Table 2.
Use or Cost Measure per Year | Controls (n = 9035) | SS Participants (n = 4766) | Adjusted Mean Differencea (95% Confidence Interval) | P Value |
---|---|---|---|---|
Health Care Use | ||||
No. (%) of people with an inpatient admission | ||||
Baseline | 825 (9.1) | 432 (9.1) | NA | NA |
Year 1 | 984 (10.9) | 454 (9.5) | –1.0% (–2.1% to –0.1%) | .05 |
Year 2 | 1129 (12.5) | 471 (9.9) | –2.3% (–3.3% to –1.2%) | <.001 |
No. of primary care visits per person | ||||
Baselineb | 4.5 (5.0) | 5.1 (4.3) | NA | NA |
Year 1 | 4.7 (4.6-4.8) | 5.3 (5.2-5.5) | 0.40 (0.27-0.53) | <.001 |
Year 2 | 4.8 (4.7-4.9) | 5.3 (5.2-5.4) | 0.26 (0.13-0.40) | <.001 |
No. of specialty care visits per person | ||||
Baselineb | 2.7 (3.2) | 3.2 (3.3) | NA | NA |
Year 1 | 2.9 (2.8-3.0) | 3.3 (3.2-3.4) | 0.22 (0.11-0.33) | <.001 |
Year 2 | 3.0 (2.9 to 3.1) | 3.4 (3.3-3.5) | 0.25 (0.14-0.36) | <.001 |
Health Care Costs per Person, $ | ||||
Total health care costs | ||||
Baselineb | 4693 (7288) | 5212 (8530) | NA | NA |
Year 1 | 5687 (5486-5888) | 5677 (5388-5966) | 2 (–341 to 344) | .99 |
Year 2 | 6742 (6480-7003) | 6155 (5843-6466) | –500 (–892 to –106) | .01 |
Inpatient admission costs | ||||
Baselineb | 1000 (4381) | 1248 (6182) | NA | NA |
Year 1 | 1391 (1268-1515) | 1346 (1130-1561) | –32 (–279 to 214) | .80 |
Year 2 | 1803 (1644-1963) | 1497 (1283-1711) | –270 (–533 to –6) | .05 |
Primary care costs | ||||
Baselineb | 788 (876) | 911 (871) | NA | NA |
Year 1 | 829 (810-849) | 962 (937-988) | 101 (70-133) | <.001 |
Year 2 | 875 (854-896) | 983 (956-1010) | 80 (46-113) | <.001 |
Specialty care costs | ||||
Baselineb | 716 (1254) | 793 (1213) | NA | NA |
Year 1 | 813 (783-843) | 825 (792-857) | –14 (–58 to 29) | .51 |
Year 2 | 890 (860-922) | 935 (895-975) | 37 (–12 to 86) | .14 |
Values are expressed as either mean (SD) or mean (95% confidence interval). NA indicates not applicable.
Adjusted mean difference is defined as the change from baseline in participants minus the change from baseline in controls. Differences were calculated using multivariate linear regression models with robust standard error estimates that controlled for age, sex, preventive services index, RxRisk (a measure of chronic disease burden and comorbidity [20]), indication of arthritis on the outpatient visit problem list, inclusion in the health plan's diabetes or heart registries, and baseline measures of health care utilization and cost.
Two-tailed t tests and chi-square tests were used for unadjusted comparisons between controls and SS participants at baseline; P < .05.