Table 3: Adjusted odds ratio and 95% confidence interval from logistic regression on any ambulatory care sensitive hospitalizations Medicare current beneficiary survey, 2000–2005.

ALLModel IModel IIModel III
OR95% CISigAOR95% CISigAOR95% CISig

Analysis I (Reference Group = No MM)

Multimorbidity
 MM/MI2.81[2.03, 3.88] *** 1.87[1.32, 2.64] *** 1.62[1.14, 2.30] **
 MM/No MI2.16[1.59, 2.94] *** 1.73[1.27, 2.38] *** 1.54[1.12, 2.11] **
No MMReference groupReference groupReference group

Analysis II (Reference Group = MM/No MI)

Multimorbidity
 MM/MI1.30[1.08, 1.57] ** 1.08[0.87, 1.34]1.05[0.84, 1.31]
MM/No MIReference groupReference groupReference group
 No MM0.46[0.34, 0.63] *** 0.58[0.42, 0.79] *** 0.65[0.47, 0.89] **

Note: Analytic sample consists of 8,963 Medicare beneficiaries who were followed for 3 years (described as panels) and were first interviewed either in 2000, 2001, 2002, or 2003 and not enrolled in Medicare Health Maintenance organizations during the observation years.
The logistic regressions also include intercept terms not presented here. Asterisks represent significant group differences compared to the reference group based on logistic regressions on any ambulatory care sensitive hospitalizations.
Model I: logistic regressions only controlled for multimorbidity categories.
Model II: logistic regressions additionally controlled for panel, gender, race/ethnicity, age, marital status, metro status, education, Medicaid coverage, private insurance coverage, health status, functional status, body mass index, and smoking status.
Model III: logistic regressions additionally controlled for provider-type variable along with all the variables included in Model II.
*** ; ** ; * .