. 2015 Sep;8(6):319–329.

Table 4.

Log-Transformed Linear Regression Model Comparing Drivers of Healthcare Expenditures Among Patients with Newly Diagnosed Diabetes in 2009 and 2011

Parameter	Estimate	SE	t value	P value
Intercept	6.187	0.04	143.79	<.0001
Newly diagnosed diabetic patients in 2009	−0.034	0.02	−1.54	.1259
Age	0.004	0.0008	4.81	<.0001^a
Utilization variable
Home health agency expenses	0.000	0.0001	2.27	.02^a
Orthodontist	0.220	0.01	15.47	<.0001^a
General dental visits	0.056	0.01	6.9	<.0001^a
Hospital discharges	0.645	0.02	39.28	<.0001^a
Nights in hospital	0.040	0.00	22.69	<.0001^a
Ambulatory optometrist	0.207	0.02	13.02	<.0001^a
Emergency department visits	0.241	0.01	17.3	<.0001^a
Outpatient physician visits	0.181	0.01	28.75	<.0001^a
Outpatient provider visits	0.025	0.006	56.11	<.0001^a
Nonphysician visits	0.023	0.003	8.18	<.0001^a
Physician visits	0.044	0.002	19.6	<.0001^a
Prescription medication refills	0.014	0.004	33.14	<.0001^a
Comorbidities
Cancer	0.162	0.02	6.7	<.0001^a
High cholesterol	0.207	0.01	16.85	<.0001^a
Obesity	0.140	0.02	6.38	<.0001^a
Arthritis	0.113	0.02	5.5	<.0001^a
Asthma	0.282	0.03	10.77	<.0001^a
Heart attack	0.294	0.02	15.52	<.0001^a
Stroke	0.158	0.05	3.36	.001^a
Insurance type
Private insurance	0.299	0.02	16.51	<.0001^a
Public insurance	−0.415	0.08	−5.09	<.0001^a
Uninsured	−0.223	0.03	−6.7	<.0001^a

Significant estimates at P <.05 only.

SE indicates standard error.