Skip to main content

Table 4 Multilevel logistic regression analysis for the probability of individual dental care utilization in the last twelve months (n = 375)

From: Contacts with general practitioners, dentists, and medical specialists among nursing home residents: a cross-sectional study in 44 German nursing homes

 

Model 1: empty model

Model 2: +individual predictors

Model 3: +contextual predictors

OR

95% CI

OR

95% CI

Individual-level variables

Sex (ref. female)

  Male

  

0.86

(0.49-1.50)

0.88

(0.50-1.54)

Age group (ref. 60-74 years)

  75-84 years

  

1.76

(0.83-3.75)

1.87

(0.87-4.00)

  85+ years

  

2.27

(1.03-4.98)

2.43

(1.10-5.38)

Long-term care grade (ref. 1/2)

  3

  

1.31

(0.64-2.66)

1.30

(0.64-2.65)

  4/5

  

0.98

(0.48-1.98)

0.99

(0.49-2.02)

Years of residence (ref. quartile 1)

  Quartile 2

  

0.90

(0.45-1.79)

0.82

(0.41-1.64)

  Quartile 3

  

0.95

(0.47-1.93)

0.88

(0.43-1.80)

  Quartile 4

  

0.76

(0.36-1.59)

0.70

(0.33-1.47)

Nursing home-level variables

Sponsorship (ref. independent nonprofit)

  Private

    

0.85

(0.34-2.08)

  Public/municipal

    

2.31

(0.36-15.07)

Number of nursing home beds (ref. quartile 1)

  Quartile 2

    

1.01

(0.30-3.37)

  Quartile 3

    

0.95

(0.25-3.68)

  Quartile 4

    

2.50

(0.68-9.22)

Transport service for contacts at a practice (ref. no)

  Yes

    

1.45

(0.55-3.85)

Medical escort service for contacts at a practice (ref. no)

  Yes

    

1.27

(0.50-3.23)

Mea’sures of variation

Nursing home-level variance (SE)

0.822 (0.392)

1.066 (0.489)

0.901 (0.426)

Proportional change in variance

  

+ 29.8%

−15.5%

MOR

2.37

2.68

2.47

Fit statistics

–2 log likelihood

461.6

453.6a

448.4a

  1. Abbreviations: OR odds ratio, CI confidence interval, ref reference, SE standard error, MOR median odds ratio
  2. Boldface indicates statistical significance
  3. aNo significant likelihood ratio test
  4. This analysis included only 375 nursing home residents because for 25 residents, no complete information was available for the nursing home-level variables