Table 3. Association of Early PCI and Hospital Spending on 180-Day Acute Myocardial Infractions Mortality Ratesa,b.
| Early PCI or Spending | Model 1, Full Sample (n=479 873)c | Model 2, Full Sample (n=479 873)d | Model 3, NSTEMI (n=331 635)c | Model 4, NSTEMI (n=331 635)d | Model 5, STEMI (n=148 077)c | Model 6, STEMI (n=148 077)d |
|---|---|---|---|---|---|---|
| Early PCI rate, No. | ||||||
| β (SE) | −0.068 (0.0094)e | −0.079 (0.0094)e | −0.080 (0.012)e | −0.0917 (0.0122)e | −0.155 (0.010)e | −0.172 (0.0099)e |
| P value | <.001 | <.001 | <.001 | <.001 | <.001 | <.001 |
| With $1000 increase in total spending | ||||||
| 180-d Mortality | ||||||
| β (SE) | −0.0002 (0.0001) | −0.0002 (0.0001) | −0.0002 (0.0001) | |||
| P value | .04 | .12 | .10 | |||
| 180-d Testing | ||||||
| β (SE) | −0.0041 (0.0027) | 0.0027 (0.0028) | 0.0053 (0.0027) | |||
| P value | .13 | .34 | .05 | |||
| 180-d Visits | ||||||
| β (SE) | −0.0043 (0.0023) | −0.0061 (0.0023) | −0.0062 (0.0028) | |||
| P value | .06 | .01 | .03 | |||
| 180-d Cardiac procedures | ||||||
| β (SE) | 0.0042 (0.0016)e | 0.0049 (0.0017)e | 0.0042 (0.0014)e | |||
| P value | .01 | .003 | .002 | |||
| 180-d Other procedures | ||||||
| β (SE) | −0.0042 (0.0017) | −0.0050 (0.0019) | −0.0050 (0.0016)e | |||
| P value | .02 | .01 | .002 | |||
| 180-d Part A | ||||||
| β (SE) | 0.0002 (0.0003) | 0.0002 (0.0003) | 0.0006 (0.0003) | |||
| P value | .40 | .50 | .04 | |||
| 180-d Postacute care | ||||||
| β (SE) | −0.0022 (0.0005)e | −0.0016 (0.0005)e | −0.0042 (0.0006)e | |||
| P value | <.001 | .003 | <.001 |
Abbreviations: NSTEMI, non-ST–elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST-elevation myocardial infarction.
Sample includes patients who visited hospitals that appeared in the data in all years and that treated at least 10 patients with acute myocardial infarction per year.
Logit model estimated; marginal effects reported using “margins, dydx() atmeans” in Stata. All regressions control for patient characteristics, hospital fixed effects, and year fixed effects. All spending measures are calculated for each hospital-year using patients who live 6 months. Includes years 1999, 2000, 2004, 2008, 2013, and quarters 1 and 2 of 2014. Standard errors are clustered at the hospital level.
P < .025 for 5% significance level using Bonferroni correction; P < .005 for 1% significance level using Bonferroni correction.
P < .0071 for 5% significance level using Bonferroni correction; P < .0014 for 1% significance level using Bonferroni correction.
Statistical significance after conducting the Bonferroni correction.