Skip to main content

Table 4 Econometric results

From: Discrete-choice modelling of patient preferences for modes of drug administration

Dependent variable: CHOICE PROBABILITY
  MNL model HMNL model EMNL model MMNL model LCMNL model
Variables/Coefficients: \( \widehat{\upbeta}(SE) \) \( \widehat{\upbeta} \)(SE) \( \widehat{\upbeta}(SE) \) \( \widehat{\upbeta_{\mathrm{s}}}(SE) \) \( {\widehat{\upbeta}}_1(SE) \) \( {\widehat{\upbeta}}_2(SE) \)
INTRAVENOUS −0.011 (0.091) 0.005 (0.057) 0.017 (0.042) −0.026 (0.110) −0.119 (0.439) 0.021 (0.123)
SUBCUTANEOUS 0.021 (0.043) 0.016 (0.027) 0.002 (0.019) −0.016 (0.049) −0.073 (0.130) 0.070 (0.060)
INTRAMUSCULAR −0.136 (0.049)** −0.090 (0.031)** −0.111 (0.037)** −0.123 (0.053)* −0.560 (0.210)** −0.219 (0.064)***
DOSFREQ −0.001 (0.000)*** −0.0005 (0.000)*** −0.001 (0.000)** −0.001 (0.000)*** −0.003 (0.001)*** −0.001 (0.000)**
NONCLINICAL_SELF 0.167 (0.037)*** 0.108 (0.025)*** 0.112 (0.038)** 0.210 (0.039)*** 0.424 (0.128)*** 0.146 (0.045)**
NONCLINICAL_SUPV −0.122 (0.028)*** −0.078 (0.019)*** −0.067 (0.025)** −0.120 (0.030)*** −0.124 (0.093) −0.098 (0.037)**
DDA_MODERATE1 0.391 (0.033)*** 0.239 (0.033)*** 0.214 (0.074)** 0.437 (0.038)*** 1.210 (0.150)*** 0.264 (0.050)***
DDA_MODERATE2 −0.267 (0.042)*** −0.171 (0.032)*** −0.167 (0.059)** −0.300 (0.046)*** −1.058 (0.141)*** −0.075 (0.061)
DDA_SEVERE −0.525 (0.039)*** −0.336 (0.044)*** −0.308 (0.103)** −0.588 (0.044)*** −1.376 (0.184)*** −0.324 (0.057)***
RAE_MODERATE 0.169 (0.034)*** 0.108 (0.024)*** 0.111 (0.041)** 0.212 (0.038)*** 0.522 (0.141)*** 0.067 (0.043)
RAE_SEVERE −0.743 (0.039)*** −0.461 (0.056)*** −0.406 (0.138)** −0.869 (0.044)*** −2.342 (0.215)*** −0.204 (0.059)***
COST −0.008 (0.001)*** −0.005 (0.001)*** −0.004 (0.001)** −0.0118 (0.040)*** −0.012 (0.002)*** −0.008 (0.001)***
A −0.054 (0.054) −0.033 (0.034) 0.002 (0.025) −0.073 (0.066) 0.370 (0.274) −0.122 (0.073)!
Entropy\( \left(\widehat{\uptheta_1},\widehat{\uptheta_2}\right) \) (1.026, 0.717)  
\( {\widehat{\alpha}}_0 \)(FEMALE) 0.170 (0.035)***  
\( {\widehat{\alpha}}_1 \)(VOCATIONAL) −0.230 (0.085)**  
\( {\widehat{\alpha}}_2 \)(GCSEs_O + A) −0.199 (0.054)***  
\( {\widehat{\pi}}_c \) 0.49 0.51
\( {\widehat{\gamma}}_c \)(VOCATIONAL) −0.892 (0.324)**
\( {\widehat{\gamma}}_c \)(GCSEs_O + A) 0.577 (0.203)**
\( {\widehat{\gamma}}_c \)(FEMALE) 0.285 (0.122)*
\( {\widehat{\gamma}}_c \)(RESPONDENTAGE)      0.024 (0.009)**  
Log-likelihood (AIC) −2855.357 (5736.713) −2812.19 (5670.381) −2837.971 (5670.381) −2782.51 (5611.019) −2652.879 (5379.757)  
  1. Notes: SE = standard error. For the HMNL and LCMNL models, we report selected effects of respondent-characteristics on the scale-parameter and latent-class membership. MMNL model was estimated using 500 Halton draws of correlated normally-distributed coefficients for the variables A, INTRAVENOUS and SUBCUTANEOUS and a log-normal distributed cost coefficient. *** p < 0.001 ** p < 0.01 * p < 0.05! p < 0.10. AIC = Akaike Information Criterion