Table 5 Conditional logit estimations for simple model (model I) and models incorporating subject-related interactions (models II-IV)
n = 114 | Model I (simple model) | Model II | Model III | Model IV | |||||
|---|---|---|---|---|---|---|---|---|---|
Variable | \(\hat{\boldsymbol{\beta}}\) (SE) | p-value (LogW) | \(\hat{\boldsymbol{\beta}}\) (SE) | p-value (LogW) | \(\hat{\boldsymbol{\beta}}\) (SE) | p-value (LogW) | \(\hat{\boldsymbol{\beta}}\) (SE) | p-value (LogW) | |
Provider | 0.1404 (0.853) | 0.1354 (0.868) | 0.1275 (0.894) | 0.1398 (0.855) | |||||
L1: Modality manufacturer | -0.0585 (0.0603) | -0.0569 (0.0607) | -0.0589 (0.0608) | -0.0573 (0.0613) | |||||
L2: RIS/PACS provider | -0.0615 (0.0631) | -0.0652 (0.0636) | -0.0653 (0.0637) | -0.0647 (0.0641) | |||||
L3: AI-software startup | 0.12 (0.0608) | 0.1221 (0.0613) | 0.1242 (0.0614) | 0.122 (0.0617) | |||||
Application | <0.0001 (11.578) | 0.0013 (2.889) | 0.0014 (2.852) | <0.0001 (4.586) | |||||
L1: Diagnostics (routine diagnostics) | 0.2984 (0.0622) | 0.2909 (0.2433) | 0.2886 (0.2387) | 0.1655 (0.2519) | |||||
L2: Process efficiency (scan time reduction) | -0.3896 (0.0579) | -0.7046 (0.2504) | -0.6696 (0.238) | -0.8704 (0.2511) | |||||
L3: Screening support (mammography) | 0.0911 (0.0642) | 0.4136 (0.2434) | 0.381 (0.2413) | 0.7049 (0.254) | |||||
Quality | <0.0001 (9.727) | 0.0524 (1.281) | 0.0521 (1.283) | 0.0519 (1.285) | |||||
L1: Same | -0.2421 (0.039) | -0.3084 (0.1583) | -0.3033 (0.153) | -0.3032 (0.1531) | |||||
L2: Better | 0.2421 (0.039) | 0.3084 (0.1583) | 0.3033 (0.153) | 0.3032 (0.1531) | |||||
Time savings | <0.0001 (13.702) | <0.0001 (14.064) | <0.0001 (14.127) | <0.0001 (14.449) | |||||
L1: Low | -0.4339 (0.0637) | -0.4414 (0.0643) | -0.4427 (0.0644) | -0.451 (0.0649) | |||||
L2: Medium | 0.039 (0.0584) | 0.0382 (0.0589) | 0.0382 (0.059) | 0.0398 (0.0594) | |||||
L3: High | 0.3949 (0.0587) | 0.4031 (0.0592) | 0.4045 (0.0593) | 0.4113 (0.0598) | |||||
Price | Price per study | -0.1607 (0.0176) | <0.0001 (20.895) | -0.1634 (0.0178) | <0.0001 (21.275) | -0.1593 (0.0178) | <0.0001 (21.185) | -0.1618 (0.018) | <0.0001 (20.554) |
No-choice | No-choice | -1.4499 (0.1198) | <0.0001 (33.422) | -1.4611 (0.1209) | <0.0001 (33.361) | -1.4762 (0.1214) | <0.0001 (33.885) | -1.4851 (0.1222) | <0.0001 (33.926) |
Subject-related interactions | Gender[M]* Application[Diagnostics] | -0.0899 (0.246) | 0.0014 (2.862) | -0.0903 (0.2414) | 0.0014 (2.850) | -0.0783 (0.2417) | 0.001 (3.009) | ||
Gender[M]* Application[Process] | 0.3435 (0.2523) | 0.306 (0.2402) | 0.3241 (0.2403) | ||||||
Gender[M]* Application[Screening] | -0.2536 (0.2453) | -0.2157 (0.2434) | -0.2458 (0.244) | ||||||
Gender[F]* Quality[Better] | 0.1605 (0.1663) | 0.0161 (1.793) | 0.1571 (0.1611) | 0.021 (1.679) | 0.1585 (0.1613) | 0.0205 (1.689) | |||
Budget responsibility[Y]* Price | -0.0347 (0.0105) | 0.001 (3.005) | -0.0366 (0.0106) | 0.0005 (3.278) | |||||
Specialization[Mammography]*Application [Screening] | 0.3873 (0.0932) | <0.0001 (4.068) | |||||||
Model fit | AICc | 2197.69 | 2186.10 | 2177.31 | 2162.70 | ||||
BIC | 2242.88 | 2261.25 | 2257.44 | 2252.79 | |||||
-2LogLikelihood | 2179.53 | 2155.67 | 2144.82 | 2126.10 | |||||
LogLikelihood | -1089.77 | -1077.84 | -1072.41 | -1063.05 | |||||