Skip to main content
. 2016 Jul 8;16:238. doi: 10.1186/s12913-016-1470-y

Table 7.

Results of the multivariate regression analysis of the impact of physician dispensing (PD) on total health care expenditures

Specification Glm, log link gamma family
PD 0.989 (0.008)
Age 1.005 (0)***
Gender = m 0.977 (0.007)**
PPO-contract 0.953 (0.012)***
Telemedicine contract 0.996 (0.011)
HMO contract 0.965 (0.008)***
High deductible 0.87 (0.008)***
Nursing home stay 3.126 (0.069)***
Outpatient hospital visits 1-5 1.783 (0.016)***
Outpatient hospital visits >5 2.427 (0.048)***
Inpatient stay in hospital 4.003 (0.058)***
Cost in past year (log) 1.117 (0.002)***
Prescriptions psychiatrists 2.252 (0.054)***
Prescriptions cardio-/angiologists 1.614 (0.109)***
Prescriptions gynaeologists 1.488 (0.026)***
Prescriptions other specialists 1.78 (0.019)***
Agricultural municipalities 1.03 (0.268)
Affluent municipalities 1.074 (0.041)
Industrial and tertiary municipalities 0.977 (0.042)
Rural commuting municipalities 1.004 (0.044)
Periurban municipalities 1.025 (0.039)
Suburban municipalities 1.022 (0.038)
Touristic municipalities 0.946 (0.088)
Urban centers 1.079 (0.04)*
22 pharmaceutical cost groups all coefficients strongly positive

Target variable per patient, first 6 months of year 2013. Exponential of coefficients [exp(^β)] displayed

Standard errors calculated by the delta method [ŝexp(β) = exp(β) ∗ ŝβ], displayed in parentheses

Significance levels: ***p < 0.001, **p < 0.01, *p < 0.05

Source: Claims data from Helsana insurance company, own calculations