAdjusted for confounders*OR (95% CI)p Value
Fixed effects
Indicator level
 1. Outcome vs process indicator2, 30.65 (0.23 to 1.84)0.4132–50.88 (0.26 to 2.93)0.836
 2. Performance score (per 10% decrease)11.54 (1.29 to 1.83)<0.0011, 4–6†1.25 (1.13 to 1.39)<0.001
 3. Benchmark category1, 21, 2, 4–6†
  Intermediate vs high23.15 (6.96 to 76.95)<0.0011.42 (0.55 to 3.70)0.472
  Low vs high12.04 (1.33 to 108.90)0.0270.64 (0.10 to 4.34)0.652
 4. Selected in previous action plan1–3, 5, 6†10.22 (4.98 to 20.98)<0.001
Feedback report level
 5. A&F iteration number1–4, 6†0.46 (0.32 to 0.66)<0.001
 6. Team size2–5†1.03 (0.87 to 1.21)0.727
Explained variance (full model‡)p ValueExplained variance (full model ‡)p Value
Random effects
 Quality indicator1.0%0.6495.8%0.004
 Feedback report7.9%0.0766.6%0.065
 CR professional16.0%0.006
 CR centre1.6%0.7600.0%0.964
Conditional R261.3%50.4%
Marginal R234.7%37.9%
  • Each analysis was adjusted for potential confounders, and limited to complete cases (no ‘grey’ indicators).

  • *Numbers correspond to the variables listed in the first column.

  • †Also adjusted for centre type.

  • ‡Also adjusted for all CR professional-level variables: gender, years of clinical experience, percentage of time spent on direct patient care, discipline, coordinating function, number of outreach visits attended in CARDSS-II (laboratory experiment part only) and the CR centre-level variable: centre type (both parts).

  • A&F, audit and feedback; CR, cardiac rehabilitation.