Skip to main content

Table 5 Marginal willingness-to-pay estimates

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

  MNL model HMNL model EMNL model MMNL model LCMNL model
Variables: \( \widehat{\mathrm{MWTP}}\left(95\%CI\right) \) \( \widehat{\mathrm{MWTP}} \)(95% CI) \( \widehat{\mathrm{MWTP}}\left(95\%CI\right) \) \( \widehat{\mathrm{MWTP}}\left(95\%CI\right) \) \( \widehat{\mathrm{MWTP}}\left(95\%CI\right) \)
INTRAVENOUS −2.12 (−4.09, −0.16)* 1.17 (−0.74, 3.08) 3.43 (1.73, 5.14)* 2.31 (−2.86, 7.49)* 5.85 (−50.38, 62.07)
SUBCUTANEOUS 2.35 (1.37, 3.32)* 2.60 (1.58, 3.63)* 0.23 (−0.52, 1.03) −8.33 (−10.31, −6.34)* 2.42 (−7.76, 12.59)
INTRAMUSCULAR −17.43 (−18.65, −16.22)* −19.19 (−20.41, −17.97)* −26.08 (−27.15, −25.01)* −9.39 (−11, −7.77)* −40.63 (−72.47, −8.80)*
DOSFREQ −0.08 (−0.09, −0.08)* −0.09 (−0.094, −0.085)* −0.12 (−0.13, −0.12)* −0.10 (−0.10, −0.09)* −0.23 (−0.30, −0.17)*
NONCLINICAL_SELF 21.45 (20.51, 22.39)* 22.41 (21.48, 23.35)* 26.35 (25.55, 27.16)* 31.01 (29.87, 32.14)* 28.57 (21.40, 35.74)*
NONCLINICAL_SUPV −15.45 (−16.06, −14.83)* −15.74 (−16.36, −15.11)* −15.46 (−15.95, −14.96)* −19.83 (−20.64, −19.02)* −15.53 (−26.04, −5.03)*
DDA_MODERATE1 50.09 (48.91, 51.27)* 49.40 (48.15, 50.65)* 50.14 (49.02, 51.27)* 58.88 (57.65, 60.11)* 84.71 (67.59, 101.82)*
DDA_MODERATE2 −34 (−35.15, −32.85)* −35.24 (−36.51, −33.97)* −38.86 (−39.89, −37.83)* −43.13 (−44.35, −41.92)* −58.84 (−72.28, −45.40)*
DDA_SEVERE −66.36 (−67.74, −64.99)* −68.64 (−70.14, −67.14)* −71.55 (−72.80, −70.30)* −73.64 (−75.15, −72.14)* −94.50 (−117.91, −71.10)*
RAE_MODERATE 22.06 (21.20, 22.92)* 22.64 (21.74, 23.53)* 26.37 (25.59, 27.14)* 30.32 (29.23, 31.41)* 23.49 (17.29, 29.70)*
RAE_SEVERE −94.65 (−96.64, −92.65)* −94.77 (−96.85, −92.70)* −94.75 (−96.58, −92.93)* −116.08 (−117.65, −114.51)* −130.32 (−165.12, −95.53)*
  1. Notes: The 95% CIs above are “standard or classical confidence intervals” calculated using 100 bootstrapped replicates of MWTP. This is because accurate, less-erratic and reliable “bootstrap confidence intervals” require replications in the order of 1000, which would have been computationally demanding and time consuming [32]. The confidence intervals reported are therefore not exact. * indicates confidence interval does not include zero