Skip to main content

Table 2 Negative binomial mixeda effects regression: Association between number of visits and patient characteristics

From: Identification of population characteristics through implementation of the Comprehensive Diabetic Retinopathy Program

 

Cohort: visited the clinic for at least 12 months

 

n = 963

Effect

Estimate

Standard Error

t Value

Pr > |t|

Age

0.00007

0.002

0.04

0.969

Female (ref = Male)

0.03

0.04

0.75

0.454

Race

 Black or African American (ref = Asian)

−0.03

0.11

−0.28

0.776

 Unknown/other/mixed (ref = Asian)

0.03

0.13

0.22

0.826

 White or Caucasian (ref = Asian)b

0.20

0.10

1.98

0.048

Diabetes type

 Type 1 (ref = Type 2)b

−0.40

0.08

−5.26

<.0001

 Undefined (ref = Type 2)b

0.21

0.06

3.28

0.001

Duration of diabetes 10 years (ref < 10 years)b

0.18

0.05

3.56

0.0004

History of kidney diagnosis Yes (ref = No)

0.08

0.05

1.74

0.082

Maximum systolic blood pressure (scaled by 10)b

0.05

0.01

4.38

<.0001

Maximum diastolic blood pressure (scaled by 10)b

0.16

0.03

6.00

<.0001

Maximum HbA1cb

−0.08

0.03

−2.52

0.012

Has Diabetic Retinopathy (ref = no)

−0.20

0.26

−0.76

0.446

Interaction between maximum HbA1c and DRb

0.09

0.03

2.75

0.006

  1. a Patients’ providers were modeled as a random effect to set apart characteristics attributed to providers rather than to patients
  2. b Denotes statistically significant effects associated with the number of visits at alpha = .05