Table 3.
Regression results main model M1 (beta, lower and upper confidence interval)
| M1 | |
|---|---|
| Log dummy | −0.059* (−0.112, −0.007) |
| Stroke unit (SU) | − 0.056** (− 0.092, − 0.021) |
| SU certification | − 0.005 (− 0.036, − 0.026) |
| THQ certification | 0.016 (− 0.019, 0.051) |
| Log stroke case volume | − 0.003 (− 0.010, 0.003) |
| Share stroke ICD/all ICD | − 0.054 (− 0.743, 0.635) |
| Hospital beds | − 0.000 (− 0.000, 0.000) |
| Teaching hospital status | 0.007 (− 0.041, 0.056) |
| CMS ICD specialization index | − 0.003 (− 0.039, 0.033) |
| Private, for-profit hospitalsa | − 0.022 (− 0.144, 0.101) |
| Public hospitalsa | − 0.011 (− 0.138, 0.117) |
| Log all OPS stroke procedures | – |
| Interaction SU and THQ certification | – |
| Constant | 0.182* (0.030, 0.333) |
| R2-within | 0.016 |
| R2-between | 0.032 |
| R2-overalll | 0.018 |
| BIC | 1084 |
| Intraclass correlation | 0.512 |
| F-statistic | 3.9 |
| Number of observations | 7376 |
QSR stroke patient volume applied as analytical weights; *p < 0.05, **p < 0.01, ***p < 0.001; time-fixed effects not displayed separately (in M1, β2008 = − 0.02, β2010 = − 0.03, β2012 = − 0.03, β2013 = − 0.05***, β2014 = − 0.08***), test for joint significance of time effects in M1 with F-statistic of 5.59
aprivate (non-profit) hospitals serve as reference category