| ACSC Hospitalization, 2008–2009 | |
|---|---|
| Primary care practice supports (extent of use) | |
| PCP communicates with patient and with other specialists about specialist care received | |
| Lower tercile (reference) | 1 |
| Middle tercile | 0.88 (0.80, 0.96)* |
| Upper tercile | 0.81 (0.74, 0.89)† |
| Extent of HIT use | |
| Lower tercile (reference) | 1 |
| Middle tercile | 1.08 (0.99, 1.18) |
| Upper Tercile | 1.07 (0.98, 1.18) |
| Registry or patient list (vs none) | 1.00 (0.92, 1.09) |
| Nurse and patient educator care management | |
| Lower tercile (reference) | 1 |
| Middle tercile | 1.03 (0.94, 1.13) |
| Upper tercile | 1.02 (0.93, 1.12) |
| Quality and performance measurement | |
| Lower tercile (reference) | 1 |
| Middle tercile | 0.98 (0.83, 1.16) |
| Upper tercile | 1.00 (0.83, 1.22) |
| Patient characteristics | |
| HCC score (already adjusted for age, sex, and dual Medicaid eligibility) | |
| Lower tercile (reference) | 1 |
| Middle tercile | 2.18 (1.89, 2.51)† |
| Upper tercile | 5.30 (4.66, 6.04)† |
| ACSC hospitalization in prior year (vs none)‡ | 5.78 (5.05, 6.62)† |
| Income (median) in ZIP code | |
| Lower tercile | 1.28 (1.16, 1.40)* |
| Middle tercile | 1.11 (1.02, 1.22)§ |
| Upper tercile (reference) | 1 |
| Patient race/ethnicity | |
| White | 1 |
| Black | 1.13 (0.97, 1.30) |
| Hispanic | 1.09 (0.85, 1.41) |
| Other | 0.87 (0.65, 1.16) |
| Practice characteristics | |
| Revenue from Medicare (%) | |
| 0–30 (reference) | 1 |
| 31–50 | 1.16 (1.05, 1.27)* |
| 51–100 | 1.17 (1.07, 1.29)† |
| Practice size (no. of physicians) | |
| 1–2 | 1 |
| 3–10 | 1.04 (0.95, 1.14) |
| 11–49 | 1.04 (0.93, 1.17) |
| ≥50 | 0.97 (0.85, 1.11) |
| Practice type | |
| Independent practice (physician owned) | 1 |
| Community health center | 0.74 (0.50, 1.08) |
| Hospital-based outpatient practice/clinic† | 0.96 (0.77, 1.19) |
Data are odds ratio (95% confidence limits). All estimates are adjusted for all the other variables listed in the first column as well as for the urban influence codes (large metro, small metro, micropolitan, rural). All analyses were conducted in SUDAAN and accounted for clustering of patients within physicians.
Data Source: Linked data from the nationally representative Center for Studying Health System Change (HSC) Physician Survey (2008) and Medicare fee-for-service claims for the years 2007 to 2009 for all beneficiaries for whom a physician from the HSC national survey was their usual source of care. Usual source of care was determined by the plurality algorithm for evaluation and management visits.
↵* P < .01 vs reference group, two-sided test.
↵† P < .001 vs reference group, two-sided test.
↵‡ Ambulatory care–sensitive conditions (ACSCs) examined included congestive heart failure, chronic obstructive pulmonary disease, asthma and diabetes. ACSC hospitalizations for any one or more of these conditions (combined rate) were calculated using claims for 2008 and 2009 (the numerator). The denominator for all analyses is 123,760 patients with one or more of these chronic conditions (identified in 2007–2008 claims).
↵§ P < .05 vs reference group, two-sided test.
HCC, hierarchical condition category score; HIT, health information technology; PCP, primary care physician.