Table 2.
Relationship between EHR Adoption and Hospital Outcomes: Full Model
| Process Compositea | Patient Satisfaction Compositeb | Efficiencyc | |
|---|---|---|---|
| EHR adoption | 0.147*** (0.0455) | 0.118*** (0.0268) | 0.00628 (0.0229) |
| CMI | 0.0891 (0.136) | 0.0287 (0.0758) | 0.535*** (0.126) |
| Market share | −0.195 (0.568) | −0.277 (0.303) | −0.877*** (0.282) |
| Market share squared | 0.0940 (0.815) | 0.177 (0.377) | 1.197*** (0.442) |
| 2009 | 0.245*** (0.0132) | 0.174*** (0.00920) | 0.0357*** (0.00753) |
| 2010 | 0.489*** (0.0153) | 0.268*** (0.0107) | 0.0808*** (0.00923) |
| 2011 | 0.639*** (0.0221) | 0.430*** (0.0123) | 0.169*** (0.0113) |
| Medium | 0.0547 (0.0793) | −0.0224 (0.0401) | −0.105* (0.0621) |
| Large | 0.115 (0.0948) | 0.0597 (0.0532) | −0.102 (0.0749) |
| Minor teaching | −0.0706 (0.0449) | 0.0280 (0.0257) | 0.0285 (0.0195) |
| Major teaching | −0.00826 (0.0925) | 0.000870 (0.0695) | −0.0171 (0.0170) |
| System membership | 0.0287 (0.0546) | −0.0527 (0.0331) | −0.00440 (0.0306) |
| Nonprofit | −0.0701 (0.0957) | 0.0638 (0.0603) | 0.00466 (0.0551) |
| For‐profit | −0.121 (0.153) | −0.102 (0.0795) | −0.0618 (0.0887) |
| Network membership | −0.0645* (0.0357) | 0.0189 (0.0227) | −0.0215 (0.0199) |
| Constant | −0.390* (0.214) | −0.203* (0.121) | −0.646*** (0.175) |
| Observationsd | 10,152 | 9,327 | 11,363 |
| R 2 | 0.229 | 0.277 | 0.091 |
| Number of hospitals | 3,798 | 3,401 | 4,135 |
The Process measure is a composite of 10 process measures drawn from CMS's Hospital Compare data that have been log‐transformed to reduce skew and standardized.
Patient satisfaction is a composite of the Hospital Consumer Assessment of Healthcare Providers and Systems, including measures of Communication with Nurses, Communication with Doctors, Responsiveness of Hospital Staff, Pain Management, Cleanliness and Quietness of Hospital Environment, Communication about Medicines, Discharge Information, and Overall Rating of Hospital.
Efficiency is measured as the total expenditures per adjusted patient day.
The number of observations for each model varies due to the number of respondents to each outcome measure.
***p < .01, *p < .10.