Table 3. Multivariable Models Examining the Association of Magnet Status With Outcomes
Inpatient MortalityaDischarge to RehabilitationaLength of Stayb
Adjusted Difference (95% CI)P ValueAdjusted Difference (95% CI)P ValueAdjusted Difference (95% CI)P Value
Instrumental variable analysisc−23.9% (−29.0% to −18.7%)<0.001−16.5% (−20.0% to −13.0%)<0.001−0.4 (−0.8 to −0.1)<0.001
Mixed effects regressiond−10.5% (−12.8% to −8.1%)<0.001−10.1% (−11.7% to −8.4%)<0.001−0.2 (−0.3 to −0.1)<0.001
Propensity score adjusted regressiond−9.7% (−11.9% to −7.4%)<0.001−9.9% (−11.6% to −8.3%)<0.001NANA
  • NA indicates not applicable.

  • a Regressions based on probit models.

  • b Regressions based on linear models.

  • c Differential distance to Magnet hospital was used as an instrument of treatment hospital.

  • d Hospital ID was used as a random effects variable.